1323 Second Edition telecommunicat 5ns. & Economic Development 4 *~~~~~~~ * & . * * .,: Robert J. Saunders Jeretmy J. Warford Bjorn Welleniits A World Bank Publication Telecommunications and Economic Development Second Edition A WORLD BANK PUBLICATION Telecommunications and Economic Development Second Edition Robert J. Saunders Jeremy J. Warford Bjorn Wellenius Published for The World Bank The Johns Hopkins University Press Baltimore and London Copyright © 1994 by the International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. The Johns Hopkins University Press Baltimore, Maryland 21218, U.S.A. The views and interpretations in this book are the authors' and should not be attributed to the World Bank, to its affili- ated organizations, or to any individual acting on their behalf. First printing April 1994 Second printing May 1996 EDITOR Elizabeth Forsyth FIGURES Kathy Rosen COVER DESIGN Joyce Petruzzelli The authors wish to thank the International Telecommuni- cations Union for its kind permission to use the cover draw- ing by Martin Widiarsono of Indonesia. The drawing was included in the ITU's 1991 "Youth in the Electronic Age" international art exhibition. Library of Congress Cataloging-in-Pubication-Data Saunders, Robert J. Telecommunications and economic development / Robert J. Saunders, Jeremy J. Waford, Bjorn Wellenius-2d ed. p. cm. Includes bibliographical references and index. ISBN 0-8018-4665-X 1. Telecommunication-Developing countries. 2. Telecommunication- Mathematical models. I. Warford, Jeremy J. II. Wellenius, Bjmrn. III. Title. HE8635.S28 1994 384'.09172'4-dc2O 93-40704 CIP Contents Preface to the Second Edition xv List of Abbreviations xviii Part 1. An Introductory Perspective Chapter 1. The Role of Telecommunications in Economic Development 3 Signals for Expanding Investment 4 Benefits from Improved Telecommunications 22 Scope and Organization of the Book 29 Chapter 2. A Perspective on Technology and Costs 37 Technological Change 37 Economies of Scale 46 Constraints on Achieving Economies of Technology and Scale 54 Combined Effects of Technology and Scale on Cost and Price of Service 56 Chapter 3. Sector Organization, Management, and Financing 64 Telecommunications Organization in Developing Countries 64 The Autonomy Issue 65 Internal Organization and Management of Operating Entities 68 The Financial Constraint 74 Part 11. Macroeconomic Analysis of Benefits Chapter 4. Aggregate Correlation Analysis 85 Correlation and Regression Studies 86 Limitations of Aggregate Analysis 94 Value of Aggregate Analysis 95 v vi CONTENTS Chapter 5. Structural Economic Analysis 99 A Cross-Country Comparison of Input-Output Coefficients 101 Information Sector Analysis 107 Value of Input-Output Analysis 114 Problems with Input-Output Analysis 116 Chapter 6. Location and Communication 121 The Spatial Framework 121 Regional Development and Balance 125 Migration between Rural and Urban Areas 128 Metropolitan Development and Balance 129 Chapter 7. Transport, Communication, and Energy Consumption 135 Improved Use of Vehicles 137 Substitution of Telecommunications for Travel 142 Conclusion 151 Part 111. Microeconomic Analysis of Benefits Chapter 8. Price Change and Best-Alternative Estimates of Consumer Surplus 157 Internal Rate of Return and Consumer Surplus 158 Estimating Consumer Surplus by the Price-Change Method 161 Estimating Consumer Surplus by the Best-Alternative Method 164 Chapter 9. Expenditure Method Estimates of Consumer Surplus 173 Estimating Consumer Surplus by the Expenditure Method 174 Conclusions from Microeconomic Analysis 189 Part IV. Telephone Access and Use Chapter 10. Characteristics of Telephone Subscribers 199 Business and Residential Telephones 200 Users of Business Telephones in Urban Areas 203 Users of Business Telephones in Towns and Villages 207 Users of Residential Telephones 212 Chapter 11. Use of Business and Residential Telephones 219 Business Telephone Use 219 Residential Telephone Use 225 Telephones and Other Means of Communication 229 Chapter 12. Use of Public Call Office Telephones 234 Characteristics of Pco Telephone Users 236 Purposes of Public Telephone Calls 240 Importance of Pco Calls and Alternatives for Communication 244 Constraints on Access to Pcos 246 Conclusions about Telephone Access and Use 249 CONTENTS vii Part V. Telecommunications Tariff Policy Chapter 13. An Introduction to Tariff Policy 255 Objectives of Public Enterprise Tariff Policy 255 Traditional Approaches to Telecommunications Pricing 258 Chapter 14. Tariff Policy for Economic Efficiency 270 Pricing 270 An Investment Strategy to Complement Efficiency Pricing 279 Chapter 15. Tariff Policy in Practice 286 Estimating Marginal Cost 286 Tariff Analysis in Practice 290 Tariff Regulation 295 Part VI. Mobilizing Resources and Promoting Efficiency: Alternative Approaches Chapter 16. Restructuring the Telecommunications Sector 303 Evolving Views on the Role of Telecommunications in Development 303 Changing Sector Structures 305 The Beginnings of Telecommunications Reform in Developing Countries 313 Closing Remarks: Looking Forward 328 Appendixes Appendix A. Applications of Telecommunications for the Delivery of Social Sciences 335 Appendix B. Posts and Telecommunications as an Integrated Communications Sector 360 Appendix C. A Survey of Rural Public Call Offices in Costa Rica 379 Appendix D. Shadow Pricing for Telecommunications 413 Appendix E. World Bank Loans and Credits 417 Appendix F. The Canadian Radio-Television and Telecommunications Commission Costing Methodology: Cost Allocation Categories 422 References 426 Index 455 Tables 1-1. Distribution of the World's Telephone Main Lines, by Region, Population, and Income, 1988 4 viii CONTENTS 1-2. Proportion of Telephone Main Lines and GNP per Capita in Selected Countries, 1987-88 6 1-3. Access to Telephone Services in Main Cities and Other Areas in Selected Countries, as of January 1, 1988 8 1-4. Growth of Telephone Main Lines in Service in Selected Developing Countries, 1979-88 10 1-5. Growth of Telex Lines in Service in Selected Developing Countries, I 979-88 11 1-6. Number of Facsimile Stations in Selected Industrial and Developing Countries, 1986-88 12 1-7. Growth in the Number of Privately Leased Circuits for Data Communications in Selected Industrial and Developing Countries, 1981 and 1986-88 13 1-8. Supply of and Expressed Demand for Telephone Main Lines in Selected Developing Countries, Various Years, 1985-89 14 2-1. Business Telephone Expenditures under Selected Model Assumptions in Selected Developing Countries 57 2-2. Residential Telephone Expenditures under Selected Model Assumptions in Selected Developing Countries 58 3-1. Trends in Telecommunications Lending by the World Bank, Fiscal Years 1986-90 78 5-1. Communications Output Distribution Coefficients, by Sector, for Data Set B 103 5-2. Communications Input Coefficients, by Sector, for Data Set A 103 5-3. Direct Input Coefficients for Each Aggregate Sector's Use of Communications Services, for Data Set B 104 5-4. Direct and Indirect Input Coefficients for Each Aggregate Sector's Use of Communications Services, for Data Set B 106 5-5. Purchase of Inputs in Dollars per $1,000 of Output from the Communications Sector, for Data Set B 108 5-6. Normalized Income Multipliers, for Data Set B 109 7-1. Successful and Unsuccessful Trips by Farmers to Major Fertilizer Distribution Centers in India, 1975-76 140 7-2. Annual Losses due to Lack of Adequate Communications Infrastructure for Transportation Routes in the People's Democratic Republic of Yemen, 1986-87 141 8-1. Internal Rates of Return from Investments in Rural PcO Telephones in Chile, 1978 166 8-2. Factory Expenditure Savings Data from Bangladesh, 1981 168 8-3. Expenditure for Communication by Bus and by Telephone for Thirty-six Villages in Andhra Pradesh State, India, 1981 169 9-1. Two Estimates of Consumer Surplus for Telephone Calls Made in a Rural Costa Rican Village, 1976 175 9-2. Delay of Ordinary and Urgent Calls Booked, by Major Route in Madras, India, 1969 179 CONTENTS ix 9-3. Estimated Average Increment for Nine Firms in the Percentage of Profits to Revenues, by Benefit Category, in Nairobi, Kenya, 1981 181 9-4. Estimated Benefits and Expansion Costs, by Firm, in Nairobi, Kenya, 1981 181 9-5. Yearly Average Frequency and Duration of Outages Affecting Industrial Consumers, with and without the New Telecommunications System, in Bahia, Brazil, 1981-2000 183 9-6. Estimated Foreign Exchange Expenditure Related to Petroleum Consumed by Three Business Organizations because of Inadequate Telephone Service, in Tanzania, 1982 184 9-7. Benefit to Cost Ratio Following Installation of Telephone Service, by Sector and Region, in the Philippines, 1984 187 10-1. Estimated Location of Telephones, by Type of Establishment, in Nepal, 1968 and 1977 202 10-2. Telephone Subscribers, by Type of [Jser, in Ethiopia: Addis Ababa and Four Small Towns, 1982 203 10-3. Number of Telephone Subscribers, by Functional Category, in Selected Towns in Burkina Faso, 1980, and Mali, 1979 204 10-4. Distribution of Telephone Subscribers, by Type of Economic Activity, in Thailand, 1980 206 10-5. Distribution of Telephone Subscribers, by Primary Occupation and Income, in Ten Small Rural Exchange Areas of India, 1977-78 208 10-6. Telephone Subscribers, by User Category, in Twelve Small Exchanges in Four Divisions (Regions) and the Dhaka Multzexchange Area of Bangladesh, December 1980 209 10-7. Telephone Subscribers, by User Category, in Four Small Manual-Exchange Towns from the Eastern, Western, Northwestern, and Central Regions of Ethiopia, 1982 210 10-8. Estimated Average Monthly Calls per Telephone, by Sector, in Two Towns in Nepal, 1979-80 211 10-9. Telephone Calling Rates for Telephone Subscribers, by Economic Category, in Small Towns and Rural Centers in Two Regions of Kenya, 1980 212 10-10. Distribution of Telephones, by Functional Location, in Eighteen Small Locales in Syria, 1976 213 10-11. Occupations of a Sample of Seventy-eight Residential Subscribers in Kathmandu, Nepal, 1980 215 10-12. Distribution of Residential Telephones, by Occupation of Subscriber, in Eighteen Small Locales in Syria, 1976 216 10-13. Emplovment Categories for 163 Residential Telephone Subscribers in Twelve Small Towns in Costa Rica, 1980 216 11-1. Proportion of Subscribers Making Local and Long-Distance Calls in Twelve Villages in Andhra Pradesh State, India, 1981 220 11-2. Most Frequent Reasons for Using Primarily Business Telephones for Two Types of Establishments in Kathmandu and Pokhara and for Waiters in Kathmandu, Nepal, 1980 221 x CONTENTS 11-3. Frequency of Reasons Given for Telephone Contact by a Sample of Telephone Users in Two Regions of Kenya, 1980 222 11-4. Reasons Given for Most Frequent Telephone Contact by a Sample of Telephone Users from Three Medium-size Towns in Ethiopia, 1982 223 11-5. Percentage of Important Telephone Contacts between Sectors, for a Sample of Telephone Subscribers, Waiters, and Coin Box Users in Two Regions of Kenya, 1980 224 11-6. Reasons for Using Residential Telephones in Several Countries 226 11-7. Preferences for Alternative Communication Media of 208 Residential Telephone Subscribers and Applicants in Thailand, 1980 230 11-8. Means of Communication Used for Economic Contact by a Sample of Telephone Subscribers in Two Regions of Kenya, 1980 231 11-9. Nfeans of Communication lJsed by 474 Male Respondents during Twelve Months in Nine Villages in Egypt 232 12-1. Interurban Telephone Calls, by Type of Subscriber, for the Manual-Exchange Service Town of Zway, Ethiopia, 1982 235 12-2. Characteristics of Persons Using Rural Pco Telephones in Five Countries 237 12-3. Distance Traveled to a Pco to Make a Telephone Call in Three Countries 239 12-4. Reasons for Using Rural Pco Telephones in Several Countries 242 12-5. Perceived Urgency of Pco Telephone Calls in India and Papua New Guinea 244 12-6. Rural Pco Calls That Would Be Substituted by Alternative Means of Communication if Telephones Were Not Available in Five Countries 245 12-7. Alternative Means of Communication Chosen by Telephone and Telegraph Users in Senegal, 1983 246 15-1. TLRIC Analysis of Existing Tariffs in an Asian Country 291 15-2. Existing Tariffs Evaluated in Terms of Policy Objectives 292 C-l. Average Estimated Price Elasticity of Demand for Rural Pco Telephone Traffic (Metered Pulses) for Selected Groups of Villages in Costa Rica, 1974-75 382 C-2. Average Duration of Outgoing Calls and Price Elasticities, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 383 C-3. Economic and Demographic Variables Available for the Cross-Village Examination of Rural Public Telephone Use in Costa Rica, 1973-74 386 C-4. Regression Equations in Which Selected Economic and Demographic Variables Are Regressed on Proxies for Village Telephone Traffic 388 C-5. The First Five Factors from a Factor Analysis of Twenty-four Economic and Demographic Variables Compiled for Sixty-four Villages in Costa Rica, 1973-74 390 C-6. Regression Equations in Which Economic and Demographic Variables Selected through Factor Analysis Are Regressed on Proxies for Village Telephone Traffic 394 CONTENTS xi C-7. Selected Characteristics of Three Villages in Costa Rica in Which Surveys of Public Telephone Use Were Made, 1974 396 C-8. Employment Characteristics of Callers from a UJser Survey of Three Rural Villages in Costa Rica, 1974 396 C-9. Summary Distribution of Calls, by Income Class of Caller, for Three Rural Villages in Costa Rica, 1974 397 C-10. Distribution of Calls and Income of Caller, by Village, for Three Rural Villages in Costa Rica, 1974 398 C-l 1. Percentage of Calls Made by the Same Caller for Three Rural Villages in Costa Rica, 1974 399 C-12. Average Monthly Income of Callers According to the Number of Calls Made during the Sample Week for Three Rural Villages in Costa Rica, 1974 399 C-13. Alternative Communication Preferences Jor Callers in Three Rural Villages in Costa Rica, 1974 400 C-14. Primary Purpose of Pco Calls Made in Three Rural Villages in Costa Rica, 1974 401 C-15. Percentage of Calls Generated in Eleven Concessionaire-Survey Villages in Costa Rica, Grouped by Characteristics of the Area, by Purpose of Call, 1974-75 402 C-16. Mean Distance That Callers Traveled to Make Calls, by Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 404 C-17. Percentage of Calls to San Jose, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 404 C-18. Mean Duration of Calls, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 406 C-19. Percentage Distribution of Calls per Day, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 407 C-20. Percentage Distribution of the Times of Day during Which Calls Are Made, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 408 C-2 1. Summary Characteristics of Calls for Three Rural Villages in Costa Rica, 1974 408 E-l. Loans and Credits by the World Bank and the International Development Association for Telecommunications Projects, 1962-89 417 Figures 2-1. Comparative Unit Investment Costs for Three Intercity Transmission Technologies in Canada, 1988 48 2-2. Average Occupation of Circuits as a Function of the Number of Circuits for Two Levels of Probability (p) of Lost Calls 49 xii CONTENTS 3-1. Sources of Funds for Telecommunications Investment in the Developing World, 1980s 75 4-1. Density of Telephone Lines as a Function of GDP per Capita for a Cross-Section of Countries, 1965 87 4-2. Density of Telephone Lines as a Function of :iDP per Capita in Sweden, 1900-65 89 4-3. Telephone Index as a Function of the Development Support Index 91 8-1. Estimating Consumer Surplus from the Effect of a Price Change 162 12-1. Occupations of PcO Telephone Users in an Urban and a Rural Area of Senegal, 1983 241 13-1. General Relation between Price and Cost for Local and Long-Distance Traffic 264 15-1. Revenue/Cost Results for 1987, by Service Category, According to the Bell Canada Tariff Analysis Methodology 294 15-2. Revenue Surplus/Shortfall Results for 1987, by Service Category, According to the Bell Canada Tariff Analysis Methodology 295 16-1. Driving Forces of Reform in the Telecommunications Sector 306 B-i. Inland Letter and Telephone Traffic in the United Kingdom, 1950-76 362 Examples 1. cciTT's Cross-Sectional Analysis 86 2. ccrr's Time-Series Analysis for Individual Countries 88 3. ccrn's Utilization Factor and Other Studies 90 4. Proxies for GDP and Telephone Density 90 5. Telex Density 93 6. Telecommunications Traffic 93 7. Communications Output Distribution Coefflicients 102 8. Communications Input Coefficients 102 9. Backward Linkages 107 10. Spatial Order and Regional Telephone Traffic (Chile 1971) 122 11. Hierarchical Level of Contact (Kenya 1980) 123 12. Rural Village Pco Benefit Regression Analysis (Costa Rica 1976) 128 13. Increase in Call Charges (Costa Rica 1976) 161 14. Changes in Real Tariffs because of Inflation (El Salvador 1977) 163 15. Derivation of Demand Functions Based on Cost Savings for Communications (Pakistan 1968) 164 16. Cost Savings Jor Communications in Two Rural Areas (Chile 1978) 166 17. Demand for Telephone Calls Based Partly on Costs of Postal Communication (India 1969) 167 CONTENTS xiii 18. Cost Savings for Transport at Two Rural Factories (Bangladesh 1981) 168 19. Bus Transport Cost Savings for Telephone Users (India 1981) 169 20. Transport Cost and Time Spent to Reach a Telephone (Chile 1975) 174 21. Transport Cost and Time Spent to Reach a Rural Telephone (Costa Rica 1976) 175 22. Rents and the Availability oJ a Telephone (Egypt 1977) 176 23. Value of Time Wasted in Call Attempts (Egypt 1977) 177 24. Cost of Unauthorized Transfers (Myanmar 1978) 177 25. Value of Time Saved by Introducing Subscriber Trunk Dialing (India 1969) 178 26. Foreign Exchange Impact of Business Telecommunications (Kenya 1981 and 1987) 179 27. Efficiency Gains by Business Firms (Kenya 1981) 180 28. The Role of Telecommunications in Improving the Reliability of the Power System (Brazil 1980) 182 29. Gasoline Consumption Resulting from Inadequate Telephone Access (Tanzania 1982) 183 30. Reduction of Wasteful Agricultural Travel and Transport (Uganda 1982) 185 31. Value of Improved Telecommunications Infrastructure at the Microeconomic Level, Extrapolated to the National Level (the Philippines 1984 and Costa Rica 1986) 186 32. Social and Economic Benefits of Rural Public Call Offices (Vanuatu 1988) 188 Boxes 16-1. The Precursors: Liberalization and Privatization in the United Kingdom 307 16-2. Privatization in a Large Developing Country: The Case of Mexico 314 16-3. Catching up with the Industrial World: Central and Eastern Europe 316 16-4. Restructuring without Privatization: Sri Lanka 319 16-5. The Beginnings of Reform in Africa 320 16-6. Regulation and Competition: Chile Five Years After 324 Preface to the Second Edition WHEN THIS BOOK FIRST APPEARED IN 1983, it quickly became the standard reference on the analysis of telecommunications in developing coun- tries. Translations into Japanese and Spanish were published in 1985 and 1987, respectively. By 1988, although the book continued to elicit much interest, the English edition was out of print. Since 1983, the approach to telecommunications taken by policy- makers, operating enterprises, and international development organi- zations has changed dramatically. Telecommunications, earlier re- garded as a minor component of infrastructure, became in the 1980s a strategic factor of development at all levels, from individual firms to regions and countries. Traditional telecommunications sector struc- tures, based largely on state monopolies, began giving way to more complex and flexible arrangements featuring reduced monopoly privi- leges, diversified and increasingly competitive supply of services and networks, increased participation of private capital and private enter- prise, and a shift of government responsibilities from ownership and operation to policy and regulation. In developing countries, efforts to build up the telecommunications sector had, for decades, focused on securing a larger share of funds from government and aid agencies and on improving the internal organization and management of state enterprises. Now this strategy is turning increasingly toward improving the policy and regulatory framework so that operating enterprises can enjoy the freedoms and incentives, as well as be subject to the market disciplines, of a modern business. Given these major changes, we debated the extent to which this book should be revised for a second edition. Surely a book on tele- communications and economic development written in the early 1990s would bear little resemblance to one written in the late 1970s and early 1980s. We decided, however, that a limited revision, rather than a major overhaul, would serve readers best. An updated version xv xvi PREFACE would meet the needs of readers interested in, for example, the role that telecommunications play in economic development, especially re- garding benefits, costs, and pricing, because much of the original ma- terial is still relevant. Moreover, the issues and quantitative tech- niques discussed in the first edition mainly from the viewpoint of developing countries are now receiving attention in industrial coun- tries as well, where initiatives are being undertaken to redevelop rural areas and inner cities and to create new competitive advantages in states or provinces where traditional industries are declining. As for the rather different issues of policy, structure, and regulation of the sector, a whole new body of literature is quickly developing. The second edition that we now offer has, therefore, the same structure of, and contains much of the material included in, the origi- nal English edition. All statistical tables, however, have been updated, and new ones added. The case material was partly replaced and con- siderably expanded to reflect more recent experiences. Several new ex- amples were added. The three chapters on pricing (chapters 13, 14, and 15) were extensively rewritten to emphasize practical approaches to setting tariffs and to reflect the shift of regulatory attention from costs to prices. The chapter on technology and costs (chapter 2) was rewritten to reflect the rapid innovation that continues to occur. The new version of the last chapter (chapter 16) now provides a link to the rapidly growing literature on telecommunications restructuring. The notes and references were extensively revised and updated. All re- maining original material was edited, some of it substantially. We would like to thank the many colleagues and friends who pro- vided assistance, advice, and material for this volume. Greta Nettleton, as consultant to the World Bank, was responsible for re- searching the large volume of literature, collating and updating statis- tical data, writing new material, and substantively editing the text cover to cover. At the World Bank, Bj6rn Wellenius was responsible for preparing the revised manuscript; he also wrote new versions of chapters 2 and 16 and various smaller parts of the whole. Peter Smith rewrote chapters 13 through 15 on pricing and tariffs. Dinshaw Joshi and Hugh Lantzke gave generously of their experience in engineering and management and contributed substantially to the revisions, espe- cially in chapter 2. Additional assistance in engineering, management, finance, and economics was provided by other colleagues at the Bank, especially Gerald Buttex, James Cowie, Alberto Cruzat, Rogati Kayani, Philippe Lecharny, David Lomax, Cecile Ramsay, Syed Sathar, A. Shanmugarajah, Mark Tomlinson, and Eloy Vidal. Many individu- als outside the Bank also contributed to the project, including James PREFACE xvii Alleman, Jerry Cohen, Christina Hennig, Bridger Mitchell, Mitchell Moss, Lionel Nicol, Edwin Parker, Kirsten Pehrsson, Joseph Pelton, Michael Tyler, and Jeffrey Rohlfs. Heather Hudson, of the University of San Francisco, rewrote and expanded appendix A for this edition. Douglas Conn, of the Center for Telecommunications and Informa- tion Study (part of the Columbia University School of Graduate Studies) provided access to extensive research material. We thank them all. Acronyms and Abbreviations AID Agency for International Development (United States) AT&T American Telephone and Telegraph Company (United States) CCITT International Consultative Committee on Telephone and Telegraph CFE Central and Eastern Europe CTSC Community teleservice centres, or telecottages (Scandinavia) EC European Communities GDP Gross domestic product GNP Gross national product IRR Internal rate of return ISDN Integrated Services Digital Network ITL International Telecommunication Union OECD Organization for Economic Co-operation and Development OFTEL Office of Telecommunications (United Kingdom) OPEC Organization of Petroleum Exporting Countries OPT Office of Posts and Telecommunications P&T Posts and Telecommunications department, board, or corporation rco Public call office PTT Posts, telegraph, and telephones (administration) TLRIC Total long-run incremental cost VSAT Very small aperture terminal xviii Part I An Introductory Perspective Chapter I The Role of Telecommunications in Economic Development TELECOMMUNICATIONS ARE INCREASINGLY RECOGNIZED as a key component in the infrastructure of economic development, yet telecommunica- tions services in most developing countries continue to fall far short of needs. The result is that development is constrained significantly throughout these economies. The imbalance between telecommunica- tions development in industrial and in developing countries was the focus of global attention in the 1985 Maitland Commission report, which noted that "in most developing countries, the telecommunica- tions system is not adequate even to sustain essential services."' How- ever, despite increasing awareness of the connection between telecom- munications and economic development, governments in developing countries and development agencies rarely pay attention to these shortfalls and the measures required to redress them. Can the economic value of the benefits derived from investments in telecommunications be demonstrated and quantified? Which seg- ments of the population derive these benefits? What is the impact on the development of other economic and social sectors? How do the benefits compare with the economic costs of expanding and improv- ing the telecommunications system? What investment and pricing policies would increase benefits and contain costs? What level of in- vestment can be justified by economic analysis? These central issues of telecommunications economics in developing countries are the main themes of this book. 3 4 AN INTRODUCTORY PERSPECTIVE Signals for Expanding Investment The developing countries have about 75 percent of the world's pop- ulation and 16 percent of its product, but only 12 percent of the total number of telephone main lines.2 Of the more than 435 million lines in existence in 1988, about 50 million were in the developing coun- tries of Africa, Asia, and Latin America, while 386 million were in the industrial and newly industrialized countries of North America, Europe, Asia, and Oceania (see table 1-1). Table 1-1. Distribution of the borld's Telephone Mainl Lines b's' Region, Population, and Income, 1988 GNP Main lines Population Billions Region Millions Percent Millions Percent of dollars Percent Industrial countries Africa South Africa 2.5 0.6 30 0.6 63 0.5 Asia Japan 49.9 11.5 123 2.6 1,925 14.1 Others" 17.9 4.1 74 1.6 209 1.5 Europe Former U.S.S.R. 27.7 6.4 285 5.9 - - Others 149.7 34.4 421 8.8 4,146 30.4 North America Canada 11.4 2.6 26 0.5 390 2.9 United States 118.4 27.2 246 5.1 4,486 32.9 Oceania Australia 6.8 1.6 16 0.3 176 1.3 New Zealand 1.4 0.3 3 0.1 27 0.2 Total 385.7 88.6 1,223 25.6 11,422 83.9 Deteloping countries Africa 3.9 0.9 568 11.9 287 2.1 Asia 22.6 5.2 2,597 54.3 1,188 8.7 Latin America 23.0 5.3 393 8.2 725 5.3 Total 49.5 11.4 3,558 74.4 2,199 16.1 World total" 435.2 100.0 4,782 100.0 13,621 100.0 - Not available. a. Israel, Republic of Korea, Hong Kong, Singapore, and Taiwan. b. A small percentage of nonreporting countries are not included. Source: American Telephone and Telegraph Corporation, The World's Telephones (various years); World Bank, World Bank Atlas 1988; ITI, Yearbook of Public Teleconimunication Sta- tistics (1990); and World Bank data. THE ROLE OF TELECOMMUNICATIONS 5 In early 1988, the industrial countries had on average about 32 main lines per 100 inhabitants.3 Canada, France, Scandinavia, Swit- zerland, and the United States all had a density of about 50 main lines and between 78 and 89 telephone sets (equipment on the cus- tomer's premises) per 100 inhabitants. Most other industrial coun- tries had over 30 lines per 100 inhabitants, while the developing countries, in contrast, averaged only about 1.5 lines per 100 inhabi- tants (see table 1-2). Telephone density also varies widely among and within developing countries. A ratio of more than 50:1 exists between countries such as Argentina and Uruguay, which have more than 10 main lines per 100 persons, and countries such as the Central African Republic, Ethio- pia, Madagascar, Nigeria, and Tanzania, which have just 0.1 to 0.2 lines per 100 persons (table 1-2). In 1988, the density of telephone main lines in Africa (excluding South Africa) was 0.7 per 100 per- sons; it was 0.9 in Asia (excluding Japan, and the newly industrialized countries of Hong Kong, Singapore, South Korea, and Taiwan) and 5.9 in Latin America. Also, in developing countries, telephones are concentrated in a few large cities, and much of the population lives in areas with little or no service. In industrial countries, the telephone density is fairly uniform across each country; in the developing world, in contrast, telephone density is several times greater in the main cities than in provincial towns and rural areas (see table 1-3). The gap that exists between industrial and developing countries in basic telephone service, although still large, is narrowing. Between 1969 and 1988, developing countries almost doubled their share of the world's telephone lines, from about 7 percent to 12 percent, while their share of world population and product remained roughly unchanged. Despite such improvements, the disparities remain extreme. Level of Investment Although a decade ago developing countries invested only about 0.3 percent of gross domestic product (GDP) in telecommunications, by the end of the 1980s this level had doubled to about 0.6 percent, or some $12 billion a year.4 This proportion is similar to the level of investment found in industrial countries, which remained relatively steady during the same period, at about 0.6 percent of GDP. In 1989, industrial countries invested about $100 billion in their public tele- communications sectors.5 Table 1-2. Proportion of Telephone Main Lines and GNP per Capita in Selected Countries, 1987-88 Number of main lines per 100 persons GNP per capita Region and country (January 1, 1988) (1987 U.S. dollars) World 9.2 2,848 Industrial countries 31.5 9,339 North America 47.7 17,927 Canada 51.2 15,080 United States 48.1 18,430 Asia and Oceania 35.1 10,819 Australia 42.8 10,900 Hong Kong 35.4 8,260 Japan' 40.8 15,770 Korea 20.7 2,690 New Zealand 41.6 8,230 Singapore 33.3 7,940 Taiwan 25.0 Europe 25.1 9,848 Austria 38.4 11,970 Belgium 34.5 11,360 Czechoslovakia 13.0 Denmark 55.1 15,010 Finland 47.9 14,370 France 44.7 12,860 Germanyb 34.2 14,460 Greece 34.7 4,350 Hungary 7.7 2,240 Ireland 22.5 6,030 Italy 33.3 10,420 Netherlandsi 42.4 11,860 Norway 46.4 17,110 Portugal 16.1 2,890 Spain 26.2 6,010 Sweden 65.1 15,690 Switzerland 52.9 21,250 United Kingdom 42.4 10,430 Yugoslavia 12.9 2,480 U.S.S.R. 9.7 _ Developing countries 1.5 618 Africa 0.7 505 Algeria;' 2.7 2,760 Benin 0.3 300 Botswana" 1.2 1,030 Cameroon 0.3 960 Central African Republic 0.2 330 C&te d'lvoire) 0.6 750 Egypt 2.2 710 Ethiopia" 0.2 120 Gabon 1.8 2,750 Kenya 0.7 340 Liberia 0.4 440 Madagascar 0.2 200 Malawi 0.3 160 Mauritius" 4.5 1,470 Morocco 1.1 620 Namibia 3.7 Nigeria' 0.2 370 Rwanda 0.1 310 Senegal 0.4 510 Seychelles 8.2 3,180 6 THE ROLE OF TELECOMMUNICATIONS 7 Table 1-2 (continued) Number of main lines per 100 persons GNP per capita Region and country (January 1, 1988) (1987 U.S. dollars) Developing countries (continued) Africa (continued) Sierra Leone' 0.4 300 Sudan 0.2 330 Tanzania 0.2 220 Togo 0.3 300 Tunisia 3.0 1,210 Zambia 0.7 240 Zimbabwe 1.4 590 Asia 0.9 458 China 0.5 300 Fiji 4.7 1,510 India 0.4 300 Indonesia 0.4 450 Iran 2.9 Iraq 4.8 d Jordan 6.7 1,540 Malaysia 7.2 1,800 Oman 5.1 5,780 Pakistan 0.6 350 Papua New Guinea 0.9 730 Philippines" 0.8 590 Saudi Arabia 12.4 6,9301 Sri Lanka 0.5 400 Syria 4.3 1,820 Thailand 1.7 840 Turkey 7.0 1,200 Latin America 5.9 1,845 Argentina' 9.6 2,370 Bolivia 2.2 570 Brazil 5.6 2,020 Chile 4.9 1,310 Colombia 7.2 1,220 Costa Rica 8.6 1,590 Cuba 3.0 Ecuador 4.4 1,040 El Salvador 2.1 850 Guyana 2.5 380 Haiti 0.5 360 Honduras 1.1 780 Mexico 4.9 1,820 Nicaragua 1.3 830 Panama 8.6 2,240 Paraguay 2.4 1,000 Peru 2.3 1,430 Suriname 7.8 2,360 Uruguay, 10.6 2,180 Venezuela 9.2 3,230 -Not available. a. Population estimated. b. Estimated for Germany after unification. c. Includes Asian U.S.S.R. d. As of January 1, 1985. e. According to Siemens data. f. As of January 1, 1986. Source: American Telephone and Telegraph Corporation, The World's Telephones (various years); World Bank, World Bank Atlas 1988; ITU, Yearbook of Public Telecommunication Sta- ristics (1990). 8 AN INTRODUCTORY PERSPECTIVE Table 1-3. Access to Telephone Services in Main Cities and Other Areas in Selected Countries, as of January 1, 1988 (number of main lines per 100 persons) Region and countrY National Main cities, Other areas Industrial countries Austria 38.38 54.20 31.32 Canada 44.49 59.20 43.54 Denmark 55.13 59.58 52.36 France 44.68 47.98 29.27 Germany" 39.27 50.20 35.98 Italy 33.28 41.48 30.65 Japan 40.81 56.13 37.48 Norway 46.41 55.81 41.89 Spain 26.18 31.84 21.02 Switzerland 52.87 65.45 46.73 Developing countries Africa Algeria 2.70 7.13 1.58 Ethiopia 0.24 3.39 0.04 Kenya 0.66 4.95 0.19 Malawi 0.28 2.20 0.07 Morocco 1.14 3.17 0.42 Sudan 0.24 1.32 0.04 Togo 0.28 1.27 0.00 Tunisia 3.01 7.00 0.79 Zambia 0.73 1.36 0.17 Zimbabwe 1.45 6.39 0.41 Asia Iran-' 3.15 6.31 1.10 Malaysia 7.21 22.65 5.17 Pakistan 0.61 2.69 0.19 Papua New Guinea 0.91 5.91 0.22 Sri Lanka 0.54 1.12 0.29 Thailand 1.67 6.94 0.45 Turkey 7.01 7.46 6.56 Latin America Brazil 5.59 10.17 4.14 Colombia 7.20 13.26 1.83 Costa Rica 8.62 15.28 2.57 Ecuador 4.41 8.27 1.91 Peru 2.30 4.90 0.52 Uruguay 10.61 16.05 5.24 Venezuela 9.19 16.20 5.08 a. Defined by the national administration; population thresholds, and consequently the number of cities included, vary widely among countries. b. Estimated from combined Federal Republic of Germany (January 1987) and German Democratic Republic (January 1988) data. Source: American Telephone and Telegraph Corporation, The W0orld's Telephones (various years); ITL, Yearbook of Telecommunication Statistics (1990); and World Bank data. THE ROLE OF TELECOMMUNICATIONS 9 Given the low base from which they begin, basic telephone systems in developing countries can grow at relatively high rates, averaging more than 9 percent; from 1979 to 1987, regional growth rates ranged between 7 and 11 percent for Asia, Latin America, and Africa.6 Some countries have sustained rates of 15 percent or more for years. These growth rates are well above the 3 to 6 percent typically found among industrial countries, which already have relatively large telecommuni- cations infrastructures in place (see table 1-4). Telex has grown even faster than voice telephone service (see table 1-5). More advanced services introduced in recent years, such as fac- simile and data transmission, have also expanded at record rates, as il- lustrated in tables 1-6 and 1-7. Growth in the use of facsimile machines has been explosive in countries such as Brazil, Colombia, Malaysia, and Singapore, and strong in many countries throughout Asia, Latin America, and Sub-Saharan Africa.7 Growth in the num- ber of leased circuits for data communications has been more uneven, reflecting differences in government policies. Between 1987 and 1988, the number of circuits increased more than 1,000 percent in Thailand but remained close to or at 0 percent in most Sub-Saharan African countries. Lack of demand does not explain the historically low level of in- vestment in the telecommunications sector in most developing coun- tries. Throughout the developing world the demand for telephone and more advanced services typically far exceeds the supply, and the num- ber of unmet applications for telephone connections often exceeds the number of existing lines (see table 1-8). New applicants frequently wait two to five years-and sometimes more than ten-to obtain service. In these situations, a large propor- tion of the potential demand for telecommunications services remains unrecorded and emerges only when the system is perceived to be ex- panding rapidly.8 Also, applications are rarely accepted in areas where service is neither available nor planned in the near future. Besides, in towns or villages where most of the population is not familiar with telephone service, significant demand arises only after service is intro- duced at the initiative of the government, an operating company, or a special interest group and the public has gradually gained experience using telephones and begun to appreciate the cost savings and other benefits that can be obtained from their use. The introduction of more advanced services also tends to expand demand as the modern business sector finds that it needs these services to compete in the global marketplace. When the shortage of telephone lines is acute, the proportion of 10 AN INTRODUCTORY PERSPECTIVE Table 1-4. Growth of Telephone Main Lines in Service in Selected Developing Countries, 1979-88 (thousands of main lines unless otherwise indicated) Annual As of As of growth rate Region and country January 1, 1979 December 1, 1988 (percent) Africa Algeria 260.4 697.2 11.6 Benin 7.8 13.6 6.4 Botswana 6.4 18.2 12.3 Egypt' 392.0 1,118.0 14.0 Ghana 36.4 40.4 1.2 Kenya 69.7 157.4 9.5 Malawi- 10.6 22.5 9.9 Mozambique 30.6 40.8 3.2 Senegal 17.3 28.9 5.9 Tanzania 38.3 66.1 6.2 Zimbabwe 92.9 118.4 2.7 Asia China 2,460.0' 5,550.0 9.5 Fiji 21.3 35.7 5.9 India-, 1,868.0 3,487.9 8.1 Indonesia 317.9 828.8 11.2 Iran 1,233.9 1,803.6 4.3 Malaysia 325.0 1,247.7 16.1 Myanmar (Burma) 28.2 74.7 11.4 Pakistan 286.0 636.6 9.3 Papua New Guinea 22.8 31.0 3.5 Thailand 332.0 1,005.9 13.1 Asia (newly industrialized countries) Hong Kong 1,173.0 2,153.8 7.0 Korea 2,341.0 10,486.2 18.1 Singapore 464.8 924.0 7.9 Taiwan' 2,630.0 4,909.0 11.0 Latin America Argentina 1,797.0 2,747.5 4.8 Brazil 3,829.0 9,081.6 10.1 Chile 351.0 625.5 6.6 Colombia 1,022.3 2,070.4 8.2 Costa Rica 133.6 256.5 7.5 El Salvador' 66.1 104.5 5.9 Mexico" 2,494.0 3,774.0 5.3 Peru` 287.0 461.9 6.1 Venezuela 701.0 1,457.8 8.5 a. From January 1, 1979, to December 31, 1987. b. From January 1, 1981, to December 31, 1987. c. Estimated from World Bank data. Source: ITU, Yearbook of Public Telecommunication Statistics (1990), and World Bank data. THE ROLE OF TELECOMMUNICATIONS 11 Table 1-5. Growth of Telex Lines in Service in Selected Developing Countries, 1979-88 Number of telex lines Annual January 1, December 31, growth rate Region and country 1979 1988 (percent) Africa Algeria 2,400 8,244 14.7 Botswana" 140 741 23.2 Burundi, 85 191 17.6 Cameroon` 563 1,940 16.7 Ethiopia 372 864 9.8 Ghana 172 563 14.1 Mauritius 215 784 15.5 Niger" 179 297 8.8 Rwanda 60 138 9.7 Swaziland 193 353 6.9 Tanzania 426 1,384 14.0 Asia Bangladesh" 256 1,090 27.3 Fiji 235 655 12.1 India,' 16,500 34,044 9.5 Indonesia 3,612 15,441 17.5 Malaysia 2,908 9,930 14.6 Myanmar (Burma) 47 138 12.7 Nepal 161 405 20.3 Oman 460 1,319 12.4 Papua New Guinea 999 1,129 1.4 Philippines 5,733 12,199 8.8 Sri Lanka 350 1,535 17.9 Syria 699 2,695 16.2 Latin America Brazil 27,362 121,200 18.0 Chile 2,157 11,648 20.6 Colombia 3,800 6,452 6.1 Costa Rica 900 1,500 5.8 Ecuador' 1,292 3,152 11.8 El Salvadort' 610 906 5.1 Uruguay 897 2,236 10.7 Venezuela' 7,848 18,800 13.3 a. January 1, 1980, to December 31, 1988. b. January 1, 1979, to December 31, 1985. c. January 1, 1980, to December 31, 1986. d. January 1, 1979, to December 31, 1987. e. January 1, 1982, to December 31, 1987. f. January 1, 1981, to December 31, 1988. Source: ITU, YeaTbook of Public Telecommunication Statistics (1990). 12 AN INTRODUCTORY PERSPECTIVE Table 1-6. Number of Facsimile Stations in Selected Industrial and Developing Countries, 1986-88 Country 1986 1987 1988 Industrial countries Australia 60,000 100,723 Canada 55,000 84,000 Finland 6,339 9,866 16,759 France 60,000 96,000 185,000 Germany, Fed. Rep. of 44,453 85,295 291,213 Hong Kong 12,673 32,076 59,876 Hungary - 22 1,096 Italy 25,203 48,269 92,813 Singapore 5,349 12,069 18,986 Spain 6,515 18,549 53,678 Switzerland 3,858 19,246 40,285 Developing countries Botswana 122 - 812 Brazil - 1,736 20,090 Burkina Faso - 4 15 Cyprus 356 872 1,858 Colombia - - 17,090, Egypt 3 767 - Ethiopia - 34 128 Fiji I - 612 Ghana I 1 49 Guatemala - - 2,503 Indonesia - 2,453 4,255 Iran 150 131 620 Malavi 13 89 Malaysia 1,415 4,707 13,702 Mauritius - 220 268 Oman 450 900 1,500 Paraguay - - 158 Rwanda - - 158 Tanzania - 89 172 Thailand 1,521 - 5,453 United Arab Emirates 1,101 2,254 6,604 Zambia - 111 198 Zimbabwe - 232 387 - Nor available. Note: Figures are as of December 31. a. Estimated. Source: ITtU1, Yearbook of- Public Telecommunication Statistics (1990). THE ROLE OF TELECOMMUNICATIONS 13 Table 1-7. Growth in the Number of Privately Leased Circuits for Data Communications in Selected Industrial and Developing Countries, 1981 and 1986-88 Country 1981 1986 1987 1988 Industrial countries Australia 40,252 86,076 96,000 - Finland 9,800 29,900 37,500 95,800 Germany, Fed. Rep. of 3,203 4,593 4,229 4,807 Hong Kong - - 68,940 67,642 Hungary 342 616 698 772 Ireland 1,098 4,822 5,450 7,827 New Zealand 833 2,919 18,887 25,233 Poland 400 570 570 888 Portugal 1,315 5,600 6,700 8,390 Singapore 417 24,184 30,380 36,000 Spain 28,700 44,600 53,389 58,261 Switzerland 11,879 18,383 - - Developing countTies Botswana - 11 - 351 Brazil 58 12,000 16,177 20.821 Burkina Faso - 5 5 - Cyprus 18 235 292 724 Colombia 150 14,427 - 16,314 Ethiopia - 1 13 13 Fiji 13 210 230 233 Ghana - 17 17 15 India - 907 1,268 - Indonesia - 934 1,132 1,628 Malawi 291 362 245 - Malaysia - 4,544 6,875 8,206 Mauritius 11 - 75 112 Oman 12 156 480 559 Paraguay - 4 - 53 Rwanda 3 3 3 3 Senegal - 65 79 512 Tanzania 11 716 715 715 Thailand 38 - 951 11,913 United Arab Emirates 271 364 393 385 Zambia - 277 296 - Zimbabwe - 644 589 816 - Not available. Note: Figures are as of December 31. a. Estimated. Source: ITU, Yearbook of Public Telecommunicaion Statistics (1990). 14 AN INTRODUCTORY PERSPECTIVI Table 1-8. Supply of and Expressed Demand for Telephone Main Lines in Selected Developing Countries, Various Years, 1985-89 (thousands unless otherwise indicated) Main Total Percentage of Region, country, lines in Unmet expressed expressed and date service applications demand demand met Africa Algeria. December 1988 697 524 1,221 57 Botswana, March 1989 18 3 21 87 Egypt, December 1987 1,118 1,137 2,255 50 Ethiopia, December 1988 113 88 200 56 Ghana, December 1988 40 30 70 58 Kenya, December 1988 157 64 221 71 Lesotho, December 1988 11 5 17 68 Malawi, December 1987 22 6 28 80 Mali, December 1988 9 4 13 68 Mauritius, December 1988 50 42 91 54 Morocco, December 1988 286 205 491 58 Senegal, December 1988 29 8 37 79 Sudan, December 1985 58 22 79 73 Swaziland, March 1987 10 3 12 79 Tanzania, December 1988 66 76 142 47 As a Fiji, December 1988 36 12 47 75 India. March 1987 3,488 1,125 4,612 76 Iran, March 1989 1,804 314 2,118 85 Malaysia, December 1988 1,248 69 1,317 95 Myanmar (formerly Burma), December 1988 75 300 375 20 Nepal, July 1987 30 41 72 42 Oman, Decenmber 1988 88 9 96 91 Pakistan, December 1988 637 531 1,168 54 Papua New Guinea, December 1988 31 2 33 94 Philippines, December 1988 591 582 1,173 50 Sri Lanka, December 1988 104 51 155 67 Syria, December 1988 478 1,339 1,817 26 Thailand, December 1988 1,006 377 1,383 73 Latin America Argentina, December 1988 2,747 789 3,537 78 Chile, December 1988 625 236 862 73 Colombia, December 1988 2,070 660 2,730 76 Costa Rica, December 1988 257 14 270 95 El Salvador, December 1986 96 11 107 90 Mexico, December 1986 3,821 757 4,577 83 Peru, December 1986 445 273 718 62 Uruguay, December 1988 345 93 439 79 Venezuela, December 1988 1,458 453 1,911 76 a. The number of lines in service plus the number of outstanding applications for ser- vice. In developing countries, actual demand is generally much higher than expressed demand. Source: T, Yearbook oJ Public Telecommunication Statistics (1990). THE ROLE OF TELECOMMUNICATIONS 15 subscribers who are very intensive users (mainly business and govern- ment) tends to be high, as does the number of users for each tele- phone line.9 Hence, the average number of calls per line is high, and the traffic of local and long-distance calls is frequently congested. This congestion results partly from the inability of major components of the system, especially telephone exchanges and trunks, to handle call traffic and a tendency in some developing countries to provide in- sufficient private branch exchange facilities. To a large extent, how- ever, it also results from the high proportion of time that the telephone called is engaged; thus subscribers repeatedly attempt to place calls, which, when added to the repeat calls resulting from equipment congestion, further strains the network's capacity. Conges- tion at major centers or long-distance routes tends to propagate across the national system.'0 When call traffic is congested, the cost per completed call is high in terms of both the system's use and the users' time, many of the potential benefits of investing in telecommu- nications do not materialize, and maintenance problems increase markedly. Returns on ltnvestniexlt In developing countries, inadequate investment in the telecommu- nications sector is not caused by telecommunications entities that lose money or require government subsidies, any more than it is caused by lack of demand (some of the more binding constraints on expansion are discussed further in chapter 3). In general, reasonably well-managed telecommunications entities can generate large finan- cial surpluses in local currency. In the usual situation, where demand greatly exceeds supply, it is easy, with proper pricing policies, to re- cover the full cost of providing telecommunications services (includ- ing the cost of capital) from tariffs, even if internal inefficiencies often result in high costs. It is likewise feasible to generate internally a large proportion of the funds required for improving and expanding the system in the future. On average, in thirteen programs recently funded in part by the World Bank, telecommunications investments are expected to yield internal financial rates of return between 13 and 25 percent, averag- ing about 20 percent. Approximately 60 percent of the new construc- tion funds required will be generated internally. The operating entities are conservatively estimated to average more than a 15 percent annual rate of return on overall net plant in service, revalued to current prices, over the five years after the project is appraised by the Bank.'' 16 AN INTRODUCTORY PERSPECTIVE Furthermore, many of the telecommunications entities will make sub- stantial net contributions to the government, which will tend to be used in sectors less able to mobilize domestic resources."2 Telecommunications investments also yield high returns to the economy as a whole. After correcting for general distortions of the price system and pure transfer payments, the economic rates of return attributable to the thirteen projects approved for World Bank support are expected to range between 15 and 30 percent. Where it has been possible to quantify meaningfully some of the benefits accruing to the country apart from the revenues realized by the telecommunications entity, the economic rates of return are expected to range between 20 and 40 percent (examples of such exercises are presented in chapters 8 and 9). These figures are higher than the usual 10 to 14 percent rate of return used as the threshold for acceptable investments in de- veloping countries. Very large economic returns also result from the telecommunications components of investment programs in other sectors (for example, railways, power, tourism, banking, and rural development). Hence, in developing countries a large excess demand exists for telecommunications services, and the measurable private and social returns on the investment required to provide those services are rela- tivelv high. Were economic efficiency the sole goal of development, such evidence of market forces would be sufficient to justify expand- ing the sector rapidly. Bentefits from Investment Other arguments are sometimes used to support increasing tele- communications investment in developing countries. Proponents of such investment contend that as economic development takes place, some form of telecommunications gradually becomes the most cost- effective means of communicating for increasing proportions of the population. Telecommunications services can substitute for other forms of communication (mainly postal service and personal travel) and are often more effective and more efficient than those forms in their use of time, energy, and materials and in their effect on the quality of the environment. Further, there is some evidence that a re- liable telecommunications system generates new communication and builds stronger, more complex, and more productive patterns of com- munication, partly because it interacts directly and indirectly with numerous production and distribution functions. THE ROLE OF TELECOMMUNICATIONS 17 It is also argued that having accessible and reliable telephone ser- vice removes some of the physical constraints on organizational com- munication in various sectors of the economy, permitting increased productivity through better management in both the public and pri- vate sectors, making it possible to adopt different structures and loca- tions, and aiding the evolution of increasingly complex organizations. Markets become more effective as communication improves, more rapid responses to market signals become possible, and access to mar- ket information is extended at village, town, city, regional, national, and global levels. Also, household operations become more efficient as telecommunications improve access to goods and services and sup- port forms of work that require some integration of workplace and residence. ' 3 In the late 1970s, as fuel prices escalated, a particularly timely ar- gument in favor of increasing investment was that telecommunica- tions could partly substitute for transport and also bring about the more efficient use of transport facilities, thus reducing the amount of energy required to sustain a given level of communication (see chap- ter 7 for a more detailed discussion). On the one hand, transport costs, led by increased petroleum prices and petroleum-related steel and materials costs, had risen significantly. On the other hand, dra- matic changes in telecommunications technology had lowered the real cost of providing telecommunications services. Hence, increased in- vestment in telecommunications was seen to create opportunities for reducing the costs of transport and transport-related materials and supplies and thus deserved serious attention. Although the energy cri- sis has since abated, technological innovation has further reduced the cost of telecommunications, and energy conservation, transportation efficiency, and related environmental matters continue to be of major concern to national policymakers (some of the recent changes in technology and their effect on costs are discussed in chapter 2). Mlore general discussions sometimes emphasize that virtually every aspect of economic and social development in any one sector of an economy interacts with development in other sectors.'4 For example, most development activities require services from the economy's infra- structure networks: roads, water, power, and-important, but not well researched or understood-telecommunications. In particular, as agri- cultural development increases the marketable agricultural surplus, it gives rise to widespread trade in agricultural commodities, seeds, fer- tilizers, fuels, and other goods and services, which, to be efficient, inevitably requires reliable means of communicating rapidly and over 18 AN INTRODUCTORY PERSPECTIVE long distances. A need also exists for reliable and rapid information about weather conditions, disease outbreaks, and new agricultural techniques. The utility of telecommunications services is also apparent for com- nmerce and industry. Industrial development requires coordination of numerous activities: acquisition of supplies, recruitment and coordina- tion of labor, control of stocks, processing of materials, billing, record- keeping, delivery of goods to buyers, and general market search activities. Commerce, however, is inherently an information-processing activity. Effective buying, selling, brokerage, and transport require a continuous supply of up-to-date information on the availability and price of numerous goods and services. In the absence of accessible and reliable telecommunications services such activities suffer a variety of inefficiencies, including the creation of markets in which a few information-rich individuals are able to gain significant advantage over the majority of individuals who are information poor. This is particu- larly important as new telecommunications technologies emerge; glo- bal business patterns increasingly require access to new services, and regions where such services are not available risk being shut out of ef- fective competitive trade. Finally, as is examined further in chapters 10, 11, and 12, advo- cates of telecommunications investment stress that the direct contri- bution that telecommunications make to the welfare of individuals and families cannot be dismissed by derogatory references to "merely social" telephone calls. The well-being of the family is assisted by tele- communications, which provide rapid access to services often needed to preserve life, health, and property and enhance contact with kin, friends, and special interest groups. Telecommunications contribute to the development of a shared environment that reaches a country's most remote areas and can facilitate political, cultural, economic, and social integration. Although telecommunications still appear to command a lower order of priority in many developing countries than is justified in eco- nomic terms, the high economic and financial returns from telecom- munications investment and the benefits they imply are being increasingly recognized. In particular, countries often assign high pri- ority to investment in the sector when, for one reason or another, they overcome some of the perceived constraints hindering the devel- opment of their telecommunications networks. For example, in the 1970s and 1980s, oil-surplus developing countries (such as Indonesia, Iran, Kuwait, Nigeria, Saudi Arabia, and Venezuela) undertook large investment programs in telecommunications once foreign exchange THE ROLE OF TELECOMMUNICATIONS 19 became readily available. Once-developing countries such as Hong Kong, Korea, Singapore, and Taiwan used telecommunications as a key part of their overall economic strategy to build up what is now a highly competitive position in the world market for high-technology industries and services. Likewise, a number of countries that did not recognize the importance of telecommunications, such as Egypt and India, have embarked on large investment programs to overcome what they now regard as a significant constraint on their economic growth and the decentralization of economic activity outside major urban areas. The most striking example of this approach among industrial countries is the telecommunications investment program of France since 1974.'5 Moreover, driven largely by technological innovations, related changes in cost structures, and growing and diversified demands of business, many industrial countries, including some of those with the most advanced telecommunications services, undertook major changes in the 1980s in how the sector was structured and reg- ulated. Liberalization and the breakup of the American Telephone and Telegraph Company (AT&T) in the United States were followed by privatization and the introduction of competition to the telecommu- nications sector in the United Kingdom and Japan, and this trend has gradually extended to most of the countries in the Organisation for Economic Co-operation and Development (OECD). These events, cou- pled with the widespread globalization and increased information content of economic activity, led developing countries from the mid- 1980s on to consider a range of options for reforming the telecommu- nications sector with the aim of helping overcome long-standing constraints on development. By 1990, some forty developing coun- tries had completed, were embarked on, or were actively preparing major reforms of the telecommunications sector.'6 More than half of the world's main lines were operated by privately held companies by the beginning of 1989.'7 These alternatives to having state monopo- lies operate telecommunications systems will be examined again in chapter 16. Vieus about Telecominunicatio?1s Detelopment Views concerning the most desirable rate of development for the telecommunications sector can roughly be grouped into three general categories. First, one group feels that telecommunications investment should be held well below what would be indicated by the market forces outlined above. Second, another group contends that telecom- 20 AN INTRODUCTORY PERSPECTIVE munications should grow mostly as indicated by the market, with operating entities behaving in most respects like commercial enter- prises with relatively unhindered access to capital markets for invest- ment funds, subject to some government controls to ensure wide access to basic services and to protect the public's interest. More and more, this point of view is coming to predominate in policymaking circles and in the literature. And finally, a more activist technology- oriented group in part promotes rapidly advancing telecommunica- tions technology as a prime means to achieve a wide range of social and economic goals in numerous socially oriented sectors, including the delivery of education and health services. More generally, this last approach would not only implement the growth of a telecommunica- tions system that is called for by market forces but would also in some instances take that growth in a different direction with respect to the types of services provided, the types of users served, pricing and con- trol of access, network planning, technology, and other matters. In some cases representatives of this group also advocate government subsidies for selected telecommunications services. The first, restrictive, view, in one of its forms, argues that expansion of telecommunications services does not deserve priority in resource allocation, partly because it has too little measurable economic effect and partly because in some countries rapid two-way communication among the population could facilitate political instability. The con- tention is that this service is used for the most part for economically and socially worthless purposes. Surprisingly, although this view is often superficially expressed, it rarely appears in print.18 A variant of the restrictive philosophy contends that telecommunications serve economic and social needs, but in an undesirable way: "Telephony de- velopment ... is generally by and for the elite groups ... primarily confined to the more modern and urban areas of society ... By creat- ing an urban-based communications infrastructure, which is only ac- cessible to a limited segment of society, economic opportunity becomes further concentrated in urban settings, and hence urban mi- gration is encouraged."'19 Some analysts caution that introducing tele- communications services, particularly advanced technologies, in developing regions will not necessarily bring social benefits and eco- nomic development. Reducing the costs for high-volume users may in- crease the costs for small-volume users, reducing the affordability of access and possibly creating structural dependencies between more- developed and less-developed regions.20 One question arising from such views is that even if they were true of some forms or mixes of telecommunications investment (and that has not been demon- THE ROLE OF TELECOMMUNICATIONS 21 strated), would they be true for all? And if not, how can preferred strategies and policies be identified that would preclude these adverse effects? The second, market-response, view suggests that the demonstrated market demand for telecommunications should be met and that new technical applications should be provided when they are the most cost-effective way to meet market demand and to provide minimum telephone access nationally as needed for broader development pur- poses. The position taken by the International Telecommunication Union (ITU), an entity that will be quoted on several occasions in this book, has, for many years, been generally along these lines.2" In con- tradiction to those who argue that telecommunications focus eco- nomic opportunity on only a few urban areas, the ITU contended that telecommunications form a vital part of the national economic infra- structure and that, when provided, they produce widespread benefits. The ITU argued through the years that failure to recognize the full range of benefits from telecommunications leads to clhronic underin- vestment: "a permanent shortage of investment funds ... leads to capital rationing which can be quite arbitrary in effect. This often results in the treasury of the country allocating a given sum to tele- communications, even though a much larger amount would be war- ranted." More recently, the ITU's Advisory Group on Telecommunica- tions Policy took a vigorous pro-market approach and recommended adjusting current policies and structures to increase private participa- tion in the sector, develop competition, and shift the government's role from ownership and operation to policy and regulation. The ITU'S plenipotentiary conference of 1989 endorsed these recommendations and brought them to the attention of all developing countries.22 The third, more activist, point of view is represented in early writ- ings such as those of Philip Okundi, in Kenya, and the late Ithiel de Sola Pool, Manfred Kochen, Edwin Parker, and Melvin M. Webber, in the United States. Okundi was one of the first advocates of an Afri- can domestic satellite system and vigorously supported the expansion of the Pan-African Telecommunications Network.23 Pool argued, "It would seem clear that international telecommunication is of great im- portance to the developing world. It can bring deficient information facilities (e.g., for access to scientific data bases) at a leap up to the best, or its lack can lead to slipping further and further behind."24 Kochen contended that computer conferencing facilities should be made available throughout the developing world and that thev would facilitate knowledge transfer and more efficient resource allocation.25 Parker argued for accelerated development of rural satellite ground 22 AN INTRODUCTORY PERSPECTIVE station installations with applications in health, education, and gen- eral rural development, as did Hudson and others.26 Finally, Webber suggested that modern telecommunications has the potential for per- mitting developing countries to organize their urban areas on a more spatially dispersed basis, which he hoped would allow people to live closer to their work and would reduce the need for large investment in mass transit and in high-density and high-cost city centers.27 Videotex in France is a prime example of activism by an industrial country's government. By 1990, several million Minitels (basic data communi- cation terminals) were in place in French homes and businesses, pro- viding a market base for the development of electronic mail, home banking, and a host of new telecommunications and information ser- vices. This was the result of a policy decision made in the mid-1970s, as part of a drive to modernize and expand the telecommunications system and to replace the printed telephone directory with an elec- tronic version and provide the necessary terminals free of charge.28 Who is right? The importance of answering this question correctly can scarcely be exaggerated. If a strong telecommunications infra- structure is indeed essential for rapid and efficient development to take place, its neglect may severely hinder the success of development efforts in both directly productive and social sectors, impose ineffi- cient spatial settlement patterns on the rapidly growing urban areas in the developing world, and reduce the ability of these countries to participate in the world economy. If, however, the present level of tele- communications service in developing countries is sufficient (al- though in many towns, villages, and informal urban settlements it is virtually nonexistent), then massive investment in the premature ex- pansion of a major infrastructure system would not only misdirect re- sources but also create a serious burden of unnecessary administra- tion, training, and maintenance.29 Benefits from Improved Telecommunications Many telecommunications professionals or development experts have observed individual cases in which specific benefits of improved communication have been associated with changes in telecommunica- tion infrastructure. The following examples of such instances are grouped somewhat arbitrarily under four convenient (but not inclu- sive or mutually exclusive) headings.30 THE ROLE OF TELECOMMUNICATIONS 23 Market ilnformationi for Buying and Selling A distributor of industrial spare parts and machinery in Nairobi, Kenya, found that after additional lines were installed to his office from the local exchange, his business expanded 35 percent. This per- mitted him to hire six new employees and add three light trucks to his company's fleet. In Greece, the packet-switching network only serves central, urban regions, leaving businesses located in peripheral regions (termed less fa- vored regions by the European Communities) isolated from business networks that are important for their operations. Without packet switching, banks in less favored regions must pay much higher costs to install electronic banking systems. This weakens the national banking system overall, since many banks cannot afford to participate in the national banking network. Similarly, travel agents and insurance com- panies in these regions have trouble competing because they do not have access to network-based quotation and booking systems that de- pend on packet switching.3' . A bank manager in a small town in India was able to reach a pro- spective customer in the regional market city 60 kilometers away rap- idly after a rural automatic digital exchange was installed in his town. The manager maintained that if direct dialing had not been available, the resulting delay in communications would have caused the deposi- tor to take his substantial business to another bank.3" * The introduction of telephone service to several rural towns and villages in Sri Lanka allowed small farmers to obtain, among other things, current and direct information on wholesale and retail prices of fruits, coconuts, and other produce in Colombo, the capital city. As a result, the farmers began to demand and receive higher prices for their products. Before they had telephone service, they sold at prices averaging 50 to 60 percent of the Colombo price. Following telephone access they regularly sold at 80 to 90 percent of the Colombo price. * A small grocer in Rosario, Uruguay, who sold and delivered gro- ceries to homes, was able to serve a large clientele beyond his immedi- ate neighborhood primarily because residential telephones were available locally and customers could order his goods by telephone for home delivery. . When the Paraguayan National Development Bank began to up- grade and extend its activities in rural areas, it found that telecommu- nications links enhanced the effectiveness of its rural branches. In addition, without access to a telephone, local farmers were unable to 24 AN INTRODlJCTORY PERSPECTIVE relate effectively with markets, were not knowledgeable about current market prices (many farmers sold their production to truck owners at prices substantially below prevailing market field prices), and had dif- ficulty procuring fertilizers and other supplies in a timely manner. * A Ministry of Communications survey showed that public call of- fices in rural areas of Korea (Republic of) averaged 85 local and 160 long-distance calls per month and that the calls helped remove feelings of isolation. They improved villagers' income by providing access to timely market information and by enabling them to bypass middlemen. The local government administration was also able to disseminate infor- mation to the villagers more quickly, as well as to save time and funds by reducing the personal visits that government staff made to the villages. Transport Efficienc'v and Regional Development . In peripheral neighborhoods of metropolitan Lima, Peru, more than 100 Community Telephone Centers have been installed to serve low-income users; each center serves, on average, more than 640 cus- tomers, who receive messages left at the center's number, use the cen- ter to make outgoing calls, and are listed in the Lima telephone directory under the center's number. The typical subscriber is underemployed and travels extensively by bus to search for temporary work in centrally located districts. Subscribers can save up to ten times the cost of their subscription in a year by using the telephone to look for work instead of making daily bus trips. Subscribers also use the centers to report fires, crimes, and health emergencies and thus gain a heightened sense of security.33 * Cameroon is experiencing a severe shortage of properly trained managers who serve the needs of the telecommunications authority and carry out strategic development. This shortage hampers their effi- ciency. Sending managers abroad to study incurs significant costs and exacerbates personnel shortages. Recently, seven managers took part in a training course in project management. Offered at a distance, this course used audio conferences and facsimile communications via telephone lines, along with traditional course materials, to link stu- dents with their instructors overseas. The local course administrator found that using the distance option was much cheaper than sending the same seven managers away from their jobs to study overseas. Par- ticipants learned as much as they would have abroad, without having to leave their jobs and families.34 . In Sri Lanka, a survey of 176 new telephone subscribers (77 busi- ness and 99 residential) in the country's two major cities, Colombo THE ROLE OF TELECOMMUNICATIONS 25 and Kandy, found that almost 70 percent increased the number of contacts they had with other people after getting their phones and nearly half increased the volume of their business. Moreover, 40 per- cent increased their profits or incomes, about 33 percent increased the regional extent of their business, and a little over 10 percent in- creased the variety of their business.35 . Brazil has a national fleet of over 180,000 trucks that must cover extremely long distances, often through rural, remote areas with no telephone service. A major problem for drivers is that their vehicles sometimes break down on the road. When this happens, several days may pass before they can contact their company headquarters for help. This poses a major security risk in addition to slowing deliveries, since cargoes are frequently stolen off the truck while the driver is seeking help. In 1986, a very high-frequency, manually switched radio-telephone system was introduced to serve truck drivers and their company headquarters. It allows a driver to access the national phone network from anywhere on the road and summon assistance without leaving the truck, and it allows dispatchers to locate any truck and relay instructions or other information. - The Road Transport Corporation in Myanmar (formerly Burma) operates more than 2,000 freight trucks out of more than twenty branch stations. In the past, telecommunications facilities were not reliable, and trucks often returned to base empty even though nearby cargoes were waiting to be picked up. Breakdowns could not be quickly reported, and there were delays in analyzing the mechanical problem and sending a repair team with suitable parts and tools. Fi- nally, transport station managers relied heavily on messengers to carry simple communications; at the Yangon (formerly Rangoon) station, from thirty to forty messages a day, which could have been conveyed by telephone, had to be sent by messenger. - According to state government officials in India, the availability of telecommunications services is the primary reason that industrial projects progressed rapidly in some locales in Maharashtra state. They observed that industries were reluctant to move to places inadequately served by telecommunications and that a lack of reliable telephone service in some areas was curtailing the state-supported program to decentralize industry. Furthermore, when telephone service was in- terrupted for technical reasons, businessmen frequently traveled by automobile to nearby locales to communicate with suppliers and cus- tomers, thus using expensive fuel and taking time away from their businesses. . After disturbances in Ogaden, which blocked the road between 26 AN INTRODLJCTORY PERSPECTIVE the port of Assab and Addis Ababa, the Ethiopian government estab- lished a series of checkpoints linked by radio along the road to speed the flow of supplies and to improve security. The progress of individ- ual trucks was monitored, and, in case of a breakdown, spare parts were ordered and brought on the next truck. The radio-linked system cut the average journey time in half, and in a matter of weeks the port of Assab was cleared of the goods that had accumulated there. A Ministry of Agriculture official in Tanzania, who is responsible for project implementation, requires daily information from parasta- tals and ministry departments. Inadequate telecommunications links necessitate frequent visits by car; the official averages three hours of travel each day. Without adequate long-distance telephone connec- tions several other officials must make between eight and twelve costly and time-consuming safaris by automobile each year. In Myanmar, the Inland Water Transport Corporation operates on the country's principal navigable waterways, carrying all of the oil produced in the country and much of the cement and rice. The cor- poration controls the movements of its boats from its headquarters in Yangon through key stations. In the past, the lack of telecommunica- tions links meant that it was often not possible to divert a boat for an unscheduled stop to pick up extra cargo. Unexpected changes or emergencies were frequent, however, and the delay in providing infor- mation made it difficult for management to make alternative arrange- ments, notify shippers and customers, arrange for spare parts, and take other steps to minimize damage and losses. In Uganda. an estimated 2,000 local government officials and senior staff make about 40,000 extra trips a year from district (provin- cial) headquarters to Kampala (an average distance of about 250 kilo- meters), from county to district towns (50 kilometers), and between county and subcounty towns (20 kilometers). These trips are made to handle administrative matters that could be dealt with effectively by mail or telephone if these services were available and reliable in the approximately 700 mostly rural centers. An estimated 250 man-years of qualified government manpower is wasted every year, at a direct cost (salaries, allowances, and public transport) of about $0.6 million. Isolationl and Enierggeicy Security . Newly installed direct dialing service in a small town in India makes it possible to summon medical care rapidly. Before this service was available, snakebite victims often died before help could arrive. THE ROLE OF TELECOMMUNICATIONS 27 Recently, a man bitten by a cobra was saved because the doctor was quickly contacted by phone.36 . In the remote province of San Martin in Peru, health workers in small villages work in primitive medical facilities and have inadequate training; they must try to help their patients in isolation, without consulting other doctors and medical workers. As part of a pilot pro- gram in rural telephony, the Peruvian state telecommunications enter- prise, ENTEL, installed radio and satellite telephone links between the larger town of Juanjui, where there was a small hospital, and four small villages. As a result, rural health workers and their patients were able to consult directly with doctors at the hospital via audio teleconferences. Doctors gave direct advice not only to the workers in charge of a specific case but also to health workers in nearby towns who listened in and thus enlarged their own experience. Although the program boosted the morale and confidence of isolated health work- ers, technical problems ended the service, which is sorely missed.37 Other well-known telemedicine programs exist in Alaska and Guyana, where doctors and health workers use conference calls to work together.38 Appendix A discusses telecommunications and the health sector in more detail. In Ghana, a forest-resources management program operates in rural areas, far from conventional telephone service. A private net- work of radios held by hand and located in vehicles and base stations was set up to tie workers in the field with field stations and park headquarters. Users can coordinate their activities much more effec- tively than before, and the mobile Wildlife Protection Squad, which enforces game and wildlife regulations, can respond much more quickly to reported incidents of poaching and other violations. - In Fiji, during a hurricane, one of the smaller islands suffered se- vere damage with some loss of life and many injuries. The only tele- communications link with the main island was an old radio, which was destroyed during the storm. One week passed before assistance was sent to the distressed island because the authorities on the main island were unaware of the severity of the problem. A properly engi- neered long-distance telecommunications system would have with- stood the hurricane and could have been used to summon immediate assistance. . In Rwanda, the political head of a region has no access to a tele- phone. He relies on passing travelers to hand carry messages to the capital. There is no assurance that his messages will get through or that he will receive a reply. Often he must wait up to two weeks before finding someone who will carry his messages. 28 AN INTRODUCTORY PERSPECTIVE .When a small Alaskan village received a telephone, the number of applications for government grants-in-aid increased dramatically; previously with the once-a-month mail barge that served the village, the turnaround time between legal authorities in Anchorage (the state's largest city) and the village was usually too long for grant ap- plications to arrive on time. Also, the relative slowness of the mail barge and the lack of other access to the outside world were more widely perceived only after a telephone link was established. Hence, the villagers began actively to lobby for building an airplane landing strip near the village.39 . Local leaders of the fairly isolated native population of northern Canada used their newly installed village telephones to plan, discuss priorities for, and coordinate political and economic strategies. Previ- ously, the native leaders had no means by which to coordinate their negotiating positions before meeting with representatives of govern- ment or commercial agencies. After the introduction of telephones, however, the native leaders were able to present their positions at meetings in a much more coordinated manner.40 . Leticia, in the Amazon area of Colombia near the borders of Peru and Brazil, has good timber and fishing resources. Supplies ar- rive only by river and air, and most workers can only visit their fami- lies at most once every two years. The installation several years ago of telephone facilities made regular personal and business communica- tion possible. Morale improved, and the supplies received corre- sponded more closely to short-term variations in needs. Coordination of Inter icntiotitol Activity D A copper mining company in Papua New Guinea is able to ad- minister its operations directly from a location close to the mine on the island of Bougainville primarily because access to telecommunica- tions is good. Among other activities, the company manages its in- vestment portfolio by international telex on a continuous, twenty- four-hour basis. * In Nepal, the tourist industry has the potential to be a major source of foreign exchange. Unfortunately, however, the operations of airlines, hotels, and travel services have been significantly constrained by inadequate and unreliable international communication. There have been frequent problems with room and airplane reservations and with overbooking. By improving international communication links, as well as the domestic telecommunications network, Nepal is now substantially increasing the capacity of its communication-dependent tourist industry to generate foreign exchange for the country. THE ROLE OF TELECOMMUNICATIONS 29 . Political problems are sometimes exaggerated by poor communi- cations among countries. The Organization of African Unity stated in 1980 that "The political backlash that plagues inter-country dis- putes in Africa can be attributed to the lack of adequate communica- tion facilities between the capitals of neighbouring countries."41 According to the OAU, relying on parties outside the continent, which often have preconceived notions of African conflicts, threatens the peace and stability essential for development. Vested interests often distort minor incidents, which could be solved reasonably and speedily if clearly defined positions could be communicated. Having the means to do so would alleviate the fears of neighboring countries in times of crisis. . In Jamaica in 1966 a tourist hotel opened before telecommunica- tions facilities could be provided. The hotel was about 20 miles from the nearest sizable town, but it was attractively located with excellent amenities and a first-class beach. Until telephone and telex facilities were provided, however, this hotel faced very low occupancies because visitors with business, political, and social interests or with health problems were not prepared to be cut off from communicating rapidly with the rest of the world. New arrivals who had not checked on the availability of communication facilities in advance often left after a day or two for this reason. The hotel itself was also unable to handle bookings and cancellations directly. During the peak tourist season when similar hotels with communication facilities were operating at 95 to 100 percent capacity, the communication-deficient hotel had difficulty exceeding 65 percent. Scope and Organization of the Book The review of selected literature on telecommunications investment and economic development, which is summarized in the main body of this book, suggests a paucity of sound analytical material and relevant empirical data on which to base policy decisions about investment in telecommunications.42 Many of the questions that are the most im- portant for analyzing and justifying investment have received little at- tention from economists, statisticians, and engineers. Furthermore, telecommunications are not an intrinsic part of the development de- bate. The study of telecommunications and development has been undertaken mostly by telecommunications specialists, and dissemina- tion of the findings has been largely confined to their peers. This is in stark contrast with, say, transportation and energy, which are regularly covered in the literature on economic development.43 30 AN INTRODUCTORY PERSPECTIVE Scope of the Book Given this shortage of analytic material, the scope of this book is limited to two basic objectives: a review of the available evidence on the role of telecommunications in economic development and an out- line of the principles and techniques of economic analysis that yield information on the telecommunications sector useful for governments seeking to improve their allocation of resources and formulation of policy. The primary questions posed are the following: a. What are the principal constraints on the development and per- formance of the telecommunications sector? What measures can be taken to help overcome these constraints? b. What evidence is currently available to help policymakers assess the effects that investments in telecommunications have on eco- nomic development? c. To what extent can the established tools of economic analysis be used to quantify the benefits of telecommunications and thus determine more accurately the appropriate composition and level of investment in telecommunications needed to allocate national resources efficiently? d. What is known about how telecommunications services are used by different categories of users and about how the benefits are shared by them? How can available evidence be used to assess the distributional effects of expanded or improved services? e. What pricing policies are appropriate for telecommunications services in developing countries? To what extent can pricing en- able the market mechanism to yield correct signals for the use and expansion of telecommunications networks? How effective are current pricing practices in this respect? f. How much could traditional state telecommunications monopo- lies improve? What changes in policy and structure could further promote efficiency and responsiveness and mobilize capital for investment? Or,ganizatioti ofS the Book Organized to address these questions, this book proceeds in prag- matic steps by examining a succession of approaches ranging from the study of sets of countries, single countries, and regions to the exami- nation of specific sectors and the detailed analysis of specific projects. A traditional cost-benefit framework has not been followed throughout. This is partly because the effects of a developing coun- THE ROLE OF TELECOMMUNICATIONS 31 try's telecommunications investment program are likely to be diverse and widespread. Hence, only an extensive and assumption-dependent general equilibrium analysis could begin to identify all of them.44 The primary reason is, however, that existing empirical evidence on the ef- fects of telecommunications investment does not fit neatly into a classic cost-benefit framework. Traditionally, benefits have been viewed in the cost-benefit literature as belonging either to a group variously labeled as primary, direct, or internal (benefits accruing to the users of the service) or to a group variously labeled as secondary, indirect, or external (benefits accruing to others).45 However, a review of the literature on the effects of telecommunications indicates that an attempt to categorize existing analytical work according to a primary/secondary or direct/indirect dichotomy, although appealing in a partial equilibrium project-related sense, has several problems. Such a review would indicate a degree of empirical precision that does not exist and would not provide a suitable framework for much of the material reviewed (for example, cross-country correlations, input- output analysis, and some literature on the energy sector). The book has six parts. Part I gives a background on investment and benefits (presented earlier in this chapter) and explains why tele- communications is, in general, an industry with declining costs (chapter 2). It also identifies the main constraints on the expansion and performance of the telecommunications sector in developing countries and discusses measures that help overcome these con- straints within the limits of a state monopoly of supply (chapter 3). The core of the book deals with the benefits to be gained from telecommunications. Part 11 examines the relationship between tele- communications and the economy from several macroeconomic per- spectives: statistical analysis of a series ot country data (chapter 4), structural analysis of the economy of individual countries (chapter 5), and studies of the spatial organization of economic activity (chapter 6) and of the transport and energy sectors (chapter 7). Part III takes a less aggregate view, presenting microeconomic approaches to mea- suring benefits and using them to analyze individual telecommunica- tions investment projects (chapters 8 and 9). Part IV addresses the question of distribution of benefits by characterizing telecommunica- tions users (chapter 10) and usage (chapters II and 12) as revealed by surveys around the world. Part V looks at the decision to invest in telecommunications from the standpoint of pricing policy in general (chapter 13) and examines how tariffs can contribute to economic efficiency (chapter 14). In ad- dition, it covers the practical application of marginal cost pricing, tar- 32 AN INTRODUCTORY PERSPECTIVE iff analysis in developing countries, and regulatory issues such as price-cap mechanisms and rate-of-return mechanisms in the context of the rapidly changing regulatory and policy environment (chapter 15). Lastly, Part VI discusses the limits to overcoming constraints on telecommunications development when traditional state monopolies supply services and outlines alternative policies and structures that may result in considerable gains in performance, responsiveness, and capital mobilization (chapter 16). This theme is further developed in two separate books.46 Limitations of Scope To kIeep this book a manageable size, three further limitations have been adopted. First, in discussions of the economics of telecommuni- cations infrastructure, the somewlhat contentious subject of radio and television broadcasting policy has been avoided, except incidentally as part of comments on the role of general telecommunications infra- structure, which permits national reception and networking of broad- cast programs.47 Second, the book focuses primarily on telephone service. Despite the dramatic growth of facsimile, data transmission, and other ad- vanced services, telecommunications in both the developing and the industrial worlds are dominated by telephone service and use of the telephone network for transmitting text, image, and data; this use typically accounts for more than 90 percent of telecommunications investment, traffic, and revenues. Although some advanced services are now well established in a number of countries, they still operate on a very small scale compared with telephone service, even in indus- trial countries. More important, however, most of the essential issues and concepts surrounding the voice, text, image, and data transmis- sion aspects of telecommunications policy can be understood through a discussion and analysis of telephone service.48 A final limitation of scope concerns the economics of telecommu- nications technology. The increasingly complex technical decisions facing developing countries are susceptible to economic analysis. The analyst will find relevant material in this book, but not an analysis of technological choices themselves. Many of these are country-specific choices, and the correct solution will also change as technology ad- vances and relative costs vary. THE ROLE OF TELEC OMMUNICATIONS 33 Notes 1. Internationial Telecommunication Union (1984). The Maitland Commission, formally known as the Independent Commission for World Wide TelecommUnications Development, was set up by the ITi in 1982 to examine global development of telecommunications. The comm[ission's final report established the general goal for world policymakers as being to bring "the whole of mankind within easy reach of a telephone ... by the early part of the 21st Century." The 1989 report of the Advisory Group on Telecommunications Policy to the IT reiterated some of the points made by the Maitland Commission, noting that al- though "there is a widespread belief that a relationship exists betwveen investing in tele- communications and boosting the overall economic health of a country," there is also "con- siderable skepticisnm about the benefits to the economy . . . in comparison with other urgent needs." IT (1989). 2. In this book, the word "telephone" is used to designate telephone sets or instruments connected to the public telephone network. The expression "telephone line" (or "main line" or "line" for short) is used to mean a connection between a local telephone exchange and (a) a subscriber telephone or other equipment with a distinct calling nunmber, (b) such a telephone with one or more in-house extensions sharing the sanme number, or (c) a larger subscriber's private branch exchange or key systenm, in which several lines are shared by many otherwise independent telephones. This corresponds to what is also called "exchange access line" or "main station" by the International Telecommunication Union, "line" in the United States, "direct exchange line" in Great Britain and Commonwealth cotintries, and "raccordement principal" in France and some African countries. 3. Until the mid-1980s, the statistics describing telecommunications infrastructure com- monly measured the number of telephone sets. CLrrent practice is to count the number of main lines instead, because the number of nonvoice terminal equlipment such as facsimile machines and computer modems has proliferated rapidly, as has the number of extensions and private branch exchanges. Each telephone main line may be associated with more than one telephone. National averages normally fall in the range of 1.2 to 1.5 telephones per main line and occasionally reach 2.0 or more. Thus, for any particular country at a gtven time, the density of main lines per 100 persons will be less than the density of telephone sets per 100 persons. This edition of the book gives nsain line densities, which are lower than the telephone densities given for the same countries in the first edition appearing ten years ago. The two densities are not directly comparable. 4. World Bank (1989). All dollar ansounts are U.S. dollars. A billion is 1,000 milillion. 5. Wilson and Inan (1989). Investment in developing countries generally expands the system, while investment in industrial countries usainly replaces outdated equipment and provides new services. 6. Luhan (1989). 7. Statistics for facsimile niachines often understate the actual nunmber of pieces of equipment in use, since in many cases, sets do not have to be registered officially with the telephone authority. Where licenses are required, machines are often installed without regis- tration to avoid fees and import restrictions. 8. For example, in 1967, the unrecorded demand for new' residential telephone main lines in greater Santiago (Chile) was approximately three times higher than the number of outstanding applications. See Wellenius (1969a). In nunmerous instances, the waiting list in different countries actually grew after a large project to expand the system made new lines available. 9. For example, when a single telephone nmust be shared by many employees or when in- sufficient lines link a private automatic branch exchange to the public netwvork. 34 AN INTRODUCTORY PERSPECTIVE 10. For instance, repeated attempts to call from a provincial town to the capital city's congested local network may, in turn, congest equipment and call traffic in the provincial town's network. 11. In this book, the World Bank refers to the International Bank for Reconstruction and Development and its affiliate, the International Development Association. Operations by the two other affiliates, the International Finance Corporation and the Multilateral In- vestment Guarantee Association, are not included 12. For example, about 50 percent of the profits from telecommunications operations in India between 1979 and 1983 were passed on to the government, and 30 percent of tele- communications profits, or about $150 million annually, were used to meet losses incurred by the postal service. In 1977, about 20 percent of the postal service's net capital and oper- ating expenses were financed by telecommunications surpluses. More recently, starting in 1989, the government of Uganda began to collect 60 percent of the overall income of the Uganda Posts and Telecommunications Organization. DLiring the five-year period 1989-94, the government will receive about $10 million in direct tax revenues, along with payments from the Posts and Telecommunications Organization of 8 percent interest on loans total- ing $4 million. The net flow of funds from telecommunications to the treasury (taxes, du- ties, and dividends minus government contributions to investment) was about $600 million in Brazil in 1984 and about $350 million in Mexico in 1987. Both amounts wvere equivalent to between 30 and 40 percent of operating revenues. 13. Wellenius (1978). 14. See, for example, Webber (1980). 15. The French government passed legislation that enabled special financing for tele- communications. The installation of new telephone lines reached 2 million a year in 1979, up from 350,000 lines in 1968. Telecom France (1981), pp. 29-32; and Thery (1977). 16. Wellenius and others (1989). 17. American Telephone and Telegraph Corporation (1990). 18. On occasion, variations of this view have been attributed to multilateral development banks and some sources of bilateral aid. In fact, between 1949 and 1974, lending for tele- communications by the Inter-American Development Bank amounted to S87.2 million, or only I percent of total lending. The amount of funds allocated for telecommunications lending by the World Bank averaged about 2 percent of total lending between 1960 and 1990. For the period 1983-89, telecommuniications represented only I percent of total lending. Also, for a considerable period during the mid- and late 1970s, several major bilat- eral donors virtually ceased to support projects for telecommunications infrastruIcture. 19. Clippinger (1977). 20. Melody (1987). 21. The ITl is an intergovernmental agency representing virtually all countries in the world. It is responsible for setting international technical standards, coordinating the use of the radio spectrum, and handling development assistance. The cIrTT (from the French acro- nyns for International Consultative Committee on Telephone and Telegraph), one of the IT's permanent bodies, became involved in the early 1960s in examining issues of telecom- munications and economic development. The other permanent bodies of the ITL are the (n IR (International Consultative Committee for Radio Communications), the IFRss (Interna- rional Frequency Registration Board), and the Secretariat. Following decisions of the 1989 plenipotentiary conference, a fourth permanent body, the BnT (Bureau for Telecommunica- tions Development) was established, taking over and expanding functions that were earlier vested in the ITI's Technical Cooperation Department, as was a high-level committee to make recommendations for a broad reorganization of the IT!. 22. (CITT (1976), chap. 1, sec. 2.1; and ITl' (1989) chap. 5. THE ROLE OF TELECOMMUNICATIONS 35 23. Okundi and Evans (1975) and Okundi (1975). The idea of using satellites for re- gional communications in Africa revived in the late 1980s, and by 1990, the ITL had com- pleted a feasibility study for k.Ascom, a regional African satellite communication project. 24. Pool (1976). 25. Kochen (1982) pp. 230-58 and Center for Science anid Technology for Development (1986). 26. See Block and others (1984); Hudson (1990); and Tietjen (1987). Despite a few outspoken advocates, the question of whether such satellite applications are now the least- cost way to provide such rural services in most developing countries has not been sorted out. Many contend that in niost instances they are not. For discussion of more general ap- proaches to the role of telecommunications in rural developnment, see Hudson (1984) and Parker and others (1989). 27. Webber (1980). 28. Key related decisions were (a) to open to private competitive supply all information services except the electronic directory; (b) to develop a national packet-switched data net- work for cost-effective connection of users and service providers; and (c) to adopt a simple tariff structure and billing procedure. A lively debate has been raging for years on the extent to which cross-subsidy is involved in getting Minitel off the ground, on whether France's Telecom is now recovering all costs, and on whether there have been benefits to the econ- omy over and above what users pav for the new services. Whatever the answers, it is clear that the French videotex systenm is bv far the largest and fastest growing in the world, with hundreds of profitable independent service providers. The French system stands in stark contrast to that of other industrial countries, such as the United States, where mosc videotex initiatives by individual businesses have failed to develop a sufficiently large market base. 29. This would be similar to one aspect of the development of railways, which in nineteenth-century Europe was a favorite sector of infrastructure for both public and pri- vate investment. One cause of the generally poor economic performance of several Euro- pean countries in the nineteenth and early twentieth centuries is thought to have been the niassive diversion of investment into the railway system, in imitation of countries such as England, before the rest of the economy was ready to make full productive use of it. 30. Except where explicitly noted, these examples are drawn from unpublished reports or memoranda on specific projects in telecommunications and other sectors, undertaken mostly in the decade from 1980 to 1990. 31. Pye and Lauder (1987). 32. Malgavkar and Chebbr (1988). 33. Castilla and others (1989). 34. Nettleton (1990). 35. Kojina, Hoken, and Saito (1984). 36. Malgavkar and Chebbi (1988). 37. Mayo and others (1987). 38. Examples are cited in Goldschmidt, Hudson, and Lynn (1980). 39. Goldschimiidt (1978). 40. The example is cited in Hudson and others (1979). 41. Organization for African Unity (1980), pp. 10-11. 42. This view is not unique. See Moss (1981), which summed up by stating that, "over- all, the papers in this volume suggest that we know relatively little about the influence of telecommunications on productivity.... We are still a long way from discerning the impor- tant links and their magnitude." Also see Baer (1981). 43. That this continuLes to be the case sonme twenty-five years after the first published 36 AN INTRODUCTORY PERSPECTIVE studies on telecommunications and development was noted by Jan Bjerningen of the Swed- ish International Development Authority during the consultative meeting with development financing agencies convened by the ITu in Geneva in January 1990. 44. Prest and Turvey (1965) p. 685. 45. Prest and Turvey (1965) pp. 683-90 and Weisbrod (1968) pp. 257-62. Benefits are also classified as tangible and intangible. 46. Wellenius and others (1989) and Wellenius and Stern (1991). 47. Although the sharp distinction between broadcasting and two-way communication media is rapidly breaking down with the advent of new types of telecommunications ser- vices, at present this is important mainly in limited segments of the more advanced econo- mies of industrial countries. 48. All this is not to argue that innovative services may not be important in some devel- oping countries. The very early stage of expansion reached by the telecommunications net- work in many developing countries and the correspondingly low degree of commitment to established sector structures present an opportunity for innovation. See Hobday (1986) and Mody (1989). To seize this opportunity successfully will require creative but realistic appli- cation of criteria for the suitability of technology. Of course, many of the telecommunica- tions innovations currently appearing in the industrial countries are unlikely to deserve priority in many developing countries for several vears. Video telephone and cable television might be put in that category, for example. Other innovative services and technologies, however, deserve more immediate consideration. These include modern electronic text com- munication services, which are superior to telex in most aspects and which may cost no more to implement in a new system; the use of data communications for such purposes as remote access to technical information; and the use of simple audio and graphic teleconferencing facilities as a partial substitute for and complement to travel. Data com- munications and specifically computer-based conferencing systems in developing areas are discussed in Balson, Drysdale, and Stanley (1982). See also National Research Council (1990); Oeffinger (1987); and Paine (1986). Chapter 2 A Perspective on Technology and Costs COSTS IN THE TELECOMMUNICATIONS INDUSTRY are declining throughout the world, and this global trend is projected to continue in the fore- seeable future. The real costs of providing and maintaining telecom- munications facilities and services in both industrial and developing countries have fallen in recent years. This is particularly important to developing countries, which, with their small base of existing facilities and potential for high rates of growth, could achieve rapidly declining unit costs. Two related factors account for this trend: technological change and economies of scale.' Technological Change Until the 1960s, the analog technologies used in telecommunica- tions were evolving and improving at a fairly steady rate.2 During the 1970s and 1980s, however, with the development of digital electron- ics, the increasing application of computer technologies to telecom- munications, and the development of very wide band systems, the pace of change accelerated sharply. This lowered equipment costs, sig- nificantly increased the maximum capacity of individual equipment units, improved reliability, lowered the amount of power used by most types of equipment, and significantly reduced space requirements. New concepts in network management and structure allowed greater flexibility and more efficient use of assets. And finally, a wide range of new facilities and services was developed. Rapid changes are ex- 37 38 AN INTRODUCTORY PERSPECTIVE pected to continue, and Further developments, especially new radio technologies, are likely to affect the cost and performance of systems. Three interdependent series of events contributed to these changes. First, innovations altered basic electronic technology. Some of these originated within the telecommunications industry (for example, the transistor in 1950), while others were exogenous to it (for example, microprocessors for computers in the 1970s). Second, new industrial processes suited to large-scale mass production were introduced in the manufacturing of telecommunications equipment, increasing reliabil- ity and lowering costs. In particular, the advent of transistors followed by printed circuit boards and large-scale integrated circuits shifted the emphasis in production from mechanics to electronics and the em- phasis in basic production elements from simple components to com- plex subsystems. The development and use of large-scale integrated circuits lowered power requirements and led to further miniaturization of discrete elements accompanied by sophisticated, automated printed circuit boards and system assembly. Third, a large spurt in demand for telecommunications services made innovations economi- cally feasible in both basic telecommunications technology and equip- ment manufacturing processes. Hence, to a large extent, the effects of new technology and of scale are inseparable. Eqtipmernt Costs Innovations, especially in microwave, satellite, and fiber-optic sys- tems, have dramatically reduced the unit costs of medium- and long- distance transmission. These cost reductions resulted from both lower equipment costs and greatly increased bandwidth (which determines the number of simultaneous voice and other signals that the equip- ment can carry). The cost of switching also decreased, but not as rap- idly as that of transmission. Considerable improvements and some cost reductions have been achieved in junction and local cable networks. Transmission equipmiicnit. During the 1970s, the cost per transmit- ted voice channel for analog microwave radio terminals (which ac- counts for two-thirds of a microwave route's costs) was cut in half in real terms.3 This was mainly the result of increases in equipment ca- pacity, which grew from 1,200 voice circuits per radio channel in the 1960s to 2,700 voice circuits in the 1970s and to 6,500 voice circuits in the 1980s. When microwave systems and other portions of the net- works were digitalized in the 1980s, terminal costs per channel were A PERSPECTIVE ON TECHNOLOGY AND COSTS 39 further reduced by more than half again. Although these cost reduc- tions were largely the result of increased bandwidth on major routes, important breakthroughs were also achieved in some of the low- capacity transmission systems used to extend service to small commu- nities with a small volume of traffic. Compact and robust integrated analog or digital multiaccess, ultra high-frequency or very high- frequency radio systems now provide high-quality rural communi- cations at a cost well below that of both conventional terrestrial transmission systems and satellites.4 In the late 1980s, for example, digital multiaccess, ultra high-frequency systems cost about $3,000 per channel compared with $10,000 per channel for conventional an- alog ultra and very high-frequency radio systems.5 Such systems are also increasingly cost-effective on routes with no access to electricity. In Australia, for example, the cost of solar power for a repeater sta- tion dropped from $A30 to $A2 over the past decade.6 The merging of multiaccess radio with cellular technology, fixed use of equipment designed for mobile systems, and new generations of radio technology were developed for use in urban areas of industrial countries and promise to provide cost-effective solutions in rural areas of developing countries as well. Improvements in coaxial cable equipment increased the capacity per pair of coaxial tubes from between 960 and 1,200 channels in the 1960s to 10,000 channels in the 1970s and thus reduced the unit costs of transmission at a pace comparable to that of microwaves. The introduction of optical fiber cable systems in the early 1980s further reduced costs. By the early 1990s, optical fibers that transmit from 560 to 2,000 megabits per second (Mb/s) were physically smaller and cost less per pair-kilometer than coaxial tubes with only a fraction of the capacity.7 Furthermore, although the cost of coaxial cables is largely determined by the price of copper in commodity markets, in- novations both in the technology for manufacturing optical fiber and in transmission equipment have steadily reduced the cost of optical fiber systems at about 70 percent annually.8 Other features result in further savings. Unlike copper cables, optical fiber cables are impervi- ous to electrical disturbances and can therefore be installed at low cost on existing power transmission lines and along electrified rail tracks. Optical fiber repeaters can be spaced at 50 kilometers or more compared with 10 kilometers for coaxial cables. Driven by all these cost advantages, optical fiber technology is gradually displacing cop- per wires and cables throughout telecommunications networks.9 The first commercial optical fiber cable systems were installed in industrial countries in the late 1970s to serve routes with heavy long-distance 40 AN INTRODUCTORY PERSPECTIVE traffic. Since then, rapidly decreasing costs in both industrial and de- veloping countries have made these systems the standard choice for new long-distance cable transmission routes and for interexchange junction routes in multiexchange urban networks. In industrial coun- tries, optical fiber cables are also being used increasingly in local loops for customers requiring broadband capability (such as high- speed data or video); by the end of the 1990s, they are expected to cost close to what copper cables do and thus to be widely used in local networks. Satellite systems, in addition to the international telecommunica- tions for which they were initially designed, now provide a variety of cost-effective solutions to domestic long-distance communications in a growing number of developing countries.'0 The unit cost of trans- mission by satellite has decreased markedly, largely because the capac- ities of satellites have increased. For example, the typical price for leasing international circuits from INTELSAT dropped from $32,000 per half-circuit a year in 1965 to $8,500 in 1975, $4,700 in 1981, and $4,400 in 1987." In real terms, the price dropped three-fourths every ten years. The price of full-size earth stations with 30-meter antennas has remained roughly unchanged in current dollars, which reflects a further reduction in real price of about one-third every ten years. The increasing power of modern satellites has also allowed low-cost, small earth stations to be used for low-traffic routes. Improved encoding for voice channels (which reduced the required bandwidth per channel) and improved earth station amplifiers have helped reduce the size of earth stations. VSAT (very small aperture terminal) antennas that are 0.6 meter in diameter are now being used either for television and other signal reception only or for transmission and reception of data and voice by small users (the application of satellites for delivering so- cial services is discussed in appendix A). As of the early 1990s, VSAT terminals are available for about $3,000 for receive-only and for be- tween $6,000 and $10,000 for receive-and-transmit capabilities.'2 Switching equipment. The costs of switching equipment declined from an average of $300-$400 per connected line in the early 1970s to $200-$300 per line in the late 1980s.'3 Thus the real cost per line was cut roughly in half each decade. The cost per line of electronic analog exchanges, initially higher than that of electromechanical ex- changes, tended to fall, reflecting general trends in the electronics in- dustry. Prices for older equipment remained roughly constant in real terms, being tied mainly to the price of metals and labor and to the limited scope for further gains in manufacturing productivity. The in- A PERSPECTIVE ON TECHNOLOGY AND COSTS 41 troduction of digital technology accelerated the decline in unit prices. By the early 1980s, digital long-distance telephone exchanges were available at about half (or less) the price of analog units with similar capacity, and prices continued to fall. Then digital local telephone ex- changes became available at prices comparable to those of electro- mechanical and electronic analog equipment, while offering other significant cost and operational advantages. By the late 1980s, the global production of analog switching equipment had been largely phased out.'4 Nonetheless, whereas the price of most advanced electronic equip- ment, especially computers, fell precipitously, that of digital telecom- munications switches declined rather slowly. One important factor keeping costs up has been the high cost of developing specialized soft- ware. First introduced in the early 1970s with stored program control analog exchanges, software today accounts for a large and still grow- ing proportion of the total cost of digital switching systems. For ex- ample, 45 percent of the initial cost of AT&T'S ESS I exchanges was in software.'" Another factor is that although switching systems are in- creasingly modular in structure, the modules of different systems are not interchangeable. Thus different manufacturers compete with one another only when selling complete systems. Having chosen a system, users can only obtain spares and expansions from the initial supplier. Also, since the modules are not standardized among systems and use proprietary technologies, the benefits of competitive global sourcing of subsystems (which plays such a major role in innovation and cost reduction in the field of computers) have not materialized for tele- communications switching. Jninction and local cables. During the 1970s and early 1980s, the introduction of pulse code modulation systems expanded the capacity of existing interexchange junction cables in large multiexchange local networks by a factor of six.'6 By the early 1990s, optical fiber systems had become the standard solution for expanding junction circuits in urban networks. These systems vastly increased the capacity of inter- exchange circuits in existing duct space, eliminating the need for ex- panding the cable duct networks and saving the attendant costs and disruption of the urban environment. When combined with digital switching equipment, which enables more efficient use of junction circuits, the use of optical fiber cables for junction circuits can also reduce the overall costs of local networks. Technological improvements in telephone sets and other terminal equipment on the customer's premises, and the new switching equip- 42 AN INTRODilCTORY PERSPECTIVE ment's capacity to work over higher resistance subscriber loops, have also reduced the cost of materials used in the local cable networks. For example, in the 1940s, standard telephone sets required the use of 6.5 or 10.0 pounds of copper conductors per mile between tele- phone exchanges and the customer's premises. By the 1970s, the norm was reduced to 2.5 to 4.0 pounds of copper conductors per mile. Since then, advances in telephone set technology and sophisti- cated loop extenders have enabled even longer subscriber loops. Tone dialing has extended signaling limits beyond those possible with dial pulse systems, increased the reliability of telephone sets, and enabled more facilities to be provided within sets. Energy and Space Requtiremenits and Network Performnance The technological changes that helped decrease the real costs of telecommunications equipment also had a significant impact on the costs of the wider system. The most important gains have been re- duced consumption of energy and building space requirements, im- proved network reliability, and gains in efficiency that are the result of better management and positive changes in the structure of the network. Energy and space requiremenits. Semiconductor technology and large-scale integration have reduced the amount of energy consumed by many types of equipment. In the 1970s, for example, a typical 960- channel microwave repeater required a power source of about 1,000 watts. By the 1980s, this had been reduced to about 100 watts. Cur- rently, 300-channel repeaters are available that require less than 10 watts, which can be supplied by solar cells in sites without electrical power or access to fuel for diesel generators. In contrast, the advent of electronic switching has not improved the amount of energy required by telephone exchanges per line. Unlike electromechanical telephone exchanges, whose energy consumption was roughly proportional to the volume of traffic actually flowing through the exchange at a par- ticular time-very high during the busy daytime hours, low during the night-energy used by digital exchanges is mostly defined by total ex- change capacitv rather than by actual traffic flow. As a result, power consumption is constantly high. Electronic components also have limited ranges of acceptable operating temperature, so that air condi- tioning is required, thus raising energy consumption. Progressive miniaturization and integration of electronic compo- A PERSPECTIVE ON TECHNOLOGY AND COSTS 43 nents have produced large reductions in the space required by all tele- communications equipment: switching, transmission, and customer terminal equipment. For example. in the early 1950s, a standard mul- tiplex transmission equipment bay held 24 channels; the bay capacity gradually increased to 240 channels in the 1960s, 600 channels in the 1970s, and 2,000 channels by the late 1980s. The amount of space required for digital switching equipment, excluding complementary items (such as the main distribution frame or a power plant), has also declined. A digital exchange of four to five times the line capacity of an old electromechanical exchange can be installed today in the same amount of space. Significant reductions are also being achieved by eliminating the amount of space occupied by maintenance personnel in each exchange. This follows from the enhanced reliability of equip- ment and the introduction of centralized maintenance and opera- tional capabilities, including facilities for monitoring exchange performance and for extending failure alarms to distant, centralized maintenance locations. Network performanice. Technological changes have also improved reliability. The higher reliability and greater stability of solid-state components, which gradually replaced vacuum tubes in the late 1960s, virtually ended the need for preventive maintenance of most types of transmission equipment and also greatly reduced the inci- dence of repairs. For example, for a typical microwave repeater, the mean time between failures increased from about 15,000 hours in the 1950s to 400,000 hours or more in the 1980s. Use of ducts for major cables, and use of modern plastic-sheathed, jelly-filled cables that ef- fectively prevent the ingress of water without costly gas pressurization, has reduced both the damage that can occur during transport and construction and the incidence of faults. However, equipment using advanced electronic technologies has also made local repair of some parts more difficult and more costly because specialized repair and test facilities are needed. Consequently, although advanced technol- ogy requires fewer maintenance staff, it also makes developing coun- tries more dependent on suppliers. This increases turnaround time for some repairs and means that local stocks of costly spare subsystems must be maintained to overcome failures that could shut down the system altogether. Also, whereas the incidence of faults has been re- duced, complex software and centralized operations make both indi- vidual pieces of equipment and total networks more vulnerable to extensive breakdowns of the system through just a single fault.'7 44 AN INTRODUCTORY PERSPECTIVE The introduction of stored program control analog techniques in switching equipment, followed by digitalization, has made equipment more flexible and improved operational facilities so that the overall system is more efficient. Call metering facilities, which required addi- tional equipment in electromechanical exchanges, is now included in digital exchanges at no extra cost, enabling the adoption of tariff sys- tems that more accurately reflect the time and duration of calls (pric- ing is discussed in chapters 13 through 15). Likewise, the automatic generation of traffic data expedites and facilitates computerized cus- toiner billing. Centralized supervision and network management facil- ities may also make network management more feasible for developing countries. The management of modern digital networks has become so complex that many functions surpass human ability and have had to be assigned to computer robots. Although this equipment is very expensive, it enables fewer skilled managers to run a given system than would have been possible with analog technology. Technological innovations and associated changes in relative costs are also altering the structure of networks. Digitalization has virtually eliminated the traditional boundary between switching and transmis- sion, reduced the interface costs, and enabled more efficient and flexible use of equipment. In the days of all-analog equipment, trans- mission and switching equipment required hybrid connections. As digital transmission technology was introduced, analog-to-digital con- version required interface relay sets between switches and transmis- sion equipment. Now that digital switching is becoming the norm, the system is becoming much more transparent. For example, in an ana- log system, the transmission system would premultiplex outgoing voice circuits in a digital format to at least the DS 1 level (that is, 24 voice channels at a total bit rate of 1.5 megabits per second); in a to- tally digital system, this function would occur within the switch itself. The very low cost of bandwidth now makes it possible to locate more of the switching function closer to the users, which would result in substantial savings in local loops that traditionally account for about 40 percent of the total cost of telecommunications investment in de- veloping countries. The use of optical fiber in local cable networks, which are being installed initially for business users requiring large bandwidths, is likely to extend gradually to individual users. It will allow the introduction of integrated services digital networks (ISDN) providing end-to-end digital connectivity and universal network inter- faces for a wide range of voice, video, and data communications.'8 New generations of cellular and other radio technologies (such as per- sonal communication networks) are expected to offer increasingly A PERSPECTIVE ON TECHNOLOGY AND COSTS 45 cost-effective technical choices to substitute for, or at least comple- ment, the expansion of wired subscriber loops. Serv'ices Developrnent Technological changes have made it possible to extend basic tele- phone service in developing countries to small populations and re- mote areas, as well as to introduce new services increasingly required by the modern sectors of the economy. Satellite communication can reach otherwise inaccessible areas where use of cable or line-of-sight radio communications would otherwise be prohibitively expensive. Current applications include connecting widely separate points in large or archipelagic countries, distributing bulk data, and networking televisions. 19 The introduction of stored program control analog techniques in switching equipment followed by digitalization made possible the pro- vision of new facilities and advanced services that would have been too difficult or too costly using electromechanical switching equip- ment. Enhanced facilities and sophisticated new services designed to meet the requirements of business customers in industrial countries are now more or less built into the designs of new digital switching systems. They are thus inherently available in developing countries even though such a high level of sophistication may be required only in parts of the network. For example, although ISDN is only in the planning stage in most developing countries, privately leased circuits and data terminals connected to dedicated public data networks are already fairly widespread, and international packet-switched links are becoming available in major cities. Facsimile transmission has spread rapidly throughout the developing world. It is particularly relevant in countries that use a script other than the Roman, where use of tele- graph and telex facilities becomes problematical unless costly special machines are designed and produced to suit the local script. Mobile communications are also growing. Many developing countries have li- censed private companies to install and operate cellular telephone systems, which provide telephones primarily to business subscribers willing to pay the high costs of cellular telephones in order to over- come shortages of wired telephone service. More broadly, technology has been the primary driving force be- hind the wave of change in the market structure, policy, and regula- tion of telecommunications that has swept the world since the mid-1980s. This is discussed briefly in chapter 16 and is covered in more depth in other publications.2" 46 AN INTRODUCTORY PERSPECTIVE Economies of Scale It is generally accepted that economies of scale apply to the costs of telecommunications systems. This phenomenon is particularly impor- tant to developing countries, which start with a small base of existing facilities and have large unmet demand, both of which offer an oppor- tunity to reduce costs by accelerating growth of the telecommunica- tions system. Little explicit information is available, however, to help estimate the extent to which accelerated expansion can reduce unit costs in devel- oping countries. Studies of economies of scale have been undertaken based on econometric analysis and on engineering cost data, but only in industrial countries with large, well-established networks, where the supply of telecommunications has kcept pace with demand.2" In econometric analyses carried out mainly in Canada and the United States, aggregate production functions have been estimated using historical operating costs and the book value of existing systems taken as a whole. According to these studies, economies of scale have been present to the extent that a 1 percent increase in input to a large, highly industrial telecommunicationis system has been associ- ated with an increase in output of an average 1.05-1.15 percent. This is, however, a very rough estimate, since results of econometric studies vary significantly (some are much higher, and a few are lower) and, in some instances, different conclusions are reached from the same data.22 Furthermore, the effects of historical changes in scale and technology cannlot be readily separated in these analyses.23 The engineering cost approach to quantifying economies of scale uses data on costs relating to currenit best-practice technology.24 These studies have largely focused on the costs of investments in ter- restrial long-distance transmission, where there are large indivisibili- ties of plant. Estimates show that a I percent increase in scale leads to a 0.6 percent decrease in the average unit cost (a I percent in- crease in input being associated with a 2.5 percent increase in out- put). However, for large multiplex systems and to some extent for switching svstems, the potential economies of scale are much lower. When multiplexing is included in transmission systems, the effect of a I percent increase in scale is closer to a 0.3 percent decrease in unit costs (a I percent increase in input being associated with a 1.4 per- cent increase in output); when both multiplexing and switching are included, the estimates approach the upper end of the range of those obtained in the econometric studies.25 A PERSPECTIVE ON TECHNOLOGY AND COSTS 47 Overall, for countries with large, well-developed telecommunica- tions systems, econometric and engineering studies both suggest that some economies of scale do exist in the provision of telecommunica- tions services but that these are not uniform throughout the system. However, the potential for economies of scale is substantial in most developing countries, which have smaller networks, are usually less en- cumbered with outdated technology, and have significant scope for rapid expansion. These cost savings relate to increasing the efficiency of equipment and networks, procurement, and organization and man- agement of the enterprise. Equipmenit and Network Efficiencv Large economies of scale can be obtained in transmission systems. In long-distance radio transmission networks, the costs per channel and kilometer of investing in 1,200-channel microwave radio equip- ment (excluding multiplex) are about one-seventh the costs of invest- ing in 120-channel equipment over an equivalent distance. For the multiplex equipment, in one African country, the average cost per channel was $960 for a 60-channel set, $460 for a 240-channel set, and $390 for a 600-channel set.26 Similar economies of scale occur in optical fiber cables. An Australian study estimated that the cost per channel for a 30-kilometer optical fiber link was more than $600 for 34 Mb/s capacity, $200 for 140 Mb/s, and $100 for 565 Mb/s. More- over, as the capacity increases, the cost increases less with distance. For example, the cost per circuit of a 565 Mb/s system covering 60 ki- lometers was onlv 10 percent more than the cost of the same system covering 10 kilometers. Another Canadian study compared unit investment costs for three intercity transmission technologies: analog radio, digital radio, and fiber optics (see figure 2-1). In all three cases, high start-up costs made the cost per circuit and kilometer very high for systems with less than 5,000 voice channels. Costs continued to fall until each system's maximum capacity was reached, that is, the exhaust point; after this point, further expansion required an additional system. The research- ers drew two conclusions from this comparison; first, for a transmis- sion system of more than 10,000 voice channels, a fiber-optic system costs less today than any of the earlier transmission systems did in their day, and second, the exhaust point for fiber systems is consid- ered to be virtually unlimited in today's context. Fixed costs, such as the costs of cable, cable installation, and the right of way, so domi- 48 AN INTRODUCTORY PERSPECTIVE Figure 2-1. Comparative Unit Investment Costs for Three Intercity Transmission Technologies in Canada, 1988 Dollars per circuit-kilometer 20 18- 16 14- Trans Canada fiber-optic system 12- 10 8 6 Analog Radio 4 Digital Radio 2 - >Montreal-Toronto fiber-optic system 0 l I 10,000 20,000 30,000 40,000 Voice Channels Note: All dollars are Canadian dollars. The analog radio curve is represented in dollars of the early 1960s and assumes that additional systems do not share an existing right-of-way. The digital radio curve is based on dollar values of the mid-1980s and ends at its exhaust point. Installing larger capacities would produce a curve similar to the analog radio curve. The Montreal-Toronto fiber-optics curve is in 1987 dollars. The Trans-Canada fiber-optic curve is still theoretical and is shown only for purposes of comparison. Source: Federal-Provincial-Territorial Task Force on Telecommunications (1988). nate the overall cost of a fiber-optic link such as the Montreal- Toronto link that the total cost of installing 12,000 voice channels was found to be less than 4 percent higher than the cost of installing the same system for 2,000 channels.27 The costs of transmission can be reduced further when the average use of circuits increases with route capacity. Consider, for example, subscriber-dialed, long-distance telephone calls made from city A to city B. Calls from A are generated at random. To ensure that all calls can be connected as they are generated, the number of circuits from A PERSPECTIVE ON TECHNOLOGY AND COSTS 49 A to B should equal the number of subscribers in A. However, since these subscribers call city B only occasionally and for a limited time, the average use of the circuits is very low, resulting in high cost per call. The average circuit occupation can be increased by sharing a smaller number of circuits among all subscribers, who are automati- cally connected to any available circuit as they make a call. This, how- ever, results in a certain proportion p of unsuccessful call attempts since a free circuit is not always available. Values of p between 0.001 and 0.005 (that is, between 1 and 5 calls out of every 1,000 fail at the first attempt for lack of circuits) are performance targets commonly used in telephone traffic engineering. For a given level of p, the aver- age occupation of circuits increases with the number of circuits in the route. Figure 2-2 shows, for example, circuit occupancy that is 40 per- cent on a 10-circuit route but 81 percent on a 100-circuit route, both Figure 2-2. Average Occupation of Circuits as a Function of the Number of Circuits for Two Levels of Probability (p) of Lost Calls Average occupation of circuit 1.000 _ 0.500 - 0.300 p 0.005 0.200 - 0.100 - 0.050 - / 0.001 0.030 - 0.020 - 0.010 0.005 - 0.003 - 0.002 0.001 1 2 3 5 10 20 30 50 70 100 Number of circuits Note: Based on Erlang's formula for lost-call systems with full availability. 50 AN INTRODUCTORY PERSPECTIVE designed so that 0.5 percent (p = 0.005) of calls are lost. Figure 2-2 also shows a tradeoff betveen quality (the probability of losing calls) and cost: the lower the proportion of calls lost, the lower the circuit occupancy and, hence, the higher the investment required to cater to the same volume of traffic.28 Switching equipment also exhibits economies of scale, although they are less pronounced than those achieved in transmission. For ex- ample, in an African country in 1990, the prices paid per line for ex- changes with a capacity of 5,000, 23,000, and 40,000 lines (including power plant, test equipment, and installation) were $420, $340, and $240, respectively. Because of their fundamental design and architec- ture, however, digital switching systems produce fewer economies of scale than electromechanical exchanges, and the prices paid per line for digital exchanges may not vary much with total capacity. For ex- ample, in a recent tender in a Latin American country, the bid prices offered for individual digital remote line units (small exchanges par- ented on main digital exchanges) with capacities between 500 and 2,000 lines ranged between $150 and $250 per line, while the bid prices for main local exchanges with larger capacities, between 3,000 and 10,000 lines, were within a similar range of $150 and $280 per line.29 Cable distribution networks, typically accounting for about 40 per- cent of the total investments in the telephone svstem in a developing country, also offer considerable potential economies of scale. In an Asian country, average bid price per pair-kilometer of cables, includ- ing the cost of ducts and installation, was $176 for 50-pair cables, $35 for 300-pair cables, $21 for 600-pair cables, and $11 for 2,400- pair cables. Although the unit cost of cable decreases with size, these economies of scale in the urban cable network result mainly from the relatively high costs of ducting and cable installation, which are mostly independent of the size of cables being installed. Procurenln Efficiency Important cost savings can also be obtained by using appropriate procurement practices. These may overshadow the savings obtained from the economies of scale discussed above. Packaging equipment into large orders of relatively few units nor- mally reduces bid prices. For example, in a Latin American country in 1990, the price quoted for 22,000 lines for three telephone exchanges A PERSPECTIVE ON TECHNOLOGY AND COSTS 51 averaged $198 per line, while on a separate order of only 13,000 lines for nine exchanges of the same type, the price quoted was $349 per line.30 At the same time, a total of 68,000 lines of equipment, similar to that of the larger Latin American order, were quoted in an African country at a much higher average of $298 per line, which is largely the result of the African purchase being fragmented into three, sepa- ratelv bid packages. The price is also strongly affected by the modality of procurement and financing. Hence, a vendor may simultaneously quote very differ- ent prices to different customers. Procuring equipment through inter- national competitive bidding or shopping extensively among suppliers can significantly reduce the costs of procurement compared with pro- curing equipment from a single source. A rule of thumb sometimes quoted is that prices under international competitive bidding are about 30 percent lower than those negotiated with a single supplier. Export development credits and most bilateral aid are tied to procure- ment from suppliers in a single country, which often means a single supplier for each type of equipment; this leads to negotiated prices that are often well in excess of those at which the same equipment is sold in competitive markets.3" There is also a tradeoff between cost and uniformity. In general, having a large variety of types of equipment from different manufac- turers is not desirable, since this increases the cost of spares, training, and maintenance. However, locking onto a single type or source tends to escalate costs over time, despite declining costs worldwide. This oc- curs because a common marketing strategy is to offer a potential cus- tomer a very low price on the first order in anticipation of developing a captive customer who will later be forced to buy additional equip- ment, at much higher prices, from the same supplier. This is particu- larly true with switching equipment, which was discussed earlier in terms of technology. For example, in one developing country in the mid-1980s, unit prices for add-ons and new equipment were between 50 and 60 percent higher than prices for the initial order. Placing price adjustment clauses in the original contract sometimes helps control prices in the future. This approach is often used for cables; manufacturing costs are largely determined by prices in world com- modity markets, so a price formula is relatively simple to establish and apply. Ultimately, however, a balance needs to be struck between maintaining a continuous supply from established sources and open- ing the market to competition, in order to lower prices as well as facil- itate technological innovation. 52 AN INTRODUCTORY PERSPECTIVE Organizational Efficiency Size also affects the efficiency of the operating enterprise. Every telecommunications entity, whether it serves 1,000 or I million tele- phone lines, has to carry out a similar number of operation and man- agement functions.32 It is not uncommon for large entities in industrial countries to have five or fewer employees per 1,000 tele- phone lines whereas those in developing countries with well-managed but much smaller networks have staff levels several times higher.33 Be- cause of the frequency and size of individual purchases, large entities also tend to develop greater strength as buyers in the international market and to improve their ability to mobilize financing; on both ac- counts, they can obtain lower costs than those incurred by smaller entities. Organizational economies, as well as equipment and network econ- omies, mav also result from the integration of fragmented services within a given service area and, to some extent, among areas. Provid- ing telex, data, or packet-switching services adds relatively minor ad- ditional administrative costs for a telephone company that already manages 80-90 percent of the total telecommunications investment in an area. Integration can reduce interconnection costs; equipment can be readily standardized, which reduces both the inventory of spare parts and the investment in training staff to maintain and oper- ate it. Furthermore, facilities can be shared among services, thus re- ducing duplication of plant.34 Facilities sharing is especially impor- tant, because the same digital electronic technologies are being used to provide an increasing variety of telecommunications services.35 Disecoriomies of Scale Networks can, however, become so large that they encounter dis- economies of scale.36 Several factors account for this. First, the traffic generated per line may increase with the size of the network because more people can be called.37 The higher the traffic per line, the greater is the quantity of switching equipment needed to process the traffic. Second, as the demand is met in the inner business districts of the main cities, further expansion takes place as new subscribers are added in the less dense outer areas, requiring longer and costlier local loops.38 Third, growth in the number of subscribers increases the quantity of long-distance plant per subscriber.39 This is because the ever-widening communitv of interests, both business and social, fos- tered by the expanded reach of communication and transport sys- A PERSPECTIVE ON TECHNOLOGY AND COSTS 53 tems, results in significant increases in demand for long-distance service. This, combined with the fact that long-distance telephone traffic often grows faster than local traffic, results in increases in both circuit-kilometers and the quantity of long-distance switching equip- ment needed to process the increased flow of traffic. Finally, in most developing countries, a high proportion of subscribers is concentrated in large cities that require multiexchange networks. The requirement of interexchange circuits increases more than proportionally to the number of exchanges, since all exchanges must be interconnected with each other either directly or through intermediate (tandem) exchanges. Operating entities can also become too large. For example, as net- works expand, the existing organizational structures may become in- creasingly ineffective and unresponsive to the needs and circum- stances of the regions or provinces served. Overgrown headquarters become unduly costly to house and service in the expensive urban centers where they are usually located, while regional staff have inade- quate authority and incentives to deal with local matters and limited prospects for developing their careers. Limited authority to retire or reassign personnel made redundant by technological changes may slow the gains in labor productivity and aggravate problems of space, support facilities, management, and welfare. Managers preoccupied with meeting large outstanding demands for basic telephone service and extending the network throughout the country have little inter- est, energy, and resources to deal with emerging needs for more ad- vanced services. Given these potential diseconomies of scale, average cost per line may or may not decrease as the system grows verv large.40 The most likely situation is, however, that even if reductions in unit costs were to taper off along major routes and costs eventually began to increase in certain large centers, costs would continue to decrease in most parts of a developing country's telecommunications system (moreover, as networks become totally digital, the possibility that the costs per line will increase in the future becomes even less likely). The inherent potential for reducing the unit costs of individual pieces of equipment and network components through economies of scale can continue to be realized if this potential is kept in view while equipment and net- works are being technically designed and engineered and while the procurement procedure is being determined for specific items. And, even if at some point the average cost per line begins to increase, the average cost per unit of traffic will almost certainly continue to decline. 54 AN INTRODUCTORY PERSPECTIVE Constraints on Achieving Economies of Technology and Scale A number of factors constrain the pace of technological innovation and system growth and hence limit the extent to which related cost savings can be achieved. Lack of timely attention to developing human resources can more than offset the advantages of technologi- cal innovation. New technology is generally more complex than the technology in existing equipment. For example, digital telephone ex- changes incorporate advanced electronic hardware and specialized software that require sophisticated skills to install and maintain and, consequently, need higher-caliber staff than did electromechanical equipment. Management of software and changes in network design have become more difficult and require higher levels of skill and train- ing. While telecommunications enterprises in developing countries are often overstaffed with poorly trained, low-level personnel, these countries have a universal shortage of highly qualified technicians and engineers capable of specifying, installing, and servicing modern equipment. This has been a major constraint on system performance. The adoption of modern technology is also often constrained by a shortage of capital. Insufficient funds for investment make it difficult for enterprises in developing countries to expand their facilities to nmeet large unmet demands and replace old equipment at the same time. Under pressure to extend services, managers often defer replace- ment of obsolete equipment, despite deteriorating performance and the increasing scarcity and high cost of spare parts. The accelerating pace of technological change in recent years, and the consequent re- duction of the economic (as distinct from the physical) life of most new types of equipment, has exacerbated this dilemma. A related issue is whether developing countries can realize substan- tial savings by purchasing the used equipment that is in good condi- tion and available from industrial countries. To take advantage of rapid technological change, operating companies in industrial coun- tries are replacing equipment well before the end of its physical ser- vice life. Some then attempt to sell the recovered equipment, particularly analog switches overtaken by digitalization, to developing countries, often at prices well below those of new equipment. Because such equipment appears to be less expensive than new modern equip- ment based on initial capital cost, countries short of capital funds may find procurement of such old-technology equipment attractive at first sight. However, the recovered equipment often needs to be re- engineered, at considerable cost, to meet new, very different traffic A PERSPECTIVE ON TECHNOLOGY AND COSTS 55 patterns. It also requires much more building space than new equip- ment. Most important, old equipment imposes heavy penalties on fu- ture development of the network. The developing country will very quickly face exactly the same problems that caused the industrial country to replace the old technology: high costs of operation and maintenance, difficulties integrating into an expanding and techno- logically advancing network, inability to provide the modern services and facilities required by business and government customers, and difficulty and high cost of obtaining adequate spares for maintenance and components for expansion. The timing of technological innovation can involve considerable uncertainties and risks, especially in the context of major shifts. Al- though the direction of technological change is foreseen years in ad- vance, the pace at which such changes will materialize is harder to predict. The cost of telecommunications equipment embodying new technology is uncertain. This is partly because the cost of equipment and its components is highlv sensitive to scale and to the rate at which new-technology equipment is produced, which depends, in turn, on demand. The price of initial supplies of equipment incorpo- rating new technology is high. In the early stages of marketing such equipment, competition among suppliers is limited; also, to speed their recovery of the high costs of developing new technology and equipment, manufacturers attempt to load such costs onto the price of initial supplies of the equipment. The latest technology is often the best choice for developing countries choosing among the technology commercially available in world markets. However, there is a tradeoff between adopting the latest products quickly to avoid further invest- ments in obsolescent equipment and waiting for these products to be proven in extensive field applications to avoid costly failures caused by equipment that has not been tested sufficiently.4' The factors that constrain technological innovation can also pre- vent developing countries from reaping the potential benefits of econ- omies of scale. The extent to which economies of scale reduce the unit cost of providing telecommunications services depends on the speed with which the network expands. This is limited, in turn, by the availability of investment capital for large-scale development, the re- sponsiveness of managers to customer demands, and the cost- effectiveness of procurement practices. In the least-developed coun- tries, the ultimate constraint on the pace of expansion often is the operating entity's technical and managerial ability to implement large development programs and thereafter effectively maintain and operate the assets created. 56 AN INTRODUCTORY PERSPECTIVE Combined Effects of Technology and Scale on Cost and Price of Service Technological change and economies of scale have produced sus- tained reductions in investment costs in developing countries. The av- erage investment cost per line added under telecommunications programs partly financed by the World Bank declined, in current terms, from about $2,000 in the 1970s to between $1,400 and $1,700 in the 1980s. Thus the real costs were cut roughly in half every decade. Cost reductions can be dramatic if the networks are properly de- signed and engineered and the procurement of required equipment is properly organized. Taiwan is an example of what an efficient tele- phone administration can achieve through rapid expansion, purchase of equipment at competitive prices in the world market, and adoption of correct least-cost technical and organizational solutions. In the twenty-one years between 1956 and 1977, the number of telephones increased from 37,000 to 1.2 million lines; by 1989, this figure had risen to 7.5 million. Between 1956 and 1977, the average annual growth rate was 19 percent; between 1967 and 1977, it was an as- tounding 25 percent a year. Strong growth has continued, registering over 10 percent a year for the period from 1985 to 1990. At the same time, labor productivity, measured as the number of employees per 1,000 lines, improved from 114 in 1956 to less than 8 in 1981 and then to less than 5 in 1989. Rapidly decreasing cost per line kept tar- iff increases well below the general level of inflation. The 1956 tariffs, which were already low by world standards, dropped by half in real terms by 1977, while the quality of service greatly improved and the number of subscribers who could be reached by telephone in the country increased by a factor of thirty.42 Tariffs are still being reduced; in 1989, local service rates were effectively cut by about $12.5 million when the areas falling into the local basic rate were enlarged, local in- stallation fees were reduced from $623 to $467, and prices charged for other services, such as international telex, fell also. Notwithstand- ing these low and falling tariffs, the rapidly growing telephone system generated internally more cash than the operating entity needed to invest, and the telecommunications sector transferred increasing pro- portions of revenue to the government in the form of taxes. In 1989, out of total net revenues of $862.7 million, about 63 percent, or $531.4 million, was returned to the national treasury. Because of economies of scale and the adoption of new technology, there can be little doubt that an operating entity with a large and rap- A PERSPECTIVE ON TECHNOLOGY AND COSTS 57 idly growing system should be able, all other things being equal, to offer its subscribers more service for a given amount of money than an entity with a smaller system and a network that is growing more slowly. Table 2-1 shows that in the 1980s the annual telephone bill of a hypothetical business subscriber (making 5,000 local calls and 200 long-distance calls a year) was roughly the same in Togo, which had only about 10,000 subscribers and where the longest call did not ex- ceed 500 kilometers, as the bill of a similar subscriber in India who Table 2-1. Business Telephone Expenditures under Selected Model Assumptions in Selected Developing Countries (U.S. dollars unless otherwise indicated) Annualized connect Annual Total Thousands and line 5,000 200 expenditure of niain deposit rental local trunk under the lines in Country charge' charge callsb calls' model service Bangladesh 45.76 21.89 258.40 231.65 557.69 130' China 84.82 38.88 41.04 97.20 261.94 5,550 Colombia 67.06 18.00 70.00 42.00 197.06 2,070 Cote d'lvoire 26.97 60.32 754.00 301.60 1,142.89 38 Ecuiador 26.23 14.40 25.00 36.00 101.63 450 Guatemala 67.96 144.00 74.67' 90.00 376.62 138 Hungary 363.86 63.94 295.26 142.04 865.10 858 India 87.87 132.66 316.58 753.75 1,290.85 3,488 Indonesia 40.91 23.06 205.88 370.58 640.42 829 Jordan 89.42 108.00 270.00 180.00 647.42 203 Laos 242.92 1,200.00 5,000.00 2,400.00 8,842.92 7 Mexico 77.20 59.05 570.00 205.79 912.04 3,774 Morocco 6.29 29.97 277.50 133.20 446.96 286 Nepal 39.33 4.52 184.73 180.96 409.55 30 Pakistan 17.88 21.60 180.00 259.20 478.68 637 Papua New Guinea 8.57 74.75 805.00 386.40 1,274.72 31 Senegal 35.41 62.37 742.50 178.20 1,018.48 29 Togo 139.04 108.00 750.00 180.00 1,177.04 10 Uganda 3.78 10.80 350.00 504.00 868.58 28 Note: Tariff schedules date from 1982 to 1990. a. Sum of connection charges and deposit, if any, amortized over a ten-year period at 8 percent interest. b. Three minutes if timed. c. Priced as three-minute calls of distances between about 200 and 300 kilometers. d. Estimated. e. Free calls: In Guatensala, rental includes 400 local call impulses a month; in India, 1,650 free calls a year: in Jordan, 2,000 free six-minute calls a year; in China, 60 free calls a month for residences and 100 for buIsinesses. Source: World Bank data and ITL Yearbook of Public Telecommiinicatori Statitics (1990). 58 AN INTRODUCTORY PERSPECTIVE could reach 3.5 million subscribers and call places several thousand kilometers away. Similarly, table 2-2 shows that a hypothetical residen- tial subscriber (making 1,500 local and 10 long-distance calls each year) would have to pay annually $870 in Uganda, with 28,000 lines, less than $200 in various countries with several times as many lines, and only about $100 in the People's Republic of China, with over 5 million lines. Differences among countries in costs and tariffs generally cannot, however, be attributed only to differences in technology and scale. Table 2-2. Residenitial Telephone Expenditures uinder Selected Model Assumrrlptionis in Selected Developing Countries (U.S. dollars unless othervise indicated) Annualized connect Annual Total Thousands and line 1,500 10 expenditures of main deposit rental local trunk under the lines in Country charge' charge calls" calls, model service Bangladesh 45.76 21.89 77.52 11.58 156.75 130' China 84.64 24.62 8.42' 4.86 122.55 5,550 Colombia 37.26 6.00 21.00 2.10 66.36 2,070 C6te d'lvoire 26.97 60.32 226.20 15.08 328.57 38 Ecuador 13.11 3.60 4.50 1.08 22.29 450 Guatemala 67.96 48.00 0.00' 4.50 120.46 138 Hungary 36.33 21.31 66.60 5.33 129.57 858 India 87.87 132.66 65.66 24.12 310.31 3,488 Indonesia 40.91 23.06 61.76 18.53 144.26 829 Jordan 44.71 72.00 0.00Q 9.00 125.71 203 Laos 242.92 1,200.00 1,500.00 120.00 3,062.92 7 Mexico 77.20 55.04 166.50 10.29 309.03 3,774 Morocco 6.29 29.97 83.25 6.66 126.17 286 Nepal 39.33 4.52 52.78 9.05 105.69 30 Pakistan 17.88 21.60 54.00 12.96 106.44 637 Papua New Guinea 8.57 74.75 241.50 19.32 344.14 31 Senegal 24.34 62.37 222.75 8.91 318.37 29 Togo 60.80 108.00 225.00 9.00 402.80 10 Uganda 3.78 10.80 350.00 504.00 868.58 28 Nore: Tariff schedules date from 1982 to 1990. a. Sum of connection charges and deposit, if any, aniorti7ed over a ten-year period at 8 percent interest. b. Three minutes if timed. c. Priced as three-minute calls of distances between about 200 and 300 kilometers. d. Estimated. e. Free calls: In Guatemala, rental includes 400 local call impulses a nmonth; in India, 1,650 free calls a year; in Jordan, 2,000 free six-minute calls a year; in China, 60 free calls a month for residences and 100 for businesses. Source: World Bank data and ITt, Yearbook of Publici Telecommiiitnication Statistics (1990). A PERSPECTIVE ON TECHNOLOGY AND COSTS 59 Many other factors complicate cross-country comparisons, including organizational structure and management efficiency, configuration and stage of network development, effectiveness of operation and maintenance, economy and efficiency in procurement, cost and pro- ductivity of human resources, and government policies on tariffs, tax- ation, financing, and foreign exchange. In particular, no consistent relation exists between the number of lines in a given system and the price charged for service. For example, for eighteen countries analyzed in 1982, the correlation coefficients between the number of lines in service and the annual bill of hypothetical business and residential subscribers were not significantly different from zero.43 The following chapter discusses the major reasons for this surpris- ing result: sector management, organization and financing, and gov- ernment policy toward the sector. Notes 1. Strictly speaking, economies of scale are said to exist when proportional increases in all inputs yield an increase in output that is larger than proportional. 2. It is assumed that the reader has at least an elementary knowledge of the technical characteristics of telecommunications systems and the main types of general solutions and equipment used. For an introduction to these topics, see Green (1986); Chapuis and Joel (1990); and Martin (1991). For an excellent history of the early evolution of telephone switching systems, see Chapuis (1982). 3. "Real terms" refers to prices expressed in a currency of constant value, for example, after discounting for world inflation. The real cost reductions cited in this chapter assume an average dollar rate of inflation of about 5 percent a year. For example, over a ten-year period, a price that remains unchanged in current dollars decreases in real terms to about 60 percent of the original price. 4. One advantage is that these systems permit ctistomers to be served farther from ex- changes, which can, in turn, be located in more populous areas with better access to power and skilled staff. Basic systems provide capacity for 100 or more subscribers, who share about twenty-eight speech channels. Larger systems, with up to sixty channels, allow even greater service coverage at lower costs per subscriber. 5. Implementation and Management Group Party, Ltd. (1980). 6. Mack and Lee (1989). Besides solar cells, low-cost power can be generated for rural telephony using alcohol-burning generators that convert energy and compact wind turbines. 7. At 2,000 Mb/s, a pair of fibers (one that transmits and one that receives) can accom- modate about 30.000 voice frequency channels. The actual bandwidth of a particular cable installation depends on its length, since capacity is reduced as the cable is lengthened. See Green (1986). 8. Flamm, (1989) p. 20. 9. Fiber is more expensive than copper in, among others, the area of terminal equipment and installation. For a long, high-capacity trunk system, this disadvantage is offset by some cost advantages; for a small system with many users, such as a wired office building, using fiber still costs much more than using copper. See Simpson (1990). 10. For a comparative cost analysis of satellite, optic, and microwave transmission media, see Polishuk, Guenther, and Lawlor (1987). 60 AN INTRODUCTORY PERSPECTIVE 11. A (COMSAT study on the future of satellite communications predicts that the net an- nual costs of circuits for satellite use will continue to decline, falling to somewhere around $2,000 by the year 2000. See Crockett (1989). 12. A VSAT satellite distribution system also requires the installation of a hub; either elec- tronic equipment costing approximately $250,000 must be added to an existing master sat- ellite antenna or a new antenna costing at least $1 million must be installed. 13. These are rough ranges of prices paid for switches purchased through international conmpetitive bidding. Prices vary considerably depending on the method of procurement and financing, the size of the exchanges (discussed later under economies of scale), and other factors. 14. Despite being obsolete, some electromechanical equipment was still being manufac- tured in the early 1990s in a few developing countries, including India and the People's Re- public of China. 15. For a more detailed discussion of the effects of computerization of telecomnmunica- tions switching on cost trends, see Flamm, (1989). 16. A digital technique invented in the 1940s, pulse code modulation only became com- merciaily viable with the advent of integrated circuits. 17. This happens as well in industrial countries with sophisticated installations and highly skilled staff. In June of 1991, for example, a rash of large-scale breakdowns occurred in various locations throughout the United States; in one instance, 5 rnillion business and residential telephone subscribers in a congested exchange in the Washington, D.C.. metro- politan area were left for several hours without service due to a software malfunction (caused by a few lines of code buried within millions of lines of programming). Damaged fiber-optic cables have also shut down service recently, affecting increasingly large numbers of users in highly congested areas. For exanmple, in January 1991, a fiber-optic cable was ac- cidentally cut by a maintenance worker, and this cut all long-distance service provided by a major carrier between New York City and the rest of the country. This event nearly shut down all three New York City airports along with niany other critical operations for an en- tire day 18. NIN is the result of an international agreement among cooperating countries, spon- sored by c( IT in Geneva in 1984. It sets electronic parameters and standards and provides a guide for subsequent investment in existing national networks so that within the next ten to twenty years, national networks could become iSU,N networks. The goal is to allow con- nected users to transmit and receive the full variety of communications services (ranging from voice telephone to cable television to data communications) over what used to be just a telephone network. The cost of converting to IsL)N is enormous, and universal implementa- tion for all customers in the public network is by no means certain or even feasible. Many feel that it would only benefit large users and affluent individuals, at the expense of those unable to pay its high costs. See Green (1986) chap. 25 and Krechmer (1989). 19. Despite the expectations of advocates, satellite technology has not become a cost- effective, general alternative to conventional wire, cable, and radio technologies for provid- ing basic domestic services. Several studies have conmpared the costs of satellite systems in rural areas with those of other technologies, such as radio and cellular, that provide mobile service. In Alaska, Hills (1988) found that when compared with existing options of single sideband and microwave/very high-frequency radio systems, a proposed mobile satellite serv- ice was generally more costly, but offered functional advantages. Another study (Khadem, 1987) looked at the cost of cellular and mobile satellite systems for applications in develop- ing countries and found that the cost per subscriber of a hypothetical cellular system would be about $2,500 (not counting administration and maintenance), while that of an equiva- lent satellite system would be $5,400. The main difference lies in the range of coverage, which was qulite limited for the cellular system. Investment sufficient to give the cellular sys- A PERSPECTIVE ON TECHNOLOGY AND COSTS 61 tem as much range as the satellite system would make the costs of the two systems comparable. 20. Wellenius and others (1989) and Wellenius and Stern (1991). 21. More detailed summaries of several relevant econometric studies can be found in Meyer and others (1980), pp. 125-34 and app. D and Littlechild (1979) pp. 49-52. The engineering studies are described in Meyer and others pp. 199-213 and Littlechild pp. 53-73. 22. For example, Meyer and others (1980) report that the Federal Communications Commission in the United States used a study prepared by a firm called T&E and concluded that the case for economies of scale had not been conclusively proven. Then, AT&T em- ployed the Stanford Research Institute to review the T&E study and produce its own conclu- sions, which were that the economies of scale were probably in a range of between 1.1 and 1.25. See Meyer and others (1980) p. 131. 23. One of the problems encountered in econometric studies of scale economies is the difficulty of netting out the effects of technological change. Meyer notes that "studies where comparisons are possible suggest that about one-third of the economies of scale which would be estimated if the impacts of technological change were not taken into ac- count could in fact be due to technological change." See Meyer and others (1980) p. 126. Another study on productivity gains at Bell Canada for the period 1958-80 found annual increases of about 3.5 percent overall. In this case, one-quarter of the gain was judged to be the result of technological change, and three-quarters w-ere the result of economies of scale arising from growing demand. See Kiss (1983). 24. The numerical estimates presented are taken partly from Meyer and others (1980) and Littlechild (1979). 25. The allocation of costs to local and to long-distance facilities is an important factor in determining economies of scale where possibilities exist for bypass. The high costs of local exchange facilities are common to the supply of local and long-distance calls and value added services. Assigning these facilities entirely to local telephone services can raise the price so high that access is placed beyond the reach of many users. Conversely, assign- ing them entirely to long-distance and value added services can stimulate large users to set up private networks to bypass high tariffs. Bypass threatens the economies of common sup- ply in a single system. See ITU (1989) p. 20. 26. Even larger economies of scale apply when the costs of initial civil works (such as roads and sites) and power supply are included. 27. Federal-Provincial-Territorial Task Force on Telecommunications, Canada (1988). 28. Figure 2-2 is based on Erlang's formula for lost-call systems with full availability. 29. Also, whereas the unit cost per termination of a purely long-distance electro- mechanical analog trunk exchange could be as much as three times higher than the cost per termination of a subscriber line in a local exchange, digital exchanges essentially do not differentiate among tvpes of circuits. 30. This price difference is too large to be attributed wholly to the smaller average size of the exchanges in the second lot. 31. The cost of such financing can, however, be considerably lower than that of multilat- eral and commercial sources, so the net effect on total cost has to be examined case by case. One way to deal with this uncertainty is to invite bids for price (of equipment) and terms (of financing) and to award the contract to the bid that reflects the lowest present value of total cost including the cost of financing. 3Z. Many instances demonstrate the organizational economies that can result from in- creased size of the system. An East Asian telecommunications entity with about 300,000 main telephone lines has set up a fundamental planning unit, whose functions revolve around long-range planning, technology monitoring, land acquisition, and ascertaining that 62 AN INTRODUCTORY PERSPECTIVE telecommunications investment programs and pricing policies complement national devel- opment goals and policies in other sectors. The unit is staffed by about ten senior profes- sionals (five engineers, two financial analysts, and three economists) and support personnel. This same unit will, however, be able to handle the entity's planning needs when the system increases to more than I million main lines. 33. Staff ratios as high as 100 and more employees per 1,000 telephone lines are found in some developing countries, including some with very large systems. Whereas this ex- tremely low labor productivity relates to antiquated technology (for example, the number of operators and clerical staff is excessive because switching and billing are not automated), it also reflects employment obligations, constraints on relocating or terminating redundant personnel, lack of performance incentives, and other factors. 34. For example, part of the telephone cable network could connect telex and data sub- scribers, and long-distance circuits could carry telegraph, telex, data, radio, and television program signals. 35. The integration of services and facilities within and among areas involves economies of scope in addition to economies of scale. Although taken for granted in most cases, inte- gration within areas is by no means universal. An extreme example existed in the late 1970s in an East Asian country and has subsequently been only partly remedied by policy reform. At that time, two of the country's more than sixty telephone companies overlapped in six exchange areas in the capital city, each providing completely separate exchanges and distri- bution networks, with only limited facilities to communicate among them; the same two en- tities had duplicate long-distance networks throughout part of the country, with no interconnections, and two other companies also operated long-distance service, partly du- plicating these two networks. 36. Some evidence of diseconomies of scale has been observed in a comparison of the average total costs per telephone line in sixteen urban districts in India. Costs tended on average to decrease up to about 20,000 lines and then to level off up to about 40,000 or 45,000 lines. Of the four telephone districts with more than 45,000 lines, three had higher costs than those with betuween 20,000 and 45,000 lines, while the largest, Bombay, showed little difference. See Kaul (1980). 37. This does not apply, of course, to situations in which expanded investment in an underprovisioned and highly congested system actually reduces the amount of traffic per line. 38. In recent years this problem has been partly offset through the use of digital remote stand-alone line units, which can be located in the center of subscriber clusters located far from the main telephone exchanges. 39. It has also been argued that, at least in North American systems, full economies of scale on long-distance transmission routes are attained at a level of output that is small compared with the market. Waverman (1975) chap. 7. 40. Cross-country comparisons of the costs per line of telecommunications equipment yield meaningless results. Dividing estimated total gross investment in telecommunications plant by the total number of telephones for several countries yields values ranging from nearly $100 to well over $2,000, with a random spread in between that is unrelated to the size of the system. What these numbers reveal, in part, is nothing more than differences in accounting practices among countries, particularly with regard to depreciation and asset re- valuation. For the raw data, see Fargo (1982). Of course, management, planning efficiency, and procurement practices also affect costs per line. For example, in one large developing country in Africa, the cost of equipment per installed line runs into many thousands of dollars. 41. The latest proven technology is, however, designed and manufactured mainly for the markets of industrial countries. Equipment is rarely aimed specifically at the service and traffic requirements, climate, power supply, human resource skills, anid other features of de- A PERSPECTIVE ON TECHNOLOGY AND COSTS 63 veloping countries. Thus developing countries miay have to pay for unnecessarily sophisti- cated equipment and for facilities that will not be used extensively for quite some time. Furthermore, this equipment may require a supporting physical and human infrastructure that is not readily available. The high and still growing cost of research and development, however, and the relative size of markets make it unlikely that this situation will change for major equipment in the near future. Niches may be identified, and manufacturers in newly industrializing countries such as Brazil, India, and Korea are, however, beginning to turn out products especially addressed to the conditions of developing countries. See Ivanek, Nulty, and Holcer (1991) and Mody (1989). 42. For example, the monthly rental of a residential telephone in a large exchange area including a number of local calls, increased between 1956 and 1977 from NT$48 to NT$100 ($2.50). With a correction for inflation (the consumer price index rose more than four times above the 1956 level), the 1977 residential rental was actually half the 1956 rental. Unit call charges changed at roughly the same pace, and in 1977, a local call cost only $0.018, one of the world's lowest call charges. 43. The correlation coefficient measures the strength of association between changes of two (or more) variables. A correlation of 0.0 means that the variables move independently of one another, whereas a correlation coefficient of plus (or minus) 1.0 means that both variables move exactly together (or exactly in opposition). The level of significance mea- sures the probability that a correlation coefficient as high as or higher than the coefficient measured could be obtained if the variables moved purely at random. Correlation analysis is an important tool when it is not self-evident whether the changes in variables are related or not. Chapter 3 Sector Organization, Management, and Financing IT IS DIFFICULT TO DISCUSS the organization and management of national or international telecommunications activities without reference to the or- ganization or functioning of other communication activities and the rules and regulations under which they operate. Telephone, telegraph, telex, data and facsimile transmission, postal messages, postal packages, money orders, messenger and other services, and physical travel serve to varying degrees as both complements to and substitutes for each other. Numerous factors affect the combinations and types of communication networks that emerge both within and between countries, such as the extent to which these activities are regulated by governments; are under- taken primarily by governments or by private enterprise; are limited by shortages of foreign exchange; are supported by other local infrastructure (roads and electric power); are supported by national training or educa- tion programs, which result in literate or skilled local labor; and are en- couraged or held back by direct or indirect government policies, duties, subsidies, competition, and so forth. Nevertheless, given the scope of this book, this chapter focuses primarily on only a portion of the highly diverse communication sector-the organization, management, and fi- nancing of telecommunications services. Telecommunications Organization in Developing Countries In most developing countries, the telecommunications sector com- prises (a) one, two, or occasionally more operating entities, which 64 ORGANIZATION, MANAGEMENT, AND FINANCING 65 provide monopoly public and leased telecommunications services; (b) various networks that meet the needs of large government and other public entities (such as the armed forces, police, railways, and public utilities); (c) private networks that have been set up to meet specialized user needs or to suppiv additional service because the pri- mary telecommunications operating company cannot provide ade- quate coverage or quality of service; and (d) one or more bodies that perform technical and economic regulatory functions.' Although a trend toward liberalization and privatization can be ob- served globally, most public telecommunications services in develop- ing countries are still provided by entities that are partly or wholly owned by the state. In a number of instances, postal and telecommu- nications services are grouped under the same organizational struc- ture. These government-owned telecommunications operating entities are organized in various ways; some form part of a conventional gov- ernment department (as in Algeria, Cameroon. and India), some are semi-independent, state-owned enterprises (as in Ecuador, Indonesia, Jordan, Malaysia, Thailand, Togo, and Uganda), and some are state- owned corporations organized under company law (as was the case in Argentina, Chile, and Mexico before privatization occurred between 1987 and 1990). Under these arrangements, three important constraints limit the development of telecommunications: the operating entities' lack of autonomy from the government, their inadequate internal organiza- tion and management, and a shortage of funds for investment.2 The Autonomy Issue Telecommunications are a rapidly evolving sector, in which signifi- cant technological change is taking place, costs are falling, and the variety of options available for new and innovative services is increas- ing. Three additional factors are particularly relevant for developing countries: (a) the excess demand for even basic telephone service is large, (b) government policymakers are increasingly aware that the economic development of provincial and rural areas requires at least a minimum level of public access to telecommunications facilities, and (c) advanced data, facsimile, electronic mail, and information systems hold the potential for facilitating large productivity increases in the modern sectors of the economy. This situation means that the man- agement of telecommunications entities must be professionally disci- plined as well as flexible, responsive, and alert to evolving needs and 66 AN INTRODUCTORY PERSPECTIVE opportunities. Hence, a certain degree of financial and management autonomy is essential for public telecommunications entities to per- form well. Indcpenidence frorn GoverTnlewt Organizing such entities as corporations with their own board of di- rectors and separating telecommunications from the postal service may appear to be a straightforward way to achieve significant auton- omy. Experience shows, however, that adequate autonomy can some- times also be attained even when telecommunications entities are closely tied to government, such as being part of a government de- partment, provided appropriate organizational and financial measures are implemented.3 Such measures must free the telecommunications entity from day- to-day government interference and should provide for continuity of management despite political changes in the government. To operate efficiently, the entity should not have to secure government approval for normal technical, procurement, and expenditure decisions; top corporate management, following national development policy and regulatory guidelines, should be able to make these decisions. Specifi- cally, this situation could be brought about by (a) streamlining the re- lations with the few government agencies that should be legitimately interested in influencing the sector's long-run plans; (b) expediting procedures for approving investment programs, which should have a multiyear horizon and be only generally specified to allow manage- ment flexibility in the details of implementation; (c) allowing freedom to operate under general guidelines, which specify tariff levels that allow full recovery of costs, generate substantial funds for new invest- ment, and promote efficient allocation of resources; (d) simplifying the procedures for periodically revising tariffs where general price in- tlation requires frequent adjustments; (e) giving the operating entities authority to collect bills from all users and to disconnect nonpaying subscribers (this should explicitly include all government subscrib- ers); (f) requiring that accounts be maintained on a commercial basis; (g) ensuring that internally generated funds are available for in- vestment under approved medium-term development programs, with the operating companies either retaining these funds or easily recover- ing them when required;4 and (h) allowing operating entities to set salaries, wages, and other benefits so that they can be competitive employers and attract and retain qualified staff at all levels. The government should expect management to pursue the objec- ORGANIZATION, MANAGEMENT, AND FINANCING 67 tives and follow the policies laid down by the government through the autonomous operating entity's board of directors and to manage the operating company in a protessional and cost-conscious way. Govern- ment should also expect the entity to pay normal taxes and duties, market rates of interest on all new debt (including government debt), and dividends on government equity at a rate that would attract eq- uity capital to private enterprises with similar risk profiles. Indepenidceicc from the Postal Scrvice In many countries, the provision of telecommunications services is organizationally linked with that of postal services, because of the early history of the sector. In the middle and late 18Os, when tele- graph and telephone services were first being developed, many Euro- pean countries treated them as an extension of postal service communication. Several important colonial powers then passed on the combined Posts and Telecommunications (P&T) Department or Office des Posts et Telecommunications (OPT) type of organization to their colonial administrations. In a more modern context, however, several arguments can be made for separating the management of postal and telecommunications ser- vices: (a) although advanced technology and management systems play a big role in the provision of postal services, in general, postal operations are highly labor intensive, whereas telecommunications operations are capital intensive with rapidly decreasing labor inputs; (b) the provision of telecommunications services requires a much higher proportion of skilled labor and professionals than does the provision of postal services-lumping them together in a developing country often contributes to telecommunications salaries that are too low to retain qualified staff or to personnel regulations that are overly rigid; and (c) the contrasts between the technologically dynamic and rapidly evolving telecommunications sector and the more mature labor- and physical-transport-dominated postal sector are so great that the same organizational structure and management style cannot plan for and implement both services efficiently. Given such consider- ations, a growing number of countries have partitioned, or are begin- ning to consider partitioning, their postal and telecommunications organization and are establishing each service as an independent en- tity. The breaking up of the British Post Office in 1981 into an inde- pendent British Telecom and the Post Office is a classical example. Of course, in a developing country there may still be some advan- tages to an organization that combines posts and telecommunica- 68 AN INTRODUCTORY PERSPECTIVE tions, particularly in more isolated areas and small towns in which the local post office building may house not only postal mail services but also savings bank facilities and a telephone public call office (rco). However, the P&T does not have to be combined organizationally for a combined facility to function effectively. Several countries provide rco services in rural areas without undue problems by having the separate telecommunications entity pay the post office or small local shop or cafe a commission for housing and operating a Pco. It could be ar- gued that in the future, when sophisticated electronic text systems are more widely available in developing countries, postal and telecommu- nications services should at least partly be administered by the same organization, since they will converge technically. Although this sug- gestion has some validity, it does not seem desirable to tolerate the major operational inefficiencies apparent in many developing coun- tries in anticipation of an uncertain development in the future. Also, when electronic text is introduced outside large urban areas, at least initially, the electronic transmission function could be handled by the telecommunications entity and labor-intensive distribution handled by the postal service. It is even possible that some competition be- tween the two entities providing text services could stimulate more ef- ficient management and service responsiveness (for a more detailed discussion of electronic mail and the interaction of postal and tele- communications services, see appendix B). All in all, at least in deve- loping countries, the case can be made that separating posts from telecommunications could, organizationally as well as financially, streamline the management structure, and hence increase the long- term efficiency, of both the postal and telecommunications sectors.' If using telecommunications revenue to subsidize postal services was thought to be desirable, this could still be done; it would just have to be done on a more explicit basis. Internal Organization and Management of Operating Entities A second major problem for the telecommunications sector in de- veloping countries is inadequate organization, management, and staff- ing of the operating entities. The following weaknesses are often observed. The entity's organizational structure is frequently inade- quate for the size of the telecommunications development effort re- quired, or for the nature of the business (as noted above this is sometimes the result of combining the functions and control of postal ORGANIZATION, MANAGEMENT, AND FINANCING 69 and telecommunications and the inherent organizational and opera- tional problems associated with such an arrangement). Job descrip- tions and requirements, service standards and staffing norms, separation of responsibilities, and lines of delegation of authority are not well defined. Administrative and financial controls are loose or nonexistent. Management is timid and lacks objectives, goals, and ac- countability. Financial management has little influence on planning and day-to-day decisions. Sufficiently comprehensive management in- formation systems are not in place. No one unit is responsible for long-range planning and economic analysis. Commercial forms of ac- counting often are not used or do not produce timely signals for decisionmaking and for assessing performance. Billing and collection of receivables are slow and not well monitored, which places an un- necessary financial burden on the organization. Little thought is given to maintaining strong, central engineering planning or to coordinat- ing and supervising projects. Finally, decentralization of administrative work and technical, operational, and billing and collection responsi- bilities is unduly delayed as the size and complexity of the organiza- tion increases. Staffing Problems The key to building a capable operating entity is adequate staffing. Many telecommunications organizations in developing countries have yet to develop and maintain an effective and expanding system for training all levels of engineering, financial, and administrative person- nel. Effective training programs increase productivity and in the long term consolidate a stable and qualified cadre of senior staff who are supported by middle-level personnel from which future executives can be promoted. Good telecommunications training centers exist in several develop- ing countries, and many of them were set up with technical assistance from the ITU and financed partly by the United Nations Development Programme. Typically, technical personnel up to the level of techni- cian are trained in national or regional centers and on the job. Junior professional engineers are also educated in these centers or in local university programs. Education abroad (other than in regional cen- ters) is typically reserved for personnel at selected professional levels or with narrow technical specialties.6 Training problems generally in- clude inadequate coordination between training centers and the oper- ating companies and lack of attention given to training in the important areas of administration, finance, and accounting. 70 AN INTRODUCTORY PERSPECTIVE An especially difficult problem is faced by some of the smaller, poorer developing countries, which depend on foreign personnel and have virtually no sources for recruiting new staff with the level of gen- eral education needed to undergo training. Nevertheless, it is crucial to transfer all posts to nationals as soon as sufficiently competent re- placements are available. Dependence on foreign staff has on occa- sion led to unnecessary overstaffing at the most senior levels and to acute personnel crises in several countries, sometimes resulting in the virtual collapse of telecommunications operations. A comprehensive strategy for recruitment, training, further education, and career devel- opnient of staff is needed in such countries, often with a ten- to fifteen-year time horizon. With such a plan the rapidly rising demand for middle-level personnel could soon be met with adequate local staff, who in turn gradually take over all posts held by temporary for- eign experts.7 A Management Checklist The following is a checklist of selected items to which management and policymakers might refer in identifying problem areas and evalu- ating steps that might be taken to improve the internal management of monopoly operating entities. Organization. Does the entity have adequate autonomy? Is it sepa- rated from the postal service? Does it operate on a commercial basis? Does it have appropriate medium- and long-term objectives or guide- lines relating to the services offered, access to services, quality of ser- vices, finances, tariff policy, and so forth? Management and control. Does the operating entity have an ade- quate management information system that facilitates effective and responsible control of current operations, stores, maintenance prob- lems, financial affairs, and long-term planning and programming? Are adequate statistics collected on traffic, faults, outages, equipment per- formance, management efficiency, and so forth? Does management have difficulty controlling or coordinating among departments, divi- sions, or units? Are subscribers connected at a reasonably high rate when exchanges are ready to be cut over? Do adequate procedures exist for maintaining, testing, and checking the quality of work per- formed? Are there enough qualified professionals, managers, techni- cians, and accounting and finance officers? Are job descriptions adequate at all levels? Can the existing staff cope with the additional ORGANIZATION, MANAGEMENT, AND FINANCING 71 responsibilities of implementing an expansion program and managing the expanded entity? Are salaries adequate to retain qualified staff? Are policies relating to staff advancement and promotion appropriate, are the accompanying rules and regulations sufficiently flexible, and are incentives for achievement built into the system? Is the level of general staff excessive? What use is made of expatriates, of consul- tants? Are present arrangements suitable for an expanded work load? Is technical assistance needed in the short run? Does management make an adequate effort to keep the public informed of major devel- opments and to promote public support for the sector? Plannig. Are economic analyses of expressed and hidden demand and of the telecommunications needs of both urban and rural areas undertaken, and are the results of such analyses adequately conveyed to officials at the finance and planning ministries and to the public? Are the telecommunications development plans closely coordinated with government plans and programs in other sectors as well as with private plans? Do updated national numbering, charging, switching, signaling, and transmission plans exist? Are they judged to be sound in light of the country's requirements and its technological and finan- cial resources? Do regional development plans exist? Do physical de- velopment plans exist for the capital city and major metropolitan areas? Are the service targets reasonable given current service levels and national goals for access and quality of service? How are the cur- rent (one to four years) programs prepared? Are the development of services and service priorities balanced properly? Is exchange equip- ment coordinated adequately with external plant works? What is the basis of costing in the works programs, and are there accurate and up- dated records of unit costs? Are the costs of various works reasonable? Is preparing civil works a bottleneck? Have the major works been sub- ject to economic comparisons of alternatives (for example, coaxial cable versus microwave and optical fiber for a major trunk bearer)? Is the discount rate appropriate? Does it (or should it) reflect capital scarcity? Do plant practices specify the forward provisioning periods for the various classes of plant, for example, main, secondary, and ter- tiary distribution cable, junction cable, exchange equipment, multi- plex equipment? Are the provisioning periods reasonable with regard to growth rates, availability of capital, and prevailing costs of equip- ment and labor? Accounting system. Is the accounting system on a commercial basis, and, if relevant, are telecommunications accounts kept separate from 72 AN INTRODUCTORY PERSPECTIVE postal accounts? Is the commercial accounting system adequate to keep management informed and able to foresee trends and possible problems? Are there adequate internal financial controls (control of cash receipts and payments, inventories, plant retirements, sale of property, related materials, and so forth)? Is the costing or cost con- trol system adequate? What accounting practices are being followed, and are there any problems relating to depreciation, the transfer of plant investment from "under construction" to "in service," and the handling of foreign exchange losses, bonuses, and so forth? Are pen- sion funding arrangements adequate? Are there sufficient procedures for writing off bad debts? Billing and collection. Are billing arrangements satisfactory? What is the billing cycle, and is the accounts receivable position satisfactory for the length of the billing cycle? Could billing be expedited through increased computerization? How much would this cost, and what would it save in time and cash on hand? How are collections made? Are there disputed account problems, and how can these be resolved? Are there satisfactory procedures for disconnecting nonpaying sub- scribers, including all government and other public sector subscribers? For revenues not collected directly by the entity, how are charges set, and how do they reach the operating entity? Inventories and stores. Is the inventory or stores control system sat- isfactory? Are inventory levels satisfactory? Is there an adequate sys- tem for replenishing inventories quickly, for checking whether turnover is reasonable in relation to the scale of operations (and of construction), and for identifying and disposing of obsolete material? Training. Have personnel development policies and practices been recently evaluated in the light of current and forecast requirements of the operating entity and of the sector? How do training programs and facilities within the entity and within the sector compare to the needs? Has this topic been adequately examined by national or inter- national agencies or experts (ITU or consulting firms)? What action might be initiated to deal with any personnel shortages likely to affect current or future development within the sector? Audit. Are present arrangements for external and internal audits satisfactory? Are the nature and extent of the audits appropriate? Is there any reason to question the independence of the external audi- ORGANIZATION, MANAGEMENT, AND FINANCING 73 tors, the adequacy of the qualifications and experience of the person- nel conducting the audit, or the adequacy of the procedures followed? Do the audit reports contain all the financial statements and supple- mentary information needed for analysis and comparisons? Do exter- nal auditors systematically advise management of weaknesses in the accounting system that should be improved? When used, is govern- ment audit adequate, or should it be supplemented by other audits? What is a reasonable period for submitting audited financial state- ments? Are internal audits and inventories done systematically? Are internal audits adequate? To whom does the internal audit section report? Tariff policy and financial plants. Does the telecommunications operating entity generate sufficient resources from users of services to cover operation and maintenance costs as well as debt and interest payments, to generate a reasonable return on assets and pay dividends on equity, and to cover a reasonable proportion of the costs of ex- panding the system in the future? What proportion of cash is ex- pected to be generated internally for the next three to five years? Is it satisfactory? Is the amount of net internal cash generation influenced significantly by the capital structure and terms of debt financing (un- usually low or high debt-equity ratios, easy or harsh debt repayment terms, and use of supplier credits)? If the percentage of self-financing for capital requirements is below 35 or 40 percent, is this level accept- able in view of the size of the work program, recent tariff action, ex- pected tariff revisions, or expected level of cash generated internally after the full revenues from the completed work program are realized? Do special factors, such as taxes, bonuses, dividends, or other pay- ments to government or exceptional customs duties, depress the per- centage of cash generated internally? If excess demand exists, is some form of price rationing used to encourage business and government priority access or use? Are peak-period pricing (local and long- distance calls) and toll ticketing practiced? Do long-distance charges increase too rapidly with distance? Do tariffs for specific types of ser- vice reflect the cost of those services? Are any deviations from pricing according to the cost of service justified on grounds of promoting a more efficient or equitable use of, or access to, service? Has the incre- mental cost of expanding or adding specific services been studied? Are telecommunications tariffs used partlv to supplement general government taxes, to subsidize postal services, and so forth? What are the present implications of this, and what are the longer-term goals? 74 AN INTRODUCTORY PERSPECTIVE The Financial Constraint Another major factor that directly constrains the organization and management of the telecommunications sector in many developing countries is the scarcity of funds for capital investment. This con- straint is especially significant because the capital requirements of the sector are relatively high; the ratio of capital to output (measured by revenue) in telecommunications has been estimated to be about 3:1.8 Likewise, the incremental capital-output ratio, which describes the re- lation between new investment (incremental requirements for addi- tional capital) and expanded output, is large compared with the value of the same ratio for most other industries. Of course, such ratios give only the most general indications of capital requirements; capital re- quirements per unit of system expansion vary widely, depending on program composition, geographic characteristics, cost of local inputs, conditions of procurement, and a host of other factors.9 In the 1970s, developing countries invested about S3 billion annu- ally. In the 1980s, this figure more than doubled to reach about $7 billion; by the late 1980s, it had reached some $12 billion (all figures are in 1988 U.S. dollars). As a tentative estimate, developing coun- tries are likely to require about $25 billion a vear in the 1990s if they are to catch up with the unmet demand for basic telephone service by the year 2000 (a rather modest target). In real terms, this is more than three times the investment level achieved in the 1980s and eight times that of the 1970s. Additional funds will be needed to meet the still relatively small but very fast-growing demand for the advanced services increasingly required by modern economic sectors.'0 Figure 3-1 presents the approximate breakdown of telecommunica- tions financing for developing countries during the 1980s by the main source of funds.!! About 60 percent of investments were financed by the operating entities themselves through internally generated cash (mainly retained earnings and provisions for depreciation). It is clear, therefore, that despite the large capital investments required for tele- communications operations and expansion, a lack of local currency should not be a constraint on this process for developing countries. As outlined in chapter 1, monopoly telecommunications entities can easily generate large financial surpluses in the local currency. With correct pricing policies, the full costs of providing telecommunica- tions services, including the cost of capital, can be recovered from tariffs, and a large proportion of the funds required for subsequent improvement and expansion can be generated internally (the pricing ORGANIZATION, MANAGEMENT, AND FINANCING 75 Figure 3-1. Sources of Funds for Telecommunications Investment in the Developing World, 1980s Internal generation 60% Bilateral and commercial Government 5k7 25% Multilateral 5% Source: Wel.enius (1990b). of telecom-unications services is discussed in chapters 13, 14, and 15). In the absence of sufficient local capacity for manufacturing tele- communications equipment, however, an investment program will have large foreign exchange requirements. The foreign exchange com- ponents under fifteen recent telecommunications expansion programs partially financed by the World Bank, for example, ranged widely be- tween 12 and 90 percent of the total investment requirements, with the norm being between 70 and 85 percent. Hence, although the tele- communications entity can generate large financial surpluses in local currency, these can rarely be reinvested within the sector as fast as they could be absorbed, because of shortages of foreign exchange.'2 This is the case even though the sector's foreign exchange require- ments are usually a small proportion, in the range of 0.5 to 3.0 per- cent, of the total annual imports of a developing country. Although telecommunications entities in developing countries usu- ally do not directly generate enough foreign exchange to support rapid expansion programs, considerable foreign exchange earnings and savings are generated in other sectors as a result of improved or expanded telecommunications investment. For example, the foreign exchange requirements of a country may be reduced because access to 76 AN INTRODUCTORY PERSPECTIVE and quality of telecommunications services are improved, which in- creases administrative and management efficiency and operational productivity in other sectors and, in turn, reduces investment require- ments in foreign exchange as well as local funds. When petroleum prices are high, such an argument is particularly relevant for sectors such as transport. As discussed in chapter 7, a well-functioning tele- communications system in a typical developing country might substi- tute some petroleum-consuming transport and begin to use transpor- tation more efficiently. The problem, of course, is that the foreign exchange earnings and savings brought about by investments in telecommunications services, but realized outside the sector, do not accrue to the telecommunica- tions entities. They are also difficult to identify and measure. Govern- ment finance and economic planning ministries in developing countries, therefore, base their decisions about sector allocations or policy on very incomplete information about the impact of foreign exchanige. 1 3 Aside from internal generation, other options include supplier cred- its, foreign commercial bank loans, and bilateral aid. These account tor about 25 percent of total investment and often constitute the main source of finance for telecommunications imports in many de- veloping countries. The drawbacks that sometimes limit the extent to which these options are used include the normally high cost of sup- plier credits and commercial bank loans, the relatively short repay- ment periods, and the relatively high price of equipment procured through supplier credits and other forms of tied or restricted procure- ment. Also, less well-off developing nations may lack the financial strength to be eligible for these sources of funds. Multilateral develop- ment banks are another source of foreign exchange for telecommuni- cations development. Multilateral development banks, such as the World Bank and the Inter-American, Asian, and African development banks, accounted for about 5 percent of telecommunications funds invested in develop- ing countries in the 1980s. From the early 1960s to 1990, the World Bank lent about $4 billion to support over 100 projects, costing al- most $20 billion, in some forty countries. This makes it the largest multilateral source of financing for telecommunications. However, since the Bank generally lends for telecommunications only when lending from other sources is not available at reasonable terms, tele- communications amount to only 2 percent of all Bank operations; this proportion varies considerably from year to year, depending on couLntry priorities and the availability of other sources of funds (see table 3-1). For example, in the fiscal year ending June 30, 1987, the ORGANIZATION, MANAGEMENT, AND FINANCING 77 Bank extended new telecommunications loans and credits totaling more than $682 million; the following year, the figure dropped to $36 million, and in 1990, it reached more than $616 million. World Bank telecommunications loans and credits traditionally fi- nance a three- to five-year slice of the country's total public telecom- munications investment program. These operations support rehabili- tation, expansion, and modernization of local, long-distance, and international facilities in urban and rural areas. They also finance technical assistance needed to improve the operating entities' organi- zation, management, and human resources. Increasingly, however, World Bank telecommunications lending focuses on selected aspects of particular complexity or urgency, such as the rehabilitation of run- down plant, the development of cable networks, or the strengthening of financial management or project planning and implementation. Further details showing the amount of loans and credit granted by the World Bank to specific countries for telecommunications projects be- tween 1962 and 1989 are given in appendix E. In addition to loans and credits to the telecommunications sector, the World Bank supports telecommunications development through a variety of lending operations not included in the above figures. In fis- cal 1988-89, about 20 percent of Bank-financed projects in other sec- tors (mainly railways, power, agriculture, rural development, and earthquake and other reconstruction projects) had telecommunica- tions components. These components were estimated to cost about $161.6 million, of which about $64.8 million was financed by the Bank. A major area of growth since the mid-1980s has been that of policy adjustment. Telecommunications enterprises have been in- cluded in Bank-financed public sector management projects, espe- cially in Sub-Saharan African countries. Loans of up to $400 million, to help offset the costs of wide-ranging public sector reforms, were made in the early 1990s to several Latin American countries subject to progress in implementing the governments' plans to overhaul tele- communications policy, restructure and privatize telecommunications operating entities, and develop competition and public regulation (these aspects are discussed briefly here and in chapter 16).14 A brief comment is in order on the two remaining sources of fi- nancing shown in figure 3-1. Government loans and equity contribu- tions, although apparently substantial in many developing countries, are often more than offset by the transfer of operating surpluses from the operating entities to the national treasuries or to meet postal and other deficits. Overall, it is estimated that net government contribu- tions do not exceed more than 5 percent of total funding require- ments. Until about 1990, private investment was limited to the very Table 3-1. Trends in Telecommunications Lending hy the World Bank, Fiscal Years 1986-90 1986 1987 1988 1989 1990 As a As a As a As a As a percent percent percent percent percent Millions of of total Millions of of total Millions of of total Millions of of total Millions of of total Sector dollars lending dollars lending dollars lending dollars lending dollars lending Telecommunications 50.4 0.3 682.3 3.9 36.0 0.2 161.0 0.8 616.7 3.0 Agriculture and rural development 4,777.4 29.3 2,930.3 16.6 4,493.9 23.4 3,490.0 16.3 3,656.1 17.7 Development finance companies 1,449.2 8.9 2,297.9 13.0 1,712.5 8.9 2,366.7 11.1 1,271.7 6.1 Education 829.2 5.1 439.8 2.5 864.0 4.5 890.7 4.2 1,486.6 7.2 --j Energy 3,018.0 18.5 3,704.3 21.0 2,395.0 12.4 3,863.6 18.1 3,304.3 15.9 Industry 821.1 5.0 418.4 2.4 2,224.6 11.6 1,982.5 9.3 795.6 3.8 Nonproject 1,321.0 8.1 2,437.1 13.8 1,687.0 8.8 3,418.5 16.0 3,044.0 14.7 Population, health, and nutrition 419.5 2.6 54.1 0.3 304.9 1.6 623.0 2.9 933.4 4.5 Public sector management' - - - - - - - - 525.6 2.5 Small.scale enterprises 274.5 1.7 421.5 2.4 513.0 2.7 585.0 2.7 207.5 1.0 Technical assistance 137.9 0.8 103.9 0.6 95.7 0.5 175.3 0.8 141.0 0.7 Transportation 1,498.2 9.2 1,745.9 9.9 2,642.5 13.7 1,830.8 8.6 2,785.3 13.5 Urban development 1,117.5 6.8 1,469.1 8.3 1,716.3 8.9 1,188.5 5.6 1,002.1 4.8 Water supply and sewerage 604.8 3.7 969.4 5.5 535.3 2.8 791.2 3.7 931.8 4.5 Total 16,318.7 100.0 17,674.0 100.0 19,220.7 100.0 21,366.8 100.0 20,701.7 100.0 - Not available. a. First introduced in 1990. Source: World Bank Annual Report (various years). ORGANIZATION, MANAGEMENT, AND FINANCING 79 few countries that allowed the private sector to provide telecommuni- cations services in any way. Private investment in telecommunications was largely ruled out in the 1950s and 1960s in many countries, as the emerging Asian and African nations became independent and many Latin American countries nationalized the foreign-owned tele- communications companies. Only a few countries (notably the Philip- pines) retained service in private hands. Some private investment has occurred in a small number of state companies that were partly owned by subscribers (mainly through mandatory purchase of shares, which occurred in Brazil), were publicly traded in domestic stock markets (which occurred in Chile), or, even more rarely, were publicly traded in foreign markets (American Depository Receipts of shares of Telefonos de Mexico were traded over the counter in the United States in the 1980s; until 1990, 51 percent of Telefonos de Mexico had been owned by the state). Altogether, private investment (volun- tary and mandatory) probably did not account for more than 5 per- cent of total telecommunications investment in the developing world during the 1980s.'5 The following chapters examine some of the facts and contentions surrounding two other constraints perceived to limit telecommunica- tions investment in developing countries: the benefits of telecommu- nications investment are not enumerated and quantified as fully as are those of other sectors, and telecommunications investments, al- though financially profitable, are perceived to benefit directly only a relatively narrow-and privileged-portion of the population. Notes 1. Some countries also have a significant domestic industry that manufactures telecom- munications equipment and cables. These activities, although closely linked to the provi- sion of telecommunications services through the procurement policies of the operating entities, must be treated as a subsector under the rubric of manufacturing industry, not telecommunications. 2. The following three sections draw on Saunders (1982). 3. In some countries, telecommunications groups within government departments have, in practice, more management autonomy than so-called government corporations in other countries, which on paper appear to be more independent. 4. From a longer-term development point of view, one problem is that government can easily reduce planned investment in the sector when buLdget deficits occur or unexpected national fiscal crises develop. With large monolithic projects such as dams, airports, or power stations, either the facilities are built or they are not; they are never left half done since that would serve no useful function. A telecommunications investment program can, however, be cut piece by piece-fewver subscribers are connected, fewer towns are served, and less long-distance capacity is offered in the short run-without totally eliminating the 80 AN INTRODUCTORY PERSPECTIVE effectiveness of much of the investment already completed. Hence, the sector is more vul- nerable than many to last-minute cuts in its investment prograni brought about by exoge- nous national fiscal problems. 5. Related to this, a study of forty-three countries showed that government organizations with joint responsibility for both posts and telecommunications tend to be the least respon- sive in setting tariffs that reflect costs and market demand. Government entities responsible for only telecommunications tend to be somewhat more responsive, whereas privately owned entities tend to be the most responsive. See Littlechild (1980). 6. The ITL recommends a pyramidal approach to technical training, in which more em- ployees are given preliminary training at national centeTs, fewer well-prepared trainees are sent to subregional centers, and the best of these trainees are sent for specialized training to higher-level regional centers in industrial countries. See ITU (1980b). 7. Ethiopia and Burkina Faso are good examples of countries with very modest educa- tional resources that were able to avert staffing problems for several years by timely aware- ness and action. They also illustrate the long lead times required. For example, in the 1960s, the telecommunications sector in Burkina Faso (then called Upper Volta) depended entirely on foreign staff to fill senior posts. By 1976, the sector was staffed exclusively by nationals. This successful transformation mainly resulted from a decision in 1960 to engage promising school students in long-term commitments by offering them extensive education and training in foreign universities and telecommunications administrations (in Europe and, later, in African regional centers). By 1976, although the OPT in Burkina Faso still had to overcome limitations in its telecommunications, engineering, and management capabili- ties, as well as some political interference, the operating entity had become a viable organi- zation staffed entirely by nationals, with some of its key positions held by individuals who had reached high standards of education and competence. 8. Chapuis (1975); Huntly (1967). Given technological advances and declining costs, the ratio in 1990 was probably less than it was in 1967; in the specific case of AT&T (now long-distance only) in the United States, statistics of the Federal Communications Com- nission show that the ratio of total assets to total operating revenues in 1988 was only 0.64:1.00 but that for local exchanges the ratio wvas 2.14:1.00. 9. For example, in recent telecommunications programs partly financed by the World Bank, the investment cost per telephone line to be added ranged between about $1,200 (for a program mainly confined to a large metropolis and its environs) and about $4,000 (for a country whose national infrastructure had to be built from a very incomplete base). 10. These figures are based on actual and projected growth of telephone lines in coun- tries for which statistics were available in 1990, assuming an average investment of $2,000 per additional line connected. Although the numbers are subject to considerable error, they are probably accurate enough for the limited purposes of this discussion. These are updated calculations of the figures given in World Bank (1989a). 11. Figure 3-1 is based on estimates of total telecommunications investment in develop- ing countries, on the financing plans of telecommunications programs supported by the World Bank in the 1980s, corrected to reflect that many developing countries do not bor- row from the World Bank, and on rough figures for teleconimunications lending by other multilateral agencies. Again, although these figures are subject to considerable error, they are probably accurate enough for purposes of this discussion. See Wellenius (1990b). 12. Limited foreign exchange particularly affects investment in rural areas, where overall revenues are likely to be much lower than in urban areas. See Goldschnmidt (1984). 13. The limited information available is confined to literature surveys of transport and energy substitution, such as the survey presented in chapter 7; to isolated cost-benefit exer- cises, such as several of those outlined in chapters 8 and 9, which show savings in the costs of transport and gasoline in specific cases; and to a few individual cost-benefit exercises, ORGANIZATION, MANAGEMENT, AND FINANCING 81 which outline specific potential for direct savings in foreign exchange (see example 26 in chapter 9). 14. Wellenius (1990b). 15. The situation changed quite rapidly in the late 1980s and early 1990s, as sweeping telecommunications reforms in a growing number of developing countries, especially in Latin America (for example, Chile in 1988, Mexico and Argentina in 1990), attracted for- eign private operators and investors to undertake large expansion and modernization pro- grams. The limits to overcoming telecommunications constraints under the traditional scheme of public sector monopoly, and the trends toward privatization and competition, are outlined in chapter 16 and treated at length in other publications. See Wellenius and others (1989) and Bruce, Cunard, and Director (1988). Part II Macroeconomic Analysis of Benefits Chapter 4 Aggregate Correlation Analysis THE MOST WIDELY CITED EVIDENCE on the benefits of telecommunications investment compares measures of availability and use with various measures of aggregate national economic activity, such as gross do- mestic product (GDP). This chapter and the next review such macro- economic evidence and conclude that, although the evidence could be improved by more rigorous econometric analysis and increased disaggregation, it provides only general and descriptive insights into the question of the benefits and development priorities of telecom- munications. Nevertheless, this aggregate evidence helps set the scene for the more detailed sector analysis and project studies discussed in subsequent chapters. Telecommunications infrastructure may be viewed as an input to a productive process, a "factor of production" like petroleum or electric- ity.' Consequently, most economic empirical work assesses the effect of telecommunications at the macroeconomic or country level in one of two ways. The first is statistical correlation or regression analysis, which usually specifies a macroeconomic country-level model (often with only one equation embodying supposedly causal relations) and then estimates the parameters of that model from data on the provi- sion or use of telecommunications and from one or more indicators of the level of national economic activity. This approach includes both cross-sectional studies, in which the variables are compared for different countries and regions at a single time, and time-series stud- ies, in which the values of variables are traced over time for a single country or region. Several such studies are reviewed in this chapter. The second general method, which can be referred to as structural economic analysis, focuses on the structure of the economy as re- 85 86 MACROECONOMIC ANALYSIS OF BENEFITS vealed by the levels of activity in different sectors (agriculture, manu- facturing, services, and so forth). This approach, which is reviewed in the following chapter, relies primarily on the classic tool of input- output analysis, which generally describes an economy in terms of more or less stable coefficients, relating the outputs of particular sec- tors to their requirements for inputs. One such input is the use of telecommunications services. Correlation and Regression Studies In 1963, Jipp brought to public attention the strong correlation be- tween telephone density (the number of telephones per 100 persons) and what he called the "wealth of nations."2 Since then, single- equation representations of this relation have been formulated many times, using different groups of countries and different periods as well as GDP per capita or related indicators as proxy measures for the wealth of a country. Representative examples are examined below, in- cluding references to the benchmark work carried out under the aus- pices of the International Telecommunication Union's CCITT.3 In retrospect, some of the early cross-country comparisons done by the CCITT may seem naive. It is important to remember, however, that when its work began in 1964, few relevant economic studies existed on which to draw, and almost no qualified economists or econometri- cians were working in the field of telecommunications. The somewhat simplistic first efforts did throw some light on the association between telecommunications and aggregate economic activity and gradually made planners aware of the complexity of the problems at hand. Example 1. cCITT'S Cross-Sectional Analysis In the mid-1960s the ccITT used cross-sectional data from thirty industrial and de- veloping countries to examine the correlation between the density of telephone lines (d) and GDP per capita (g). Figure 4-1 shows the data and model fitted for 1965. A scatter diagram using loga- rithmic scales for both variables showed that most of the data clustered along a line, which can be roughly represented by the equation d = or equivalently log d = log a -4- h log g where a is the intercept, and b is the slope. Figure 4-1. Density of Telephone Lines as a Function of GDP per Capita for a Cross-Section of Countries, 1965 Telephone lines per 100 persons (d) 40~~~~~~~~~~~~~~~~~~ 40 I I I - g - 1 I -SL I- 3\0 _ = X - - 30~~~~~~~~~~~~~~~ 20~~~~~~~~~~~~~~~ 10~~~~~~~~~~~~ 8O-- - X _ - __ __ = 6--- __ ___ ___ 5---- 100_00_00 40 50 1,00 I00 0.8~~~~~ 0.6 - - -- - - - 0.5 - - - -_ _ _ 0.4 - - _ _ _ _ _ _ _ _ _- - - - 0.3 --- 0.2 -- - 0.1 - - - -_ _ 70 100 200 300 400 500 1,000 2,000 GOP per capita dollars; (g) Source: CCITT, CAs-5 Handbook: Economic Studies at the National Level in the Field of Tele- communications (1968). Reproduced with permission. 88 MACROECONOMIC ANALYSIS OF BENEFITS The parameters a and b were estimated separately for 1955, 1960, and 1965 by or- dinary least-squares regression of log d and log g: 1955 log d = -3.0932 + 1.444 log g 1960 log d = -3.1171 + 1.432 log g 1965 log d = -3.1329 + 1.405 log g High correlation coefficients were obtained (0.91 to 0.92), indicating a strong cross- country relation between the variables. The slope b is positive, which indicates that density increases as GDP per capita in- creases. The slope is also larger than 1.0, which indicates that density, d, tends to in- crease faster than GDP per capita, g. To the extent that a cross-section of countries can indicate the way in which telephone density will vary with GDP per capita over time for any particular country, then density can be expected to grow at approximately 1.4 times the growth rate of GlP per capita. For example, if GDP per capita grows at 5 per- cent annually, telephone density will, on average, increase 7 percent.4 The ccmrr recommended in the 1968 GAs-5 handbook that the slope of the fitted lines be used to forecast both demand for and supply of main telephones, in terms of expected growth in the country's GDP and population, and also to forecast initial levels of demand and supply. This recommendation did not explicitly claim that the ob- served cross-country correlations implied a causal link between the provision of tele- phones and the growth of GDP, although it was suggested elsewhere in the 1968 handbook that such relations can indeed indicate the benefits of telecommunications. The 1972 handbook advocated such a method only in combination with other fore- casting methods, and the 1976 version omitted all suggestions of a causal interpreta- tion and the use of the estimated equation as evidence of consequent economic benefits. Example 2. CCITT'S Time-Series Analysis for Individual Countries An alternative to examining a cross-section of many countries at a given time is to examine the density of telephone lines for a single country as its GOP per capita in- creases through time. Figure 4-2 shows an exercise of this type for Sweden, reported by the ccirr in its 1968 GAS-5 handbook. Unfortunately, reliable long-term data were not then, and are not now, available for most developing countries. The Swedish data show two trends, which represent different rates of exponential growth of telephone density in relation to the corresponding growth of GDP per capita. The ccITT interpreted the first trend, starting in about 1900, as reflecting the period in which telephone service was introduced. The second trend, starting between 1915 and 1920, shows the more gradual process of connecting large proportions of the popula- tion to the system. The parameters of the exponential equations were estimated by least-squares regression of log d and log g separately for the two periods, 1900-15 and 1920-65: 1900-15 log d -10.4106 + 3.1935 log g 1920-65 log d = -4.6445 + 1.5476 log g AGGREGATE CORRELATION ANALYSIS 89 Figure 4-2. Density of Telephone Lines as a Function of GDP per Capita in Sweden, 1900-65 Telephone lines per 100 persons (d) 50 40 1965 3 90 _ ___ ____ ____ 30 ~~~~~~~~~~~~~~1960o 1955 20 1950 1945 1940 10 _ _ _ 6 communications (1968). Reproduced with permission. 90 MACROECONOMIC ANALYSIS OF BENEFITS High coefficients of correlation were obtained (r = 0.99), indicating a strong relation between the transformed variables. The coefficient of log g for 1920-65 (1.55) is similar in magnitude to those for the cross-country models discussed in the previous example (1.41 to 1.44) and in this case implied for Sweden that telephone density increased more rapidly than GDP. Per- haps encouraged by this similarity, the ccurT also carried out correlations using mixed time-series and cross-sectional data (1954 to 1968) from both developing and indus- trial countries. This analysis yielded no findings significantly different from those ob- served previously. Example 3. CCITTI'S Utilization Factor and Other Studies The ccrTr tried several other correlations involving GDP, which were thought to hold some promise, but which in fact proved to be of little additional value. One exercise used a so-called telecommunications "utilization factor," defined by the cciTT as the number of telephones (all stations, presumably) per $100,000 of GDP. Examining the utilization factor for fifty-five economies in 1961 and 1971 led to the conclusion that whereas, in general, industrial economies had high utilization factors (10 or more for the United States, Canada, and most of Western Europe), lower factors were found among developing economies with a strong industrial sector (5 to 8 for Brazil, the Re- public of Korea, Mexico, Singapore, and Taiwan). The very lowest utilization factors (typically I to 3) tended to be associated with the poorest developing economies. As in other ccrT exercises, this simply suggests that economies become increasingly in- tensive in telecommunications services as they grow. The cCTr did, however, suggest that the observed disparities in utilization factors can be used as a general argument that developing countries should raise their priorities for investment in the telecom- munications sector. This implies an assumption of causation between telecommunica- tions investment and GDP. The cciTT also attempted to find some simple trend in the proportion of GDP or of gross fixed capital formation that goes into telecommunications investment. No mean- ingful correlations were found, although developing countries appeared to invest lower proportions of both in telecommunications than do industrial nations.5 Example 4. Proxies for GDP and Telephone Density Other exercises based on statistical correlations have been undertaken to improve the representation of economic and telecommunications activities. Gellerman and Ling, for example, observed a strong relation between the number of telephone sub- scribers in Panama's two main cities and the total consumption of electricity, as well as between the ratio of telephone to electricity subscribers and annual electricity con- sumption per subscriber., Another study, based on a cross-section of twenty-nine countries, was also published in 1976 by Bebee and Gilling.7 Since there is evidence that the secondary and tertiary sectors of an economy are the most intensive users of telecommunications, Bebee and Gilling selected the GoP per capita produced by these two sectors as the relevant indicator of a country's level of economic development. They then defined a "development support index" in terms of measures of capital expansion and quality of manpower, both of which they noted Figure 4-3. Telephone Index as a Function of the Development Support Index Telephone index 7 Least-developed countries Region of high priority for telephone network expansion o26 6 -19o C2 3 29 13 0 210 016 5 -7 17 25 0 20 5~~~1 4 10 Region of low priority for telephone network expansion 6(? 0 9 3 / o4 /3 0 2 / 01 Most-developed countries 0 0 4 5 6 7 8 9 10 11 Development support index Note: l O.United Kindom 20. South Africa 1. United States 11. Italy 21. Bulgaria 2. Sweden 12. Israel 22. Trinidad and Tobago 3. Canada 13. Poland 23. Colombia 4. Switzerland 14. Ireland 24. Turkey 5. France 15. Argentina 25. Korea 6. Australia 16. Venezuela 26. Paraguay 7. German Dem. Rep. 17. Spain 27. Malaysia 8. Czechoslovakia 18.Uruguay 28. Malawi 9. Japan 19. Mexico 29. Ethiopia Source: Bebee and Gilling (1976). Reproduced with permission. 91 92 MACROECONOMIC ANALYSIS OF BENEFITS were essential ingredients of development. Under the capital expansion category they included the proportion of CDP used in fixed capital formation and the value of gross fixed capital formation per capita. For the quality of manpower variables they included the literacy rate of the population over fifteen years of age, the median number of years of education of the population (as a proxy for the quality of educated man- power), per capita daily intake of protein and calories (as a proxy for health), per capita expenditure on education, the proportion of population that is urban, and the population growth rate (low rates being considered favorable to development). A "telephone index" was constructed from three measures of telephone availability and use: number of telephones per 100 literate persons over fifteen years of age (as an indicator of the availability of telephones to the population who can most effectively use them), number of business telephones per 100 nonagricultural persons (as a mea- sure of the penetration of telephones in the telecommunications-intensive sectors of the economy), and average annual number of telephone calls per telephone (as a mea- sure of the intensity of use of telecommunications services). Those three indexes were calculated for 1970 for twenty-nine countries for which adequate data had been published. Thirteen of the countries were developing coun- tries. Through multiple linear regression, several analytical models were fitted to linear transformations of the data categories, using level of economic development (GDP pro- duced by the secondary and tertiary sectors) as the dependent variable. The authors gave the following as a typical result: Y = 5.928 - 9.078 (1 / X) - 7.093 (I / X,) where Y = economic development indicator, XI development support index (com- posite variable), and X, = telephone index (composite variable). Following an examination of the related simple and partial correlation coefficients, this equation was interpreted to show that the availability and use of telephones have an important effect on other factors that support development by increasing the out- put of the economy's secondary and tertiary sectors. This, according to the authors, upheld their basic hypothesis that economic development requires a balanced mix of telecommunications and other support factors. To examine the specific mix, least-squares regression was used to correlate the tele- phone index (X2) and the development support index: X2 = 7.8 - 18.0 (1 / X). The result is shown in figure 4-3. Since the curvature is greatest near the origin, that is, for the higher-income industrial nations, the authors concluded that as countries become more developed, telephone availability and use appear to increase faster than other development support factors. This is consistent with what was shown by the ear- lier ccnTT correlations between simple telephone density and GDP per capita and in ccrrT utilization factor exercises. Although Bebee and Gilling make an interesting addition to the literature on the relation between economic development and telephone availability, they go a little too far in arguing that government policies should be directed toward attaining the mix of telephone and other development support described by the curve derived from their model. In their view, countries lying above the curve should give priority to telecom- munications, whereas those below the curve should emphasize other development sup- port factors. They also considered that the derived curve offered a method for ranking countries by priority for telecommunications expansion. It is not at all clear that such prescriptive conclusions can be legitimately drawn from this kind of approach. AGGREGATE CORRELATION ANALYSIS 93 Example 5. Telex Density Telex density has also been compared with several economic variables.8 A cross- section of eighty-seven industrial and developing countries was used to estimate pa- rameters for a model to explain variations in telex density (number of telex lines per 1,000 inhabitants). Seven variables were initially specified to reflect various proxies that could be associated with the use of telex service in a country: GDP, value of ex- ports, value of imports, bank deposits, telephone lines, telephones, and energy con- sumption, all on a per capita basis. Since the independent variables showed significant correlation among themselves, stepwise regression was used to select a simple model with high explanatory power. The result was the following: d = 8.9324 i 10g3 ,705 I )440D_ where d = telex lines per 1,000 persons, g = GDP per capita in 1974 dollars, and i = value of imports per capita, in 1974 dollars. The correlation coefficient was r = 0.92. Hence, as would be expected, telex density, like telephone density, was posi- tively correlated with GDP per capita across countries. In addition, since telex is used extensively in international trade, it is no surprise that telex density was also corre- lated with imports. Example 6. Telecommunications Traffic Correlation analysis has also been used to explain variations in telecommunications traffic across countries and through time. For instance, in the 1968 and 1972 editions of the GAS-5 handbook, the ccirT reported on several traffic studies based on time- series data for individual countries, which essentially followed well-established tech- niques used to forecast traffic in' industrial countries.' The exercises consisted of correlations between telephone calling rates and GDP, GDP per capita, value of imports, value of exports, and other measures of economic activity. Results were mixed. The ccir concluded that, with the data available, local telephone call traffic could not be shown to be influenced by the normal course of economic development, whereas long- distance telephone traffic increased at roughly double the rate of increases in real GDP. Since wide variations occurred among countries and from one year to the next, how- ever, this rate could not be used as a universal benchmark. Not surprisingly, interna- tional telephone traffic through time was most closely associated with changes in the volume of international trade. With regard to cross-sectional exercises, Yatrakis published a study based on an ex- amination of forty-six industrial and developing countries.', He analyzed the correla- tions between international telephone, telex, and telegram traffic and twenty economic indicators specified as dependent variables. The number of tourists per year, GDP per capita, and size of the country's population were found to be the main vari- ables explaining international telephone traffic among countries; all three were posi- tively associated with traffic. Telex traffic was the most closely correlated with tourism and trade. The ccirT, in reviewing this study in the 1976 GAs-5 handbook, concluded that it showed that good and reliable telecommunications are a stimulus to a viable tourist trade, which in turn has a positive effect on the host country's economy. 94 MACROECONOMIC ANALYSIS OF BENEFITS Limitations of Aggregate Analysis As outlined above, since the early 1960s, statistical regression and correlation analysis at the aggregate level has frequently been used to show how telecommunications services are linked to economic devel- opment. The simplicity of the approach and its modest requirements for data made it an attractive tool in the initial stages of economic analysis of the effect of telecommunications in developing countries. However, such analysis is subject to important limitations. Among the most constraining of these are the inadequacy of using a single equa- tion to represent highly complex relations, the heterogeneous nature of the data used, and the inability to attribute causation to any of the observed correlations. With regard to the first point, the relations between telecommu- nications and economic activity are far too complex to be usefully represented by a single-equation model with only one or a few inde- pendent variables. The large aggregation of factors involved in the variables specified in these single-equation exercises by definition ob- scures most of the relations that could be meaningful to decision- makers who are concerned with investment needs in many sectors, and who probably intuitively understand at least some of the inter- dependencies and tradeoffs involved. Second, the data used for both the cross-sectional and time-series correlation exercises are generally not comparable. The highly devel- oped countries at the upper end of the telephone density range tend to have efficient, reliable, and modern telephone service, wide tele- phone coverage throughout the country, and only minimal (if any) waiting lists and waiting times. These countries are compared with those at the lowest end of the telephone density spectrum, which have very limited national coverage, a large unsatisfied demand, lim- ited access to subscriber-dialed, long-distance calls," and major call traffic congestion during business hours on both local and long- distance networks. 12 Also, in the center of the telephone density spectrum there are cer- tain anomalies.'3 The OPEC countries (Organization of Petroleum Ex- porting Countries) tended to have lower telephone densities than would have been expected on the basis of a cross-country correlation between income and telephone density. The reason, of course, is that a telephone infrastructure takes time to develop, and telephone ser- vices in the OPEC countries were generally unable to keep up with the rapid increases in GDP. Also, telephone densities for the former East- ern European countries tended to fall below the aggregate cross- AGGREGATE CORRELATION ANALYSIS 95 country income-density regression lines. In these centrally planned countries, with tightly controlled markets, the planners had evidently concluded that some types of telephone communication (most promi- nently residential) were not necessary or desirable or, alternatively, that in many instances telephones were not necessarily the least ex- pensive way to communicate, as they were in market economies with similar income levels. The third and perhaps most significant problem with the GDP- telephone density correlations is that, even if the data were homoge- neous and the aggregation problems less severe, the movement of two variables through time, or their association across a set of countries, does not imply that changes in one of the variables causes changes in the other; correlation is a test for joint variation, not causation. This was often forgotten in the applications reviewed.'4 For instance, using GDP per capita to predict telephone density (as in example 1) assumes that a change in economic activity results in a change in telephones, which is simple one-way causation. One could also expect, however, that the economy's output to some extent re- flects the fact that telephone service is available as a production input."5 In example 6, the high correlation between telex traffic and tourism was interpreted by the authors as showing that telecommuni- cations stimulate tourism and that this, in turn, has a positive effect on the economy. Although these statements are compatible with the statistical evidence presented (which is a necessary condition for them to be true), they do not necessarily follow from it. Example 4 il- lustrates the temptation to draw unwarranted policy implications. Not only do the authors suggest that their hypothesis is proven by the sta- tistical regression analysis (which it cannot be), they also interpret what in fact is a poor fit of the model to the data as having a norma- tive value: they contend that countries placed above the curve should give priority to telecommunications investment, whereas those below the curve should expand other development inputs. Value of Aggregate Analysis Given these problems, what, if any, is the value of using correla- tions at the aggregate level to indicate the effect of telecommunica- tions in a developing economy? First, such exercises orient the thinking of analysts and planners by providing simple quantitative de- scriptions of existing patterns of telecommunications at various levels of economic development. The findings also hint that the combina- 96 MACROECONOMIC ANALYSIS OF BENEFITS tion of market prices, physical and managerial constraints on invest- ment, and government preferences seems to generate a somewhat consistent-although not necessarily optimal-pattern of evolution across a wide variety of countries.16 This has some illustrative value. Finally, by throwing some light on the complexity of the phenomenon being examined and by providing some limited empirical results with which more elaborate analyses must be compatible, such studies could direct efforts to yield more useful results. There would seem to be some, albeit limited, scope for improving the aggregate statistical modeling approach. Better techniques are available.17 Econometricians often face the task of disentangling sev- eral simultaneous causal relations of varying directions and between different variables-the identification problem.18 Also, a growing body of research on the roles of communication in the family, in the organization, and within regions is beginning to throw some light on relations between telecommunications and society. This might provide a starting point for an analysis of telecommunications in terms of spe- cific models and strategies of development. One outcome of such an analysis could be better criteria for selecting the variables to be in- cluded in the models and the shapes of the analytic functions to be filled. 19 In addition to the analytic limitations of aggregate correlation anal- ysis, other shortcomings must be addressed: the quality of the data needs to be improved, and, equally important, the data need to be an- alyzed on a more disaggregated basis. Nevertheless, the approaches outlined above at the national level are in many instances not suitable for analysis of smaller territorial units, since even the simple data specified become increasingly unreliable-and often are not avail- able-for provinces, regions, and individual cities, towns, and villages. A more manageable way to approach disaggregation might be to ex- amine the association between telecommunications infrastructure and economic activity in various sectors. This is the subject of the follow- ing chapter. Notes 1. This is particularly true in most developing countries in which commercial and indus- trial uses of telecommunications by far outweigh residential uses. 2. Jipp (1963) pp. 199-201. 3. The International Consultative Committee on Telephone and Telegraph (ccrn, from the French name) is one of the ITL'S permanent divisions. From 1964, the ccIiTT included a AGGREGATE CORRELATION ANALYSIS 97 work group on economic studies of telecommunications at the national level, designated (AS-5. See CCITT (1976, 1984a, and 1984b). 4. Let g increase by Ag, and d by Ad. Then, d + Ad = a(g + Ag)' and d = ag Dividing the first line by the second gives I + (Ad / d) = [I + (Ag / g)l' For small values of Ag / g (say, less than 0.1 or I1 percent), this is approximately equal to I + (Ad / d) I + k(Ag / g) For exanmple, if the relative increase of gross domestic product (car), Ag / g, equals 0.05 (or 5 percent), it can be expected to be accompanied by a relative increase in density Ad / d equal to about 1.4 times 0.05, or 0.07 (7 percent). The coefficient b remained fairly con- stant (at about 1.4) during the period examined by the ccirT. 5. That developing countries have tended to invest lower proportions of their C.NP in tele- communications than industrial countries was also discussed in chapter 1. 6. Gellerman and Ling (1976). The authors also noted that the demand for both tele- phone and electricity connections was fairly well met during the period of analysis (1963- 74), and thus connections could be considered generally to reflect denmand. 7. Bebee and Gilling (1976). 8. Wellenius, Budinich, and Moral (1979). 9. See, for example, Taylor (1980) chaps. 3 and 4; Drewer (1973); and Fox (1973). 10. Yatrakis (1972). 11. In several telecommunications projects that the World Bank helped to fund in the 196Qs and 1970s, the introduction of subscriber trunk dialing service almost immediately increased long-distance call traffic between two and six times. 12. Whereas in industrial countries the number of telephone lines is a fairly good nmea- sure of demand, in developing countries it represents only the level of supply-which often falls far short of demand-and therefore is not a good indication of telephone requirements consistent with existing economic activity. 13. The consistencies in deviations of orEC and Eastern European couintries were ob- served by the authors and found to be statistically significant at the 0.05 level in several correlation exercises using data from the mid-1970s for a cross-section of eighty countries. 14. In the United States, there has been a very strong positive correlation between teach- ers' salaries and total national alcohol consunmption. No one has yet, however, argued that a way to reduce national alcohol consumption would be to reduce teachers' salaries. 15. For one of the very few applications reviexved, which mentions that this should give rise to two regression lines, not one, see Shapiro (1976). 16. It is likely that over several years, as aln economy develops and becomes more com- plex, and as time becomes more important in production and distribution processes, some form of telecommunications gradually becomes the least expensive way to communicate. The correlation exercises examined above produce results that are consistent with this hy- pothesis but, of course, do not "prove" it. 17. One intercountry study ran regressions that pooled cross-sectional data into a time series, lagged the regressors by one year, and examined the results in terms of standardized regression coefficients. Unfortunately, the results of the numerous regressions seemed on occasion inconsistent and provided little insight into causality. The overall conclusions of the study-that 'the first role of the telephone is as a channel of information' and "the 98 MACROECONOMIC ANALYSIS OF BENEFITS second role . . . is in allowing a number of alternative communication patterns to occur in business and social activity-did not necessarily follow from the numerous regressions. See Hardy (1980), pp. 1-146. Subsequently, anocher study drew on Hardy's correlations and ac- tially claimed (incorrectly) to be able to estimate the effect on national GDP of telephone installations in regions of low telephone density. See Hudson, Hardy, and Parker (1982). 18. An introduction to the problem and approaches to its solution can be found in Walters (1968) and in the multitude of more recent econometric texts. 19. An interesting study initiated by O'Brien uses principal component analysis. See O'Brien and others (1977). See also Thomas (1984). The latter examined the temporal priority of communications in relation to other factors commonly associated with develop- ment; these included a spectrum of social modernization indicators, such as literacy and school attendance, and were subject to a number of control variables, such as type of colo- nial heritage and political regime. Although this analysis produced some interesting findings, it did not convincingly resolve the issue of causality. More recently, Heymann (1987) used correlations between national products and selected development indicators (established using regression calculations) to determine quantitative relationships between telephone density and other aspects of the national economy but explicitly avoided the en- tire issue of causality. Chapter 5 Structural Economic Analysis AS INPUTS TO PRODUCTION PROCESSES, telecommunications services are not equally important to all productive sectors of an economy. Pri- mary sectors make relatively little use of telecommunications, tertiary sectors are the most intensive users, and secondary sectors lie some- where in between.' Several studies have illustrated this. The cci-rr reported that in 1950 in the former Federal Republic of Germany, where 25 percent of the economically active people were employed in agriculture, the agricultural sector had only 7 percent of the country's telephone lines and accounted for a bare 4 percent of telephone revenues. In contrast, commerce and transport, which em- ployed 16 percent of the work force, accounted for 39 percent of tele- phone lines and 41 percent of revenues. Data from sixty-nine German cities with more than 100,000 inhabitants in 1962 showed that cities predominately engaged in mining or heavy industrial activities had far lower telephone densities than cities of similar population that were mainly active in other sectors of the economy.2 Lonnstrom, Marklund, and Moo found that Sweden's telephone density has varied closely through time with the proportion of the work force in industry, com- merce, transport, and communications.3 Finally, a cross-sectional cor- relation analysis of twenty-two states in Brazil showed telephone density to be much more closely associated with the gross product of the service sector than with either industrial or agricultural output.4 The case of France referred to in chapter 1 also illustrates the rele- vance of questions about the sector's structure. Until the early 1970s, telecommunications development in France lagged significantly be- 99 100 MACROECONOMIC ANALYSIS OF BENEFITS hind that of most other advanced industrial countries, including those with a similar level of per capita income. This had little obvious adverse effect on the aggregate rate of economic growth, which was more than 6 percent a year in real terms for most of the period from 1955 to 1970. It has been suggested, however, that France was able to enjoy rapid economic growth in the 1960s despite a highly deficient telecommunications infrastructure partly because of certain structural features. During that period particularly strong growth occurred in the primary and secondary sectors, with emphasis on agriculture, con- struction, petrochemicals, and steel. At the same time, growth in the tertiary sector clustered in only a few dominant centers, led by Paris. The massive effort first launched in 1974 to overcome the relative backwardness of France's telecommunications system reflected judg- ments made by the government that a structural shift in the growth process would occur in the late 1970s and in the 1980s, with empha- sis on communications-intensive sectors and with increased regional dispersion of economic activity.5 If the use or benefit potential of telecommunications services is likely to be sector specific in developing countries, a further step in examining the relation between telecommunications and economic development would be to disaggregate an economy's structure, to the extent available data permit, and to examine the use of telecommuni- cations services sector by sector. A useful starting point for such an exercise would be to consider how much of the output of the telecom- munications service sector is sold to final consumers, whether house- holds or public agencies, that are not themselves production units, and how much is sold as an intermediate good that contributes to the production of other goods and services in the primary, secondary, and tertiary sectors. Very general information relating to this question can be obtained by examining telephone revenues by class of subscriber. For example, in the 1976 version of its GAS-5 handbook, the CCITT noted that nearly 90 percent of the expenditure on telephone services in developing countries was accounted for by subscribers in industry, banking, transport, and government. The proportion was lower among the more developed of these countries, which had a higher proportion of residential subscribers.6 Similar but more detailed evidence on sector- specific usage is available from a study that examined the 1968-69 input-output transaction table for the Indian economy.7 In that exer- cise, the consumption of postal and telecommunications services was partitioned roughly into three groups: 43 percent of the consumption was by households and nonprofit institutions, 42 percent by interme- STRUCTURAL ECONOMIC ANALYSIS 101 diate business consumers, and 15 percent by government departments and agencies.8 Of the intermediate business consumption of postal and telecommunications services, 95 percent fell in the tertiary sec- tor; of that amount, trade accounted for 68 percent; banking and in- surance, 14 percent; education and research, 9 percent; and transport, 8 percent. A more recent study on Korea in 1980 found that 6 percent of communications output went to exports, 24 percent went to private consumption, and 71 percent went to intermediate consumption. Of the output for intermediate consumption, only I percent went to agriculture, forestry, and fisheries (15 percent of GDP), 23 percent went to mining, power, and industry (33 percent of GDP), 4 percent went to building and public works (8 percent of GDP), and 72 percent went to commerce and services (44 percent of GDP).9 Although this example closely conforms to traditional expectations about how tele- communications input and output are distributed among sectors, re- cent evidence from Italy and Greece does not follow the expected pattern. In Greece, for example, a large survey undertaken by the Greek telecommunications authority, OTE, found that 80 percent of telephone lines were connected to households, which would indicate a mature system with high residential penetration, but that only 7 percent of communications output was consumed by intermediate de- mand, which would indicate a poorly developed system in the areas of business and administration.10 A Cross-Country Comparison of Input-Output Coefficients Suitable data for more elaborate analysis for both developing and industrial countries are limited, primarily because, as in the Indian example, readily available national input-output data rarely separate telecommunications from other communications services (mainly the postal services).33 Nevertheless, input-output tables still provide use- ful insights. For example, in one comparative analysis of input-output tables for several developing and industrial countries, sets of data were collected and grouped so that three general types of indicators could be calculated.12 a. Communications input coefficients, that is, the amount of com- munications services purchased by each sector per unit of output (sales) of that sector 102 MACROECONOMIC ANALYSIS OF BENEFITS b. Communications inputs to each sector as a proportion of total purchases by that sector c. Communications output distribution coefficients, that is, the proportion of total output from the communications sector pur- chased by each other sector. Communications input coefficients and output distribution coeffi- cients were derived from a United Nations input-output data base or- ganized around twenty standardized sector definitions for 1960 or the nearest year available (data set A), and tables of interindustry flows were obtained primarily from World Bank files for seven countries varying from 1961 to 1970 (data set B).13 As noted above, data for postal services and telecommunications are rarely separated; hence the data sets examined contain only a single communications cate- gory. Indeed, even the United Nations interindustry tables did not distinguish between transport and communication; in a few instances, however, a separation was approximated by going back to the original raw data. Despite reservations about the suitability and accuracy of the data available for interindustry input-output comparisons, some significant patterns were discerned as outlined in the following three examples. Example 7. Communications Output Distribution Coefficients The communications industry is one of the few whose product serves as an input to nearly every other industry in the economy. That is, in most input-output tables, few cells in the communications row are empty. Table 5-1 shows communications output distribution coefficients for the countries in data set B. It is apparent from the table that most intermediate communications output is con- sumed by service industries, whereas manufacturing and mining generally follow in second place. Classification of the major communications-intensive industries accord- ing to the nature of their inputs and the destination of their outputs highlights the tendency of these industries to have relatively high value added and to produce goods primarily for final demand markets."4 In all countries examined, the agricultural sector uses relatively little communications; the highest agricultural coefficient was 2.7 per- cent of total communications output in the Philippines for 1961.I1 Example 8. Communications Input Coefficients Many differences in the consumption of communications among countries are partly determined by the different proportions of output that each sector contributes to the total output of the economy. For example, the high share of communications STRUCTURAL ECONOMIC ANALYSIS 103 Table 5-1. Communications Output Distribntion Coefficients, by Sector, for Data Set B (percentage of total sales of telecommunications and postal services purchased by each sector) United Colombia Japan Korea Philippines Taiwan Turkey States Output sector (1970) (1965) (1966) (1961) (1964) (1963) (1967) Agriculture 1.84 0.26 0.33 2.71 0.67 0.22 0.55 Mining and manufacturing 17.78 34.82 18.23 48.47 15.88 11.05 15.33 Services 51.59 47.43 49.88 16.60 71.54 43.92 47.30 Other - 5.09 1.49 - 2.39 - 3.97 Household consumption 28.79 12.40 30.07 32.22 9.52 44.81 32.83 - Not available. Source: World Bank data. output sold to the agricultural sector in Colombia (1.8 percent), relative to that in the United States (0.6 percent) and Japan (0.3 percent), partly reflects agriculture's con- tribution of more than 22 percent of gross output of the economy in Colombia com- pared with 3.9 percent and 6.6 percent, respectively, in the other two countries. Such distortions can be reduced by examining the consumption of communications services per unit of output in other sectors. Table 5-2, for example, shows the average expenditure on communications services by each sector per $1,000 of sales for coun- tries in data set A. A sharp contrast is found in the intensity of the use of communi- cations inputs among primary, secondary, and tertiary sectors. Likewise, large differences are observed in the countries' level of communications services and broad economic development. For a more detailed analysis, table 5-3 shows the expenditure on communications services incurred by aggregate sectors per $1,000 of output for the more disaggregated data available for the countries in data set B. These are the coefficients of direct input Table 5-2. Communications Input Coefficients, by Sector, for Data Set A (dollars of expenditure on communications services per $1,000 of sector output) Mining and Economy Agriculture manufacturing Services Belgium, Finland, France, Fed. Rep. of Germany, Netherlands, United States 1.0 4.4 11.2 Greece, Italy, Japan, Yugoslavia 0.4 3.7 9.5 Korea, Philippines, Turkey 0.3 1.7 5.8 Source: World Bank data. Table 5-3. Direct Input Coefficients for Each Aggregate Sector's Use of Communications Services, for Data Set B (dollars of expenditure on communications services per $1,000 of sector output) United Colombia Japan Korea Philippines Taiwan Turkey States Input sector (1970) (1965) (1966) (1961) (1964) (1963) (1967) Agriculture 0.64 0.34 0.08 0.28 0.16 0.03 2.55 Mining 2.37 8.67 2.79 5.33 1.36 1.89 1.85 Food 2.28 1.26 2.27 2.08 1.72 1.26 4.00 Textiles 4.50 4.06 2.82 0.39 0.95 1.44 2.99 Clothes a 4.95 2.20 2.20 3.65 1.35 6.13 Wood, paper, printing 6.99 7.12 5.84 2.07 1.19 4.15 11.18 Rubber ' 5.11 1.89 2.70 2.18 1.44 5.46 Chemicals 3.14 4.12 4.03 1.37 2.10 1.41 4.63 Petroleum and products 0.67 3.22 1.44 0.30 1.28 0.83 1.58 Cement and minerals 4.28 4.56 3.25 0.75 1.95 - 5.87 o~ Metals 5.20 3.81 1.99 0.43 1.11 2.13 3.43 4. Transport equipment 1.75 4.08 2.22 0.74 3.00 1.95 5.29 Machinery 7.05 6.75 3.36 0.89 2.82 2.08 7.43 Utilities 3.96 4.52 1.59 0.60 1.25 - 5.82 Construction 1.75 6.69 1.17 0.56 2.26 - 3.59 Trade 7.48 15.07 15.84 -- 5.32 15.31 Transport 8.35 9.11 3.60 0.05 4.89 2.13 11.26 Communicationb 99.13 3.69 0.27 - 0.67 37.53 16.24 Services 9.60 12.11 8.60 0.87 9.81 5.10 25.46 Other - 29.27 4.80 1 0.28 - - Households 5.26 4.14 4.05 1.08 0.09 3.93 18.46 - Not available. a. Included in textiles. b. This row has little significance, since the entries vary according to the P&T administrations' practices in accounting for their own use of communica- tions services. These tend to be arbitrary or nonexistent. c. This result is highly anomalous and is not thought to be reliable. Source: World Bank data. STRUCTURAL ECONOMIC ANALYSIS 105 for each sector's use of communications services. The patterns that can be observed in table 5-3 roughly correspond with those in table 5-2. In addition to the direct purchases of communications services, each sector of an economy uses other inputs, whose production also requires the use of communica- tions. Considering all relations of this kind produces coefficients of direct and indirect input for each aggregate sector's use of communications services. Such coefficients are shown in table 5-4. As can be seen, among developing countries agriculture directly uses relatively few communications inputs. Even though the agricultural sector may contribute from 20 to 30 percent of the gross output in many developing economies, the very small size of the communications input coefficients for this sector reflects a low total expenditure on communications by the agricultural sector." This was apparent from table 5-1. This is in stark contrast to the highly productive agricultural sector in the United States, where the direct use of communications is generally from ten to twenty times as high as it is in all but one of the developing countries considered (the exception being Colombia, which in 1970 used about 25 percent as much communications per unit of agricultural output as the United States). A comparison between the direct inputs on the one hand and the direct and indi- rect inputs on the other shows a greater overall dependence of agriculture on the com- munications infrastructure than is apparent from the direct coefficient alone. This effect, however, is less apparent in the least developed of the countries considered, pre- sumably because they use relatively few inputs, such as fertilizers, machinery, and tech- nical assistance, and have relatively undeveloped markets and transport and telecommunications infrastructure.'7 In connection with nonagricultural activities, the use of communications inputs seems to follow a pattern related to overall develop- ment of the economy.'t For manufacturing sectors such as chemicals and machinery, the values for commu- nications input coefficients are fairly widely dispersed among countries. Coefficients are lower for the developing countries, but there appears to be some lower limit for each sector, so that extremely low values (less than $1 per $1,000 of sector output) are rare. On average, the coefficients for the developing countries are about one-third to one-half the size of the coefficients for industrial countries. Also, the indirect com- munications requirement is important throughout. In Japan, for example (1965 data), the metals industry directly consumed V 3.8 of communications services per V 1,000 of output. It also consumed V 26.6 of utilities per V 1,000 of output. The utilities sec- tor, in turn, consumed Y 0.12 of communications to produce V 26.6 of output. Sum- ming all such indirect demands for communications by the industries supplying inputs to metals gives a total indirect communications requirement of V 14.6 per v 1,000 of metals output. This is, of course, far larger than the direct requirement alone. For nonagricultural, nonmanufacturing sectors such as services or trade, the differ- ences between coefficients for industrial and developing countries are somewhat less pronounced; few direct and indirect coefficients are below $5 per $1,000 of output. For trade, few coefficients for developing countries are less than half the size for indus- trial countries. Inspection of table 5-4 suggests that the structure of the coefficients-that is, their relative size for the various sectors-is similar in all countries examined. This, in turn, may imply that differences among countries result from factors of more general rele- vance (management and administrative practices, substitution of low-cost travel time for telecommunications, scarcity or low quality of telecommunications services), rather than from factors specific to individual sectors. Table 5-4. Direct and Indirect Input Coefficients for Each Aggregate Sector's Use of Communications Services, for Data Set B (dollars of expenditure on communications services per $1,000 of sector output) United Colombia Japan Korea Philippines Taiwan Turkey States Input sector (1970) (1965) (1966) (1961) (1964) (1963) (1967) Agriculture 1.9 4.3 1.1 0.7 1.7 0.4 14.6 Mining 5.4 14.6 4.8 5.6 3.9 3.1 11.2 Food 5.2 7.7 4.9 2.4 4.7 2.3 16.5 Textiles 10.9 13.6 6.7 0.8 5.1 2.9 14.4 Clothes 16.2 7.7 3.0 8.6 3.1 17.3 Wood, paper, printing 14.9 16.5 10.6 2.7 5.4 6.7 24.0 Rubber 14.7 6.2 3.1 6.9 3.2 15.8 Chemicals 10.5 14.1 2.0 2.2 6.4 3.4 17.3 Petroleum and products 4.2 11.6 3.9 0.8 4.1 1.4 12.4 Cement and minerals 9.1 13.3 7.1 1.5 5.7 1.5 15.5 Metals 13.8 18.4 8.2 0.8 6.6 4.2 13.7 ON Transport equipinent 8.5 14.8 8.0 1.1 7.3 4.4 16.4 Machinery 15.8 17.2 8.4 1.3 8.4 3.9 17.7 Utilities 7.7 12.0 3.5 0.9 3.4 1.6 14.8 Construction 7.5 17.2 6.0 1.0 6.7 1.8 15.0 Trade 12.5 18.2 16.8 - - 6.2 22.9 Transport 12.7 14.3 6.2 0.2 8.1 3.6 17.1 Communication" - - - - - - Services 12.8 15.7 10.6 0.9 11.7 5.7 34.9 Other - 38.4 7.6 5.3 - 17.0 - Not available. a. Included in textiles. b. This row has little significance, since the entries vary according to the P&T administrations' practices in accounting for their own use of communica- tions services. These tend to be arbitrary or nonexistent. c. This result is highly anomalous and is not thought to be reliable. Note: Indirect communications requirements were not calculated for the households sector. Source: World Bank data. STRUCTURAL ECONOMIC ANALYSIS 107 Example 9. Backward Linkages Direct and indirect input coefficients for the communications sector itself indicate the extent to which growth in that sector will stimulate growth in other sectors through the demand by communications enterprises for the outputs of other sectors. Table 5-5 shows production functions of the communications sector as they appear in the input-output tables for data set B. In terms of dollar input per dollar of communi- cations output, the most important inputs to communications are generally house- holds (labor), services, and transport. Machinery, trade, utilities, and construction are also important inputs in some countries. This concentration of inputs in the labor- intensive service sector and the high labor requirement (primarily for postal services) indicate that demand for an extra unit of communications output will feed back pri- marily into household income, with little direct effect on the major processing sectors of the economy. Also of interest is the total increase in household income that results from a one- unit expansion of the demand for an industry's output. This includes direct income (additional labor used as input to produce the additional output), indirect income (in- come generated in the supplying industries), and induced income (income generated through direct and indirect income spent by households). The ratio between these es- timates of total income generated and their direct income components gives industry income multipliers, which can be normalized by dividing each by the average income multiplier for each country. Table 5-6 shows that the normalized income multipliers are generally lower for the communications sector than for all industries in the countries examined.'" This means that if the growth of a sector is being constrained by demand, the expansion of final demand for the products of most other sectors gives rise to greater indirect income- generating effects than does the expansion of demand for communications. A general complication in interpreting such results in developing countries is that the output, not only of the communications sector but also of many others, is mainly constrained not by demand, but by insufficient supply. Information Sector Analysis Input-output tables, or some of the raw data that were originally collected to construct such tables and national income accounts, have also been subject to a different form of analysis and aggregation. Such work has addressed the concerns of some researchers that much exist- ing economic literature does not adequately consider the information- related function of administration and that therefore much of the traditional economic theory of the firm is perhaps too far removed from reality.20 The objective of this information sector analysis has been to identify the extent or the importance of directly productive activities that mainly involve handling information (as distinct from goods and materials) and processes of controlling, coordinating, mon- itoring, recording, or more generally organizing directly productive ac- tivities. These two groups of activities are referred to as the primary Table 5-5. Purchase of Inputs in Dollars per $1,000 of Output from the Communications Sector, for Data Set B United Colombia Japan Korea Philippines Taiwan Turkey States Input sector (1970) (1965) (1966) (1961) (1964) (1963) (1967) Agriculmtre - - 0.04 - - - - Mining - 0.12 0.14 - 4.42 0.03 Food - - - - - Textiles 1.87 0.50 0.42 - - - 0.08 Clothes - 2.55 0.50 - 35.17 - 0.75 Wood, paper, printing 3.74 12.33 30.13 - - 8.83 3.37 Rubber - 0.11 0.37 - - - 0.11 Chemicals 1.25 - 2.76 - 6.98 2.21 0.23 Petroleum 6.23 2.53 7.33 2.55 1.77 - 3.74 Cement and minerals - 0.01 16.42 - - 4.42 - Metals - - 3.30 - - 26.49 0.37 Transport equipment 1.87 2.51 1.64 18.75 3.63 4.42 5.52 Machinery 19.33 4.49 25.36 - 20.55 35.32 10.78 Utilities 9.98 7.99 5.83 13.60 4.19 4.42 7.90 Construction 2.49 5.99 3.80 13.10 8.20 - 21.30 Trade - 2.13 18.62 - - 11.04 8.10 Transport 31.17 29.16 18.97 - 38.74 24.28 29.40 Communications" 99.13 3.69 0.27 - 0.61 37.53 16.24 Services 53.62 12.14 9.90 206.07 28.06 15.45 106.81 Other - 4.00 4.82 - 38.13 - - Households 521.20 475.65 348.28 537.55 642.34 737.31' 775.92' - Not available. a. The figures in this row represent communications input to the communications industry itself. They have little meaning since their size depends on the accounting practices used by the communications administrations of each country. b. These household input figures include value added that is not household labor and so are overstated. Source: World Bank data. Table 5-6. Normalized Income Multipliers, for Data Set B United Colombia Japan Korea Philippines Taiwan Turkey States Input sector (1970) (1965) (1966) (1961) (1964) (1963) (1967) Agriculture 0.62 0.81 0.78 0.66 0.47 0.81 1.13 Mining 0.66 0.51 0.60 0.70 0.48 0.68 0.72 Food 1.72 1.35 1.09 1.46 1.37 2.31 1.44 Textiles 1.05 1.09 1.11 1.14 1.53 1.17 1.38 Clothes - 1.09 1.12 0.96 1.19 1.18 1.17 Wood, paper, printing 1.02 0.91 0.94 1.01 2.21 1.30 0.98 Rubber - 0.89 0.99 1.09 1.14 0.98 0.93 Chemicals 1.28 1.31 1.01 1.31 1.07 1.16 1.19 Petroleum 1.77 3.02 2.39 2.97 0.80 0.70 1.62 o Cement and minerals 0.82 0.78 1.05 0.98 0.85 0.82 0.88 Metals 1.07 2.18 1.65 1.06 2.46 1.02 1.16 Transport equipment 1.77 0.94 1.08 0.82 0.65 1.59 1.11 Machinery 1.14 0.81 0.98 0.94 1.08 1.00 0.95 Utilities 0.60 0.66 0.80 0.71 0.63 0.89 0.84 Construction 0.68 0.82 0.84 0.79 0.71 0.89 0.98 Trade 1.05 0.50 0.65 0.57 - 0.57 0.60 Transport 0.66 0.50 0.69 0.66 0.57 0.78 0.69 Communications 0.54 0.39 0.58 0.59 0.39 0.55 0.55 Services 0.55 0.44 0.65 0.59 0.40 0.55 0.69 Other - - - - - - - Average income multipliers 3.94 4.48 2.45 2.17 16.60 20.61 74.19 - Not available. Source: World Bank data. 1 10 MACROECONOMIC ANALYSIS OF BENEFITS and secondary information sectors, respectively. Thus defined, the in- formation sector includes many productive activities that are tradi- tionally counted under the tertiary economic sectors (for example, services and government) as well as organizational activities that are lumped with the organization's main outputs in the primary (agricul- ture, mining) and secondary (manufacturing, construction, transpor- tation) sectors. Empirical work relating to this objective has grown out of a litera- ture, starting in the late 1950s, on the role of information at both the market level and within individual economic units. The former is generally known as "economics of information,"21 whereas the latter falls within the fields of "decision theory" and "modern organization theory."22 A third body of related literature has also gained some prominence since the mid-1960s. It focuses on the broader role of all communications in national development and uses concepts from sev- eral disciplines, traditionally sociology, but increasingly also anthro- pology, psychology, political science, and geography. Until relatively recently such "communications" research focused almost entirely on mass media, although in recent years two-way communication has gained some attention.23 Modern organization theory primarily relates to the need to recon- cile the somewhat diverse objectives of individuals and special interest groups within organizations. In so doing it emphasizes the processes by which expectations are formed and decisions are made under con- ditions of uncertainty. The economics of information, however, is con- cerned at a more aggregate level with tvpes of information-related market failures and analyzes supply and demand decisions made under conditions of uncertainty. Both explicitly recognize that incom- plete information and uncertainty are prime factors facing individual decisionmakers. Hence, the greater the extent to which communica- tion can be used to reduce uncertainty, and thus increase the proba- bility of a correct decision, the better the chance organizations, economic units, or individuals within those units will have of achiev- ing their goals.24 The implications for the telecommunications needs of developing countries are not trivial. Communication.emerges as a crucial factor in the performance of both individual economic units and of markets. Also, communication is necessary for anv division of labor to occur, since the process of development implies specialization and greater in- terdependence. In a developing country, a subsistence farmer does not have to interact significantly with persons outside his own family. If, however, his output is to be raised above the subsistence level, com- STRUCTURAL ECONOMIC ANALYSIS 111 munication must take place to facilitate the division of labor-the specialization by function-that increases productivity.25 For a change to occur, communication, no matter how slight, is necessary for the initial coordination, reorganization, and perhaps technical innovation. In a more urban setting it has been documented that some firms and individuals use more up-to-date techniques and are more ready to innovate than are others. The innovators have, among other things, knowledge about techniques, procedures, and opportunities that the others lack. Some literature in the communications research field ex- amines the question of diffusion of innovations.26 Early neoclassical economists pointed out the innovative role that entrepreneurs play in economic development, but they were never particularly interested in the process through which such innovative entrepreneurs arose. Com- munications literature, however, examines the place of communica- tion in the process of technology transfers and investigates the characteristics of innovators and successful allocators, who in fact tend to be early adopters of new procedures and technologies.27 The presence or absence of a communications infrastructure is an impor- tant factor in their findings. In the broader context of a national economy-whether free mar- ket or centrally planned-information is required to coordinate eco- nomic activities. In a free market, producers and consumers attempt in their own self-interest to improve their decisions by reducing uncer- tainty, while in a centrally directed economy, where the division of labor between producers and decisionmakers is more marked, a rapid feedback of information can reduce the uncertainty associated with the myriad variables facing each group.28 Following from this, several information theory analysts have argued that, from a communications perspective, an economic system consists primarily of two functions: first, the technical maximization of production (given resources and resource costs) and the technical minimization of the costs of distrib- uting goods and services and, second, the exchange of information as- sociated with organizing and coordinating production and distribu- tion.29 They note that the former is associated with production and distribution costs that depend on resource costs and technology, whereas the latter encompasses transactions costs that depend partly on the mode of resource allocation and on the communications infrastructure. Empirical work based on a reaggregation of data previously col- lected for more traditional input-output or national income analysis has in fact shown that the organization and coordination of economic activity consume a substantial and increasing proportion of the re- 1 1 2 MACROECONOMIC ANALYSIS OF BENEFITS sources of industrial countries. For example, for the United States, Porat estimated that about 46 percent of 1967 GNP was associated with the information-related processes of producing, processing, and distributing information goods and services. For 1970 he found that information-related activities accounted for more than 40 percent of the labor force and 53 percent of all labor income.30 Subsequently, Jonscher attempted to estimate the proportion of U.S. resources de- voted specifically to "organizing the economy-to directing, co- ordinating, monitoring, and recording economic activity."31 Using a reaggregated version of Porat's original data, he found that roughly 43 percent of labor income in 1967 was associated with an organizing ac- tivity, with the remaining 57 percent being associated with doing the things that were being organized. Similar empirical analyses of the information sector have been un- dertaken in other industrial countries. Halina, using Porat's classifica- tions and definitions, estimated that 47 percent of Canada's GNP is associated with information sector activities.32 Exercises undertaken in Australia, the United Kingdom, and the former Federal Republic of Germany indicated the same broad pattern for the importance of the information sector labor force; the percentage of population employed in the information sector was estimated as 28 percent in Australia (1971), 37 percent in the United Kingdom (1971), and 31 percent in Germany (1970).33 For Japan, Uno has shown that "knowledge work- ers" encompassed 12 percent of the labor force in 1975 and have been increasing since 1960 at an annual rate of almost 5 percent, while other types of workers have been increasing at only 1 percent annually.34 It is therefore evident that the information sector or information- related portions of industrial economies are large and, at least until the early 1970s, were growing relatively rapidly. In recent years, growth in the information sector appears to have leveled off. In both the United States and Japan, the proportion of total GDP comprised by the information sector has remained at about 45 and 35 percent, re- spectively, since the early 1970s, although the proportion of "knowl- edge workers" in the total work force continued to grow into the 1970s.35 Information activities in the primary and tertiary sectors have also continued to grow in industrial countries, as shown by data from a 1980 study on what was then the Federal Republic of Ger- many and from studies on rural regions of four European countries undertaken by the Special Telecommunications Action for Regional Development (STAR) program.36 STRUCTURAL ECONOMIC ANALYSIS 113 During the 1980s, similar studies were undertaken in a growing number of developing and newly industrialized countries. Studies by Jussawalla and others in the Pacific Rim area have focused on back- ward and forward multipliers and linkages between the information sector (limited to the primary information sector because nontraded, internal information on organizational activities is difficult to include in the data) and all other sectors of a country's economy.37 The Pacific Rim is a natural focus of interest for this kind of re- search. It is an area where considerable investment in information- intensive industry, infrastructure, and services has occurred in developing and newly industrialized countries during the past ten to fifteen years and where the information sector is expected to provide the best opportunity for strengthening economic growth. In these studies, broad definitions of the information sector and subsectors were applied to each country's standard classification of the industry, and national interindustry accounts were then recast to highlight the information sector as separate from the traditional divisions of pri- mary, secondary, and tertiary sectors. On the basis of the recast input-output tables, calculations were made to determine the informa- tion sector's value added, its total output, and the coefficients that measure its relation to other sectors. In most of the countries studied, the primary information sector was found to have relatively high backward-linkage effects compared with other sectors in terms of derived demand for other goods and services produced in the economy and in terms of lowered input costs for other sectors of the economy."' However, in the four largest countries studied-Indonesia, Malay- sia, the Philippines, and Thailand-the information sector was found to be a smaller part of GDP than all other sectors, despite the recent, strong growth of their information infrastructure. For Singapore and Taiwan, where growth of the information sector is widely considered to have been spectacular, the information sector is still smaller than either the secondary or the tertiary sectors, although larger than the primary sector. Furthermore, in the case of Singapore, the informa- tion sector's share of GDP remained roughly constant at about 25 per- cent between 1973 and 1978, and labor productivity appeared to be no higher than average for the economy. Overall, much of what is expected about the information sector is neither supported nor refuted by the data, and some key hypotheses still need to be tested. In particular, the presumed desirability of shift- ing the economic structure from agriculture and industry to informa- 11 4 MACROECONOMIC ANALYSIS OF BENEFITS tion in the developing world should be examined in terms of broader macroeconomic considerations, and the assumption that the informa- tion sector poses exceptional opportunities for economic development must be substantiated.39 Where does telecommunications infrastructure fit into this picture? In a macroeconomic sense, the large unsatisfied demand for telecom- munications services in developing countries and the high returns on new investment are evidence that the perceived communications needs of both producers and consumers are not being met. Not meet- ing this demand may worsen the unequal distribution of information between the parties and diminish the opportunity for transferring information.40 The extent to which this retards increases in productiv- ity by slowing functional specialization and the division of labor and the extent to which communication facilities are inadequate for the efficient organization and coordination of economic activity are not explicitly known. In the context of increasing globalization and information intensity of economic activities, however, a plausible hypothesis is that al- though the gap between the telecommunications infrastructure of in- dustrial and developing countries is narrowing (see chapter 1), a veritable information gap is building. It is still not clear how impor- tant this gap might be nor how policymakers could put in practice the broader concept of informatics. Value of Input-Output Analysis Input-output studies convey a general picture of the role of com- munications services in the overall structure of an economy. Of greater interest, however, is the contribution that such analysis can make to assessing or quantifying the benefits of telecommunications investment and to planning the scale and nature of services that should be provided. Assessment of benefits depends on the functional relation between the observed communications inputs used by enterprises or organiza- tions and their economic outputs, that is, on the nature of the pro- duction function. If the observed input coefficients varied little across countries, this could indicate a stable production function with fixed coefficients. In other words, one could conclude that managerial and technological considerations make it essential to use a fixed amount of telecommunications services and that if this amount is not used, the output of other sectors will drop. From the limited data available, STRUCTURAL ECONOMIC ANALYSIS 1 15 however, this does not appear to be the case. Nevertheless, at least in the trade and services sectors, there is less difference among countries than in the other sectors examined, which suggests that a finer break- down of sectors might identify some activities (banking, foreign trade, or tourism) where a fixed-coefficients model would, in fact, fit. For sectors other than services and trade, the observed results can be interpreted in several ways. One possibility is that the variation in coefficients reflects some wasteful use of communications, particularly in higher-income countries. Another more likely possibility is that the quality of communications output varies greatly among countries; sometimes supply is less timely or the quality is relatively poor.4' Fi- nally, there may be systematic differences in sector organization, man- agement, technology, work organization, benefits, and so forth between higher- and lower-income countries, including some substitu- tion of other factors of production, such as increased use of labor or transport, for the intensive use of communications services.42 Whatever forces determine their size, input coefficients for commu- nications might nevertheless indicate the minimum levels of demand from different portions of the economy that must be met if the exist- ing allocation of resources is to remain unchanged. Consequently, input-output analysis might help to forecast demand, not so much in cases where the economic structure is changing slowly (where econo- metric methods might be more appropriate), but in cases where rela- tively large structural changes are expected. Examples that are not uncommon in developing countries would be the initial development of a large-scale tourism industry, a major expansion in mineral extrac- tion, the creation of a new manufacturing sector, or the transforma- tion of agriculture from subsistence to commercial forms. At least one attempt has been made to estimate input-output de- mand for the United States. Using the 1963 U.S. input-output table and projections of future gross outputs by each sector, Bower esti- mated aggregate investment requirements for telecommunications.43 The estimated gross outputs of each sector were multiplied by tele- communications input coefficients to arrive at the total industrial demand for telecommunications in a specified year. The final de- mand for telecommunications was estimated by assuming it to be, as a proportion of total demand, a function of the relative share of resi- dential telephone connections in total connections. Once estimated, total demand was converted to a capital investment requirement by using an estimated capital-output ratio for the telecommunications sector.44 The results observed above in which input coefficients were shown 1 16 MACROECONOMIC ANALYSIS OF BENEFITS to vary significantly across countries do, of course, imply that any such forecasting method would have to be based on local input- output data for each country. Such a method cannot be applied in the many cases where such data are not available nor in the many de- veloping countries where levels of unmet telecommunications demand are high and where current investment in the sector is clearly sub- optimal. Of course, it is not known if Bower's estimates are signifi- cantly more reliable than those obtained using other, much simpler estimating procedures. Using input-output analysis to allocate telecommunications in- vestment implies that the cost of not meeting the communications requirement indicated by the input coefficients is likely to be signifi- cant. In terms of economic theory, this assumes that the marginal rev- enue product of telecommunications services as an input to the productive process in each of the other sectors is significant and that telecommunications services are not readily substituted by other fac- tors of production; the elasticity of substitution is low. Although these appear to be reasonable assumptions, their validity can be settled only at a more microeconomic level of analysis. Problems with Input-Output Analysis Other difficulties arise in using input-output techniques to analyze the relation between telecommunications and other economic activity. First, the usual industrial classification schemes used to compile input-output tables are inadequate for examining relations between telecommunications and other sectors. As most input-output tables are used primarily to study the flow of material goods, the highest level of detail is shown in the agricultural and manufacturing sectors. Postal services and telecommunications are rarely separated; this is a serious deficiency since their input requirements (especially capital and labor) are very different, as are the nature of the demands for their output.45 Also, the service sector is often treated as a single group. Since this sector consumes 50 percent of all communications services in most countries, while manufacturing consumes about 20 percent and agriculture less than I percent, a more comprehensive analysis of communications use would require a more detailed classifi- cation within the service sector. A second problem is a deficiency inherent in all cross-country input-output comparisons. Since prices across countries vary widely and in some instances bear little relation to costs, conventional input- STRUCTURAL ECONOMIC ANALYSIS 117 output data expressed in terms of value of transactions may not accu- rately reflect the actual level of activity expressed, for example, in millions of telephone calls or tons of steel. A third problem also inherent in all cross-country input-output comparisons is that the original raw data are subject to inaccuracies. Input-output tables are only as accurate as the data on which they are based, and this accuracy varies widely, particularly among developing countries. Analysis of a sector such as communications, which in most countries is relatively small, is especially liable to statistical errors. A fourth problem is that data on the size of the information sector do not reveal much qualitative information about the sector; for in- stance, they do not address the extent to which a particular sector is information-intensive. The numbers can thus be interpreted many ways, leading to different conclusions and policy prescriptions. If an information sector were found to exhibit larger-than-average econo- mies of scale, a high proportion of GDP in this sector could signal a small economy, rather than an advanced one, and the growth of ser- vices (which are partly counted in the information sector) could be either a positive sign of economic modernization or a negative sign of a bloated public sector, which is a major target of efforts to reform economic policy in many developing countries.46 A fifth problem, which in many ways is the most limiting, would impose a major roadblock on the usefulness of cross-country or through-time input-output analysis for telecommunications, even with accurate data, comparable price levels, and finely disaggregated activi- ties. As has been noted frequently throughout this book, in develop- ing countries the volume of purchases of telecommunications services is limited in most cases by inadequate supplies. An equilibrium be- tween demand and supply cannot be assumed; developing countries typically experience acute and persistent shortages in the supply of telecommunications services, and the quality of service is often poor. Hence, existing expenditures that various sectors or economic activi- ties make on telecommunications may not have much normative significance. Notes 1. Examples of primary sectors are agriculture, fishing, forestry, and mining and quarry- ing. Secondary sectors include construction and manufacturing. Tertiary sectors include commerce (wholesale and retail), communication (posts, telecommunications, press, and 11 8 MACROECONOMIC ANALYSIS OF BENEFITS broadcasting), finances (banking, insurance, and real estate), public administration, trans- port (including storage), utilities (gas, electricity, water, and sewerage), and other services (education, health, personal, and professional). 2. cciT-n (1965) pp. 32-33. 3. Lonnstrom, Marklund, and Moo (1975). 4. This study used TELEPRJAS data for all of Brazil covering the years 1975 to 1980. More recently, this work was enlarged by a follow-up study, which found that the impact of tele- communications services on economic growth varies according to the development stage of the region involved. Telecommunications were compared with two other basic elements of infrastructure, energy and transportation, and with regional levels of revenue to determine the role that each played as economic growth occurred. The study found that the demand for telecommunications, as measured by density of installed telephone lines, was relatively strong in the early stages of developnment. As industrialization began, investment in energy was given higher priority than investment in telecommunications, but when industrializa- tion reached a certain threshold, the increase in commercial activity stimulated a second surge in demand for telecommunications services. See Guiscard Ferraz (1987). A similar re- lation between growth in telecommunications infrastructure and economic growth was found in Korea, as measured by the ratio of elasticity of main lines to GDP over time. See Gille (1986). 5. Nora and Minc (1980). 6. To the extent that the data are based on the distinction that telecommunications companies make between business and residential lines, the results may be somewhat in error. Particularly in a developing country, the borderline between home and workplace is sometimes unclear, and in some instances a tariff provides an incentive to classify tele- phones as residential. See chapter 10 for further discussion of this point. 7. Kaul (1979). 8. The proportion of postal expenditures to total postal and telecommunications expen- ditures was 30 percent in 1976-77. 9. Gille (1986). 10. Pye and Lauder (1987). 11. Input-output is used here to denote national data detailing transactions among in- dustrial sectors (inputs and outputs) and final consumption purchases made by households and government according to the linear input-output models developed by Leontief and others. For an introductory description of input-output analysis and some of its uses, see Miernyk (1965). For a specific application oi input-output analysis to the information sec- tor, see Karunaratne (1988). 12. World Bank staff member Colin Warren carried out much of the initial work on the input-output analysis discussed in the following pages. To build up a comparable data set for a larger number of countries, relatively old data had to be used. It would be possible to argue that dependence on communications services has increased since the time of the studies examined. 13. United Nations (1969). 14. Value added refers to the difference between the value of output of the industry and the cost of inptits not produced in the industry. 15. As noted in chapter 1, the small towns and rural areas of most developing countries have a large amount of unsatisfied demand for telephone access. Hence, observed usage by the agricultural sector could be considerably less than an efficiency-related optimum. 16. Of course, a small farmer making a call to the local market in a developing country is probably not counted as part of the agricultural sector. Also, the low coefficients reflect the lack of access to telephones in rural areas of developing countries. However, as noted in chapter 11, a survey in Kenya showed that although the agricultural sector may not pur- chase extensive quantities of telecommunications services, the variety of conmmunications STRUCTURAL ECONOMIC ANALYSIS 119 contacts that it does maintain is extensive and covers a relatively large number of other pro- ductive sectors. 17. It is important to bear in mind the detail that may be lost in statistical aggregation. Whereas in developing countries telecommunications inputs are small for agriculture in general, they may be large for certain activities important for general economic strategy (such as plantation agriculture for supplying export markets or food to nearby cities). 18. Some of the data from the Philippines are thought to be unreliable. 19. Except for trade in the Philippines and trade and services in Turkey. 20. Chandler (1977), p. 490. 21. One of the first papers written was Stigler (1961). Among the more recent are Jonscher (1980; Jonscher (1984); and Lamberton (1990/91). 22. See, for example, Simon (1959); Cyert and March (1963); and March and Simon (1958). 23. One of the better-known pieces of this general communications literature is Schranim (1964). For other examples, see McAnany (1980) and Hudson (1984). 24. Arrow (1980) has argued that because information costs are generally independent of the scale upon which information is used, organizations that rely relatively more heavily on information can derive economic benefits disproportionately large in relation to their size. 25. For example, in East Africa, although the smaliholder agricultural sector traditionally has not had a high potential use for telephones, the increasing production of cash crops is steadily changing this. Okundi, Ogwayo, and Kibombo (1977) p. 51. 26. Ashby and others (1980) 27. Contreras (1980), chap. 5. 28. Arrow has characterized market failure as "the particular case where transactions costs are so high that the existence of the market is no longer worthwhile." See Arrow (1970) p. 17. 29. Porat (1977) stated the sanie concept as functions involving the transformation of matter and energy from one form into another and the transformation of information from one pattern to another. Jonscher (1980) subsequently extended the analysis by explicitly specifying a model that includes both a production process and an inforniation process. 30. Porat (1977). 31. Jonscher (1980) p. A.4. 32. Halina (1980). 33. Lamberton (1982), pp. 36-59. The U.K. and German data were cited from, respec- tively, Wall (1977) and Lange and Rempp (1977). 34. Uno defines knowledge workers as persons who produce and distribute knowledge. Empirically he counted them as natural science researchers, engineers, judicial workers, ac- countants, educators, physicians, dentists, pharmacists, writers, reporters, editors, fine art- ists, designers, photographers, government officials, managers, and clerical workers with college degrees. See Uno (1982) pp. 144-58. In the United States, Machlup (1962) first attempted to measure the share of the U.S. ONP connected with "knowledge" as opposed to other kinds of activities. 35. Wellenius (1988). 36. Deutches Institut fur Wirtschaftforschung (1984) and Pye and Lauder (1987). 37. Jussawalla, Lamberton, and Karunaratne (1988). 38. Jussawalla (1986). 39. This section is based on material published in Wellenius (1988). 40. Spence (1975) has suggested that the organizational problem is to allocate resources under conditions of "imperfect and asymmetrically located information." 41. The wide discrepancies among countries in the quality and timeliness of telecommu- nications output were discussed briefly in chapter 4. 1 20 MACROECONOMIC ANALYSIS OF BENEFITS 42. In the United States, for example, the input coefficient data show that the radio and television broadcasting industry spends $4 on telecommunications services for every $100 of gross output-a result of its heavy use of wideband transmission channels. This would be a poor basis for planning in a country where television broadcasting is not widespread, where physical distribution of audio and video tapes (with increased use of labor and vehi- cles) may be cheaper than real-time transmission, and where the corresponding delay may not be considered important. 43. Bower (1972). 44. Bower's suggested method goes a step farther than the ccirT 1976 GAs-S handbook's section on input-output analysis, which notes that input-output tables might be used for telecommunications planning; Bower suggests estimating telecommunications investment requirenients using projected outputs in all sectors of the economy. 45. Little is known about substitutability or complementarity of telephone and postal services. Although some rather descriptive information relating to the subject is presented in chapters II and 12 and appendix B, few attempts have been made to estimate the cross- elasticity parameters econometrically: one such attempt was a time-series study for the United Kingdom carried out by the Statistics and Business Research Department of the British Post Office. Despite the use of acceptable econometric techniques, it failed to pro- duce statistically significant estimates of the coefficients. 46. Wellenius (1988). Chapter 6 Location and Communication THE PREVIOUS TWO CHAPTERS discussed statistical correlation and regres- sion analyses of cross-sectional and time-series country data as well as cross-country comparisons of input-output sector indicators. Such ex- ercises by definition examine telecommunications services and eco- nomic activity at a rather aggregate macroeconomic level. Several of the following chapters examine the other end of the spectrum by re- viewing selected microeconomic project- and program-related at- tempts to identify and quantify the benefits of telecommunications as individuals and groups perceive them. This chapter and the next, however, examine some of the middle ground between the global anal- ysis of previous chapters and the microeconomic exercises of later chapters. This is done primarily by, first, reviewing several issues per- taining to spatial aspects of telecommunications investment and, sec- ond, in a related sense, examining some of the existing evidence about general linkages between telecommunications investment and the transport and energy sectors. The Spatial Framework A voluminous literature addresses the question of why economic activity or specific forms of economic activity locate where they do.' The literature recognizes that an inevitable consequence of economic development is increasingly complex interdependency among individ- uals and organizations. In general. the various models formulated by 121 1 22 MACROECONOMIC ANALYSIS OF BENEFITS economists and regional scientists assume that businesses, workers, and consumers tend to locate in places that minimize the net costs of production and consumption. Various factors affect these costs, such as population skills and density, topography and natural resources, cli- mate, availability of modern technology, existing infrastructures, pric- ing policies, government controls and regulations, and political boundaries. The availability of rapid and reliable communication is, of course, another important consideration in the decision governing location, that is, a reliable telecommunications system has significant space-bridging qualities.2 In a country or a region within a country, a basic spatial order is often clearly recognizable. Centers (cities, towns, or villages) can be ordered or ranked according to the groups of central functions found in them.3 Such a ranking defines a hierarchy of central places, which is closely re- lated to a hierarchy of population size, of economic functions, of eco- nomic units, of government administrative centers, and of ways of life. Several analyses of the relation between central places and their hinterland have examined the number and distribution of existing telephones. Christaller, in pursuing the hypothesis that a hierarchy of places depends on the relative size and economic importance of cen- tral places, used the number of telephones as a proxy for a common measure of centrality in his examination of the size, number, and dis- tribution of central places in southern Germany in 1933.4 In 1982, Kilgour examined central places in Costa Rica and, using Christaller's methodology, concluded that the geographic placement of the tele- phone in national space elucidates a country's spatial structure and identifies efficiently organized areas as well as underorganized ones. She found that the spatial structure in Costa Rica conforms to the interlocking structure predicted by Christaller and argued that the telephone could be used in any country to indicate centrality and that this indicator would be comparable across countries.5 Other telephone-related analyses also support the contention that the flow of people, communication, energy, and goods that produces the interdependences among a region's centers to a large extent re- flects a basic spatial order. The following two examples illustrate this in the context of telephone traffic. Example 10. Spatial Order and Regional Telephone Traffic (Chile 1971) The relation between long-distance telephone traffic and the structure of central places was partially examined by using 1969 data on twenty-two cities and towns in LOCATION AND COMMUNICATION 123 the Aconcagua Valley region in central Chile., A high concentration of long-distance traffic was observed: the average small city or town exchanged 50 percent of all calls with just one other place, 71 percent with two places, and 81 percent with three places.7 Traffic also tended to follow the lines of the population hierarchy: as many as 90 percent of the region's smaller centers exchanged at least 60 percent of their calls with larger places.' Given this high concentration of long-distance traffic and its tendency to follow the population hierarchy, it was hypothesized that it should be possible to estimate inde- pendently the ordinal distribution of a large proportion of each place's total long- distance telephone traffic from a relative measurement of the interdependence among pairs of centers. To this end cultural, economic, administrative, and locational inter- dependence indicators were combined to rank overall interdependence between a cen- ter and each other center. Postulating that traffic between one center and all others is distributed in the same order as overall interdependences, the ordinal distribution of 80 percent or more of each place's long-distance telephone calls was estimated accurately. | Example 11. Hierarchical Level of Contact (Kenya 1980) Another study, in two regions in western and south-central Kenya, also examined the level in the hierarchy of towns or villages among which contacts were taking place.' I Telephone calls into and out of four general types of towns within the two re- gions were examined; the towns from largest to smallest were denoted as principal town, urban centers, rural centers, and market centers. In all cases more calls were re- ceived than were initiated, and, as would be expected, telephones were used most in- tensively in the principal town. Presumably this last result is partly because access to service was greater in the principal town, which therefore had more subscribers to call. Also, most calls were terminated within the region in which they originated, and the proportions were relatively higher for smaller places. The most notable result is that the called parties tended to be at either the same hi- erarchical level or a level above the calling parties. This higher level was generally the principal town at the center of the region. Furthermore, contact among places at the same level decreased as the level of hierarchy decreased. This effect spilled over into telephone communication patterns with the rest of the country. Hence, principal towns and urban centers generally had a much wider horizon of contacts. As the exercises for Chile and Kenya illustrate, a spatial analysis of the interactions between telecommunications and economic activities at the regional level can be used to describe and perhaps even predict the patterns of telecommunications traffic as a function of regional structure and change. Such studies also suggest that development tends to compel more specialization and less self-sufficiency.'2 In essence, a development process generates the need for greater con- nectivity: the more development, the more specialization, the more in- terdependence, the more need for connectivity.13 1 24 MACROECONOMIC ANALYSIS OF BENEFITS For national development policy, however, the above descriptive ma- terial is of limited value. It would be of greater interest to understand the extent to which telecommunications services can be used to influ- ence locational decisions, rather than merely to reflect them. Three of the most common spatial problems experienced by developing coun- tries are the following: a. The heavy concentration of employment, especially in modern sectors, in one or a few urban centers, and the difficulty of in- ducing modern enterprises to locate in smaller centers or the hinterland b. The rapid pace of migration from the countryside to major urban centers-one cause of the very rapid growth of urban populations-and the corresponding growth of the largely un- planned squatter settlements sometimes referred to by planners as the "transitional urban sector" c. The apparent lack of major success so far in implementing, on a large scale, the multicentered urban structures that city planning studies frequently recommend for developing countries. Work in locational analysis explains these patterns, in part, by ex- amining communications linkages among organizations, and empirical research in the industrial countries suggests that the enhancement of telecommunications infrastructure can have a significant effect on the location of economic units.'4 In recent years, several studies have ex- amined the role that telecommunications has played in the economic development of rural and less favored regions of Europe and North America. Interest in this area arose primarily in response to the struc- tural changes that accompanied the dramatic increase in information- intensive activities taking place in the growing service sector during the postindustrial phase of economic development. These changes present rural regions with new opportunities for growth and development but also threaten to increase economic marginalization if such opportunities are not aggressively pursued. In particular, the traditional objective of universal service may not be practicable in an environment of technical change and liberalized telecommunications policy, and structural changes may increase the gap between information-rich and information-poor regions. In North America, investment in advanced telecommunications in- frastructure in rural areas is seen as an important catalyst for local economic growth, and failure to invest is predicted to cut such regions off completely from the economic mainstream. Furthermore, invest- LOCATION AND COMMUNICATION 125 ment in telecommunications alone is not sufficient; complementary investment in such areas as transport are needed if the potential gains from increased productivity are to be realized.'5 For rural Europe, attention has focused on the supply of and de- mand for new telecommunications services in what the Commission of the European Council has termed less favored areas, paying special attention to the limited access afforded by private networks that are embedded within large organizations.'6 Expressed demand has gener- ally been found to be low; one study on Greece found that of 425 provincial manufacturing firms surveyed, only 19 percent declared that "adequate infrastructure" affected their locational decisions, compared with over 30 percent that specified government incentives, low-cost land, and availability of labor.17 Yet the potential benefits of increased investment are very high, particularly for the intermediate rural areas (as distinct from the extremely rural areas). One model that was developed to quantify the estimated aggregate gains in em- ployment from investment in new information technologies in rural areas of Europe calculated that if all the rural and very rural areas of the European Community invested 1 percent of their GDP in new com- inunications infrastructure, between 700,000 and 900,000 new jobs would be created, at a cost of about ECU11,600 per job."' In deve- loping countries, however, little empirical work has yet demonstrated the extent to which suitably designed telecommunications policies and investment programs can help promote a specific pattern of set- tlement and estimate the resulting benefits.'9 Regional Development and Balance In developing countries, there is a widely held perception that the existing high degree of spatial concentration of modern sector activity is undesirable.20 Following from this, government goals and policies are often aimed at dispersing development. This generally occurs at two levels: regional planning on a national scale, which seeks to achieve some form of balance between the levels of development in different regions or different cities, and, for the very large population centers, planning at the intraregional or metropolitan level, which typically calls for a pattern of several urban centers to relieve the heavy concentration in and around a congested central business district. 1 26 MACROECONOMIC ANALYSIS OF BENEFITS When looking at the problem of regional imbalance it is helpful to distinguish between structuralist and locationalist views. The former attributes slow growth to the unfavorable mix of economic activities in particular areas (a preponderance of slow-growth sectors or activi- ties). In contrast, the locationalist view emphasizes spatial and infrastructural factors, such as the physical accessibility of particular locations, the absence of services such as water supply and power, and so on. It has been argued that for effective change, the structuralist point of view may be more relevant for more mature industrial coun- tries, whereas the locationalist view may have greater relevance, at least in the near term, for many developing countries. However, some arguments attempt to integrate elements of both the structuralist and locationalist approaches.21 Nicol, for example, seeks to analyze the type and nature of effects that improvements in communications may have on economic performance and structure, and subsequently on the location of economic activities, in develop- ing countries. This argument differs from traditional economic ap- proaches, which do not treat information and communications as decision variables, in that it posits that access to information is a fun- damental determinant of economic and spatial structures and that communications weakens the ties of proximity in the distribution and organization of economic activity in space. Consequently, improved interregional communication should re- duce information and transaction costs to businesses, thereby reduc- ing the relative benefits of clustering in large urban areas. If firms have less need to cluster to capture agglomeration economies, they may relocate to smaller secondary centers where the costs of land and resources tend to be lower and where there is less congestion and fewer environmental constraints. It has also been suggested that im- proved telecommunications can ameliorate unequal income distribu- tion among regions within a country by functioning as a partial substitute for physical proximity, if it can operate in tandem with an efficient physical transportation network.22 However, the evidence is ambiguous. The opposite point of view has also been argued in the case of rural areas of Europe, where the costs of transmitting information remain exceedingly high over dis- tances, in spite of technological advances, and agglomeration econo- mies have been strengthened by the introduction of new patterns of handling information.23 If it is accepted that improved telecommunications might facilitate achievement of locational objectives along with other incentives to LOCATION AND COMMUNICATION 127 locational change, relevant policies might take one of two forms: more generalized locational inducements, which have been familiar policy instruments in several industrial countries, or a more focused ap- proach, which concentrates public investment and inducements to private investments at specific locations, sometimes designated as growth centers or growth poles. A regional type of general inducement incentive was used in Swe- den in the 1970s, where the relatively disadvantaged northern regions (Norrland) got priority in service improvements and public invest- ment, and there was a large financial cross-subsidy from the central regions through telephone rates.24 In 1973, the Swedish government reduced telephone charges for firms and individuals in Norrland by 20 to 50 percent. Calls made within the Norrland counties to the pri- mary centers-usually the administrative headquarters-were often reduced 35 to 50 percent, and local calls made within the same com- munity, 50 to 70 percent.25 The resulting financial subsidies for tele- phone service in Norrland were relatively larger than the correspond- ing subsidies adopted for freight transport. A telecommunications investment policy to consider explicitly the problems of disadvantaged regions and a shift in pricing policy to dis- criminate less against long-distance calls both seem on the surface to be relatively attractive ways to promote regional development. Also, the relatively high prices charged for medium-distance calls, which cross the boundaries of charge areas for local calls, might on occasion be mitigated by expanding local call areas to the maximum extent possible, while attempting to maintain a reasonably direct relation be- tween price and costs.26 The latter could encourage economic interac- tion and the development of local markets within the region, possibly even a more important goal than improving interregional linkages. In fact, it can be argued that effective intraregional policies should be implemented first, since without them, improved interregional link- ages might only increase the dependence of the disadvantaged regions on, and their propensity to import from, more strongly developed centers. Another approach is to stimulate economic development by more focused measures such as investing in communications or other facili- ties at designated growth poles.27 For example, evidence exists that within a developing country or region, the larger-income growth pole villages tend to derive the most benefits from the introduction of a public call office (Pco) telephone. The following example illustrates this. 1 28 MACROECONOMIC ANALYSIS OF BENEFITS Example 12. Rural Village Pco Benefit Regression Analysis (Costa Rica 1976) For a cross-section of sixty-four villages in rural Costa Rica data were gathered on public telephone use and on selected economic and social characteristics for each village." The data came from three sources: official government census data on hous- ing, population, and agriculture; rural telephone traffic data collected by the national telecommunications entity; and survey data about individual telephone users compiled by telephone concessionaires. Several models were specified and statistical regressions estimated in which tele- phone traffic or telephone use variables were used as the dependent variable. It was hypothesized that if differences in telephone use could be explained on the basis of the different economic and social characteristics of the villages, then it would be pos- sible to predict which villages, then without service, would benefit most from gaining service. Those villages with a high potential to benefit would then be placed high on a priority list for new telephone investment. The results of the regressions suggested that rc:o benefits tended to be greatest in rural Costa Rican villages that possessed one or more of the following (sometimes col- linear) characteristics: a. Per capita village income was higher than the average of all villages b. The village had a relatively large population C. The village was located relatively far from the major economic, social, and gov- ernment center of San Jose d. The educational level of the population was above average e. The population tended to be clustered more closely around the site at which the telephone was located. It was also observed that the mix of calls made for agriculture and business purposes was clearly correlated with the corresponding economic base and demographic charac- teristics of the villages. Migration between Rural and Urban Areas Migration on a large scale from rural areas to large cities, some- times through smaller regional centers, is generally acknowledged to be one of the most difficult problems facing developing countries- although some observers note that it also creates opportunities for economic and social development. This rapid flow of population often strains the social and economic overhead capital of the big urban areas, and it is widely assumed that if population flow were slowed, the big cities would be better able to absorb and generate em- ployment for new immigrants and to cope with internal development LOCATION AND COMMUNICATION 129 programs. Investment in rural infrastructure, including roads, water supply, and telephones, has been suggested as one way to slow rural to urban migration because these improvements can reduce some of the basic disadvantages of living in a rural location.29 It is difficult to disentangle the negative and positive effects on mi- gration of a PCo investment program in rural areas. Access to current agricultural prices, more efficient use of transport and movement of supplies, rapid flow of information in times of medical need or natu- ral disaster, and so forth could improve conditions in rural areas of developing countries and may in fact create employment opportuni- ties. People migrate for various reasons, however, and the relative short- or long-run lure of jobs, higher incomes, and educational and other opportunities in urban areas would not be greatly changed in the short run by providing a rural telephone connection. In fact, it might even be argued that a rural telephone program would encourage migration since the migrants could more easily keep in touch with their home village and family and would not have to endure long peri- ods without access to local information.30 Metropolitan Development and Balance Regional policy at the metropolitan level tends to mirror policies for planning location and land use at the national level. Metropolitan areas clearly have locales where the private benefits of a central loca- tion are sufficient to induce firms to pay large premiums, in spite of problems of congestion-problems that have led many planners to contend that the social costs of such spatial concentration outweigh the private benefits.3" Improved telecommunications may reduce such costs. For any given size, a city with a better system of two-way communication will enjoy a higher level of connectivity and is likely to achieve more agglomera- tion economies than one with a worse system. Also, with a fixed level of communication, larger cities will achieve higher agglomeration economies than smaller ones because they offer larger communities of interest. 32 Nevertheless, in many developing countries, governments are plan- ning more dispersed patterns of settlement within the urban region. These policymakers contend that the large and rapidly growing cities of poorer countries cannot afford to mirror the high-density spatial patterns of cities in richer countries. They argue that in developing countries jobs and residential locations should be more contiguous 1 30 MACROECONOMIC ANALYSIS OF BENEFITS spatially, partly to reduce transport costs so that more family members in the lower- and middle-income groups can have access to employ- ment opportunities.33 There are numerous reasons for the inability of developing coun- tries to disperse economic activity significantly. One of the prominent reasons for this failure is the overdependence on relatively slow and costly face-to-face communication and messengers and the related ab- sence of reliable and widespread telephone service. Webber argued that "a major contributor to transport congestion in such concen- trated centers as Caracas is the thousands of messengers who race around the city carrying pieces of paper from one office to another ... Rather than sinking its investment capital into an underground rail system, Caracas might have realized a far greater return had it in- vested that same capital in the telephone and postal systems."34 Some evidence on the relation between telecommunications access and transport is discussed in the following chapter. With particular re- gard to the use of metropolitan space, the form of the metropolitan region is largely determined by the interaction of workplace location, residential location, and commuting. Well-focused telecommunica- tions investment might modify the demand for travel, particularly re- garding the journey to work, by altering two key locational processes: the choice of location by employers and the choice of residential loca- tion by households. The latter decision could, in turn, affect the loca- tion of workplaces by changing the supply of available labor. Pool, in fact, noted that the telephone has had a profound effect on the spatial patterns of cities in the United States. He outlined the process by which, even as the centralization of business activities led to the growth of high-density centers in cities, the telephone encour- aged the scattering of particular business districts. Once the tele- phone was available, businesses could move to cheaper quarters; a firm could move out from city centers or even up to the tenth or twentieth floor of new buildings. As early as 1902 a writer in Tele- phony calculated that if business continued to rely on messengers, the elevator wells of skyscrapers in big cities in the United States would have to be about double the size they were, which would make such buildings uneconomical.35 The general tendency throughout the world is still for business and industrial activity to be concentrated spatially, whereas residential areas are widespread. As cities grow in size, the journey to work be- comes longer. This problem is particularly acute for developing coun- tries. In a city of 1 million persons, the average journey to work is approximately 3 miles; in a city of 5 million, 7 miles. Some low- LOCATION AND COMMUNICATION 131 income groups in Asia and Africa reportedly walk two to three hours every day to reach their place of employment.36 Webber contended that for developing countries a feasible way to reverse such a trend is to undertake a large and focused investment in postal and telecommunications systems. He noted that "demand for connectivity generates city expansion; indeed it is the sole source of urban growth. If connections through geographic space were not costly, there would be no cities; and there would be no central con- centrations within them either. People and organizations concentrate in cities only because the costs of connectivity are least where linked establishments are close to each other. That's also the reason they tend to concentrate in city centers, because each of them is striving for lower costs of being connected."37 Following from this, he argued that greatly increased investment in modern telecommunications and postal infrastructure would "permit spatial relocation of many estab- lishments for whom a central location is presently the only viable op- tion. Phones, post, and the parallel data-transmission and video communications channels all are inherently random-access technolo- gies. They do not rely on spatial concentration for their effectiveness. When they work (technically) well, they actually work best in spatially dispersed settings. These are systems that foster connectivity when partners to transactions are not spatially contiguous."38 There is a growing body of industrial-country literature on office location, such as Nilles's in-depth study of the possibilities for dis- persing work from a major California insurance office.39 Even for in- dustrial countries with much of their infrastructure in place, improved telecommunications might promote locational changes that are thought to be economically, socially, or environmentally advanta- geous. Representative research on industrial countries was reviewed by Pye and Goddard, and a simplified framework for assessing the costs, benefits, and effects of such locational changes was presented by Harkness.40 In general, however, these issues have not been suffi- ciently researched, and in a developing country where the greatest po- tential for innovative change may lie, little research has been undertaken. Nevertheless, even if policies to increase the attractiveness of life in small villages and towns and to promote land use planning in metro- politan areas are strikingly successful in the foreseeable future, the large unplanned or transitional urban squatter settlements typical of most cities in developing countries will continue to grow.4' Much of the population of these urban squatter settlements works in an infor- mal economy-an economy that largely provides services and that 1 32 MACROECONOMIC ANALYSIS OF BENEFITS has been characterized as having unregulated competitive markets, small-scale operations, ease of entry, family ownership, reliance on in- digenous resources, labor-intensive and adaptive technology, and reli- ance on skills acquired outside the formal educational system.42 The mechanism to enhance the economic contribution and result- ing welfare of urban squatter settlements cannot be totally divorced from communication. Production linkages between firms or group as- sociations of different sizes were a striking feature of the economic structure in countries that became industrialized in the past century. Large enterprises subcontracted work heavily and efficiently to small ones. The same pattern was followed later by Japan, Korea, and Taiwan.43 Such linkages might also be facilitated in developing coun- tries by improving the access to public telecommunications services throughout the country. Notes 1. For summaries of the substance of this literature, see Greenhut (1963 and 1970); Richardson (1969); Nourse (1968); Isard (1969); and Hagget (1965). 2. For a discussion of factors affecting the population density of cities in the context of the transport sector, see Linn (1983), pp. 90-105. The concept of the space-bridging quali- ties of telephones is examined in Kilgour (1982). 3. For instance, in the region of example 10, the main cities are seats of municipal au- thority and have a standard set of government services (for example, treasury, inland reve- nue office, and local court of justice), elementary and high schools and sometimes also technical or vocational schools and occasionally university campuses, one or more hospitals, police stations under the charge of senior officers, main lines of road and railway public in- terurban transport, branches of one or more national banks, and a fairly developed tele- phone exchange service. In less important places, education is limited to the elementary level, national health service is available only at clinics staffed by paramedics, police sta- tions seldom include officers with rank above lieutenant, and banks are rare. Small towns often have only an elementary school, a small police station, a health station, a public long. distance telephone, and intercity transport only to the next larger town. About one-fourth of the region's settlements have none of these services. 4. Christaller (1966). 5. Kilgour (1982) p. 139. 6. Empirical evidence is presented in Klein (1971). 7. All places outside the region were counted as one. 8. This is consistent with findings in other developing countries. In Costa Rica in 1976, for example, an average of 60 percent of calls from surveyed rural Pcos were addressed to the capital city. Likewise, in El Salvador, in 1977, approximately 70 percent of telephone calls from rural places were directed to the capital city. More recently, in Peru during the two-year period of 1984-85, 70 percent of calls originating from five snmall rural communi- ties in the province of San Martin went either to Peru's two major cities (Lima or Trujillo) or to the commercial center of the province, Tarapoto. See Mayo and others (1987). 9. For example, for each place the arc carrying the highest number of calls, the second highest, and so forth is identified until a large proportion of traffic is accounted for. LOCATION AND COMMUNICATION 133 10. Klein also found that the proportions of a center's traffic that were carried over its highest traffic arc, second highest, and so forth were roughly constant among centers of like characteristics. Hence, he proposed that the proportion of each center's traffic that was car- ried over each of its major arcs could be predicted by mapping standardized distribution curves on the ranking of arcs obtained from measures of interdependence. With this, the problem of predicting long-distance telephone traffic in a region would be reduced to fore- casting the total volume exchanged by each center. This might be useful, for example, for designing a regional network in a newly developing area or for predicting the changes in traffic patterns likely to follow changes in regional structure. 11. Cleevely and Walsham (1980). 12. The contention that development tends to compel more specialization is of course a form of Adam Smith's principle that division of labor is limited by the extent of che market. 13. For an elaboration of this point, see Webber (1980). 14. Thorngren (1970); Pye and Goddard (1977). 15. Parker and others (1989) and Bradshaw (1990). 16. See, for example, Gillespie and others (1989) and Mansell (1988). 17. Kottis (1986). 18. Hansen and others (1990) and Analysys, Ltd. (1989). 19. One limited but interesting study used a theory of hierarchies of places to examine the interactions between telecommunications and economic activities in Kenya's Rift valley and central provinces. See Cleevely (1979). 20. For example, costs of transport congestion in the form of high vehicle operating costs and wasted time, environmental damage, accidents, high energy consumption, and high costs for other infrastructure systems, notably water and sewerage. Research relating to the general contention that large urban areas are too large is, however, inconclusive. For a sum- mary of the optimum city size argument, see Renaud (1981) and Saunders and Warford (1976), pp. 80-82. 21. Nicol (1983a and 1983b). 22. Preece (1987). 23. See, for example, Hall (1990). 24. Ministry of Labour and Housing and Ministry of Physical Planning of Local Govern- ment, Sweden, (1973). 275. By contrast, policies to provide and price telecommunications services in many deve- loping countries probably increase the disadvantage to the less-developed regions. This is discussed in chapter 14. 26. This can have a major effect on traffic. For example, in the study of the Aconcagua Valley in Chile (example 10), extension of local area charges to some links normally priced as long-distance resulted in call traffic two to fourteen times as high as would be expected from the analysis of the region's economic and previous traffic patterns. 27. For a capsule discussion of some of the primary arguments associated with creating growth poles in rural areas of developing countries, see Saunders and Warford (1976) pp. 102-03. 28. This study is described in more detail in appendix C. 29. Other rural problems include lack of job opportunities; inferior social, economic, and health care amenities; and relatively poor-quality educational opportunities. For a summary of some of the literature on migration in developing countries, see Levy and Wadycki (1972) and Yap (1975). 30. Where migrants go is partly determined by distance and contacts, and it has been ar- gued that improvements in communications have been more important in stimulating mi- gration than has the reduction in the costs of physical movement. See Beier and others (1975). 31. One such premium is, of course, a high black market price for a telephone. It is not 1 34 MACROECONOMIC ANALYSIS OF BENEFITS uncommon in cities with long telephone waiting lists to find black market telephone prices in the central business district exceeding those in outlying areas by a factor of ten or more. 32. Nicol (1983a and 1983b). 33. Beier and others (1975) p. 63. This is in contrast to the case of several more indus- trial countries, which have achieved a more dispersed settlement within the urban region. Indeed, in some countries, notably the United States, a problem in several older industrial areas is how to stop businesses from leaving the city centers and to prevent those central areas from decaying. 34. Webber (1980) p. 9. 35. Pool (1979) p. 182. 36. Beier and others (1975) p. 62. 37. Webber (1980) p. 2. 38. Webber (1980) p. 9. 39. Nilles (1974) and Nilles and others (1976). 40. Goddard and Pye (1978) and Harkness. (1971). 41. In the late 1970s, an estimated 60 to 80 percent of the popularion in some Asian and Latin American cities (and perhaps 40 to 60 percent in African cities) lived in these settlements. 42. )ILL (1972) and Beier and others (1975). Another study estimated that this informal economy employs between 50 and 70 percent of the labor force in Asian and Latin Ameri- can cities, and between 30 and 50 percent in African cities. See Mazumodar (1975 and 1976) Of course by no means all of the urban poor are employed in informal sector activi- ties. For a discussion of this point, see Linn (1983) pp. 37-42. 43. ILO (1972). Chapter 7 Transport, Communication, and Energy Consumption THE DISCUSSION IN THE PAST CHAPTER focused on some of the spatial is- sues of developing countries from a relatively wide, multisectoral point of view. In this chapter the general contention that the spatial characteristics of a developing economy are relevant to the relation between telecommunications and development is extended further. This chapter examines the one sector that, on a spatial basis, is par- ticularly affected by distances between communicating parties, the time required for communication, and the alternative forms of com- munication available. This sector, which is both complementary to and competitive with telecommunications, is transport. Clearly there are strong interactions between the telephone and al- ternative modes of transport, such as roads, railways, waterways, the postal system, telegraph, and telex. Both complementarity and substi- tution are present: a smoothly running transport system will necessi- tate fewer telephone calls to overcome disorganization in freight, business, and personal travel but will stimulate telecommunications demand by promoting trade and other interaction. An efficient postal system depends on efficient transport and may be a substitute for cer- tain uses of telecommunications, especially in the case of overnight package delivery services. Efficient road and rail travel depends to a considerable extent on telecommunications, and good telecommuni- cations facilities enable some trips to be avoided and others to be or- ganized on short notice. Two questions arise concerning the relation between the transport sector and telecommunications services: 135 1 36 MACROECONOMIC ANALYSIS OF BENEFITS a. How far can complementarity of telecommunications and trans- port be demonstrated? To what extent can gains in the efficiency or output of the transport sector resulting from an increased use of telecommunications be identified and measured? b. Where telecommunications and travel are likely to be substitutes, can such substitutability be demonstrated? Can cases be identi- fied where the use of telecommunications is the least-cost solu- tion in meeting a specific communication need, and can the cost savings and economic rates of return on the investment associ- ated with the use of this least-cost solution be estimated? In addressing these two questions it is a simple step to move slightly beyond the association between telecommunications and transport to consider the important relation between telecommunica- tions, transport, and energy consumption. There are essentially two reasons for doing this. The first is that the transport sector typically accounts for a significant proportion of the petroleum consumption in a country. In Canada, for example, which has the world's second highest level of energy use per capita, the transport sector accounts for 58 percent of petroleum used, and 60 percent of this is used to transport people.' The second and more important reason is that because of changes in relative costs over the past decade, telecommunications services have become much less costly relative to transport. Transport costs rose significantly in the 1970s as a result of increases in the price of petroleum and the cost of petroleum-induced steel and other materi- als. Although these prices have more or less stabilized, they remain a significant component of transport costs for many countries where petroleum must be imported and paid for with hard currency. In con- trast, telecommunications has been experiencing changes in technol- ogy (electronic switching, network digitalization, multiple-access subscriber radiotelephone, satellite communications, fiber optics, solar-powered repeaters, and so forth), which have brought about, and should continue to bring about, a decline in real terms in the cost of providing communications services. (Several such changes were out- lined in chapter 2.) Hence, opportunities for realizing savings in the costs and use of transport through telecommunications continue to exist. Compounding its potential to facilitate change in the general loca- tion of economic activity, which was discussed in the last chapter, telecommunications can help to conserve energy resources in at least two other ways: by reducing waste and improving efficiency in the use TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 137 of fleets of road vehicles and by substituting telecommunications for certain types of travel. In addition to other ecological benefits, using petroleum more efficiently also creates less air pollution, which is a significant benefit in large urban areas in developing countries. Research on the above topics relates primarily to industrial coun- tries, although many of the studies have some relevance to developing countries.2 The energy saving potential of both the telecommunica- tions and transport systems in the more industrial countries is, how- ever, reasonably well established, and substitution among telecommu- nications, transport, and energy, at least in the short run, is likely to be marginal. In developing countries, however, telecommunications networks are grossly inadequate in terms of unmet demand, low na- tional penetration, and poor quality and reliability of service. Hence, improvements in telephone access, quality, and reliability could dra- matically improve the efficient use of transport as well as energy. Improved Use of Vehicles In industrial countries it is generally accepted that good manage- ment practices together with the use of telecommunications can facil- itate the efficient use of vehicle fleets. This is true for a wide range of applications, from taxi operations to delivery truck fleets or vehicles used to maintain equipment in the field. It is also apparent that bene- fits can be reaped at various levels of technology, including a one- vehicle truck business using a public telephone to locate a destination or secure a return load and a large business using a sophisticated sys- tem of radios to locate and identify vehicles automatically and to transmit posting instructions from a central control location. Some of the relevant technologies, all well known and tested in numerous field applications, are public telephones (PCOs and public coin telephones), radio paging, private mobile radio, open shared channel (citizen band) radio, switched radiotelephone service (with or without con- nection to the public switched telephone network), and data commu- nications by radio (manual or automatic data acquisition, automatic identification of vehicles and location).' Freight vehicles are costly capital assets, and replacement parts, driver time, fuel, and tires are important costs-for both the firm and the country, where imported petroleum and, sometimes, vehicles often contribute significantly to balance-of-payments deficits. Im- proved vehicle use entails full loading and minimizing idle time, misrouting, and return journeys made with an empty vehicle. Such 1 38 MACROECONOMIC ANALYSIS OF BENEFITS improvements require organization, and efficient organization requires communication. The relevant questions, however, are on what scale and how significant are the benefits? Evidence from Industrial Countries Although fragmentary, the evidence suggests that even where tele- phone, telex, and two-way data and radio systems are used intensively to organize vehicle movements-as in the United States-further in- vestments in complementary telecommunications systems remain at- tractive. One U.S. study suggested that using mobile radios to control the movement of vehicles for local pickup and deliverv can increase output per vehicle by from 15 to 25 percent, so that four vehicles could do the work of five.4 In Sweden in early 1983 more than 3,000 standard taxicab radios were replaced with computerized mobile ra- dios linked to a central computer at headquarters. The central com- puter ran a constant check on the availability and location of every taxi, so the dispatcher could instantly locate the closest taxi to a fare and relay the address through the vehicle's printer. By cutting down the time that drivers spent roaming the streets and dispatchers spent locating them, it was claimed that Swedish taxi companies would in- crease their profits 30 percent by lowering petroleum consumption and using the vehicle fleet more efficiently.5 During the 1974 energy crisis, the GTE Corporation allowed employees to take company vehi- cles home to be "home-dispatched" to their next task in the morning, saving 2.4 million gallons of fuel compared with 1973. In a 1975 study of the costs of operating freight and service vehicles in four U.S. cities, Lathey suggested that employing radio control could save 1,300 million gallons of fuel by 1980.6 Corresponding findings were reported in the United Kingdom and the former U.S.S.R.7 The Soviet studies are of particular interest since they are fairly elaborate and come from an environment rather different from the industrial western market economies examined in most studies. One 1974 study analyzed the benefits of a project de- signed to promote the use of a telephone and mobile radio dispatch system to improve shared use of agricultural machinery in sixty farms in the Rostov area." The study concluded that the system a. Reduced idle times of machines by a factor of between two and three, improving the productivity of tractors and other agricul- tural machines between 20 and 25 percent TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 139 b. Reduced the time spent transporting machines between working locations from 30 to 40 percent c. Reduced the time required for basic agricultural operations, such as harvesting, from 20 to 25 percent d. Saved from 40 to 50 percent of the time spent by managers and specialists in machine operations. By also using the dispatch system to monitor animal husbandry, feeding, and care, cattle productivity (output per ruble of expenditure on inputs of all kinds) increased 5 to 10 percent. In a second study in the former U.S.S.R., a survey of managers in various industrial enterprises showed that increased use of long- distance telephone calls for managing vehicle fleets-made possible by recent improvements in service-had improved their use of vehi- cles. For example, improved long-distance telephone service report- edly increased the efficiency of the Sedin lathe-building works in Krasnodar 10 percent.' In addition to using mobile radios and improving long-distance service in industrial countries, savings can also be achieved by using electronic devices that display up-to-date advice on traffic conditions, routes, and parking. Computer-based scheduling and data communi- cations for coordinating and documenting freight are also available. Since much unnecessary vehicle mileage results from imperfect infor- mation about routing, traffic conditions, and parking, drivers could probably make trips between the same origins and destinations with fewer vehicle miles and a consequent savings of energy, time, and cap- ital if they were better informed and advised. Several studies in the United States sought to estimate the amount of wasted vehicle mileage that might be avoided by such systems. Schoppert and others found that 17 percent of the motorists at ramps of major highways were taking an indirect route.'0 Gordon and Wood estimated that the use of electronic route guidance and parking information technologies could reduce total vehicle mileage more than 3 percent." System concepts for electronic route guidance were described by, among others. General Motors, the French Prr, and Stephens and others, for both the general user and for special pur- poses such as bus or truck tleets.'2 Such systems are obviously techni- callv feasible, and if they included a parking advisory capability, a 5 percent reduction in vehicular energy consumption would not be un- realistic in urban areas of industrial countries.'3 1 40 MACROECONOMIC ANALYSIS OF BENEFITS Evidence from Developing Countries A few studies in developing countries showed that poor communi- cation causes much unnecessary movement of empty vehicles. Kaul reported from a sample survey of fertilizer distribution in India that out of all vehicle trips made by farmers to fertilizer distribution points, the proportion of trips that failed (because the intermittent pattern of supply and demand had, unknown to the farmer, resulted in the required fertilizer being unavailable) was tvpically 10 to 25 per- cent, and for some locations it was more than 50 percent of all trips.14 The trips typically covered more than 6 kilometers. Kaul's re- sults are summarized in table 7-1. It would seem that the impressive returns on investment in fleet- control telecommunications in industrial countries could at a mini- mum be paralleled by gains obtainable from the use of more efficient Table 7-1. Successful anid IJnsuccessful Trips by Farmers to Major Fertilizer Distribution Centers in India, 1975-76 Distance from farm fertilizer Unsuccessful distribution Number Number of trips as a location of unsuccessjhil percent of Region and state (kilometers) trips trips total North Haryana 11.4 2.3 0.5 22 Himachal 13.1 1.4 0.1 7 Punjab 14.7 3.1 1.7 55 East Assamn 8.4 2 4 0.4 17 Bihar 9.9 2.1 0.6 29 Orissa 6.6 3.0 0.4 13 West Bengal 15.5 3.3 0.8 24 West and Central Gujarat 5.9 2.7 1.3 48 Maharashtra 8.7 1.8 0.5 28 Rajasthan 13.6 1.9 0.4 21 Uttar Pradesh 7.7 2.2 1.0 45 Madhya Pradesh 7.9 1.5 0.9 60 South Andhra 16.0 2.5 1.1 44 Karnataka 5.8 1.5 0.2 13 Kerala 2.7 3.4 0.3 9 Tamil Nadu 6.3 3.3 2.1 64 Souirce: Adapted from National Council of Applied Economic Research. India, "Fertilizer Survey, 1975-76." Summarized in Kaul (1978). TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 141 telephone service-including public telephones accessible by the roadside and mobile radio applications-in developing countries.15 The data from India together with the limited data from the former U.S.S.R. suggest that a transition from the nonuse or very limited use of telecommunications facilities in developing countries to the use of basic facilities such as telephones in public call offices in both urban and rural areas could substantially reduce transport costs, even before considering other benefits that might accrue from selected applica- tions of mobile radio facilities. More recently, the ITU sponsored a detailed analysis of the transpor- tation sector in the People's Democratic Republic of Yemen in order to examine the potential savings that could be obtained if telephones were installed along the country's major truck routes.'6 The study aimed to provide a quantitative basis for determining how much should be invested in telecommunications that support a transporta- tion system and to estimate how much savings might accrue if emer- gency breakdown assistance and scheduling were improved. The study interviewed many transport officials throughout the country and transport workers on one route, the Aden-Mukalla-Seiyun road. The resulting data were used to calculate the value of losses incurred due to lack of communication (but only losses that could be reduced with better communication). These figures were then extrapolated to the country as a whole (see table 7-2). Table 7-2. Annual Losses due to Lack of Adequate Communications Infrastructure for Transportation Routes ini the People's Democratic Republic of Yemen, 1986-8/7 (U.S. dollars) Aden-Mukalla- People's Democratic TYpe of 105s Seiyun road Republic of Yemen" Idle time due to transport breakdowns 883,503 11,666,094 Perished goods due to transport breakdowns 11,760 155,292 Unused return loads 17,096,606 a. The losses for the country were obtained by extrapolating data on losses for the route on the basis of information known about the volume transported: 363,136 tons (according to field survey figures) for the route compared with 4,861,000 tons for the country. An ade- quate coefficient for extrapolation could not be determined for unused return loads. Source: TI (1988). 142 MACROECONOMIC ANALYSIS OF BENEFITS The study then examined possible reductions in losses that could be obtained if a telephone system were installed at 20-kilometer inter- vals along the Aden-Mukalla-Seiyun road, where no telephones cur- rently exist. The number of full return loads was not expected to increase more than 15 to 25 percent, since the problem of empty re- turn loads was strongly linked with factors not related to communica- tions. However, improved communications were expected to reduce the time needed by management to respond to reported breakdowns, which would reduce the amount of time wasted by loaded trucks while waiting for repairs and spare parts. It was estimated that provid- ing quick access to telephones along the road would reduce the time needed for a truck driver to report a breakdown from as long as sev- eral days to just one hour. The total cost of implementing the system nationally was an estimated $600,000 a year, based on an estimated cost per terminal of $3,000, annualized with 5 percent depreciation and 6 percent servicing, maintenance, and attendance costs. The ana- lytical model showed total potential transport losses avoided annually through the use of this system to be $687,480 (YD229,160) for the Aden-Mukalla-Seiyun road; when extrapolated for the entire country, the resulting losses avoided were $9,013,962 (YD3,004,654). This gives a benefit to cost ratio for the project of 15:1; other benefits would include a gain of $1,710,000 a year from the 15 to 25 percent increase in full loads for return journeys, a reduction in the size of the fleet and related costs and expenditures for the same volume of goods transported, better scheduling that could improve performance 10 to 15 percent, and avoidance of $156,000 in losses from spoiled perishables for the country as a whole. Substitution of Telecommunications for Travel Pool reported that within three years after the telephone was in- vented in 1876, the London Spectator newspaper predicted that the new device would replace personal meetings.17 Predictions of this kind are still being made but are related, of course, to increasingly sophisti- cated technical alternatives to face-to-face communication. The Feasihility of Potential Substitution As was the case with the improved use of vehicles discussed in the previous section, little quantitative research on the potential for tele- commLinications to replace travel has been conducted in developing TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 143 countries. Research has, however, been carried out in industrial coun- tries on the extent to which business travel (and the use of the tele- phone or other media to conduct business) are substitutes in the functional sense that they produce the same result with the same user satisfaction. 18 The primary research methods used have comprised the following: a. Controlled experiments in which businessmen and civil servants carried out realistically simulated tasks (such as decisionmaking and bargaining) by audio and video communications media and (in the control group) face-to-face meetings b. Detailed monitoring and evaluation of telecommunications field trials c. Detailed surveys on the scale and content of communication ac- tivity and the extent to which respondents believe that business currently carried out face-to-face could be carried out effectively and acceptably over the telephone or by other means of telecommunications. The research findings are complex. In the United Kingdom the overall results showed that where travel substitution was feasible, video was rarely necessary to achieve the required interchange, and audio (simple telephone service where only two people are involved, loudspeaking telephone, a telephone-conference link, or a studio- based audio conference system) was almost invariably cheaper than travel. As for effectiveness in completing various business tasks pres- ently carried out in person, a U.K. Department of Transport study found that approximately 41 percent of the total activity could be ef- fectively and acceptably carried out by audio telecommunication.'9 Another study in the United Kingdom concluded that, although a substantial portion of all conversations with people outside an indi- vidual's immediate work group took place by telephone (rather than face-to-face), a much greater portion of all remaining meetings involv- ing travel could also have been carried out by telecommunications without loss of effectiveness.20 Telecommunications can also substitute for travel by linking em- ployees who work at home with their employers or entrepreneurs who work out of their homes with their clients. Sometimes referred to as telecommuting, working at home was first "rediscovered" in the 1970s as a way to avoid time and energy wasted in weekly commuting; it has caught on relatively slowly, however, despite predictions that it would revolutionize the workplace. At the end of the 1980s, about 630,000 people in Great Britain were working at or out of their home (exclud- 1 44 MACROECONOMIC ANALYSIS OF BENEFITS ing domestics, including sales personnel), or about 2.5 percent of the total work force.2' In the United States, the figure is estimated to be much higher, at about 25 million.22 No data are available on the ex- tent to which these workers depend on telecommunications. Some descriptive studies exist on telecommuting arrangements.23 In general, they have found that adoption of telecommuting has been slowed, on the one hand, by internal resistance to nontraditional forms of organization and management within large companies and, on the other hand, by external criticism from researchers and labor unions that telecommuting makes workers vulnerable to economic ex- ploitation similar to that suffered by cottage workers in the early days of the industrial revolution. One area of significant growth has been in offshore data processing. Large insurance carriers, for example, in- creasingly send information such as claims via high-speed data links to be processed abroad in countries with lower wage scales. Tele- commuting by independent workers, such as entrepreneurs, designers, salespersons, writers, software experts, and high-level professional consultants has also grown significantly, aided by widespread use of innovations such as facsimile machines and computer modems and supported by next-day package delivery services that enable finished products to be delivered in a timely manner. Empirical Evitedece Having established that a partial substitutability of telecommunica- tions and travel seems to be feasible when a full range of costs and behavioral factors are considered, the next question is how much travel is actually saved in practice? A partial answer is that only costs associated with the direct substitution of business trips by telecon- ferencing have been observed in most of the current field-trial studies.24 Teleconferencing is successfully used in the United States by orga- nizations such as NASA, Bank of America, Ohio Bell, Exxon, Procter and Gamble, Texas Instruments, Ford Motor Company, Union Trust Bank, ARCO, Aetna, Boeing, Honeywell, and IBM, which report that the use of selected audio teleconference systems has replaced much travel.25 For example, "almost all participants in the trial IUnion Trust Companyl substituted use of the system for at least 50 per- cent of their face-to-face meetings; more than one-third substituted for 80 percent or more."26 Other examples include reports that IBM's teleconferencing system reduced travel costs by $414,000 in 1979, TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 145 its first year of operation, and by $830,000 in 1980;27 that a subsidi- ary of RCA Corporation held a video conference for 450 sales people around the United States costing $85,000, compared with $555,000 for a similar conference involving travel;28 and that during a recent three-day period each of seven branches throughout the United States of a division of the 3M Company conferred for one hour using audio conference facilities with a panel of marketing, sales, and technical experts at the home office. The total cost of the audio conference was only 15 percent of what the travel bill alone would have been to bring everyone together.29 Finally, a study of audio conferencing in a U.K. civil service department, Her Majesty's Sta- tionery Office, also showed substantial travel savings.30 As have sev- eral others, this study concluded that a significant degree of additional travel substitution over and above that achieved by the basic telephone system can, even in industrial countries, be-and al- ready is-achieved with simple means such as the loudspeaking telephone.3 Overall, the findings of such studies indicate that the withdrawal of vision and separation by distance have no measurable effect on the outcome of normal conversations that exchange information and solve problems. However, in conflict situations, in bargaining, and in instances in which first contacts between strangers require interper- sonal judgments and eye contact, the telephone is sometimes less ef- fective than a face-to-face meeting.32 One reviewer of such studies concluded that the potential for substitution between travel and tele- conferencing (presumably in industrial countries) is between 20 and 25 percent.33 Somewhat less rigorous studies in the former Soviet Union also have provided evidence of travel substitution and resulting gains in work productivity. A study by the Moscow Electrotechnical Institute of Communication analyzed the extent to which intensified use of long-distance telephone communication (made possible by improve- ments in the quality and availability of telephone service) improved labor productivity. Two kinds of gains were identified: direct labor sav- ings through the substitution of telephone calls for the preparation and sending of letters and telegrams and for personal visits, and indi- rect productivity improvements associated with more efficient man- agement arrangements made possible bv improved telecommunica- tions-more operational control of production, smoother production flows, speedier marketing, and an improved supply of materials and machinery.34 Both effects were estimated from the results of a survey 1 46 MACROECONOMIC ANALYSIS OF BENEFITS of fifty-seven industrial enterprises and twenty-one construction orga- nizations, which were then entered into a cost and productivity model.35 With regard to the substitutability of telecommunications, the So- viet surveys showed that, in the opinion of operational managers, some 24 percent of letters and 17 percent of telegrams could appro- priately have been replaced by telephone service if it were of adequate quality. Although its role is not clear, price did not seem to be a very important factor compared with physical availability. Similarly, the re- sponses indicated that on average some 4 percent of the volume of business travel undertaken was replaced by telephone calls (at a rate of about three calls per trip) once a reliable automatic trunk service was introduced and that a further 12 to 13 percent could be so sub- stituted. The study concluded that the value of future increases in productivity and, hence, the reduction in production costs predicted to follow improved long-distance telephone communication would be 4.4 times greater than the annual cost of the telecommunications in- vestments needed to bring about those improvements. A particularly interesting side result was that, given the basis of pricing for trunk telephone service in the former U.S.S.R., only 5 to 10 percent of these benefits would accrue as direct revenue to the telecommunica- tions carrier itself-a result closely in accord with the low price elas- ticity and large consumer surplus found in many developing countries. There is some limited evidence from Kenya about telephone-travel substitution. In 1979 telephone calling rates and the prevalence of other means of communication were surveyed for two regions in Kenya.36 One, the Eldoret region, is farther from Nairobi and was the more industrial of the two regions examined. The other, the Nyeri re- gion, although closer to Nairobi, also had a higher density of popula- tion and settlement. The survey showed that telephone subscribers in the Eldoret region tended to make about 80 percent more telephone calls per day than those in Nyeri and that they tended to rely much less on personal visits and vehicles to facilitate communication. The researchers concluded that some relative substitution of telephones for travel was taking place in the more sparsely populated and distant Eldoret region.37 More recently, in 1983, a survey of telephone users in four rural and semirural districts in Thailand found that 64 percent of office workers who used a telephone for their jobs would have made a per- sonal visit to carry out their task if a phone had not been available and 11 percent would have sent a messenger. Only 17 percent would have written a letter or delayed the message. Although most of the TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 147 calls made during the survey period were not local (about half were to other locations within the province and about one-quarter were out- side the province, including 6 percent to Bangkok and 2 percent abroad), respondents emphasized that the distance involved did not make any difference and that they would have traveled to other prov- inces or even Bangkok to convey their message.38 The Question of Travel Generation In addition to the possible substitution of telecommunications for travel, improved and extended telecommunications systems will likely also generate telecommunications traffic. Such a possibility is also im- portant in the context of energy consumption, since improved tele- communications could actually generate new transport traffic as well, thereby to some extent offsetting the substitution effect. In the ab- sence of reliable data on users' communication choices, opinion var- ies widely: some claim that there will be substantial reductions in travel, while some believe that improved telecommunications will, at least in industrial countries, stimulate a net increase in travel.39 More- over, care must be taken in evaluating such effects. Additional traffic is not necessarily undesirable; indeed, the benefits of the additional energy consumption may well exceed the costs. This will be more likely as the relation between energy prices and the economic costs of supply grows closer. The general concern is that in some instances en- ergy prices may be less than the economic costs of supply. Little empirical data explicitly identify transport traffic that has been induced by an overall increase in rapid communication. The use of teleconference systems in the United States has, in fact, produced the general view that there will be relatively little stimulation of traf- fic. For example, the use of teleconferencing by the NASA Apollo manned lunar landing project reduced travel expenditure 25 percent per staff member.40 An independent interpretation of these data (not necessarily endorsed by NASA) concluded that this represented a sub- stitution effect of up to about 35 percent and a generation effect in the region of 10 percent, resulting in a net reduction of 25 percent.41 Also, the mix between travel substitution and travel generation var- ies in different parts of the system, both among regions and among user groups; the effect may be different in rural than in urban areas. For example, in a study in rural Central Lincolnshire, England, in which 361 respondents kept diaries of their contacts for five days, re- searchers found that the major effect of the telephone was to bring individuals closer together. They contended, however, that the tele- 148 MACROECONOMIC ANALYSIS OF BENEFITS phone did not and could not acceptably replace the face-to-face con- tacts upon which most social relations were based. Hence, they concluded that, at least in Great Britain, increased telecommuni- cations was likely to stimulate the demand for social travel in rural areas .42 From the sole view of saving energy, the travel-generation effect of telecommunications investment should perhaps be minimized. Even in the extreme case in which the introduction of telecommunications services results in increased travel, however, the overall energy effi- ciency of communication will increase, since presumably a larger share of total communication will be carried by telecommunications. Hence, in most instances, the total economic efficiency gained from the interaction of travel and telecommunications should be at least as great as the gains calculated on a simple transport-substitution cost- savings basis. The Potential for Transport Substitution iti Developing Countries None of the fragmentary evidence on the interaction between tele- communications and transport can be regarded as conclusive. It is, however, sufficient to suggest that a substantial degree of cross- elasticity or substitutability is possible and that a shift in the balance of investment between the two sectors in favor of telecommunications could be beneficial in at least two ways. First, in most developing countries, where telecommunications net- works are highly inadequate, trips are made that could be replaced by telecommunications by any standard of judgment. Deficiencies in public telecommunications facilities are made up either by the use of more costly modes of communication or by the construction of pri- vate, high-cost, special-purpose networks. Alternatively, the activity that generated the need for special communications is abandoned al- together or carried out in a second-best way, which requires less com- munication. Hence, given the relatively undeveloped state of existing telecommunications services in developing countries, improved ser- vice should decrease the overall costs of communications on a per unit basis and increase efficiencV.43 Second, in developing countries, where the total amount of travel possible is sometimes directly constrained by absolute limitations on the amount of petroleum, vehicles, and road maintenance equipment that can be imported, increased opportunities for substituting tele- communications for travel could directly increase the absolute volume TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 149 of business that can be transacted as well as reduce the amount of en- ergy consumed per unit of business activity. Travel and the use of the telephone or other telecommunications services are clearly not perfect substitutes: most of the studies cited earlier scrupulously emphasized the limits of substitutability. However, where excess demand for telecommunications is chronic-as in most developing countries-travel is often used reluctantly as a substitute for telecommunications, with high costs in terms of energy, use of capital in the transport system, and time. This last point is worthy of particular emphasis, especially if the objective is to maximize the ben- efits derived from the communication sector, as distinct from merely minimizing the costs of attaining specific communication access or penetration targets. In such a case it will almost certainly be worth addressing the issue of attaining a more efficient use of scarce and costly human capital, as represented by the time wasted by persons with professional, technical, or managerial skills. Given that telecommunications can and does substitute for travel in a wide variety of circumstances, what are the economic benefits? Few studies have explicitly calculated a rate of return on telecommu- nications investments aimed at travel substitution, primarily because causally related changes are difficult to identify. One exception is the ITU study on the People's Democratic Republic of Yemen cited earlier; in this case, the economic model focused tightly on the relationship between communications infrastructure and transport costs, with nec- essary assumptions held to a minimum because both the economy and the transport network were small. Travel Substitution and Energ' Saving Although few studies have attempted to measure systematically the full benefits of telecommunications resulting from travel substitution, several attempts have been made to measure the effect of investment in telecommunications on energy consumption. These studies have been conducted primarily in industrial countries, for example the studies undertaken in Canada and the United Kingdom by Bell Can- ada and the former British Post Office.44 Studies by these tvo entities used a similar conceptual basis, which distinguishes between direct and indirect energy use, between average and marginal rates of energy use, and between the energy dissipated by the system under considera- tion and the primary energy input (coal, oil, or nuclear-generated heat) used.45 1 50 MACROECONOMIC ANALYSIS OF BENEFITS The estimates discussed here are for direct energy consumption. They do not include energy embodied in goods or services consumed by the telecommunications system (for example, the energy consumed in refining copper for cables). They do, however, account for all of the primary energy input used to supply the system's direct energy needs. The estimates are of marginal energy consumption, that is, the incre- ment of energy consumption associated with an increment of call traf- fic. They assume that system capacity is adjusted proportionately to such changes in traffic (so that the lumpiness in providing a new air- line scheduling data network or a new microwave route is smoothed out) and that overhead energy consumption by administrative, main- tenance, or similar functions in the telecommunications and trans- port corporations is fixed; it will not increase proportionately. The Canadian and British studies yielded similar results.46 The Ca- nadian study, in projecting the comparative energy costs in 1985 for a three-hour meeting between people in Montreal and Toronto, showed that the energy used by an audio conference call was far less than one-hundredth of that which would have been used if the persons had met face-to-face after traveling by railway, automobile, or airplane.47 Overall, the estimate of the savings that might be realized by replacing 20 percent of Canadian intercity business travel by telecommunica- tions amounted to 3 percent of the total energy consumed by the transport sector in Canada. This was equivalent to 1.3 percent of na- tional petroleum consumption, or about 0.8 percent of total energy use. A British study of a substitution for travel between London and Glasgow (400 miles compared with 325 miles between Montreal and Toronto) presented its results in terms of savings per kilowatt-hour. It showed that for a three-hour meeting including four people, two of whom would have to travel from London to Glasgow if travel were un- dertaken, the relative efficiency of a studio-based audio telephone conference hookup compared with railway, automobile, and air travel was 1:225, 1:800, and 1:1,250, respectivelvy48 A second study carried out in Great Britain compared the savings per kilowatt-hour of mak- ing a one-hour telephone call rather than traveling by automobile, railway, and air for several distances and estimated that the telephone was more cost-effective than rail and automobile by, respectively, ra- tios of at least 3:4 and 3:13 for 10 kilometers, .3:18 and 3:63 for 50 kilometers, and on up to 3:233 and 3:800 for 650 kilometers. The comparison at 650 kilometers for air travel was 3:1,241.49 Exercises have also estimated the potential for realizing petroleum savings by substituting telecommunications for interurban travel for TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 151 business meetings. Lathey estimated that just under 3 percent of the total 1974 U.S. petroleum demand could have been saved by substi- tuting interurban travel by car. The estimate was made using a U.K. estimate that 40 percent of business trips could be substituted by tele- communications and then assuming that the 40 percent estimate would hold for all interurban travel in the United States.50 Similarly, Tyler, using an approach like Lathey's, estimated that 40 percent of business trips in the United Kingdom could be replaced by tele- conferencing, and that if this were true about 0.3 percent of total U.K. national primary energy consumption would be saved.5' How far would the energy consumption equations be modified by including indirect energy consumption? In the absence of detailed energy input-output analyses of transactions between telecommunica- tions common carriers, transport firms and agencies, and their suppli- ers, no definitive answer can be given. However, there are indications that including indirect energy consumption would shift the compari- son further in favor of telecommunications. Studies of motor trans- port in the United States suggest that indirect energy consumption adds between 30 and 90 percent to the energy used directly.52 A com- parison of aggregate results from input-output studies suggests that the proportion would be considerably lower for telecommunications. A 1974 input-output study by INSEE in France estimated that a 100 percent increase in the price of all energy sources would increase the costs for telecommunications only 1.9 percent as against 12.7 percent for transport.53 Conclusion This brief literature review has supported intuitive notions that there are strong linkages between the transport, energy, and telecom- munications sectors. The extent of these linkages and the relative im- portance of the three sectors in any one country depend upon many geographical and economic factors. Industrial countries with relatively mature transport and telecommunications systems seem to have some potential for short-run changes in the relative use of transport and telecommunications as well as for a longer-term spatial reorganization of economic activity, which might save energy. However, in the absence of either major government restrictions on transport, petroleum, and so forth or further major price increases affecting transport, such changes are likely to be only marginal, at least for the next few years. For developing countries, however, the picture may be different. 152 MACROECONOMIC ANALYSIS OF BENEFITS The changes in the relative costs of transport and telecommunica- tions; the relative immaturity of both sectors; the widespread ineffi- ciencies in the transport sector resulting from information deficien- cies (unnecessary trips, empty return trips, badly timed trips); and the limited spatial dispersion of trade, commerce, and industry (also partly related to information deficiencies) all indicate that potential gains from increased coordination between the two sectors and from increased penetration of the telephone network into both urban and rural areas may be relatively large. The studies reviewed were limited, however, and mainly related to industrial countries. Although such studies showed that significant transport and energy savings are possi- ble, only the ITU study on the People's Democratic Republic of Yemen proposes a ready means for determining the optimum level or mix of investments in the telecommunications and transport sectors, assum- ing that market signals can be relied upon. In many countries, where prices for telecommunications and transport services or major inputs such as energy often do not approximate their costs of supply, effi- cient solutions are exceptionally hard to determine. Notes 1. International Energy Agency (1989) and World Bank (1989). 2. Portions of this chapter draw on a review of selected literature published on the substi- tution of transport and telecommunications by Michael Tyler and Emma Bird. 3. Early sources of information on technical and operational matters include Moon and others (1977); Lathey (1975); Schoppert and others (1960); and Stephens and others (1968). 4. Moon and others (1977). 5. Wall Street Journal (1983). 6. Lathey (1975). The rapid growth in the demand for mobile communication terminals in the United States, which in large measure reflects recognition of such potential effi- ciency gains, was acknowledged by a mid-1970s decision of the U.S. Federal Communica- tions Commission to double the range of the radio frequency spectrum allocated for land-mobile uses. Savings in vehicle mileage of from 30 to 40 percent have been reported as typical, and the rate of return on investment is high. 7. Pye Telecommunications, Ltd. (1976). 8. Medinikov (1975). A model was developed to attempt to estimate the total economic benefits associated with these kinds of changes, and the Regional and Vocational College of Agriculture in Rostov reportedly carried out a program to measure the numerical parame- ters from the actual performance of the Rostov experiment. 9. Gorelik and Efimova (1977). Both of the U.S.S.R. studies relied heavily on judgmental assessments of gains rather than detailed measurements of physical quantities before and after the use of telecommunications was provided. 10. Schoppert and others (1960). 11. Gordon and Wood (1970). TRANSPORT/COMMUNICATION/ENERGY CONSUMPTION 153 12. General Motors Research Laboratories and Delco Radio Division (1962); United Nations (1977); and Stephens and others (1968). 13. Favout (1970). Before such systems are introduced in any one area, three area- specific questions would have to be addressed. (I) A behavioral question: will the availabil- ity of an enhanced information device actually lead drivers to rationalize better their choice of routes and parking places? (2) Financial or market questions: will the system generate sufficient cost savings (public and private) to justify the public and private costs involved, and can the concept be sold to the individuals who would buy in-vehicle display units and the public agencies that would provide infrastructure? (3) Broader questions of economics and energy analysis: is the direct effect of introducing an electronic route guidance system clearly to reduce energy consumption? It is likely, although not demonstrated in detail, that the energy cost of such a system would be trivial. However, the improved traffic flow result- ing from electronic route guidance might encourage greater use of private automobiles, longer journeys, and increased traffic generally. 14. Kaul (1978). 15. Several short examples relating to the potential of telecommunications for fleet con- trol in developing countries were outlined in chapter 1. 16. ITU (1988). 17. Pool (1979), p. 181. 18. Basic research was carried out at the Communications Studies Group, University College, London, and at the Johns Hopkins University, the Stockholm School of Econom- ics, the London School of Economics, Bell Canada, and elsewhere. It was comprehensively reviewed in Short, Williams, and Christie (1976); Communications Studies Group (1975); and Johansen, Vallee, and Spangler (1979). 19. Tyler, Cartwright, and Collins (1977) and Tyler (1978). 20. Pye (1974). Several objections to this conclusion spring naturally to mind. For exam- ple, although travel substitution on a significant scale may be feasible, will the preference of individual business travelers allow it? Survey research in Canada and the United States ad- dressed this question. The results tend to show that the answer is yes-most of those busi- ness travelers who travel more than the average would prefer to reduce the amount of travel, and they account for by far the greater part of the total volume of business travel. See Kollen and Garwood (1975) and Pye and Weintraub (1977). 21. Kinsman (1987). 22. Williams (1991). 23. National Research Council (1985) and Stern and Holti (1986). 24. Some scattered evidence exists of cost savings derived from telecomnmuting. For ex- ample, see Hughson and Goodman (1986). 25. Johansen, Hansell, and Green (1981) and Charles (1981). 26. Tomey (1974). 27. Satellite Communication Services (1981). 28. Wall Street Journal (September 4. 1980), as reported in Charles (1981) p. 300. 29. Telephony (1982). 30. Trevains (1978). 31. Williams and Young (1977). 32. For a summary of much of this research, see Reid (1977) pp. 386-414. 33. Charles (1981) p. 298. 34. The model and the aggregation of the results to give predictions of the economic ef- fect of telecommunications investment at the national level are described in Gorelik, Efimova, and Kareseva (1975) and in Gorelik and Kareseva (1975). 35. The survey method relied heavily on the judgment of the respondents (industrial managers and specialists) about such matters as (a) the extent of substitutability of tele- phone calls for letters, telegrams, and visits; (b) the extent to which increased use of trunk 1 54 MACROECONOMIC ANALYSIS OF BENEFITS telephone systems had already saved costs through such substitution; (c) the extent of fur- ther substitution and cost savings expected to result from planned improvements in the quality and availability of trunk telephone service; and (d) the more subtle indirect gains in productivity. The problems with such judgmental surveys are well known, and in a planned economy where respondents may have an incentive to emphasize their need for improved fa- cilities, such problems may be even more significant. There are other problems also. For ex- ample, it is not clear what benchmark of comparison is used in the published work to assess the productivity benefits of existing long-distance telephone service, although the benchmark seems to be a state of zero utilization. 36. Cleevely and Walsham (1980) pp. 21-22. 37. A World Bank project analysis in Pakistan in 1969 also indicated that substitution does indeed occur in a developing country and that even at the lower wage and salary levels, the additional staff time and fuel costs involved are often as much as three times the cost that would have been incurred by using the telephone service to do the same job. 38. Chu, Srivisal, and Supadhiloke (1985). 39. Respectively, Pierce (1962) and Hall (1969). 40. Fordyce (1974). 41. Tyler (1976). 42. Clark and Unwin (1981). 43. The development of low-cost portable audio conference terminals, such as the Video- graph developed in France or the Electronic Blackboard developed in the United States, and the use of slowscan video to transmit images over regular voice telephone lines have been proposed as options to supplement audio conferencing in developing countries. Such subscriber apparatus can transmit handwriting, computer graphics, or simple visuals simul- taneously with voice among several locations over ordinary telephone lines. Although this kind of system has been experimented with in developing countries, implementation has proven to be technically complex, and few systems are currently in operation despite nearly ten years of availability. See Goldschmidt, Tietjen, and Shaw (1987). 44. Respectively, Tyler, Cartwright, and Bush (1974) and Katsoulis (1974). See also Katsoulis (1976). 45. Since the parameters of the calculations are largely an engineering matter, the results for a developing country are not likely to be enormously different, except perhaps that given their relatively inadequate telecommunications networks and varying geographic and social conditions, the existing energy efficiency of the transport sector may be appreciably lower. 46. It is important to remember the linmitations of the existing comparisons between transport and telecommunications as a basis for estimating overall energy savings. For ex- ample, the energy studies compared the energy consumption for a single teleconference with that involved in travel to an equivalent face-to-face meeting. Business trips, however, often conmbine several meetings, especially in cases of long-distance travel. A Bell Canada survey of intercity business travel showed that the average trip included 2.7 meetings, but this ef- fect was less pronounced for the nmuch more numerous short trips. See Kollen and Garwood (1975). 47. Tyler. Cartwright, and Bush (1974). 48. Pye, Tyler, and Cartwright (1974). 49. Tyler, Cartwright, and Bookless (1974). 50. Lathey (1975). 51. Tyler, Elton, and Cook 1977). Primary energy was defined as energy available to the final consumer (secondary energy) plus conservation losses and waste in the industries that supply energy. 52. Tyler, Cartwright, and Bookless (1974). 53. Tornato (1974). Part III Microeconomic Analysis of Benefits Chapter 8 Price Change and Best-Alternative Estimates of Consumer Surplus THE RELATION BETWEEN telecommunications investment and economic activity and development is highly complex. The survey of the litera- ture and experience presented so far suggests that the benefits of a specific telecommunications project or investment program cannot readily be identified and measured by the aggregate international comparison of input-output tables, by the analysis of relations be- tween GNP and telephone availability or usage, or, apparently, by analy- sis of national transport, energy, or locational objectives. This part of the book therefore focuses on the experience of microeconomic anal- vsis in addressing the issue of telecommunications benefits. Cost-benefit analysis can be used to help determine not only the amount of resources that should be devoted to telecommunications but also the best way to allocate those resources within the sector, that is, local versus long-distance facilities, urban versus rural, public call offices (Pcos) versus subscriber telephones, alternatives for new types of service, regional choices, and so on. Traditional economic theory prescribes that investment in every sector should be increased until all projects that yield rates of return in excess of the opportunity cost of capital have been financed. The capital rationing invariably encountered in developing countries requires that the rate of return produced by expanding the telecommunications sector should be compared with the returns offered by the best-alternative investment program that would be implemented if the funds were not spent on telecommunications; this in effect represents the real opportunity cost of capital. Because of artificial restrictions on the growth of telecom- 157 1 58 MICROECONOMIC ANALYSIS OF BENEFITS munications and massive unsatisfied demand, in most developing countries, the telecommunications sector offers scope for expansion according to this criterion, that is, rates of return to investment in telecommunications tend to be relatively high.' Some argue, however (as noted in chapter 1), that these high finan- cial rates of return may reflect benefits accruing only to higher- income groups of society and that expansion of the sector merely exacerbates income inequality. To examine the validity of such argu- ments normally requires supplementing the evidence on willingness to pay for telecommunications investments with more qualitative analy- sis of the nature and likely effect of telephone usage. Hence, two kinds of information should usually be reviewed to determine the ef- fects of a particular project or program in a particular developing country. First, the estimated investment and associated operation and maintenance costs should be compared with the estimated benefits to the users of telephone and other services. These estimates should be based on their willingness to pay for access and calls and perhaps on their willingness to travel to attain services, to purchase or rent con- nections from intermediaries, or to pay more to communicate in the absence of a telephone or other telecommunications service. Second, information should be gathered on the characteristics of the persons and groups benefiting from the improved and expanded services and on how those beneficiaries use them; these data should then be qualitatively analyzed in light of that country's national develop- ment policy goals and programs. The first point is addressed in this and the following chapter; the second is the subject of chapters 10, 11, and 12. Internal Rate of Return and Consumer Surplus As noted above, identifying and measuring the economic benefits relative to the costs of a proposed telecommunications project or in- vestment program typically require observation of how much subscrib- ers and callers actually spend or are willing to spend to have access to and use telecommunications services. An assumption underlying this approach is that decisionmakers (individuals, households, firms, gov- ernment agencies, and other organizations) are rational and under- stand their activities better than outside planners, so that the amount of money they are demonstrably willing to spend on telecommunica- tions services is at least a minimum measure of the worth of those services to them. PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 159 The first step in such a benefit and cost calculation is to estimate the incremental revenues that will be generated by the additional tele- communications investment being proposed (a proxy for benefits) and then compare that amount with the investment, operating, and maintenance costs that will be incurred to generate them. Such a comparison, of course, must be Inade through time since the costs in- curred and the revenues produced by the new investment are associ- ated with both the present and the future. Hence, the principle of discounting plays a part; a given unit of revenue received today is of greater value than that same unit received five years from now; like- wise a given unit of cost incurred today involves giving up more than if that same unit of cost were incurred in five years. Given this, the economic criterion for the acceptability of a project is determined by estimating the present (discounted) value of the project's benefits net of its costs, where the benefits and the costs are defined in incremental terms compared with what the situation would be without the project. Risk and sensitivity analysis also must be in- corporated in such predictions. In the telecommunications sector this analysis is usually presented in terms of an estimated internal rate of return (IRR) on the new investment.2 The IRR is defined as the rate of interest (discount rate) that would be required to yield a net present value of zero.3 Before such a return is calculated, however, at least two types of adjustments are necessary to ensure that the data on finan- cial revenue and cost reflect the real costs and benefits of resources according to the national allocation of resources.4 First, adjustments must be made in the prices at which inputs (such as equipment purchases, foreign exchange, local currency, or labor) are valued, to compensate for any distortions in the price sys- tem. The intention is to reflect the true social cost or forgone bene- fits associated with each type of input used in terms of its scarcity and the opportunity cost of diverting it from other uses. This can be done by using shadow prices to value inputs.5 Shadow prices are dis- cussed in more detail in appendix D. Second, items that are considered to be revenues and costs by the telecommunications operating organization, but that are pure transfer payments from one entity to another and do not represent flows of real goods and services in the economy, must be removed from the cost and revenue streams. For example, government taxes and cus- toms duties are a cost to the operating entity but are a pure transfer and not a real cost according to the national use of resources.6 The IRR resulting from the above calculation. which essentially uses revenues from the operating entity as an estimate of the benefits that 160 MICROECONOMIC ANALYSIS OF BENEFITS consumers derive from telecommunications services, usually underes- timates the economic worth of those services because a. Subscribers or Pco users may value the service more highly than the amount they are required to pay for it, that is, there might be consumer surplus that is not quantified b. New telephone subscribers not only incur benefits for themselves but also increase the benefits of being connected to the system for current subscribers; that is, there are subscriber-related externalities c. The willingness to pay a given price to make a telephone call re- flects only a minimum estimate of the benefits incurred by the caller and does not reflect the benefits received by the callee or those who the caller or callee then contact, that is, there are call-related externalities.7 Points b and c are discussed in chapters 14 and 15. With regard to point a, documentation exists that in developing countries telecom- munications subscribers or Pco users are often charged less than they are actually willing to pay for telecommunications services, or they in fact pay higher prices for the services than are recorded by the finan- cial receipts of the telecommunications operating entity.8 In other words, the benefits that consumers perceive to receive from the ser- vice exceed those measured by the revenue received by the telecom- munications entity. This difference between what users are prepared to pay-the real measure of the benefits they perceive-and the reve- nues actually collected is known as "consumer surplus."9 The concept of consumer surplus has been relied upon extensively in attempts to supplement the IRR as an indication of the benefits of proposed telecommunications investment. Three methodologically different, but in practice interrelated, approaches have generally been taken at the level of a project or investment program to measure a portion of the consumer surplus: a. Methods based on observing the consumption effects of price changes b. Methods based on comparing the difference between the cost of carrying out a given activity using telecommunications and the cost of the best-alternative means of communicating c. Methods based on estimating more completely the costs that telecommunications users actually incur when communicating. For brevity, these are referred to as the price-change method, the best-alternative method, and the expenditure method. PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 161 Estimating Consumer Surplus by the Price-Change Method Observing the response of users to a change in price can provide some information on consumer surplus at that time."' For instance, if call traffic is not congested, the price elasticity of demand of tele- phone call traffic might be estimated by observing the change in traf- fic associated with a change in the charge for calls. With this information, a portion of the demand curve for calls can be esti- mated, and an impression of the consumer surplus associated with calls can be obtained. In developing countries, however, where demand is seldom fully met, observations of price and quantity rarely represent actual points on demand curves. Nonetheless, since what users actually pay for ser- vices reveals at least part of their valuation of the benefits received, even the empirical observation of points in the supply curve (al- though probably not on the demand function) can yield some insight into consumer surplus. The following two examples illustrate possibili- ties for partially estimating consumer surplus using the price-change method. Example 13. Increase in Call Charges (Costa Rica 1976) Consider figure 8-1, where DD' is the demand curve for telephone traffic. At price p, per metered pulse, calls are made that result in q, metered pulses a day. When the pulse price is increased to p,, traffic drops to q. pulses a day. Since before the price in- crease telephone users were willing to pay at least p. for each of the first q, pulses, but were only asked to pay PI, they were receiving a consumer surplus at least equal to (p. - pl)q, for these calls. For the remaining (q, - q,) calls, they received a consumer surplus equal to the area of the triangle ABC below the demand curve Dy'. The sum of both consumer surpluses is shown in the shaded area in figure 8-1. If the observed price change is small, the straight line through points A (p,q,) and B (p'q,) would closely approximate the corresponding segment of DD' and the area of ABC below the curve would therefore be approximately equal to (I / 2) (P2 - PI) (q1 - q7). The total benefits received by users making q, calls were at least equal to what they paid (p,q,) plus these consumer surpluses. Actual benefits were even higher, however, because the total consumer surplus also includes the area above p. and below DD' in figure 8-1. A better (but still conserva- tive) estimate of consumer surplus can be obtained by adding the area of the un- shaded triangle p,Ba to the shaded area described above. It can thus be shown that total benefits equaled or exceeded I - (1/2e) times the revenue (p,q), where E is the price elasticity of demand for pulses (a negative number) in the neighborhood of price p,.1 I 162 MICROECONOMIC ANALYSIS OF BENEFITS Figure 8-1. Estimating Consumer Surplus from the Effect of a Price Change D a P2 B 0~ ~ q q1 Such a method was used in 1975/76 to examine the demand fot public telephone use in rutal Costa Rica. An analysis of ttaffic data for ninety-two rca telephones in eighty-two small villages befote and after a 25 petcent increase in call charges (from p,1 12 centimos per pulse to p. 15 centimos) generally indicated that the average price elasticity of demand for pulses in the eighty-two villages was about -0.5. Hence, a village generating q1 = 100 pulses a day would have reduced its traffic to q. = 87.5 pulses a day shortly after the call charges increased. At the initial price p,, the con- sumer surplus on the first 87.5 pulses would have been at least 3 centimos per pulse or 262.5 centimos (2.63 colones) a day. The remaining 12.5 pulses would add a further 18.75 centimos of consumer surplus to make a total of 2.81 colones a day. Since at p,, PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 163 revenues were 12 centimos per pulse times 100 pulses a day, or 1,200 centimos (12.00 colones) a day, the measured revenues received by the telephone operating entity rep- resented only about 81 percent of the quantified benefits perceived by the telephone users. This, however, still understates the benefits. Using the better estimate of consumer surplus, initial telephone revenues generated by the Costa Rican village were found to represent at most only half the benefits perceived by the population. For further de- tails, see appendix C. Example 14. Changes in Real Tariffs because of Inflation (El Salvador 1977) In an economy affected by price inflation, prices in real terms paid for telecommu- nications services decrease during the periods between increases in tariff levels. Look- ing back through time, this means that during periods of price inflation telecommuni- cations users revealed a willingness to pay higher real prices than those being charged today. Looking forward, users will increasingly be charged less in real terms than they pay today, until tariffs are adjusted upward. These facts can be used to quantify part of the consumer surplus for telephone service, as in the following case relating to El Salvador's 1978-82 telephone development program.'' In 1977 El Salvador had long waiting lists of individuals and business firms wanting telephones. Telephone tariffs had remained virtually unchanged for more than ten years, during which time general price inflation was significant. Hence, telephone users had paid considerably higher prices in real terms (relative to other goods and services) than they were currently paying for telecommunications services. Given this, consumer surplus was partially estimated by tabulating the prices in real terms that ex- isting consumers had been willing to pay when they acquired telephone service and by assuming that in the existing situation of excess demand new subscribers would ex- hibit characteristics similar to those of the average existing consumer. With one minor exception, telecommunications tariffs had not changed since 1964. Between 1964 and 1977, however, domestic consumer prices inflated approximately 80 percent. As a result, subscribers using telephones in 1964 demonstrated in real terms that they were willing to pay at least 80 percent more for telephone service with lower quality and more limited access than current subscribers were paying. To estimate the prices in real terms that all existing telephone subscribers were willing to pay when they received service between 1964 and 1977, the average price paid in real terms and the quantity of new telephones supplied as well as the existing supply were taken to represent a point on the telecommunications supply curve at the end of each of the thirteen years. This point lay below the demand curves that actually existed at those times, since even at the higher real tariff, large excess demand (waiting lists and traffic congestion) was expressed during each past year. Given these tabulations of historical real prices and quantities, a weighted average price paid by present consumers when they joined the network was calculated. Given this estimated average price that pres- ent subscribers had in the past demonstrated a willingness to pay, the IRR on the future 1978-82 telecommunications investment program was estimated to be 36 percent. This rate of return might still underestimate total benefits, however, since (a) the benefits to subscribers connected in past years were only measured by what they actu- ally paid, with no allowance made for their consumer surplus at those times, (b) new subscribers to be added under the 1978-82 investment program would obtain a better 164 MICROECONOMIC ANALYSIS OF BENEFITS quality of service (less congestion, improved transmission performance, fewer faults, and so forth), and (c) new subscribers would be gaining access to a larger number of other subscribers than was the case in earlier years. Even so, the estimated 36 percent rate of return was more than double the IRR of 16 percent that was estimated if only the projected financial returns to the telephone operating company were considered.'3 Estimating Consumer Surplus by the Best-Alternative Method Instead of observing the responses of telecommunications subscrib- ers and other users to changes in price, a portion of the consumer surplus can be measured by estimating how much a user gains by using the telecommunications service rather than an alternative means of communication. Therefore, this section examines how much individuals or groups save by having a telephone, or, alternatively, how much they spend by not having ready access to a telephone.14 Five examples are reviewed. In one, a telephone demand function is derived and then tentatively verified by estimating the amount of money that at least some users spend communicating by using the next-best means in the absence of telephone service. In the second, the cost savings realized by using a telephone rather than the next- best medium are calculated for telephone users in two rural areas. In the third, consumer willingness to pay is estimated by examining al- ternative postal costs and priority access surcharges on long-distance calls. In the fourth, partial transport cost savings resulting from the installation of one telephone at each of two rural factories in Bangla- desh are estimated. In the last example, the savings realized by using a telephone rather than bus transport are examined for thirty-six vil- lages in India. Example 15. Derivation of Demand Functions Based on Cost Savings for Communications (Pakistan 1968) In this example a portion of the economic benefits associated with Pakistan's 1968-72 telecommunications investment program was measured. As a first step the demand curve for telephone services was estimated.!r To account for benefits associated with making local calls as well as those incurred solely by being connected to the network, the price variable was defined as a typical subscriber's annual expenditure on local telephone service; it comprised the annualized initial connection charges, one year's rental, and the charges for a typical 3,000 local calls. Under prevailing tariffs at the time, this amounted to Rs880. With PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 165 regard to the quantity variable, at the beginning of the investment program the de- mand for telephone connections was estimated to be 196,000, of which 98,000 were in service, 65,000 were outstanding applications, and 33,000 were estimated as un- reported demand. This, together with the annual price estimate (Rs880), fixed one point on the demand function. The demand curve itself was arbitrarily assumed to be a straight line. Based on available data from other countries and on qualitative consid- erations of the differences in demand composition between those countries and Paki- stan, a price elasticity of demand of -0.25 at the identified point of Rs880 and a quantity of 196,000 connections demanded were assumed. Since the shape and slope of the demand curve were largely based on judgment, an independently derived control was considered necessary to test the demand curve for overall validity. Hence, visits were made to several industrialists who had established factories on the outskirts of towns or in new industrial areas before 1968 and who were still waiting for telephone service. From interviews with these businessmen and information about employee salaries and transport costs, it was roughly estimated how much these industrialists were spending due directly to their lack of a telephone. The following is a composite case of what was found. A business firm located about 4 miles from the center of a large town could be reached by car in about fifteen minutes. junior executives at this establishment re- ceived a salary of about Rs2,000 a month, which, with 60 percent allowance for pen- sion, leave, overhead, and so forth, and assuming 175 working hours a month, resulted in an average cost per executive-hour of about Rsl9. Lacking a telephone, these exec- utives often had to contact suppliers, buyers, merchants, transporters, government of- ficials, and so forth in person. On average, they were out of their offices for such meetings about half a day every second day. It was estimated that, if they had tele- phones, these absences could have been cut to about half a day per week.16 An esti- mate of the direct savings from having a telephone for an executive was therefore considered to be, first, the value of the time involved in traveling downtown, moving between meetings, and waiting for individual interviews due to the difficulty of prear- ranging and verifying times, and, second, the costs of transport for the journeys. Using a conservative estimate of two and one-half hours a week for lost time, at Rsl9 an hour, the value of time saved by having a telephone to make local calls would have been about Rs2,500 a year per executive. Transport costs, which were high partly because of import duties on vehicles and gasoline as well as heavy traffic congestion on downtown streets, might have been eval- uated at anything from Rs200 to Rs500 a year depending on whether allowance was made only for the short-run marginal costs involved in these trips or for the full eco- nomic cost of transport, including depreciation of vehicles, and so forth. Hence, con- sidering the time and transport cost saved, at least some telephones that were to be added under the 1968-72 telecommunications investment program could have been valued at about Rs2,700 to Rs3,000 a year."7 This amount was much higher than the price at which demand would have been equal to current supply according to the de- mand curve that was initially estimated. Since industrial subscribers were given priority, the foregoing information was rele- vant for the major uses to be made of the expanded telephone network. The data sug- gested that the demand curve initially specified probably represented a conservative estimate of demand. Hence, the demand curve could be used to calculate a somewhat conservative estimate of the benefits of Pakistan's 1968-72 telecommunications in- vestment program. The IRR on the program was estimated at 19.1 percent. Adding the consumer surplus estimated by using the above demand curve approximately quadru- 166 MICROECONOMIC ANALYSIS OF BENEFITS pled this figure. With a shadow price of foreign exchange equal to twice the official price and with the customs duties on imports paid by the telecommunications entity removed, the estimated economic rate of return on Pakistan's 1968-72 telecommuni- cations investment program exceeded 50 percent. Example 16. Cost Savings for Communications in Two Rural Areas (Chile 1978) In 1978 the costs and benefits of telephone service were studied for two rural areas in Chile.'- One area included San Vicente de Tagua Tagua, a town with 239 subscrib- ers and three Pca telephones in neighboring small villages. The area is located in a prosperous fruit-growing region and is well integrated into the modern sectors of the national economy. The other area included Cabildo, a town with sixty-four subscribers and two Pco telephones in neighboring small places. This area is in a relatively de- pressed, drought-prone, small-scale mining and cattle-grazing region. To collect information on the perceived benefits of Pco long-distance telephones, 520 calls were immediately followed by a structured interview of the caller. From the information collected a portion of the cost incurred by each caller in making each call was estimated; this comprised the price paid for the call and estimates of the cost of round-trip transport and the time required to reach the telephone from the caller's place of residence or work.'9 The caller was also asked whether, for that particular call, he would have resorted to an alternative means of communication if a telephone had not been available.20 A portion of the consumer surplus associated with each call was then estimated as the monetary savings obtained by using the telephone rather than the least-cost alter- native means of communication. No consumer surplus was computed in the few cases in which the alternative means would have cost less than the call actually placed.' Likewise, no consumer surplus was calculated when the alternative was not to commu- nicate. Given the callers' reluctance to discuss income and the fact that in these rural areas a part of income is nonmonetary, seasonal, and difficult to estimate without de- tailed studies, no attempt was made to assign a monetary value to the estimated time savings obtained by using a telephone.'2 Three alternative measures of benefits were calculated from the PcO telephone user Table 8-1. Internal Rates of Return froni Investments in Rural Pco Telephones in Chile, 1978 (percent) With tax, shadow price, and partial With tax and consumer shadow price surplus Region Financial adjustment adjustment Cabildo 2.4 4.3 9.6 San Vicente 13.3 16.4 21.5 Source: Adapted from Nicolai and Wellenius (1979). PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 167 survey: an [RR, which measured the financial return on the investment in rco tele- phones by the operating company; an IRR, after taxes were removed from both the cost and benefit streams, and foreign exchange and unskilled labor were shadow priced; and an estimate of the economic rate of return, where, in addition to the tax and shadow price of adjustments, the estimated consumer surplus (a comparison of the sum of call charges and transport costs with the cost of the alternative communica- tion medium that would have been used) was added to the benefits stream. Table 8-1 summarizes the results of the three rate-of-return estimates for the sampled Pcos in the two regions. For both village areas the estimated economic returns on the Pco telephones were greater than the financial returns on the investment to the telecommunications opera- ting entity, and the returns on the Pco investment in the more economically prosper- ous region of San Vicente were the largest. This latter result is consistent with the findings of the cross-village correlation analysis in Costa Rica reviewed in example 12. Example 17. Demand for Telephone Calls Based Partly on Costs of Postal Communication (India 1969) This study addressed the substitutability between telephone calls and letters.23 In 1969 in India the cost of sending a message by a local letter was estimated as a mini- mum of 45 paisa (postage was 15 paisa and typing, handling, and stationery were at least 30-35 paisa). It was assumed that important telephone conversations would gen- erally substitute for at least two letters: an initial letter and a reply. It was further as- sumed that telephone users on average considered at least 5 percent of all calls made to be important enough to warrant paying as much as the cost of a letter and a reply; other calls could be replaced by one letter or other means of communication, or they would simply be eliminated. Given these assumptions, one point on the demand curve for local telephone calls was estimated to be at a quantity equaling 5 percent of all local calls and a price of 90 paisa. A second point on the demand curve was derived under the assumption that the value of the most valuable local call was at least eight times the charge for a local call. This assumption was based on the fact that, on the long-distance manual trunk net- work, some callers were opting to be placed in a "lightning call" category under which the P&T telephone operator placed the call immediately rather than putting the caller at the end of the queue where he would wait up to four hours for a connection. The charge for this service was eight times the normal long-distance rate. Since the charge for local calls was only a fraction of that for trunk calls, the assumption that the value of the most valuable local call was also at least eight times the standard charge for a local call was considered to be conservative. Finally, a third point on the demand curve for local calls was taken as the number of local calls made each year at the actual charge rate of 15 paisa. An S-shaped de- mand curve was then drawn through the three points. The curve as drawn generally implied that local calls per direct exchange line would fall by about 50 percent if the local call charge were increased threefold (demand was relatively inelastic). Given this demand curve, the consumer surplus above the call charge rate of 15 paisa was estimated. Without consumer surplus, the IRR on India's P&T 1969-72 tele- communications investment program was estimated to be 6.5 percent. With the con- sumer surplus added to the revenue stream, the economic rate of return on the program was conservatively estimated to be about 25 percent. 168 MICROECONOMIC ANALYSIS OF BENEFITS Example 18. Cost Savings for Transport at Two Rural Factories (Bangladesh 1981) In early 1980 a radiotelephone was installed at both the Nabaran Jute Mill (NJM) and at the Ghorasal Fertilizer Factory (CFF) outside Dhaka. The two telephones were linked directly to a central telephone exchange in Dhaka. Data on both telephone usage and cost savings were gathered through interviews with the general managers and several senior staff of both enterprises. Although the data are considered indica- tive, they were not confirmed through independent sample observation. Based on these data, for one year, before and after differences in direct expenditures on gasoline and on the wages of managers communicating with Dhaka were estimated. Local gasoline prices were assumed to be roughly in line with world market prices. The initial data and the estimated expenditure savings are presented in table 8-2. These differences in before and after expenditure include only direct cost estimates related to gasoline and some transport time related to management wages. A more complete cal- culation would include savings in vehicle capital, other operating and maintenance costs, and other direct and indirect expenditures. The capital cost of one very high-frequency radiotelephone installed in each factory plus a proportionate share of the base station cost of the receiver/transmitter at a cen- tral exchange in Dhaka were roughly Tkl90,000.4 It was not considered necessary to shadow price foreign exchange. Each factory reimbursed the Bangladesh Telephone and Telegraph Board for this cost when the telephone was installed. Installation cost was about TklO,000, and annual maintenance cost beginning after one year of opera- tion was assumed to be Tk2O,000. These costs were also reimbursed by the factories. Based on available data, it was assumed that ninety-six outgoing telephone calls were Table 8-2. Factory Expenditure Savings Data from Bangladesh, 1981 Item Nj OFF Approximate distance from Dhaka (miles) 20 40 Average number of round trips to Dhaka a week Before installation of telephone IS 11 After installation of telephone 2 1.5 Assumed number of working weeks a year 48 48 Average round-trip gasoline cost (taka) 350 650 Average total round-trip time (hours) 4 7 Average number of staff making trip (including driver) 3 3 Average wage a month of management staff making trip (taka) 1,000 1,200 Average number of working hours a week 42 42 Gasoline expenditure savings (taka) 268,800 296,400 Reduced wage expenditure for transport communication for two management staff (taka) 36.557 39,197 Total expenditure difference (taka) 305,357 335,597 Note: NiNM, Nabaran Jute Mill; CFF, Ghorasal Fertilizer Factory. Source: World Bank data. PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 169 made each week at TkO.75 a call (roughly reflecting the Board's operating cost), tele- phone operation costs to the factories (and the Board) averaged about Tk3,456 a year, and a similar number of telephone calls were received by each factory each week with the TkO.75 call charge being paid by the caller. The additional capacity cost that the network as a whole incurred as a result of the two additional telephone connections was assumed to be trivial. Given the above, the cost of telephone service to each factory during the first year of operation was approximately Tk203,456. The direct communication expenditure savings by NJ1i and GFF was Tk305,357 and Tk335,597 a year, respectively; in the first year direct communication expenditure savings exceeded costs by about 1.5 times at NJM and 1-6 times at GFF. During the second and each succeeding year, quantified di- rect savings exceeded costs (Tk23,456) by about thirteen times at NJM and fourteen times at GFF. Example 19. Bus Transport Cost Savings for Telephone Users (India 1981) A survey of 174 users of public telephones in thirty-six villages in six districts in Andhra Pradesh State in India found that 120 of the users regarded their most recent rco telephone call to have been so important that they would have traveled to the des- tination of the call if the public telephone had not been available.25 Hence, the cost that would have been incurred by each caller to travel to the call destination rather than use the telephone was estimated. Travel cost was taken as the fare for the cheap- est means of travel-by bus. The value of the time lost by making the journey was conservatively estimated by using the prevailing wage rate for unskilled labor. Table 8-3 summarizes the results of the analysis. For the 120 Pco telephone users who judged that their most recent call was urgent enough to warrant travel in the absence of a telephone, the estimated consumer sur- plus exceeded the cost of the telephone call between 2.5 and 5.5 times, depending on the distance involved. Table 8-3. Expenditure for Communiicationi bv Bus and by Telephone for Thirty-six Villages in Andhra Pradesh State, India, 1981 Cost of Value of Total Consumer Expenditure bus iare tinie lost bus surplus Distance Average on and bus through travel for called distance telephone terminal bus travel cost telephone (kilometers) (kilometers) calls (Rs) access (Rs) (Rs) (Rs) call (Rs) 0-20 11.24 1.37 4.53 2.00 6.53 5.16 20-50 34.57 3.54 8.45 4.00 12.45 8.91 50-100 80.54 4.56 16.19 8.00 24.19 19.63 100+ 149.00 5.44 27.69 8.00 35.69 30.25 Source: Adapted from Economics Study Cell, Posts and Telegraph Board, India (1981). 1 70 MICROECONOMIC ANALYSIS OF BENEFITS Notes 1. Examples of these high rates of return are presented in chapters I and 9, as well as in this chapter 2. In the absence of good data, estimating such costs and revenues through time and cal- culating the resulting IRR can be highly sensitive to the assumptions made. For example, one analyst may predict rapidly growing demand and falling unit costs because of new technol- ogy and may assume that the telecommunications entity has the technical and managerial capacity to expand rapidly Another analyst may predict less rapidly rising demand, constant unit costs, and a much lower implementation capacity of the telecommunications entity In such an instance, the two analysts would reach different conclusions about the size of in- vestment programs that could be justified and the extent of that justification, even though both followed the same analytical procedures; that is, the results of an IRK calculation to a certain extent depend on the assumptions used. Hence, the importance of sensitivity analysis. 3. Another way of stating this is that the PR is the discount rate that equalizes the pres- ent values of the project's cost and revenue streanis, that is, itK = r when X (R, - C) (I - r) I= where R, is revenue and C, is cost in time period t. 4. For a more general discussion of investment appraisal among alternative uses of funds and how choice depends on the decisionmaker's objectives and attitudes toward risk and time, see Littlechild (1979), chap. 7 5. If a more detailed knowledge of the size and incidence of project benefits is available, social weights might also be used to value project output. See Squire and van der Tak (1975). In telecommunications projects, however, such information is rarely available. Fur- thermore, although the concept of social weighting attracted considerable interest in the 1970s, little progress has been made in its practical applicationl to project analysis generally 6. In some instances two additional problems must be dealt with when revenues are used as a minimum measure of benefits for telecommunications projects. First, the telecommuni- cations price structure may not correspond to the real costs incurred in providing the added or improved service. Therefore, reliance on willinigness to pay as the sole justification for in- vestment may send misleading investment signals. Second, reliance on willingness to pay may not explicitly consider the achievement of other national goals, such as income re- distribution, the generation of public savings, regional developnient, and so forth. These problems are considered in following chapters. 7. In some cases such externalities have "public-good" characteristics, that is, excluding noncallers or noncallees from reaping at least some external benefits (direct or indirect) of telephone service is not feasible. However, given the generally accepted definition of public goods-that "each individual's consumption of such a good leads to no subtraction from any other individual's consumption of that good" or that they "must be consumed in equal amounts by all" or be "a good for which the resource costs are not attributable to bene- ficiaries"-it would seem that telecommunications services that are sold to individuals (telephone rental, call charge, and so forth), whose consumaption on the whole precludes others from that consumption, should be generallv categorized as private goods, with which externalities are associated in some instanices, rather than "public" goods. See Burkhead and Miner (1971), chap. 2; McKean (1968), pp. 43-47; anid Arrow (1970). Nevertheless, some advocates of government subsidies for telecommunications services have focused ex- clusively on the public-good characteristics of telecommunications investmnent externalities and have based their subsidy arguments on a contention that telecommunications services are or should be treated as public goods. See, for example, Hudson and others (1979), PRICE CHANGE AND BEST-ALTERNATIVE ESTIMATES 171 chap. 4, and Pierce (1979). Of course, on theoretical grounds the existence of externalities alone, whether or not they have public-good characteristics, can be enough under some cir- cumstances to suggest subsidies, although in developing countries where demand for tele- communications services generally far exceeds supply, financial subsidies of general telephone service are very difficult to justify on economic grounds. 8. Several examples in this and the next chapter are illustrative. 9. The concept of consumer surplus is closely related to the concept of the downward- sloping demand curve. For example, a consumer (subscriber) would generally be willing to pay a higher price for only one call in a day than for the twentieth most important call of that day. Since telephone call charges are usually fixed, however, consumers pay the same amount for all calls. Therefore, the caller receives a "surplus" of benefit (satisfaction) for all calls for which he would have been willing to pay a price greater than that which he was ac- tually charged. The same concept, of course, holds for connection fees and monthly rental charges. 10. Of interest to all telephone administrations is the question of how changes in the number of calls made or connections requested (Aq) relate to changes in price (Ap). To avoid possible confusion associated with different units of measurement and the absolute sizes of the variables, this responsiveness of quantity demanded to price is usually repre- sented in terms of a price "elasticity" coefficient defined as e = (dq / dp) (p i/ q). For a simple approximation, an arc price elasticity of demand can be calculated by dividing the percentage change in the number of calls made (or connections requested) by the per- centage change in price, defined as E = (Aq / Ap) -(p / q). Given this estimation, if the unit call charge (or connection fee) is increased 50 percent and the number of calls made (or connections requested) then falls 25 percent, the price elasticity coefficient would be -0.5. Demand is said to be inelastic if the absolute value (that is, not taking into account the sign) of the coefficient is less than 1, elastic if the co- efficient is greater than 1, and of unitary elasticity if it is equal to 1. 11. In figure 8-1 let the straight line AB be p = a - bq (a and b are constants). Differen- tiating both sides gives dp / dq = -b. Substituting dp / dq = (I / E) (p, / q) (from the defi- nition of E) yields -b = (I / E) (p1 / q,). Substituting this expression of b in the equation of line AB then gives a = p II - (I / c)J. The consumer surplus CS at price p. estimated from the area of the triangle p,Aa is CS -Ž q, (a - p1); substitution of the above expression for a yields CS Ž p1 q, / 2E. Revenues R are given by R = p, q,. Hence the ratio of total bene- fits to revenues is (CS + R) / RŽ - I / (2E). This ratio equals 3.0 for E = -0.25, 2.0 for £ = -0.5, and 1.5 fore = -1.0. Under the (usual) assumption that the demand function is convex toward the origin, the tangent will lie wholly below DD', and benefits will continue to be underestimated by some margin; this is acceptable since the calculation of an eco- nomic rate of return is usually used in project analysis to test against low returns. 12. The consumer surplus exercise outlined for El Salvador in this example wvas originally undertaken with Mihkel Sergo. 13. A second estimate of the economic rate of return was obtained bv assuming that new subscribers to be connected during 1978-82 would be willing to pay at least what was being charged in real terms to subscribers connected in 1977. This alternative estimate of the rate of return was betveen 31 and 33 percent, depending on whether the connection fee for new subscribers was taken to be that prevailing when they applied for service or that prevailing when they were actually connected. 14. In concept, this approach is only useful for estimating consumer surplus under fairly strict conditions. For example, if the user has no preference between a telephone call and 1 72 MICROECONOMIC ANALYSIS OF BENEFITS some other means of achieving a given result, for example, a personal visit, then the differ- ence between the cost of the telephone call and that of the alternative can (assuming the alternative is more costly) indicate the size of the consumer surplus. 15. This example is based on a study undertaken by Christopher Willoughby. At the time, the official exchange rate for the Pakistan rupee was PRsl,00 = $0.21. 16. Letters were not believed to be a relevant alternative to the telephone and face-to- face meetings for local communication, that is, within a metropolitan area. They were pri- marily used to convey documents and confirmatory notes, which had to be in written form. In Pakistan, messages borne by a private messenger may be, in some cases, a poor but par- tially relevant substitute for telephone communications. However, messages were not be- lieved to be an important substitute in the type of composite business dealings described in this example. 17. Actually, some telephones would be valued higher, since the figures used reflect something of a group average, not extreme values, and a single telephone would benefit more than one executive. The original demand curve assumptions were also checked using information on private transfers. 18. Nicolai and Wellenius (1979). 19. Like the previous exercise, this one involves a mix of the best-alternative method and the expenditure method of estimating consunier surplus. The exercise up to this point in- volves the "expenditure method" of benefit analysis. 20. More than 80 percent of calls were made by local residents or workers. Use of ac- cepted alternative means thus involved an additional trip somewhere out of town, which could in many cases result in carrying out several other activities besides communicating. Hence, not surprisingly the preferred alternative means svas sometimes not the one with the lowest cost for the communication alone. 21. About 3 percent of the calls could have been replaced by letters, according to the callers, which would have cost less. The fact that the callers actually used the telephone, de- spite its being more costly. indicates that they attached a value to the qualitative differences and to the faster response timles (these benefits were not quantified). 22. Since the alternative means generally involved considerable additional travel or re- sponse times, the time saved by using the telephone was on average quite high. 23. This example is based on a study undertaken by Christopher Willoughby. At the tinme, the official exchange rate for the Iidian rupee was Rel = 100 paisa = $0.13. 24. At the rimie the official exchange rate was Tkl6 = $1.00. 25. Economics Study Cell, Posts and Telegraphs Board India (1981), pp. 48-50. Chapter 9 Expenditure Method Estimates of Consumer Surplus THE EXPENDITURE METHOD of estimating a portion of the consumer sur- plus associated with telecommunications services attempts to measure more completely the costs that users actually incur in the process of communicating. In many situations it is possible to observe the amount of telecommunications services demanded at a given time by users who, for one reason or another, willingly incur additional costs while using the service.' For example, people traveling to a nearby public call office (Pco) use time and may incur transport costs in ad- dition to paying the call charge.2 Hence, if it is assumed that public telephone users in and around a rural village in a developing country are relatively homogeneous, with similar income and tastes, but that they must travel different distances and incur different costs to use the telephone. a rough demand curve can be estimated by comparing the number of visits to the telephone with telephone tariff charges plus costs associated with accessing the telephone. ft is also possible to estimate partial consumer surplus using the ex- penditure method without observing the costs of travel and travel time. In many countries, existing telephone installations are sold pri- vately, either on the black market or legallv, with the new user paying the old subscriber a substantial price in addition to the connection or transfer charges levied by the operating telephone company. Also, highier rental and purchase prices are sometimes paid for residences or offices with installed telephones than for those without. Finally, be- cause call traffic is frequently congested during business hours, tele- phone users spend potentially valuable time in repeated attempts to make a call. Circumstances such as these provide further opportuni- 173 1 74 MICROECONOMIC ANALYSIS OF BENEFITS ties to estimate the value of telecommunications services over and above payments made to the telecommunications authority itself. Estimating Consumer Surplus by the Expenditure Method The following examples illustrate how indications of users' willing- ness to incur costs in excess of official telephone tariffs can be used to estimate at least part of their consumer surplus. Example 20. Transport Cost and Time Spent to Reach a Telephone (Chile 1975) In Chile, the costs and benefits of providing 200 towns and villages with a rco tele- phone were estimated.' The 200 places were categorized into twenty-three groups ac- cording to locational characteristics and income level. A control village with similar income and locational characteristics, but which already had a rco telephone, was as- signed to each group. Linear demand functions for telephone calls were then estimated separately for each of four income strata in each village without a telephone. One point on each demand curve was arbitrarily fixed, assuming that, were a telephone available, traffic in the town would be approximately equal to that currently observed in the associated con- trol place if similar tariffs applied. Since observed calls were not homogeneous, an "equivalent price" was calculated as the weighted average price paid for calls in all of the control places, and the number of "equivalent calls" was calculated for each con- trol place and for each income stratum by dividing actual revenues for calls by this equivalent price. A second point on each of the 800 demand curves was estimated using interview in- formation on the calls residents made on the nearest telephone outside the village. An "equivalent price" was worked out for each village and income stratum by adding aver- age transport cost and estimated value of the time used to get to and from the tele- phone to the price calculated for the set of control places. The number of equivalent calls was calculated as the total expenditure incurred by callers from a given place and income stratum divided by the appropriate equivalent price per call. The estimated ad- ditional revenue and consumer surplus that would be generated by a public telephone were then estimated for each place by calculating the areas under the assumed linear demand curves, together with what were thought to be reasonable assumptions about the growth of call traffic through time. Assuming an equipment life of fifteen years (twenty years for open wire installa- tions), a discount rate of 12.6 percent, and a shadow price for foreign exchange equal to 1.3 times the official price, the present value of the total benefits expected to result from providing rcos in the 200 villages averaged 4.8 times the present value of the forecast incremental revenues alone. For seventy-seven of the villages the present val- ues of the estimated net benefit streams were positive. Hence, the immediate provision of telephones could be justified in terms of expected effect on the economy as a whole. In contrast, from the viewpoint of the telephone company, adequate financial returns would be obtained in only five of rhese seventy-seven) places. For the remaining 123 places, investment was not recommended at the time; however, given forecast in- EXPENDITURE METHOD 175 creases in traffic, it was estimated that twelve additional places would have a positive net present value of benefits if the rco installations were postponed three or four years. Example 21. Transport Cost and Time Spent to Reach a Rural Telephone (Costa Rica 1976) During one week in 1970, everyone who used the only telephone in the village of Puerto Cortes in Costa Rica was interviewed. Information was assembled on the price paid for each call, the caller's income, travel time to the telephone, reason for the trip, transport cost, and hours worked during the week.' Assumptions were made about how the time of the callers should be valued and how the separate and joint costs of various means of transport should be allocated to the telephone call. The underlying premise of the exercise was that persons who lived closer to the PcC and therefore used it more than persons living farther away obtained a consumer surplus for at least some of their calls. Summary results of the exercise are shown in table 9-1. Given the somewhat arbi- trary nature of the assumptions involved in the analysis, two sets of assumptions were used to estimate transport cost and time value. As can be seen, the resulting estimates of consumer surplus are highly sensitive to those assumptions. Table 9-1. Two Estimates of Consumer Surplus for Telephone Calls Made in a Rural Costa Rican Village, 1976 (colones) High Low Item estimate estimate 1. Average call charge 4.41 4.41 2. Average transport cost per call' 19.02 2.54 3. Average travel time (hours) 0.59 0.59 4. Average monthly income of caller 422.90 422.90 5. Average number of hours worked during week 39.31 39.31 6. Opportunity cost of time" 1.51 0.38 7. Consumer surplus per call 20.53 2.92 8. Ratio of consumer surpltus per call to average call charge 4.65 0.66 a. Of the 404 calls used to tabulate the information in this table, only fifty-four of the callers were able to offer what was considered to be reliable infornsation on transport costs. It was therefore assumed that, for the high estimate, transport costs for the missing obser- vations were equal to the average for the persons responding and that, for the low estimate, nonrespondents incurred no transport costs. b. For the high estimate, it was assumed that the value of time for all callers was equal to the average income for that time in the village, that is, (6) = (3) . (4) / 1(5) 4.21. For the low estimate, it was assumed that many of the calls were made during periods when real re- source costs in terms of time were minimal, and therefore the low estimate was arbitrarily set equal to one-fourth of the high estimate. c. (7) = (2) + (6). Source: See appendix C. 1 76 MICROECONOMIC ANALYSIS OF BENEFITS Example 22. Rents and the Availability of a Telephone (Egypt 1977) If telephone connections in offices, houses, or apartments can be transferred when the property is sold or leased, the relation between the value of real estate property and the presence of a telephone line can provide information on the willingness of users to pay for telephone service. In 1977, Egypt had an acute shortage of telephones, and local businesspersons and middle-income residents often obtained telephones by renting furnished offices or res- idences in which, legally, the telephone could be transferred as one of the furnishings. Rental advertisements in local newspapers commonly listed the availability of tele- phone service as one of the prominent attributes. Differences in rent of from £E50 to £E150 a month5 were observed for offices or residences that were similar except for the presence of a telephone.' Places with a tele- phone that was only one of several extensions on a shared line or places that were in areas known to suffer from heavily congested telephone traffic tended to command less of a premium than those with private lines or located in areas with less call traffic congestion. The larger and better located offices or apartments tended to carry the highest telephone premiums (£E150 and more), probably reflecting the high value that large businesses and higher-income residents attached to the telephone. Given these findings, the consumer surplus for the monthly rental of a telephone was estimated using two assumptions. First, the demand curve for telephone rentals was downward sloping, reflecting a situation in which a relatively small proportion of the total population was willing to pay high monthly rentals for telephone service, whereas most of the population was willing to pay only much lower telephone rental charges. Hence, the demand curve was assumed to be convex to the origin of the price and quantity axes; mathematically, the demand curve was specified as a rectangular hyperbola with elasticity equal to -1. Since empirical estimates of the price elasticity of demand for telephone service are generally between 0 and - 1, this was a conserva- tive assumption.7 Second, two points on that demand curve are 15,400 telephones (less than 3 percent of the total number of telephones in service) at a price of fE51.50 a month (the official monthly fee of fEI.50 for automatic message rate ex- change service plus £E50 for the typical difference in rent) and 530,000 telephones at the official price of fEl.50.4 The bounded area under the demand curve above the official monthly price of fEI.50 and below a monthly price of fE51.50, and between the quantities 0 and 530,000 telephones was equal to fE2,807,669. This amounted to an average of fE5.30 per telephone a month or £E63.60 per telephone a year. Hence, for estimating bene- fits, the value of renting a telephone averaged £E63.60 a year above the average price actually paid to the telecommunications operating entity (fF18.00). It was also assumed that this estimate of average consumer surplus would continue to be valid for users added to the system during the next four years. This assumption was considered reasonable since, despite rapid expansion of the system, a large propor- tion of the excess demand would not be satisfied-and hence most new users would be high-value users-for several years to come. Thus, for the following four years, the benefits arising from having a telephone were estimated to average at least four times the revenues paid to the telephone operating entity for telephone rentals. EXPENDITURE METHOD 177 Example 23. Value of Time Wasted in Call Attempts (Egypt 1977) In addition to benefits derived from being connected to a telephone system, bene- fits also arise when calls are completed. In Egypt call charges paid by telephone users underestimated the users' costs of completing a call since the high proportion of fail- ures to complete a call, time spent waiting for a dial tone during peak periods, and the necessity to repeat calls because of poor transmission quality and interruptions wasted a significant amount of otherwise productive employee time.' The following assumptions were believed to reflect conservatively the situation in Egypt's two largest cities, Cairo and Alexandria, in 1976-77: a. In 1976 the average urban caller earned an estimated fE6.00 a week.") Assum- ing the caller worked an average of forty-five hours a week, each working minute was worth approximately fE0.0022. b. A minimum of one and one-half minutes per failed call was spent during busi- ness hours trying to complete a call that was unsuccessful because of technical problems or call traffic congestion. c. Of calls made in business hours, 95 percent were related to business or government. d. Unsuccessful calls resulting from technical faults or call traffic congestion dur- ing prime business hours constituted approximately 75 percent of all attempts to make local calls in Cairo and 55 percent in Alexandria. e. An average of about 1,820,000 and 560,000 calls were completed, respectively, in the two cities during the five or six peak calling hours each working day. f. During peak hours an unsuccessful call attempt rate of 25 percent would have been acceptable. An estimate was made of the value of time wasted by government and business call- ers in Cairo and Alexandria when more than 25 percent of their attempted calls were unsuccessful. (No attempt was made to extend the calculation nationally.) Using the above estimate of a portion of the costs that consumers willingly incurred in making a telephone call (official call charge plus value of time wasted) and the esti- mate from example 22 of the yearly value of having a telephone connection (average monthly telephone rental charge plus the estimate of consumer surplus due to rent differential), a stream of project benefits was estimated for Egypt's 1978-82 telecom- munications development program. With program costs appropriately shadow priced, the internal rate of return in real terms on the proposed new investment was esti- mated to exceed 25 percent. This was about two and one-half times the rate of return that was calculated by using only revenues from tariffs as a proxy for benefits. Example 24. Cost of Unauthorized Transfers (Myanmar 1978) In many developing countries, the private transfer of telephone connections is either legally permissible or informally tolerated.'' A good example is provided by the experi- ence of Myanmar (formerly Burma), where during several years, numerous telephone subscribers revealed a willingness to pay a price in excess of that prescribed in the offi- cial tariff schedule to acquire a telephone.'' In Yangon (formerly Rangoon), in particu- lar, an unofficial private market existed in which telephone connections were bought 1 78 MICROECONOMIC ANALYSIS OF BENEFITS and sold. In 1977 the national telephone organization began an intensive effort to trace unauthorized connections and, as a result, found about 520 cases in which telephones had been privately transferred at prices averaging K10,00O.' All the unauthorized sub- scribers identified were fined K2,400 each but were allowed to keep their telephone after they paid the fine and the official connection charge (K450) and deposit (K300). When word spread that the unofficial connections were being legalized, about 470 addi- tional unofficial subscribers in Yangon voluntarily turned themselves in, paid the fine, connection charge, and deposit (K3,150), and had their telephone status legalized. Given this information and the fact that in 1977, 15,545 potential subscribers with- out telephones were registered on waiting lists, thus demonstrating a willingness to pay at least the official connection charge of K450, the consumer surplus associated with obtaining a telephone in Myanmar was estimated using the following two assumptions. First, the demand curve for telephone connections was shaped so that a relatively small proportion of the total population was willing to pay high initial fees to obtain a telephone, whereas most of the population was willing to pay only relatively lower fees. Hence, the demand curve for telephone connections was assumed to be convex to the origin of the price and quantity axes and to have a price elasticity of -1.14 Since empiri- cal estimates for the price elasticity of demand for telephone service are generally be- tween 0 and - 1, this was a conservative assumption. Second, one point on that demand curve was 1,522 telephone connections (the number of recently legalized telephones in Yangon in fiscal 1977 plus the estimated number that were expected to be legalized in a forthcoming drive against unauthorized connections in large places other than Yangon) at a price of K12,850 (KIO,000 private market price plus K2,400 fine plus K450 con- nection fee). This point, together with the slope of the curve implied in the assumed rectangular hyperbola with a price elasticity of -1, specified the demand curve. Calculating the bounded area under this demand curve above the official connec- tion fee of K450 and below K12,850 and between the quantities of 0 and 17,067 (1,522 plus the 15,545 waiting applications in Myanmar as of March 1977) and divid- ing that by 17,067 gave an average consumer surplus of K3,466 per telephone connec- tion. Hence, to estimate benefits, the value of acquiring a telephone in Myanmar during 1977-78 and for the five-year period thereafter, during which a large excess de- mand was forecast to continue, was estimated to be the sum of the connection fee of K450 plus a one-time consumer surplus of K3,466. Example 25. Value of Time Saved by Introducing Subscriber Trunk Dialing (India 1969) In portions of India before the introduction of subscriber trunk dialing service, con- sumers wanting to make long-distance calls during business hours normally had to wait lengthy periods for their call to be completed. They could, however, move up in the queue by agreeing to pay more, from twice the ordinary charge for so-called "urgent" calls to eight times the ordinary charge for "lightning" calls. With the introduction of subscriber trunk dialing service, call waiting times were projected to disappear on some routes. In an attempt to estimate the value of the time that would be saved by introducing this service, outgoing trunk traffic was analyzed on the principal routes serving the re- gional center of Madras. As shown in table 9-2, the correlation, irrespective of dis- tance and hence of call charge, was high among three factors: the percentage of calls booked as urgent, the average delay in being connected for ordinary calls, and the av- erage reduction in the delay for calls booked as urgent rather than ordinary. EXPENDITURE METHOD 179 Table 9-2. Delay of Ordinary and Urgent Calls Booked, b-y Major Route in Madras, India, 1969 Average difference Average delay in delav: Destination Percent oJf on ordinary urgent vs. oj call, calls booked calls ordinary by area "urgent" (minittes) (minutes) Bombay 39.8 80 50 Calcutta 25.3 32 11 Coimbatore 35.0 70 31 Delhi 23.2 30 Hyderabad 25.9 41 20 Madurai 27.5 50 17 Salem 10.8 16 Tiruchi 28.4 46 28 Vellore 14.6 32 12 Vijayawada 11.5 13 Zero or negligible. Source: Unpublished data from India l',T. An analysis of the raw data indicated that telephone users were willing to pay a pre- mium of about 25 percent of the base charge to save twenty minutes' delay in obtaining an ordinary call, 35 percent to save thirty minutes, 40 percent to save forty minutes, and 50 percent to save one hour. The average difference in delay between urgent calls and or- dinary calls was about twenty-five minutes. Hence, with the introduction of subscriber trunk dialing, which was supposed to eliminate these delays, the time saved was assumed to be worth at least 30 percent of the base call charge to the average caller. Example 26. Foreign Exchange Impact of Business Telecommunications (Kenya 1981 and 1987) The Kenya Cashew Nuts Company was located between Mombasa and Malindi as part of a government program to decentralize industry to rural areas. It served as a collection and processing point for cashew nuts and marketed them through five agents located in centers throughout the world. Several of the eighteen grades of ca- shew nuts sold by the company occasionally experienced significant price movements on a day-to-day basis and through time. For example, one primary grade declined in price 67 percent in three months. Telex was the accepted means of communicating with the company's five market agents and keeping abreast of global prices and contracted quantities for cashew nuts. In mid-1981 the nearest telex connection was located in Mombasa 35 miles distant; the company used this connection periodically with the agreement of the subscriber, which rents the telex from the Kenya Posts and Telecommunications Corporation. This arrangement resulted in delays for the company of up to two days in the receipt of, analysis of, and transmitted response to market information. Given such information delays, together with the sometimes rapid price movement 180 MICROECONOMIC ANALYSIS OF BENEFITS and slight difference in price among markets, managers estimated that if they had ac- cess to two telex lines at the processing plant, which would allow them to reduce mes- sage turnaround time to less than half an hour, they could on average secure 5 to 10 percent better prices for the cashew nuts they market. Managers cited several instances in which, because they lacked timely information or response, they sold low in a rising market or missed a sale in a falling market. Given the company's current foreign ex- change earnings of about $15 million a year, a 5 percent increase in the average sale price of their cashew nuts would amount to an additional $750,000 a year of foreign exchange earnings for the Kenya Cashew Nuts Company and for Kenya. Management planned to open a branch marketing office in Mombasa, which would have access to a shared telex line and would maintain some contact with the processing plant through a magneto telephone exchange connection. The company estimated that the Mombasa branch office would cost the equivalent of $1,000 a month to operate and maintain. This was a small fraction of the increase in sales revenue that was estimated to result from this somewhat improved but still cumbersome communications arrangement. Additional microeconomic analysis of the role that telecommunications plays in foreign exchange savings in Kenya was carried out by Communications Studies and Planning International with the support of the ITU in 1987. Twenty companies, repre- senting nearly 18 percent of Kenya's total export earnings, were selected from the agri- cultural, industrial, and service sectors and were interviewed in depth about how telecommunications affects their exports and imports. The benefits were calculated using improvements in infrastructure that were expected to be the result of a forthcom- ing investment by the World Bank of $3.26 million in Kenya's telecommunications sec- tor. The twenty businesses interviewed were calculated to save over $2.0 million in foreign exchange as a result of this project. Extrapolated for the export sector as a whole, savings in foreign exchange were calculated at $11.6 million, yielding a benefit to cost ratio for foreign exchange of 3.6:1.0 for a major telecommunications project.'5 Example 27. Efficiency Gains by Business Firms (Kenya 1981) A survey of how telephones were used by nine business firms in Nairobi examined the hypothesis that some businesses could not exist in their present form without tele- phones, whereas others could, but only on a smaller scale, with lower productivity, and with higher costs."' Three means were examined to improve a business firm's perfor- mance by improving communications services. These were a cost reduction effect (lower unit costs through lower-priced inputs and increased management efficiency), a selling price effect (higher selling prices through better market information), and a business expansion effect (sales gains through better and wider area marketing). Two of the nine firms surveyed were engaged in manufacturing and processing (East Africa Industries and House of Manji), five provided services or transport (Alliance Hotels, Industrial Distributors, Interfreight, Pan African Travel, and Standard News- paper), and two were engaged in agriculture or horticulture (Kenya Horticulture Ex- port Corporation and Kenya Nurseries). Within the above three categories the study documented telecommunications benefits relating to business expansion, sales price effects, purchase price effects, vehicle use, production stoppages, distribution costs, labor time, and managerial time. Table 9-3 shows estimates of the average increment in the percentage of profits to revenues for each form of benefit that the nine firms would experience if access to the national telecommunications system were extended and if the quality and reliability of the service were improved. Overall, the study iound that the single most important potential benefit to the EXPENDITURE METHOD 181 Table 9-3. Estimated Average Increment for Nine Firms in the Percentage of Profits to Revenues, by Benefit Category, in Nairobi, Kenya, 1981 Estimated average percent increase in profits Benefit category to revenues Business expansion 2.8 Sales price effects 0.4 Purchase price effects 0.2 Inventory costs 0.0 Vehicle use 0.1 Production stoppages 0.0 Distribution costs 0.3 Labor time 0.9 Managerial time 0.3 Mean for the nine firms 5.1 Source: Adapted from Communications Studies and Planning International (1981). nine firms would be an increase in sales made possible by better access to markets for buying and selling. It was estimated that the major part of the efficiency benefits gained from expanded sales would come from cost reductions obtained through im- proved capacity utilization rather than from greater economies of scale; that is, dimin- ished machine downtime, better allocation of managerial time, greater regularity of sales and purchases, and similar effects. Table 9-4 shows the estimated gains or benefits to the nine firms compared with Table 9-4. Estimated Benefits and Expansion Costs, by Firm, in Nairobi, Kenya, 1981 Estimated Estimated telecommu- Benefit/ benefits to nications cost Firm firm (K Sh) costs (K Sh) ratio East Africa Industries 11,930,000 125,460 95.1 House of Manji 4,417,900 14,760 299.3 Alliance Hotels 870,000 13,480 64.5 Industrial Distributors 2,380,000 17,080 139.9 Interfreight 5,565,000 57,900 96.1 Pan African Travel 275,400 2,320 118.7 Standard Newspaper 6,035,600 31,060 194.3 Kenya Horticultural Export 2,627,500 31,460 83.5 Kenya Nurseries Corporation 27,400 2,560 10.7 Total or average 34,128,080 296,080 115.3 Source: Adapted from Communications Studies and Planning International (1981). 1 82 MICROECONOMIC ANALYSIS OF BENEFITS what the study authors acknowledged was a very rough indicative estimate (likely to be on the low side) of the average cost per telephone line and per call if telecommunica- tions services were available throughout Kenya. On the basis of an analysis of the in- terview data the study concluded that telephone calls made by employees of these nine Kenyan firms were on average worth over ten times the cost; that is, a large con- sumer surplus was associated with expanded telephone service for businesses in Nairobi. Example 28. The Role of Telecommunications in Improving the Reliability of the Power System (Brazil 1980) In 1980 an analysis attempted to determine the effects of an improved telecommu- nications network on the operation and maintenance of the COELBA power system in the Brazilian state of Bahia.', It was hypothesized that an improved telecommunica- tions network would improve the operation and control as well as maintenance of the existing COELBA power system and that, as a result of such improvements, the reliability of the supply of electricity in the state would improve. Specifically, it was contended that improvements in operation and control would allow the COELBA system to operate more efficiently with fewer interruptions in supply (outages) resulting from overloaded circuits, burned-out transformers, and so forth. The principal economic benefits of improved operation and control were measured by the decreases in outage costs result- ing from reductions in outage frequency. With regard to maintenance, it was assumed that improvements in telecommunications would allow failures within the COELBA SyS- tem to be identified and corrected more quickly. The major economic benefits of such improvements were measured by the projected decreases in outage costs resulting from shorter and less-frequent outages. The estimation process involved the following: (a) the effects of outages on impor- tant industries in Bahia were determined from a survey of seventeen industrial firms around Salvador, the central city; (b) the effects of outages on residential consumers in Bahia, and estimates of consumers' willingness to pay to avoid such outages, were determined from a survey of residential consumers; this survey covered 182 consumers in five economic regions in Bahia; (c) data on value added according to industrial sec- tor for 1970-75 were obtained from the state government and served as a basis for es- timating industrial value added over the lifetime of the telecommunications project; (d) data on the number of residential electricity consumers in the COELBA system in 1980 and estimates of the growth rate of such consumers were obtained from COELBA; (e) information on the effect of the telecommunications system on the quality of elec- tricity supply within the COELBA system, that is, estimates of resulting reductions in the frequency and duration of outages, were obtained from interviews with officials from L:OELBA; (0 estimates of project capital and operating costs were obtained from the var- ious telecommunications project feasibility studies; and (g) data on the number of street lights used in Bahia were supplied by COELBA. Table 9-5 presents the final estimates for the frequency and duration of power outages in the state of Bahia with and without the proposed telecommunications sys- tenm (point e, above). Given these assumed estimates and the derived estimates of the costs of power outages to both industrial and residential users in the state (points a through d. above) as well as the project costs, the internal rate of return on the invest- ment in the proposed telecommunications project was estimated to be 98 percent. EXPENDITURE METHOD 183 Table 9-5. Yearly Average Frequency and Dutration of Outages Affecting Industrial Consurners, with anid without the New Telecommunications System, in Bahia, Brazil, 1981-2000 Outages without netw Assumed outages with new telecommunications (hours) telecommunications (hours) Time of day Frequenc'y Duration Frequency Duration 0800-1800 96.00 0.80 90.00 0.60 1800-2400 58.92 0.80 54.00 0.60 0000-0800 36.00 0.85 30.00 0.64 Source: COELSA and industrial survey. Example 29. Gasoline Consumption Resulting from Inadequate Telephone Access (Tanzania 1982) As of January 31, 1982, there were 17,718 connected telephone lines in Dar es Sa- laam and 17,816 people registered as waiting for telephones. Some of the effects on gasoline consumption of such an imbalance in telephone supply and demand were in- dicated through interviews with the management of three local business organizations in April 1982. The three organizations were the Sanitary Appliance and Hardware Company, the Board of External Trade (a government agency), and H. J. Stanley and Sons, Ltd. In reviewing communications needs and patterns, managers of the three or- ganizations discussed ways in which inadequate telephone access and quality of service forced them to rely on motor vehicle transport for the simple function of soliciting and exchanging information. The Sanitary Appliance and Hardware Company sold and distributed a variety of appliances and hardware and, beginning in 1978, manufactured locks, hinges, and other small hardware. The main retail and wholesale store in the center of Dar es Sa- laam had a telephone; the factory about 6 miles distant was enrolled on the waiting list for a telephone in January 1979. Frequent contacts between the main store and the factory, and between the store, the factory, and other places in Dar es Salaam, were necessary to check stocks, coordinate deliveries, order supplies, arrange transport, contact banks, customs officials, and buyers, and report problems and occasional emergencies at the factory. The Board of External Trade was a government marketing information and coordi- nation agency that served exporters and buyers. It had two telephone lines to contact ministries, banks, customs officials, exporters, buyers, and so forth. The Board at- tempted to maintain contact with about 200 exporters in the Dar es Salaam area and with another 200 suppliers, transporters, and exporters throughout the country. Al- though Board staff attempted to complete thirty to forty local contacts each working day, they were able to contact an average of only fourteen exporters by telephone. As- suming thirty daily contacts, an average of sixteen exporters either drove a vehicle to the Board office or a Board staff member traveled to them (an average of 10 miles one way in the Dar es Salaam area). Assuming a fifty-fifty split and that some of the trips were multipurpose and might therefore more properly be counted as four net Board- 1 84 MICROECONOMIC ANALYSIS OF BENEFITS related trips rather than eight, the Board was associated with about twelve local vehi- cle trips approximately 20 miles long each working day. H. J. Stanley and Sons, Ltd., was a multifunction business firm, which among other activities extracted and marketed salt, operated a coastal vessel, and dealt in coconuts and coconut processing. The Stanley group offices had three telephone lines with sev- eral extensions; office space was rented to a fourth firm, which also used the Stanley telephones. Many of the business contacts of the Stanley group complained that they could not complete calls into the Stanley offices since the Stanley telephone lines, when not out of order, were busy most of the time. In reviewing their daily business contacts and total miles driven, the two senior partners at H. J. Stanley and Sons cal- culated that they could each reduce their personal vehicle travel by an average of 225 miles a week if they and their salt and coconut facilities, some port facilities, and se- lected suppliers and wholesalers had access to reliable telephone service. The Stanleys also noted that if telephone service in the Dar es Salaam area became so congested as to be totally unusable during business hours, not only would they have to increase their own personal driving, but they would also at a minimum have to add two senior purchasing clerks at a salary of 4,800 Tanzanian shillings a month ($8,861 a year). Each clerk would on average drive at least 35 miles a day to contact suppliers, transporters, banks, distributors, and so forth. Table 9-6 summarizes the value of the direct petroleum consumption that managers of these three establishments estimated could be saved given improved and expanded telephone service in the greater Dar es Salaam area. Only net mileage savings were tabulated, and the Tanzanian shilling was shadow priced at T Sh13 per $1 rather than the official rate of T Sh9.28. One year was assumed to consist of forty-eight working weeks. The price of gasoline at retail petrol stations-$2.54 per imperial gallon-was assumed to reflect the foreign exchange cost of gasoline to Tanzania, and an average Table 9-6. Estimated Foreign Exchange Expenditures Related to Petroleum Consumed by Three Business Organizations because of Inadequate Telephone Service, in Tanzania, 1982 Assumed Gasoline net round Working expenditures trips saved days per per year Item per day Miles week (dollars) Sanitary Appliance and Hardware Owner/manager 3.5 124 5.5 Assistant manager 2.0 12, 5.0 - Total 3,421 Board of External Trade 12.0 20' 5.0 11,697 H. J. Stanley and Sons, Ltd. Partner - 41' 5.5 - Partner - 41" 5.5 - Total - - - 4,386 Two senior clerks - 70' 5.0 3,412 - Not available. a. Miles per trip. b. Miles per day. Source: World Bank data. EXPENDITURE METHOD 185 vehicle mileage figure of 12.5 miles per gallon was used to reflect the variety of motor transport used, including small cars, four-wheel-drive vehicles, and trucks. The managers of all three organizations, however, emphasized that they considered the savings in gasoline consumption to be quite unimportant compared with the sav- ings in scarce managerial time. The two H. J. Stanley and Sons partners estimated that they spent at least two hours each day in motor vehicles, which would be unnec- essary if adequate telephone service were available (a total of 22 hours of management time lost each week). The Sanitary Hardware and Appliance Company manager and assistant manager estimated that they spent at least 12.8 and 6.7 hours, respectively, a week of unproductive time in motor vehicles, which they could spend on managerial tasks if better telephone service were available. The owner/manager of the company contended that, given current local demand for hardware and several new products on which he would like to work, the company could grow at least 25 percent a year more rapidly if he could trade time wasted driving for time spent on management and prod- uct development. Example 30. Reduction of Wasteful Agricultural Travel and Transport (Uganda 1982) The Busoga Growers' Cooperative Union was Uganda's largest producer of cotton and the fourth largest producer of coffee. It comprised about 200 farmer cooperatives, 7 cotton ginneries, 2 coffee factories, and a small dairy farm. Telecommunications fa- cilities were limited to three telephone lines at headquarters and one line each at three of the other places. During the annual four-month cotton marketing season a typical ginnery's manager traveled about 100 kilometers to the union's headquarters three times a week to follow up on orders for production imports, make arrangements for transport, enquire about likely pay dates, and take care of other routine administrative matters, which could all have been handled by telephone if such service had been available. During the remain- ing eight months of the year, the manager traveled for these purposes about once a week. Also during the cotton marketing season, the secretary of a typical cooperative rode a bicycle or hitched a ride three or four times a week to the ginnery, an average of some 13 kilometers away, for reasons similar to those requiring travel between the ginnery and headquarters. He also went twice a week to the nearest bank (an average of 65 kilometers away) to collect or enquire about payments for past deliveries of cot- ton. During the remaining eight months of the year, the secretary traveled about one- third as much. Each trip required a full day of the manager's or secretary's time. Assuming that the value of one day's time equaled one-twentieth of monthly salaries plus allowances, and using average costs of public transport for the distances traveled, these trips cost the union the equivalent of $0.1 2 million a year for all units and co- operatives. This understated the true cost, since it did not include losses related to unit and cooperative performance resulting from repeated absence of managers and slow administrative communication. The union had a fleet of twenty-nine 8-ton lorries and hired fifty-five lorries from private transporters, mainly to carry the cooperatives' produce to the union's units (for example, cotton to the ginneries). During the four-month marketing season, cot- ton was carried three times a week between the typical cooperative and ginnery. About 20 percent of the time the lorry returned empty, however, because a lack of timely communication prevented effective coordination with the cooperative's management. 186 MICROECONOMIC ANALYSIS OF BENEFITS Thus, for each ginnery, about twelve empty return trips were made during each cotton marketing season. For all 200 cotton growers' cooperatives, these empty trips cost the equivalent of $0.13 million each season, considering only vehicle depreciation, main- tenance, and fuel. Moreover, 150 of these cooperatives also produced coffee, which was collected three times a week during nine months of the year; about 10 percent of the coffee-related lorry trips were ineffective for the same reason, at a total cost of about $0.22 million a season. Management contended that with adequate telephone service, these excess travel and lorry movements could be mostly eliminated. Furthermore, management argued that both administrative communication and fleet movements could be managed with greater flexibility; empty lorries could be diverted to pick up produce at alternative co- operatives nearby. The benefits of providing all cooperatives with good-quality telephone service were estimated to be the sum of the excess costs incurred in travel and freight transport as calculated above, or about $0.5 million a year. Extension of the existing high- frequency public radio call service to the 200 cooperatives would meet the low call traffic requirements of administrative and transport coordination; the total cost of such investment could be recovered by reducing wasteful travel and freight in about one year. A facility giving higher-quality and more reliable service (for example, very high-frequency multiaccess radio) would cost about four times as much; such costs would still be more than offset by the direct savings in three to four years, which is much less than the life of properly maintained equipment. Example 31. Value of Improved Telecommunications Infrastructure at the Microeconomic Level, Extrapolated to the National Level (the Philippines 1984 and Costa Rica 1986) In 1984, in-depth interviews were conducted with the managers of 250 businesses in Northern Luzon and Northern Mindanao. These interviews asked managers how lim- ited telephone service affected the performance of their company and sought to ascer- tain the possible impact of expanding and enhancing that service." The survey results were used to estimate aggregate benefits to the economy of a planned telecommunica- tions project. At the time, the government of the Philippines was beginning two major telecommunications programs, the Rural Telecommunications Development Project and the National Telephone Program (phase 1). The study took into account the spe- cific expansion being planned for each geographic area. The data were based on the assessment of benefits from the changes planned, not on actual results. Benefits as estimated by the businesses interviewed were examined according to three categories. Direct benefits included cost reductions that were the direct result of improved telecommunications services in the course of normal opera- tions (such as reduced travel and use of messengers or more efficient use of manage- ment time). Consequential benefits represented the cost of opportunities lost due to poor-quality telecommunications; they could be realized only if normal operations changed somehow. Business expansion benefits accrued to firms that could expand more rapidly if telecommunications were improved; they were calculated according to improved economies of scale in production. Costs were calculated based on the capital costs of improving the network as planned by the two government projects. The results were examined by sector for each region and are summarized in table 9-7. Some businesses were not likely to benefit from improved telecommunications EXPENDITURE METHOD 187 Table 9-7. Benefit to Cost Ratio Following Installation of Telephone Service, by Sector and Region, in the Philippines, 1984 Proportion of Number in Benefit to sample with Sector and region sample cost ratio net benefits Agriculture 90 44.6 45 Farming, forestry, fishing 40 18.0 28 Agricultural manufacturing 23 69.0 57 Agricultural trading 27 40.8 52 Health 47 33.4 45 Public and private health care delivery 42 35.5 38 Distribution of health care products 5 26.7 100 Other businesses 94 13.1 64 Manufacturing 54 10.5 57 Construction 2 39.4 100 Wholesale and retail trade 20 7.4 80 Transportation 6 31.6 83 Financial services 5 21.1 60 Other services 7 13.6 43 Northern Luzon Agriculture 47 35.7 47 Health 27 6.4 41 Other business 63 12.1 75 Northern Mindanao Agriculture 44 56.3 39 Health 20 63.1 50 Other business 51 12.1 53 Source: Communications Studies and Planning International (1986). (those with a benefit to cost ratio less than I). For the purposes of the study, it was assumed that these firtns would probably not choose to subscribe, and therefore all benefit to cost ratios included only those businesses showing positive benefits. Overall, the study found that the economic benefits of the planned expansion clearly exceeded the costs; the average business subscriber would obtain yearly benefits equal to thirteen times the annualized cost per line of the project in Northern Luzon, where the investment was targeted toward the countryside, and equal to twenty times the cost in Northern Mindanao, where the investment was targeted toward towns. Al- though the benefits to primary agricultural producers were moderate, the highest cost to benefit ratios were for agricultural industries and traders (rice and grain mills and commodity exporters). The study also found a direct relation between increased busi- ness size as measured by number of employees (ranging between one to four employees and twenty to forty-nine employees) and increased telephone-related benefits (ranging between $120 and $13,700 a year for these same categories). Since telephone service 1 88 MICROECONOMIC ANALYSIS OF BENEFITS was estimated to have a total econoniic cost of $200 a year, using telephones was not viable for very small businesses. Aggregating the microeconomic data involved two steps: first, extrapolation from the firms studied to obtain regional results and, second, conversion of benefits initially expressed as financial profits into benefits expressed as consumer and producer wel- fare costs incurred by the economy for the total costs of constructing and operating the network. On the basis of lines per business, the benefit to cost ratios were 13.5 and 20.1 for Northern Luzon and Northern Mindanao, respectively. Looking at the telecommunications system and including only business line benefits, the respective benefit to cost ratios were 1.0 and 1.2. This means that the benefits accruing to busi- ness users were sufficient to pay back the entire investment cost of the system; the benefits to the other 90 percent of subscribers would be added to this, making the total benefit to cost ratio even more favorable. A similar study was carried out by the same researchers in Costa Rica in 1986. This study compared economic benefits that were estimated by nonsubscribers, by recent subscribers, and by persons or businesses that had subscribed for five years. Nonsubscribers estimated their potential gains to be $480 a year, which was very close to the gains estimated by recent subscribers, $500 a year. This gives some support to the accuracy of nonsubscribers' estimated benefits. However, long-standing subscribers realized benefits of $720 a year, which suggests that long-term changes in the way a business is run are required in order to take full advantage of the potential benefits and that these changes take time to learn and implement. Based, in broad terms, on the results of all three studies, a new telecommunications project was expected to yield benefits to its business customers of approxiniately 5 percent of the total revenues of all establishments they served.'' Example 32. Social and Economic Benefits of Rural Public Call Offices (Vanuatu 1988) Although urban areas and large businesses in rural areas are adequately served by telecommunications services in Vanuatu, the majority of the rural population does not have easy access to telephone service. About sixty rural telephones (forty-four of which are public) serve about 80 percent of the total population of 130,000. In the expectation that expansion will occur, a study was done to determine the extent and optimal spatial layout that would yield the highest benefits to the rural users of public telephones for a given cost." Using econometric methods, the study directly measured the losses resulting from the lack of adequate telecommunications. Data were collected on round-trip distance traveled (either on foot or by vehicle) to all existing public phones in rural areas. Demand for domestic calls per 100 inhabit- ants a year was then expressed as the cost incurred by the individual callers, based on the minimum subsistence wage and the cost of paid transportation by vehicle. The valuation of travel distance was determined to be V30 per kilometer, which was, in turn, used to calculate consumer surplus. The number of domestic calls (combined domestic and business) made annually per 100 inhabitants in rural areas was inversely proportional to the distance of the caller from a telephone and declined roughly from 290 calls a year at 0-1 kilometer to 101 calls a year at 2 kilometers and to 40 calls a year at 5 kilometers. At distances between 5 and 20 kilometers, calling rates were clustered around 40 calls a year and then fell EXPENDITURE METHOD 189 further to about 16 calls a year at distances over 20 kilometers. Government calls fell much more sharply, from 176 calls a year at 0-1 kilometer to about 15 calls a year at 2 kilometers. Two spikes at 9 and 18 kilometers represented an unusual frequency of calls in locations where a government office did not have its own phone. A spatial model was then developed to determine the benefit of extending rural telecommunications by reducing the distance rural dwellers had to travel to reach a telephone. The country was divided into ninety-six zones, and the model was used to compute the call rate per 100 inhabitants, volume of telephone traffic, distance to the nearest telephone, zone in which the nearest telephone was located, and consumer surplus for each zone and for the entire study area in aggregate. Expanding the num- ber of telephone locations from forty-four to sixty-four was calculated to yield an in- cremental benefit of V2.6 million a year (56,000 kilometers of travel saved a year); if 100 zones received a phone, the incremental benefit would be VII.4 million a year (125,000 kilometers saved a year). The average annual benefit realized per location from the addition of ten locations (increasing the number of telephones from forty- four to fifty-four) was V94,100 and from the second addition of ten locations (from fifty-four to sixty-four) was V78,200. Overall, the economic benefits gained from the telephone network increased as the number of locations increased, but at a decreasing rate; most of the benefit was obtained in the first 100 locations, since as it increased, the traffic tended to include more lesser-valued calls. One limitation of the study is that benefits were not evaluated net of costs, since relevant information was not avail- able on the costs of providing more service. Conclusions from Microeconomic Analysis The microeconomic or project-specific analysis of telephone bene- fits examined in this and the previous chapter revolves around the as- sumption that total benefits of telephone access and use can be perceived and assessed best by those who actually have or are de- manding access to and use of a telephone. As such, the analysis does not usually consider benefits external to the telephone subscriber or caller, the indirect or secondary benefits, or the distributional effects of the benefits. Within this limited framework, the analysis focuses on case-by-case attempts to quantify the willingness of consumers to pay for tele- phone service not only by observing the official tariffs that are paid but also by attempting to derive at least minimum estimates of bene- fits that consumers actually accrue, over and above tariffs, in other words, consumer surplus. Specific examples involving three proce- dures to estimate consumer surplus were reviewed. To varying degrees, some of the examples encompassed more than one estimation tech- nique or method of analysis; considerable subjectivity was involved in placing some examples in the expenditure method group rather than the best-alternative method group and vice versa. More fundamen- 190 MICROECONOMIC ANALYSIS OF BENEFITS tally, however, the benefit estimation exercises themselves generally suffered major limitations. The fundamental limitation of the exercises reviewed, and in fact of much project analysis, is that adequate data are not readily available to assure accurate results.21 In the absence of adequate inputs, some- what arbitrary assumptions must often be made. Given a sufficient number of assumptions in any one exercise, a rate of return or other numerical benefit total can be obtained. In some instances, however, the results depend more on the assumptions used in the analysis than on the data reviewed. Generally, in the examples reviewed in this and the previous chapter, dependence on two types of assumptions was critical: assumptions about the analytical framework or concept and assumptions about the interpretation of the collected data. With regard to the framework, examples of critical assumptions in the studies cited included "other things remained equal," "the obser- vations were homogeneous," and "the demand curve had a particular shape and elasticity." In abstracting a field situation, it is often as- sumed that variables not directly included in the analysis remain un- changed. In many instances, this assumption may not be entirely realistic. For instance, consider the determination of price elasticities of demand for telephone calls in example 13. If, as is common, a sec- ular growth trend underlies short-term variations in telephone traffic, elasticities determined from traffic measurements before and after a tariff increase could be underestimated and lead to an overestimate of consumer surplus. Also, as countries become more able to adapt to changes in relative prices, the possibility of substituting other expen- ditures for telecommunications usage grows; that is, in the long run, demand becomes more elastic. However, a consumer's reaction to a price change after one month might be more severe than it would be after three months. Hence, an estimate of price elasticity of demand made one month after a price change could result in a much lower es- timate of consumer surplus than one made after three or four months. Example 14 illustrates the problem in another context. Clearly, many relevant factors besides the real level of telephone tar- iffs changed between 1964 and 1977 in El Salvador, and the econo- metric problem of disentangling the effect of this variable from the host of other forces at work was probably insurmountable. In a grow- ing and increasingly modern economy and telephone network, ne- glecting these factors probably resulted in an underestimation of consumer surplus.12 Assumptions that communication events are homogeneous are also rarely valid. Telephone calls vary in duration, distance spanned, pur- EXPENDITURE METHOD 191 pose, and transmission quality, and these differences produce different values for the persons making and receiving the calls. Likewise, users differ in their occupation, education, income, and other factors, and these differences affect the way in which they value calls and on their calling patterns. Although these issues may be partially addressed (by averaging a number of heterogeneous calls to a more aggregate "equiv- alent call," as in example 20), they usually could benefit from more definitive analyses.23 A related problem is that of establishing analogies among towns or villages (example 20), residences or offices (example 22), and the dis- tribution of future subscribers (example 24). There is the obvious dif- ficulty of matching things that are different by definition. In example 20 the very possibility of finding meaningful control places can be questioned on grounds that the villages with telephones probably were already significantly different (had closer economic ties wvith cities, as well as higher average income and educational levels) when the decision was made to provide them (and not others) with tele- phones. Thereafter they probably evolved differently because they had telephones.24 A third critical assumption about the analytical framework, which directly affected the size of benefits calculated in several of the exam- ples, relates to the shape and elasticity of the demand curves. In all cases reviewed, this choice was of necessity made on a somewhat arbi- trary basis. In examples 13, 15, and 20, the demand curve was as- sumed to be linear, an assumption unlikely to be realistic particularly for large changes in quantity. In examples 22 and 24, the demand function was specified as a rectangular hyperbola, whereas in example 17 it was taken to be an S-shaped curve. Although such shapes may be more consistent than straight lines with what might be expected given the usual pattern of income distribution in a developing coun- try, and they usually yield a more conservative estimate of consumer surplus, they are still choices with little empirical support. Four different price elasticity estimates were used in the examples: -0.5 for call charges in example 13, -1.0 for rental charges in exam- ple 22 and for connection fees in example 24, -0.25 for a composite of telephone service charges in example 15, and roughly a -0.17 for local telephone calls in example 17. Except for example 13, none of these elasticities had much empirical basis in the particular situations in which they were used. Finally, using an implied single demand function to estimate price, quantity, and benefit values through time, as was done in several of the examples, involves critical assumptions about other things remaining equal and does not explicitly consider 192 MICROECONOMIC ANALYSIS OF BENEFITS the fact that demand for telecommunications services in developing countries has grown, and will probably continue to grow, at a fast pace even when tariffs are increased in real terms. Besides the above problems relating to the formulation of the gen- eral framework for the analysis, a general class of problems is associ- ated with actually interpreting or extrapolating the data collected. Examples observed include placing a value on time or transport and sorting out multiple trip objectives. With regard to trip objectives, when individuals journey to a town or village center to use a tele- phone, they frequently perform other tasks in addition to making the telephone call. The decision on whether and how to communicate is not independent of other actions. Hence, the travel-related data used in examples 15, 16, 18, 19, 20, 21, and 32 are by definition affected by the assumptions about time allocation that were used. Furthermore, placing a monetary value on the total cost of travel (examples 15, 16, 18. 19, 20, 21, and 32), on the value of time (ex- amples 15, 18, 19, 21, and 23), and on postal communications (ex- ample 17) is a complex undertaking. Even if time saved could be valued at the users' average or marginal income (often this would be too high), measuring income at a sufficiently disaggregate level is in many instances not possible. As shown in the studies from which ex- amples 16, 21, and 32 were drawn, this is particularly so in rural areas of developing countries, where (a) a large part of income is received in kind rather than in money and hence is not included in the basic sources of economic information (census or national accounts), (b) income is subject to large seasonal variations and thus difficult to es- timate at any particular time, and (c) as occurs in some countries, rural inhabitants are far more reluctant than their urban counterparts to give accurate information on income during interviews. Example 21, which provides two divergent but equally plausible estimates of the value of time and the cost of transport, illustrates the problem well. Also, given the many variables and accompanying assumptions involved in example 20, it is likely that the authors of that exercise could have rationalized at least half a dozen equally plausible outcomes. Other problems relating to data interpretation were apparent in the examples that were based on interpretations or forecasts made by in- terviewers or managers with potential vested interests. Examples 26, 27, 28, 29, 30, and 31 rely heavily on interpretive analysis or unverifiable claims about future expertise or telecommunications- related savings. As such, the conclusions reached in these examples are perhaps more dependent on assumptions than are most. In exam- EXPENDITURE METHOD 193 ple 31, some confirmation of data was obtained in the Costa Rican study, where projected estimates were compared with actual first-year benefits. A final point remains to be noted. The calculations in the exam- ples, to the extent they are valid, probably yield only partial measures of consumer surplus. To some extent, the systematic biasing of calcu- lations toward low or conservative estimates is what makes these sim- ple methods useful. Since cost-benefit comparisons are generally used to identify and eliminate low-return projects, an underestimation of benefits, if understood, helps to ensure that overprovisioning of ser- vices does not occur. It can be argued, of course, that underpro- visioning is an equally serious problem; in most developing countries it is certainly the most prevalent of the two."5 Although their limitations are formidable, the studies described serve two useful functions: they assist in examining project justifica- tion, and they assist in making decisions about project composition. With regard to project justification, such limited analytical exercises at the project or program level have often provided, at low cost, ap- proximate but reasonably credible quantifications of some portion of consumer surplus. This has been especially useful in examining poten- tial projects that are only marginally acceptable when judged by purely financial criteria. Such exercises also have attracted the attention of decisionmakers, have furthered an understanding of how users value telecommunications services, and ultimately have highlighted oppor- tunities for additional investment. With regard to project composition, estimates of price elasticities have helped shape tariff reforms, and the more detailed analyses of telephone usage in rural areas of developing countries have on occa- sion served to define such programs better. In some instances they have also assisted in selecting the particular rural locations for early investment. Such analysis may, in fact, be more useful for determining priorities within the telecommunications sector than for assessing the merits of the total program. There is much scope for improvement and innovation in micro- economic analyses to assist investment decisionmaking. Some of the experience outlined in the transport economics literature might be useful in dealing with problems of valuation of time and allocation of cost in multipurpose trips. Computer-assisted traffic and user analysis could provide operating entities with better estimates of demand, usage by type of subscriber, value by function, and so forth. Opera- tions research approaches to inventory management might be applied to the analysis of telecommunications benefits in specific sectors or 1 94 MICROECONOMIC ANALYSIS OF BENEFITS for large entities (food storage and distribution). And, as has been done in the power sector (example 28), the responses of users to major outages of telecommunications service might yield information on the benefits derived from these services. Notes 1. The expenditure approach has been extensively used in other sectors to evaluate recre- ational benefits, following the pioneering work of Marion Clawson. A typical example is Kavanagh and Smith (1969). 2. Several approaches to the evaluation of time have been developed in the literature of transport economics. See, for example, Beesley (1965). 3. Baeza, Bunster, and Schenone (1975). 4. The data for this exercise were collected by Instituto Costarricense de Electricidad and the University of Costa Rica. See appendix C for details. 5. One Egyptian pound (LE) was equivalent to $1.43. 6. This exercise used a small sample identified initially from newspaper advertisements. Similarity of accommodation was judged on the basis of advertised description and the opinion of real estate brokers. 7. For estimates of the price elasticity of denmand for telephone connections and calls, see Littlechild (1979), pp. 35-37; Alleman (1977), pp. 67-82; Meyer and others (1980) app. C; Taylor (1980), chaps. 3 and 4; and Taylor (1983). 8. The difference of £E50 a month in rent was considered representative for modest one- or two-bedroom second- or third-floor walk-up apartments or offices with three-party lines occupied by visiting and middle- or upper-middle-income Egyptian businesspersons. 9. For estimates of the value of time wasted in the United Kingdom in 1974 and earlier because of network congestion, see Short (1976). 10. This estimate was derived by adjusting the 1968 national average weekly wage upward by changes in the cost of living. This was considered to be a conservative adjustment, since it was thought that wages generally grew more rapidly than the cost of living over the per- iod, that wages in large urban areas were higher than the national average, and that wages of telephone users were on average higher than the average urban wage. I1. For example, in Chile, Costa Rica, India, Thailand, and Venezuela, among others, telephone connections were bought and sold in the middle or late 1970s through advertise- ments in local newspapers. In Brazil's state of Bahia in the 1980s, the telephone company facilitated the private transfer of telephone connections by acting as a broker between buy- ers and sellers and by publishing the current average price of such transfers by exchange areas. In addition, the company collected fees when telephones were moved from one loca- tion to another. 12. This example was originally cited in Saunders (1979). 13. In 1977 the exchange rate for the Myanmar (Burmese) kyat was approximately Kl.00 = $0.149. 14. As in example 22, the demand curve was assumed to be a rectangular hyperbola. 15. ITn (1988b). 16. Communications Studies and Planning International (1981). 17. This exercise was initially undertaken by Mark W. Gellerson for Robert F Gellerman at the Inter-American Development Bank, Washington, D.C. The methodology for valuing savings brought about by less frequent or less lengthy power outages is outlined in Munasinghe (1979) and Munasinghe and Gellerson (1979). EXPENDITURE METHOD 195 18. This study was carried out by Michael Tyler and Charles Jonscher, with the support of the I[L. See Communications Studies and Planning International (1986); Booz Allen and Hamilton (1984). A similar, complementary study was carried out in Costa Rica in 1986. See Booz Allen and Hamilton (1986). 19. Communications Studies and Planning International (1981) and Jonscher (1987). 20. ITU (1988c). 21. To obtain data through surveys, interviews, observation, or other methods normally requires a large amount of field work. In many instances, however, the results are not worth the effort required to prepare and evaluate a telecommunications project that has a high in- ternal financial rate of return. 22. In particular, the technical quality of service increased during the period, as did the number of subscribers that would be called by anyone with access to a telephone. 23. Also, communications media, which under given conditions might be acceptable sub- stitutes of one another, are not necessarily valued equally; for instance, it was found in ex- amples 16 and 26 that a high value was often attached to the speed of the communication and response that are obtained by using the telephone. This was also found in several of the surveys reviewed in chapters 10 through 12. 24. There is also the problem of grouping places on the basis of judgment-this, besides yielding rather arbitrary association, is not a practical proposition when the number of enti- ties to be analyzed is large. 25. Also, it is difficult to compare different program components (long distance, urban, and rural) or alternative programs with different compositions, since the extent to which benefits have been underestimated is unknown and probably varies from case to case. I Part IV Telephone Access and Use Chapter 10 Characteristics of Telephone Subscribers AS DISCUSSED IN PREVIOUS CHAPTERS, unmet demand, high economic re- turns on investment, a wide diversity of effects related to develop- ment, and declining unit costs suggest that telecommunications should be accorded a relatively high priority for new investment in de- veloping countries. In fact, were economic efficiency the only crite- rion, such evidence would typically be sufficient to justify a rapid expansion of the sector. Despite such evidence, however, planning au- thorities in developing countries often assign other needs a higher priority and thus postpone investment in the telecommunications sec- tor. Such decisions are often explained by the view that, although telecommunications investments are profitable in a financial sense, only a relatively narrow and privileged sector of the community bene- fits directly from them. Such opinions are, however, usually based on intuition rather than substantive analysis. Little is known about the extent to which lower-income groups in developing countries actually need to communicate rapidly beyond their immediate environment.' Also, the extent to which they cur- rently have effective access to communications facilities and under- stand how to use them is not well documented. Access is not only a question of the physical location of facilities relative to that of popu- lation but also a matter of service quality, pricing, and social, cultural, and educational constraints on use. In this and the following two chapters, evidence primarily from de- veloping countries is examined on the distributional aspects of tele- phone benefits and access to service. The following questions are 199 200 TELEPHONE ACCESS AND USE addressed: What is known about the characteristics of business, resi- dential, and public telephone users? What are telephones used for? What is known about telephone use by different employment, educa- tional, or income groups? And, are there consistent differences in rea- sons for use, or use patterns, by type or location of user? Although a full analysis of the distribution of telecommunications access and use in developing countries would be complex, even if reliable data were available, several limited studies have been reported. A review of these studies gives some useful indications. Many of the data examined, particularly on telephone use, were de- rived primarily from surveys of subscribers or other users. As such, they are subject to the limitations of interpretation and potential bias normally associated with survey techniques. Business and Residential Telephones For tariff purposes some telecommunications entities divide tele- phone subscribers and applicants into residential and business (non- residential) categories, and the latter sometimes into various other groupings (government, commerce, and professional). Although user categories are labeled differently by different entities, across the array of countries, the proportion of residential telephones clearly tends to increase with telephone density, that is, with the size of the telephone system relative to that of the population. This is predictable since in the early stages of system development, business users (including gov- ernment) demand a high priority and are willing to pay a high price in the expectation that reliable and rapid communications will reduce costs and increase efficiency. Such is the case in many countries in Asia and Africa, which typically have about one telephone main line per 100 inhabitants, and where 45 to 65 percent of existing lines are used by business or government subscribers. Furthermore, many of the telephones officially classified as residential are, in fact, provided to meet needs related to business or government. At higher levels of national telephone penetration, business de- mand tends to be more adequately met, and residential users receive a larger share of expanded telephone services, when the demand for telephones tends to increase with education and income. For exam- ple, in South America, where most countries average between three and ten telephone lines per 100 persons, the proportion of residential telephones is generally in the range of 60 to 80 percent. This level is equal to that in countries like the Sweden, Switzerland, and the CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 201 United States, which have more than fifty telephone lines per 100 persons, and where residential subscribers also account for 70 percent or more of all telephones. Published data on the relative proportions of business telephones or telephone lines do not necessarily reflect the primary use to which the telephones are put.2 In many of the developing countries that dif- ferentiate between types of subscribers, business subscribers pay a higher monthly rental fee than residential subscribers. One result is that when a building is used for both business and residential pur- poses (small shops, farms, and so forth), subscribers and potential subscribers have an incentive to list their telephones as residential even though they are used mostly for business.3 A few countries have at least partly overcome this problem by classifying telephones located in shared accommodations as business irrespective of where they are specifically placed; in other countries the practice is ambiguous.4 The general tendency observed among countries-that the propor- tion of residential telephones tends to increase as national telephone penetration and density rise-holds within countries as well. In Ma- laysia the proportion of residential telephone lines increased from 24 percent in 1956 to 40 percent in 1964, to 47 percent in 1978, and to 71 percent in 1988 as density increased from 0.6 telephone lines per 100 inhabitants to 0.9, 2.0, and 7.2, respectively.5 Table 10-1 shows that the proportion of residential telephones also increased in the two largest towns in Nepal, Kathmandu and Biratnagar, as the number of telephones increased from 1,947 in 1968 to 5,620 in 1977.6 The proportion of nonresidential telephone lines also tends to be larger in towns and villages than in bigger centers, and in cities with more than one exchange the proportion varies greatly with location. For example, in 1977 the proportion of lines connected to govern- ment and business subscribers in Kathmandu (Nepal's capital and largest city) was 43 percent, whereas in Biratnagar (the second largest city) it was 58 percent, and in the rest of the country it averaged 71 percent. In Ethiopia in 1982 the proportion of subscribers registered in the government, public service, business, and commercial catego- ries was 28 percent in Addis Ababa, the capital city (see table 10-2). This contrasted with 39 percent in the nation as a whole and 69 per- cent in a sample of four small manual-exchange towns. In Uruguay, in 1980 the proportion of lines officially registered as nonresidential was 20 percent in Montevideo (the capital and largest city) and about 33 percent in the rest of the country. In Thailand in 1980, the propor- tion of nonresidential telephone lines in Bangkok varied between about 10 and 80 percent depending on the location of the exchange Table 10-1. Estimated Location of Telephones, by Type of Establishment, in Nepal, 1968 and 1977 1968 1977 Ten other Ten other Type of establishment Kathmandu Biratnagar towns Kathmandu Biratnagar towns Percent of telephones located in primarily residential quarters 43 33 0 57 42 29 Percent of telephones in government sector 15 18 0 11 5 12 Percent of telephones in business, commerce, industrial, service, and mining sectors 42 49 0 32 53 59 Total telephones 1,643 304 0 4,737 883 1,671 Percent of national total 84 16 0 65 12 23 Note: Kathmandu is the capital and largest city; Biratnagar is the second largest city. Source: Adapted from Integrated Development Systems (1980). CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 203 Table 10-2. Telephone Subscribers, by Type of User, in Ethiopia: Addis Ababa and Four Small Towns, 1982 Ethiopia Addis Ababa Four small townsa Tvpe of subscriber Number Percent Number Percent Number Percent Business and commerce 17,542 23 4,630 11 51 27 Public service, government, and parastatals 12,203 16 7,225 17 81 42 Registered as residential 45,000 59 30,416 70 60 31 Other 1,525 2 995 2 0 0 Total subscribers 76,270 100 43,266 100 192 100 a. See table 10-7 for a finer breakdown of subscribers in the four small towns. Source: Ethiopian Telecommunications Authority. area, averaging 61 percent; in the provinces, where there is rarely more than one exchange per city, the proportion varied only in the range of 60 to 80 percent.7 Finally, a similar relation holds in Burkina Faso and Mali. As shown in table 10-3, in Burkina Faso in 1980, 19 percent of the tele- phones in three larger provincial towns were in the hands of individ- ual subscribers without a readily identifiable occupational listing compared with only 9 percent in six smaller towns. In Mali in 1979 the proportion of telephones in the hands of individual subscribers fell from 60 percent in Bamako, the capital city, to 11 percent in the very smallest places with some telephone service. The low proportion of only 9 percent of telephones classified as residential in small towns in Burkina Faso corresponds closely to the proportion of telephones classified as residential in twelve small exchange areas in various parts of Bangladesh. There, only 8 percent of the telephones were identi- fied as residential (see table 10-6). Users of Business Telephones in Urban Areas On a national basis, the input-output analysis reviewed in chapter 5 suggested that the most frequent users of business telephones in de- veloping as well as industrial countries are persons employed in trade, services, government administration, and to some extent transport. In other words, the tertiary sector of an economy (rather than the pri- mary and secondary sectors) generally purchases the largest quantity of telecommunications services. This finding also tends to be re- 204 TELEPHONE ACCESS AND USE Table 10-3. Numnber of Telephone Subscribers, by Functional Category, in Selected Towns in Burkina Faso, 1980, and Mali, 1979 Burkina Faso Three larger Six smaller towns with towns with 200-line smaller automatic manual exchanges exchanges Subscriber Total Percent Total Percent Public service activities Government administration' 120 n.a. 69 n.a. Education" 15 n.a. 6 n.a. Health, 18 n.a. 9 n.a. Law and order' 12 n.a. 10 n.a. Churches and missions 5 n.a. 8 n.a. Total 170 37 102 67 Economic activities Commercial banks 23 n.a. 6 n.a. Shops, hotels, garages, and gas stations 77 n.a. 17 n.a. Other business establishments, factories, and managers, 91 n.a. 5 n.a. Transport, 7 n.a. 3 n.a. Total 198 43 31 20 Other Individual subscribers, 86 n.a. 14 n.a. Public call office 3 n.a. 6 n.a. Total 89 20 20 13 Total telephonies 457 100 153 100 n.a. Not applicable. a. Central and local government offices and officials; quasi-government organizations and enterprises, such as electric power company and posts and telecommunications (offices and officials); and international and aid organizations. Government offices include special programs such as forestation, agricultural extension, meteorological service, and so forth. b. Secondary and technical schools, colleges, adult training programs, and teachers. c. Hospitals, clinics, dispensaries, physicians and nmedical workers, and drug stores. tlected in data showing the mix of business telephone subscribers by economic category and by the call traffic they generate. Information available for Thailand in 1980, and partly summarized in table 10-4. shows that business and government subscribers in Bangkok, which comprised 61 percent of all telephone lines in ser- vice, accounted for 83 percent of all calls made and a similar propor- tion of telephone revenues. Eighty-nine percent of business and government lines were connected to subscribers in the tertiary sec- tors, mainly wholesale and retail trade, finance, transport, communi- CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 205 Mali Three Six towns Sample of Bamako, towns with rural the with small services capital automatic manutal in Sikasso city exchanges exchanges region Total Percent Total Percent Total Percent Total Percent 212 n.a. 66 n.a. 84 n.a. 47 n.a. 57 n.a. 17 n.a. 10 ma. 4 n.a. 77 n.a. 28 n.a. 26 n.a. 13 n.a. 78 n.a. 22 n.a. 28 n.a. 14 n.a. 16 n.a. 11 n.a. I n.a. 0 n.a. 440 17 144 28 149 36 78 55 44 n.a. 8 n.a. 3 n.a. I n.a. 430 n.a. 74 n.a. 86 n.a. 18 n.a. 46 n.a. 12 n.a. 4 n.a. 16 n.a. 44 n.a. 14 n.a. 12 n.a. 5 n.a. 564 22 108 21 105 25 40 29 1,531 n.a. 266 n.a. 155 n.a. 15 n.a. 10 n.a. 4 n.a. 7 n.a. 8 n.a. 1,541 61 270 51 162 39 23 16 2,545 100 522 100 416 100 141 100 d. Police and courts of justice (offices and officials). e. Could include some shops. f. Railway and truck terminals and agencies. g. Telephones listed under individual nanies (could include some shops and businesses). Source: Data compiled by Gerald Buttex, Hurman Ruud, and Robert J. Saunders with assistance from the respective Office of Posts and Telecommunications, Burkina Faso and Mali. cation, utilities, and other services.8 Tertiary sector subscribers accounted for about 92 percent of all revenues generated from busi- ness lines. With regard to changes through time, it is generally accepted that as an economy develops, the tertiary sector becomes an increasingly important source of output and employment. This relative growth through time of the tertiary sector and its heavy reliance on commu- nications have incited efforts to define a fourth sector, the informa- tion sector. As outlined in chapter 5, research into the size and effect 206 TELEPHONE ACCESS AND USE Table 10-4. Distribution of Telephone Subscribers, by Type of Economic Activity, in Thailand, 1980 (percent) Bangkok rnetropolitan Subscriber area Provinces Thailand Government (excluding state enterprises) 5.2 5.1 5.2 Services Wholesale and retail trade 24.5 34.2 26.9 Banking, insurance, and real estate 3.8 2.7 3.5 Transport and communications 2.4 1.4 2.2 Electricity and water 0.2 0.5 0.3 Other services 18.1 14.6 17.1 Total services 49.0 53.4 50.0 Manufacturing 5.4 4.2 5.5 Construction 1.1 0.6 1.0 Agriculture, fishery, forestry, mining, quarrying, and others 0.2 0.9 0.3 Residences 39.1 35.8 38.0 Total 100.0 100.0 100.0 Source: Adapted from Telephone Organization of Thailand (1980). Based on a stratified random sample of 1,993 lines in nine Bangkok exchanges and 4.669 lines in twenty-four provincial exchanges. Classifications were made using information available at the local ex- changes supplemented by telephone and personal interviews. of the information sector has been undertaken in several industrial countries. In conjunction with such work, it has been observed that within each sector of an economy the level of demand for telecommu- nications services is generally a positive function of the proportion of employment that works on information-related tasks rather than on physical production itself.9 Some evidence from developing countries also supports such a conclusion. For example, data on local telephone calls and employment were collected and analyzed in a 1971 survey of 2,000 business establish- ments in the metropolitan area of Santiago, Chile, of which 1,100 had telephones.'0 The results showed that for each firm, when the total number of persons employed was correlated with the total num- ber of local telephone calls made during a typical day, three statisti- cally significant relations were found when the firms were categorized as being manufacturing industries, wholesale and retail trade, or white-collar office companies. For all three categories, the number of CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 207 daily calls increased with the number of employees, and, for any given size of firm, the white-collar office companies made more telephone calls than commercial establishments, and commercial establishments made more calls than manufacturing industries. Furthermore, as the number employed by the three types of firms increased, the number of calls made by the office firms increased faster than those of the other two, The number of local telephone calls made by the firms surveyed were then compared with the number of white-collar workers and owners." The result was that no statistically significant difference in telephone calls per white-collar person was found among the three categories of firms. Further, the observed relation between total num- ber of telephone calls and the number of white-collar workers and owriers was virtually identical to that previously observed when tele- phone calls and total employees were compared for only white-collar office firms. Such a result is consistent with the conclusion that busi- ness telephones are most important to persons employed in the ter- tiary sector or as administrators and managers in other sectors, that is, persons working in the information sector. Users of Business Telephones in Towns and Villages Several exercises provide information on the characteristics of tele- phone subscribers in towns or villages of developing countries: two each from India and Egypt, and one each from Burkina Faso, Mali, Bangladesh, Ethiopia, Nepal, Kenya, and Syria are discussed here. Table 10-5 summarizes the occupations and income levels of a sample of forty subscribers in ten small-town exchange areas in five states in India in 1978.12 Of those subscribers sampled, 92.5 percent were self- employed, and most were businesspeople with incomes that were higher than those of the majority of the population in those mostly rural areas. In a 1981 study from India in which sixty telephone sub- scribers in twelve exchange areas in the Andhra Pradesh State were sampled, 67 percent of the subscribers were engaged in business, whereas 20 percent were engaged in agriculture."3 Thirty-five percent of the subscribers earned annual incomes of less than Rsl5,000 an- nually, whereas 40 percent had annual incomes of more than Rs25,000, which, even by urban standards in India, would be consid- ered relatively high. Of the sampled subscribers, 72 percent had passed through high school, and 20 percent had taken at least some college courses. 208 TELEPHONE ACCESS AND USE Table 10-5. Distribution of Telephone Subscribers, by Primary Occupation and Income, in Ten Small Rural Exchange Areas of India, 197 7-78 (percentage in category) Annual income (rupees) Business Agriculture Professional Private Official Total 4,800-6,000 2.5 2.5 0.0 2.5 2.5 10.0 6,000-7,200 7.5 0.0 0.0 2.5 0.0 10.0 7,200-10,000 12.5 0.0 2.5 0.0 0.0 15.0 10,000+ 52.2 7.5 5.0 0.0 0.0 65.0 Total 75.0 10.0 7.5 5.0 2.5 100.0 Note: Individual percentages in each cross-tabulated cell do not merit much attention due to the small size of the total sample. Source: Adapted from National Council of Applied Economic Research, India (1978). Likewise, in a pilot study of nine rural villages in Egypt, the more educated villagers had access to and used the available telecommuni- cations and postal services for business, family, and emergency pur- poses more than lesser educated villagers. Also in the Egyptian study the relation was positive between the proportion of the population in nonagrarian occupations and telephone penetration in the villages.'4 A follow-up study surveyed 2,000 individuals in 143 Egyptian vil- lages, representing a national sample. The results confirmed those ob- tained earlier. Farmers in Egypt generally marketed their produce and purchased their inputs through government cooperatives, which tended to accentuate the shift of phone ownership from farmers to government administrators. The study found that approximately 63 percent of village omdahs (mayors or headmen) had telephones, as did 5 percent of individuals in commerce or professions but only 0.2 percent of farmers. Although each nonagrarian user averaged about eight calls a month, the population as a whole averaged 0.08 calls, or slightly less than one call a year. This means that one-quarter of 1 percent of the population made 25 percent of all calls.'" Table 10-3, referred to earlier, showed that in Burkina Faso in the six smaller towns with manual telephone exchanges, at least 82 per- cent of all telephones were located in tertiary sector offices or estab- lishments or were used primarily by persons involved in tertiary sector activities. The corresponding figure was 59 percent for the three somewhat larger towns with 200-line automatic exchanges. This was primarily because government telephones were less important. Table 10-3 also showed that a similar relation holds for Mali. Table 10-6, CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 209 Table 10-6. Telephone Subscribers, b-y User Category, in Tuelve Smiiall Exchatiges in Four Divisions (Regions) and the Dhaka Multiexchange Area of Bangladesh, December 1980 Dhaka Twvelve small multiexchange exchanges area Subscriber Number Percent Number Percent Public service Government administration' 562 30 8,392 18 Quasi-government and other Schools and colleges 61 Doctors and nurses 46 Hospitals and dispensaries 12 Clubs and associations 6 Total quasi-government and other 125 6 4,9241 11 Total public service 687 36 13,316 29 Commerce, business, and industry Shops 130 Business offices 115 Banks 100 Industrial factories 77 Lawyers 33 Agricultural farms 30 Hotels and restaurants 16 Cinema and theater 10 Business office and shop telephones listed under owners' names 540 Total commerce, business, and industry 1,051 56 - - Residential telephones 155 8 Total business and residential 1,206 64 33.067 71 Total telephones 1,893 100 46,383 100 - Not available. a. Includes some government residential telephones. b. Includes eighty-seven working connections for the railway board and connections for electric power board, water board, and so forth. Source: Telephone and Telegraph Board, Bangladesh. which summarizes a survey by the Telegraph and Telephone Board in Bangladesh, shows that for twelve small exchange areas (three in each of the four regions of Bangladesh), approximately 86 percent of the telephones were located in or used by tertiary sector establishments or agencies. Finally, in four small towns in Ethiopia, with manual- exchange service in 1982, almost all of the telephone subscribers not registered as residential were providing tertiary sector services (table 10-7). 2 10 TELEPHONE ACCESS AND USE Table 10-7. Telephone Subscribers bv User Category, in Four Small Manual-Exchange Towns from the Eastern, kVestern, Northwestern, and Central Regionis of Ethiopia, 1982 Subscriber Number Percent Business and commerce 51 27 Commercial bank offices 4 Contractors, shoe factories, tailors 4 Groceries, bakeries, hotels, bars 17 Pharmacies, gas stations, other retail shops 24 Distribution and warehousing 2 Public service and government 81 42 Municipal government administration 4 Central government administration 19 Police 6 Public utilities and public construction 15 Agriculture and forestry administration 12 Schools 10 Health service clinics 3 Social, political, and special-interest associations 12 Registered as residential 60 31 Total telephone subscribers 192 1O0 Note: Of the four towns, Bedele had fiftv-six subscribers, Butagira had fifty, Gursum had thirry-six, and Worota had fifty. Source: Ethiopian Telecommunications Authority. The Nepal study referred to earlier examined the specific use of subscriber telephones by different economic sectors and residential users in the capital city of Kathmandu and the small town of Pokhara. Table 10-8 shows that the three largest users-tourist facilities, finan- cial institutions, and government-were clearly involved in tertiary sector activities.'6 Of the registered nonresidential parties on the wait- ing list (waiters) in those two towns in early 1980, almost 65 percent were from the five clearly tertiary sectors of tourism, financial institu- tions, government, services, and trade. In a sample of fifteen residen- tial waiters, fourteen were employed in the services or trade sectors or were schoolteachers. Another example of telephone calling rates is available from a sam- ple of subscribers in several small towns and market centers in two re- gions in western and south-central Kenya. Table 10-9 shows that unlike the two somewhat larger principal towns in Nepal, calling rates in these more rural places were highest for the transport and storage CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 211 Table 10-8. Estimated Average Monthly Calls per Telephone, by Sector, in Two Towns in Nepal, 1979-80 Subscriber Kathniandu Pokhara Tourism sector, 1,043 219 Financial institutions 688 - Government 659 - Mining and quarrying 658 - Agricultural agencies and corporations 601 - Construction sector 424 - Services' 422 - Wholesale and retail trade 365 73 Transport and communications 271 - Residential 196 72 Electricity 160 - Total average 335 102 Average excluding residential 496 122 Number of observations in twelve-month stratified random sample 147 20 - No subscribers. a. Refers only to hotels, tour groups and guides, airlines, and so forth. b. Refers to hospitals, social organizations, international agencies, diplomatic missions, and so forth. c. Factors leading to the lower average in Pokhara include the fact that Pokhara is served by one small manual exchange connecting about 125 subscribers. Call traffic is also congested. Source: Adapted from Integrated Development Systems (1980). sector. The two regions sampled contain mostly smaller towns and market centers, and they are located a considerable distance from Nairobi. Hence, these towns relied heavily on vehicle transport to move supplies and goods. Financial institutions had the second high- est calling rates in both the Kenyan and Nepalese studies. Finally, in a study initiated in Syria, information was collected on the location of telephones by type of establishment in eighteen small localities.'7 The communities had an average of 116 telephones, with the smallest having 33 and the largest 242. Telephone waiting lists ex- isted in some of the communities. Table 10-10 shows that, just as in the other studies reviewed, businesses operating in the tertiary sector, even in the microcosm of these small Syrian communities, had the greatest access to telephones. The top five location-occupation activi- ties, which can all clearly be categorized as tertiary, accounted for more than 80 percent of total business and government telephones.)8 2 1 2 TELEPHONE ACCESS AND USE Table 10-9. Telephone Calling Rates for Telephone Subscribers, b-v Economic Category, in Small Towns and Rural Centers in Two Regions of Kenya, 1980 (number of calls per day per respondent) Telephone Telephone calls Subscriber calls made received Total Transport and storage 38.1 43.0 81.1 Banking and finance 22.6 23.6 46.2 Soft industry, such as food processing, textiles, paper products 17.6 27.3 44.9 Public utilities 20.1 22.8 42.9 Public administration and security 16.5 21.9 38.4 Telecommunications, broadcasting, and printing 12.0 17.5 29.5 Mining, metal chemical, and machinery industries 14.3 14.5 28.8 Wholesale trade 9.0 19.2 28.2 Education, health, and other services 10.4 12.7 23.1 Construction 10.6 11.0 21.6 Retail trade 10.0 11.1 21.1 Plantation agriculture 5.4 14.9 20.3 Other agriculture 4.7 5.3 10.0 Mean 12.9 16.5 n.a. n.a. Not applicable. Source: Adapted from Cleevely and Walsham (1980). Users of Residential Telephones A study of urban residential telephones in Chile found that income was the most important variable influencing both the demand for telephones and local and long-distance calling rates.'9 About 80 per- cent of the sampled families in the upper decile of the income distri- bution demanded a telephone, whereas only about 10 percent in the fifth decile and 2 percent in the lowest decile requested telephone service. What is particularly interesting is that local calling rates in- creased sharply with income (from four calls a day at the lowest in- comes to fourteen calls a day at the highest) even though there was no charge for local calls. Such a result suggests that urban residential telephone use was influenced not only by purchasing power but also by the extent to which different segments of the population partici- pate in community life. In other words, income (which was also found CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 213 Table 10-10. Distribution of Telephones, by Functional Location, in Eighteen Small Locales in Syria, 1976 Total number of Percent of Location of telephone telephones telephones Government 372 17.8 Business Shops 333 16.0 Traders 118 5.7 Doctors and pharmacies 49 2.4 Auto service facilities 42 2.0 Unions and clubs 42 2.0 Public telephones 39 1.9 Restaurants and cafes 29 1.4 General offices 27 1.3 Small industries 26 1.2 Cinemas 13 0.6 Education and religious groups 12 0.6 Societies 9 0.4 Hotels 8 0.4 Farms 7 0.3 Other businesses 4 0.2 Total business and government 1,130 54.2 Residential location 954 45.8 Total 2,084 100.0 Source: Data compiled by Syed Sathar and Yuji Kubo, with assistance from the Syrian Telecommunications Establishment. to be closely correlated with education and occupation) may be re- lated to urban residential telephone use not only as a constraint on total expenditure but also as a factor in, and indicator of, the inten- sity, complexity, and range of the family's communication pattern.20 Two complementary findings in the Chilean study are also of inter- est. First, at a given income level, the aggregate demand for residen- tial telephones was significantly higher among urban households that were already intensively using other means of communication (mainly travel, post, and telegraph) than among those that were not. Second, families with residential telephones averaged two to four times as many communications with kin, friends, and special-interest groups as did families with like incomes but without telephones. Somewhat similar results were found in an examination of residen- tial telephone subscribers in Guatemala City. In July 1981 a sample of 2,360 residential subscribers was classified into low-, middle-, and 2 14 TELEPHONE ACCESS AND USE high-income categories on the basis of a field survey in which the quality and value of their housing were judged by external visual in- spection. On this basis, it was estimated that about 13 percent of cur- rent residential telephone subscribers in Guatemala City lived in housing that would be typical of relatively low-income inhabitants, 49 percent middle-income, and 38 percent high-income. As in the Chil- ean example, on average, the telephone calling rates for Ibcal, long- distance, and international service also increased with the proxy for income level. Occupants of the relatively high-income residences gen- erated an average telephone bill almost three times that of middle- income homes and four times that of lower-income subscribers.21 Nevertheless, even though income (and education) is a factor in urban residential telephone demand and use, at least in some of the relatively high-income Latin American countries, telephones seem to be demanded by and reasonably well dispersed among the middle- and lower-income groups. For example, in Montevideo, Uruguay, in 1980 only 7 percent of residential telephone lines (and about 6 per- cent of outstanding applicants) were in areas of high income, 15 per- cent in upper middle income, 56 percent in lower middle income, and 22 percent in low income. In Medellin, Colombia, about 70 percent of the planned 1980-90 investment in residential telephones was pro- posed to meet the demand of middle- and low-income families.22 Evidence is also available on the characteristics of residential tele- phone subscribers in lower-income nonmetropolitan communities. The Nepalese study, noted earlier, provided information on the educa- tional level and occupations of a sample of residential subscribers in the capital city of Kathmandu.23 It showed that 87 percent of the resi- dential subscribers had a general education and 13 percent a profes- sional education (engineering, medicine, law, and so forth). Of the total, 23 percent could be categorized as "only literate," with the re- maining 77 percent having attained higher levels of academic certifi- cation. The conclusion is that residential telephone ownership goes hand in hand with higher education. Table 10-11 shows that busi- nesspersons and government officials dominate residential telephone access in a sample of residential subscribers in Kathmandu. In Sri Lanka, most residential respondents in a survey of new sub- scribers in Colombo and Kandy were relatively well off. Even so, ex- cept for the 20 percent of respondents in the highest income bracket (about Rs5,000 a month) who spent a little over 3 percent of their income on telephone usage, most subscribers (in income brackets ranging from Rs2,000 to Rs5,000 a month) were willing to spend as much as 5.5 to 7.6 percent of their income to use the telephone. CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 215 Table 10-11. Occupations of a Sample of Seventy-eight Residential Subscribers in Kathmandu, Nepal, 1980 Occupation Percent Businesspersons 21.4 Government officers 18.6 Politicians 14.3 Engineers and doctors 10.0 Teachers 7.1 Social workers 4.3 Lawyers 2.9 Not mentioned 21.4 Total 100.0 Source: Adapted from Integrated Development Systems (1980). Given that the connection fee was RslO,000, even well-off users had to make a substantial expenditure to pay for their new phones. Busi- ness subscribers, by contrast, spent an average of only 0.38 percent of sales on their telephone usage.24 A general survey of four countries in Southeast Asia (Malaysia, Singapore, Sri Lanka, and Thailand) used demographic information and found that where residential telephones were available, all income groups above a certain threshold (occurring in the middle class) were subscribers. In urban areas, as many as one-third of all residential phones belonged to the "poorer" half of all households, but this same type of household had only one-fifth of telephones in rural areas.25 With regard to residential telephone users in even smaller, more rural communities, the survey of eighteen small localities in Syria, re- ferred to earlier, also provided limited information on the occupations of residential subscribers. From the data presented in table 10-12 it can be inferred that in these Syrian villages the better-educated mem- bers of the communities wvere also well represented among the resi- dential telephone subscribers. Furthermore, given that more than 70 percent of the subscribers were farmers, drivers, teachers, or other public officials and employees, most of the residential subscribers were probably from the middle-income group. Finally, the respondents in a survey of selected residential tele- phone subscribers in twelve small towns in Costa Rica were employed as outlined in table 10-13. Of these residential subscribers, 29 percent did not complete primary school, 19 percent completed only primary school, 24 percent were educated at the secondary level, and 28 per- cent were educated at the postsecondary level. The conclusion is that 2 1 6 TELEPHONE ACCESS AND USE Table 10-12. Distribution of Residential Telephones, by Occupation of Subscriber, in Eighteen Small Locales in Syria, 1976 Total Occupation of residential number of Percent of subscriber telephones telephones Public officials and employees 292 30.6 Farmers 214 22.4 Teachers 93 9.7 Drivers 92 9.6 General workers 73 7.7 Merchants 43 4.6 Owners of shops 40 4.2 Doctors 28 2.9 Owners of small industries 22 2.3 Lawyers 16 1.7 Contractors 13 1.4 Cafe owners 11 1.2 Pharmacy owners 11 1.2 Engineers 6 0.6 Total 954 100.0 Source: Data compiled by Syed Sathar and Yuji Kubo, with assistance from the Syrian Telecommunications Establishment. Table 10-13. Employment Categories for 163 Residential Telephone Subscribers in Twelve Small Towns in Costa Rica, 1980 Employment category Percent Not working 5 Government Professional 24 Nonprofessional 13 Total 37 Private Bars, restaurants, and cantinas 6 Retail stores and shops 14 Small industries and nonprofessional services 10 Professional services 4 Transport and communications 6 Total 40 Agriculture and livestock 18 Total 100 Source: Adapted from Kilgour (1982), p. 209. CHARACTERISTICS OF TELEPHONE SUBSCRIBERS 217 the subscribers roughly represented a cross-section of the rural middle class. 26 Notes 1. The distributional issue in telecommunications, of course, has many other dimen- sions, of which distribution betveen different inconme groups is only one. The distribution between urban and rural areas, and between the more modern sector and the "informal sec- tor" in urban areas, are also important, as is the overall balance of provision among regions, between subscribers and rco users, between local and long-distance facilities, among differ- ent types of services, and so forth. 2. Such inaccuracies are probably partly responsible for some of the questionable results found in the analysis of statistics within and among countries done by Hardy (1980). 3. This is also sometimes the case in industrial countries. Mayer, in presenting data on the frequency of residential calls in the United States, explains some of the subscribers with abnormally high calling rates by noting that some "are undoubtedly businesses conducted from the home, and they pay a residential rather than a business rate." Mayer (1977), pp. 246-61. 4. One East Asian country abandoned the dichotomy of business and residential classifi- cation primarily because of the problems with subscriber relations, which the telecommuni- cations entity was having with small shop owners in the capital city. These small business owners lived in the back of their shops and usually requested that their telephones be in- stalled on the residential side of the wall separating their living quarters from their shop. It was estimated that during business hours the telephones were used 90 percent of the time for business purposes, but the shop owners insisted that the lower residential tariffs should be charged because the telephone was located in their residence. 5. American Telephone and Telegraph Corporation (1980, 1981, 1982, 1989, 1990) and World Bank project documents. 6. In the Nepal survey, "residential" was to a certain extent defined as "ownership of dwellings." 7. Telephone Organization of Thailand (1980). 8. In mid-1980 when these data were collected, telephone service was limited in rural areas in Thailand. Aggregate national data thus essentially reflect the situation in Bangkok and several smaller urban places. 9. For one example, see Pierson (1979). 10. Wellenius (1971). Original analysis carried out by Richard Meunier. I 1. Owners were included in this category to accommodate small shops or establish- ments with only one or two employees and in which the owner may have been the only per- son serving in a primarily white-collar function. 12. National Council of Applied Economic Research, India (1978). Two exchange areas were surveyed in each of the states of Harvana, Punjab, Madhya Pradesh, Rajasthan, and Uttar Pradesh. 13. Economics Study Cell, Posts and Telegraphs Board, India (1981). 14. Kamal, Dessouki, and Pool (1980). 15. Pool and Steven (1983). 16. In 1979, 95 percent of the lines in Nepal's beginning telex system (with only fifty-six lines in operation at that time) were connected to entities in the tertiary sector, especially tourism (23 percent), commerce (23 percent), and transport and communication (16 percent). 2 1 8 TELEPHONE ACCESS AND USE 17. Initial data for this exercise were compiled by Yuji Kubo and Syed Sathar with the co- operation and assistance of the Syrian Telecommunications Establishment. 18. Part of the reason for this, of course, may be that tertiary sector activities tend to be clustered near the spatial center of communities, where telephone access is usually available first. Hence, farnis. small industries, and so forth, although they may desire service, are often located outside existing local service areas. 19. Wellenius (1978), chap. 9 and Wellenius (1969b). 20. Also some of the higher-income, higher-calling-rate residential telephone users could have been professionals working at least part-time from their residences, as is common in developing countries. 21. World Bank data. 22. World Bank data. 23. Integrated Development Systems (1980). 24. Kojina, Hoken, and Saito (1984). 25. Engvall (1986). 26. Kilgour (1982), pp. 209-20. Of further interest is the finding that 58 percent of the professionals surveyed considered themselves the primary user of the home phone; this compares with 52 percent of nonprofessionals and 41 percent of farmers. Chapter II Use of Business and Residential Telephones IN VIEW OF THE EVIDENCE in the previous chapter it is not surprising that telephones in developing countries are used-as distinct from the cat- egory of users-overwhelmingly for purposes related to government and business. For example, in 1980 in Bangkok, Thailand, although 39 percent of the lines were residential, an estimated 85 percent of all calls related to business and government activities. Estimates of tele- phone use in large urban areas in Egypt, El Salvador, and India showed that more than 90 percent of all calls made during the peak business hours were for purposes related to business. Business Telephone Use Even where most lines are residential, business use tends to domi- nate. In Montevideo, Uruguay, where almost 80 percent of the tele- phone lines are residential, survey data suggested that about 60 percent of all telephone calls and 70 percent of telephone revenues related to business functions. In Guatemala City, partly because tar- iffs paid by residences are lower than those paid by business and gov- ernment subscribers, 67 percent of all subscribers were categorized as residential. As in the above instances, however, business and govern- ment subscribers are the most intensive users of telephones. From a sample taken in July 1981 of 4,400 lines registered in Guatemala City, business and government lines averaged 3.2 times as many outgoing international calls, 2.3 times as many long-distance calls, and 1.4 219 220 TELEPHONE ACCESS AND USE times as many local calls as did residential lines. Overall, subscribers officially categorized as being business and government accounted for more than 60 percent of all telephone company revenues.' The 1971 survey of 1,100 business telephones in Santiago, Chile, showed that 88 percent of those telephones were used most frequently for work-related communication.2 Of these business telephones, 10 percent were, however, used most frequently for social purposes, that is, to communicate with members of the family, kin, and friends. The remaining 2 percent were used for unspecified purposes, which appar- ently included household-related calls (to order goods to be delivered at home and so forth). Studies in several other Chilean cities showed the proportion of business telephones used most frequently for social purposes to vary between 9 and 33 percent.3 In the 1978 study of ten small-town exchange areas in five states of India, referred to in the previous chapter, 72 percent of the most re- cent calls made by the (mostly business) subscribers in the sample were to conduct business, 18 percent were to contact family members, and 10 percent were to reach friends or undertake public service, health, or other purposes.4 Of particular interest is that callers per- ceived 25 percent of their most recent calls to be of an urgent nature and a further 50 percent to be calls that could not be delayed until the following day. ln the 1981 study of twelve small-town exchange areas in the Andhra Pradesh State of India, the purpose of calls varied greatly de- pending on whether the call was local or long distance.5 Of the sam- pled subscribers, 70 percent frequently used the telephone for long-distance communications, and, as table 11-1 shows, calls relating Table 11-1. Proportion of Subscribers Making Local and Long-Distance Calls in Twelve Villages in Andhra Pradesh State, India, 1981 (percent) Local Long-distance Purpose of call calls calls Agricultural business 18 40 Nonagricultural business 23 23 Both agricultural and nonagricultural business 13 22 Professional needs and contacting governmnent offices 12 13 Emergencies and social calls 34 2 Source: Adapted from Economics Study Cell, Posts and Telegraphs Board, India, (1981). USE OF BUSINESS AND RESIDENTIAL TELEPHONES 221 to some form of agricultural, business, government, or professional activity involved roughly 98 percent of all long-distance calls, al- though they involved only 66 percent of local calls. With regard to the urgency of the most recent call made, 73 percent of the subscrib- ers sampled stated that the last long-distance call made was of imme- diate necessity concerning their work and business; 27 percent felt that the call could have waited. Finally, two-thirds of the subscribers sampled made most of their long-distance calls during off-peak peri- ods. This was probably partly because off-peak tariffs were lower and partly because long-distance and large urban local networks tended to be highly congested during business hours, greatly reducing the chance of getting through. In the Nepalese study, a sampled group of trade and tourism estab- lishments in both the capital of Kathmandu and the smaller town of Pokhara also responded that their most frequent reason for calling was for so-called productive purposes. As table 11-2 shows, these calls were even more important among persons who had to use someone else's telephone, and emergency calls constituted a significant 6 to 14 percent of all calls made or received.6 Finally, business applicants waiting for telephone service to become available were asked to rank the reasons why they needed their own telephone. On average they ranked business promotion as number 1, saving transport and postal Table 11-2. Most Frequent Reasons for Using Primarily Business Telephones, for Two Types of Establishments in Kathmsandu and Pokhara and for kVaiters in Kathrnandu, Nepal, 1980 (percent) Tourism Registered establishments: hotels, g Wholesale and retail tour guides, waiters in trade establishments airlines, etc. wholesale and retail trade, Purpose of call Kathmandu Pokhara Kathmandu Pokhara Kathmandu, Social 29 34 32 26 6 Productive 50 45 50 60 81 Emergency 8 14 7 7 6 Unspecified 13 7 11 7 7 Note: Calls include both incoming and outgoing. a. Waiters do not have telephone connections but use the telephones of friends or other private sources. Source: Adapted from Integrated Development Systems (1980). 222 TELEPHONE ACCESS AND USE costs as number 2, and saving time as number 3. Public relations were ranked fourth and "social prestige" last.7 Somewhat similar results emerged from the sample of (mainly busi- ness) telephone subscribers in towns and market centers in the two regions in west and south-central Kenya. Of a total of 493 sampled telephone subscribers, 81 percent used their telephone only or mainly for business. Of the remainder, 11 percent used it for "some busi- ness," whereas 8 percent used it only for social purposes.' Table 11-3 gives a more detailed breakdown of the results. When 574 subscribers, applicants on waiting lists, and public coin box telephone users were asked to list and describe their most important telephone contacts, the dominant reasons noted were placing and receiving orders, giving or receiving information (which presumably takes place in almost all calls), and taking care of administrative matters.9 Together these three highly interrelated categories accounted for 69 percent of all contacts. Table 11-4 shows the results of a very similar survey undertaken in 1982 in three medium-size towns in Ethiopia. There, when 996 tele- phone subscribers and public call office users were given the summary list of reasons for making telephone contact that emerged from the Table 11-3. Frequency of Reasons Given for Telephone Contact by a Sample of Telephone lJsers in Two Regions of Kenya, 1980 Times Number Average times mentioned of times mentioned per 100 Reasons for telephone contact mentioned per respondent respondents Placing or receiving orders 803 1.4 33 Giving or receiving information 653 1.1 27 Handling administration 210 0.4 9 Ordering spares or repairs 129 0.2 5 Handling payments or finances 121 0.2 5 Arranging transport 109 0.2 4 Arranging meetings 98 0.2 4 Soliciting business 54 0.1 2 Conducting urgent business 13 0.0 I Negotiating 8 0.0 0 Other 255 0.4 10 Total 2,453 n.a. 100 n.a. Not applicable. Note: Users comprised subscribers, registered applicants, and public coin box telephone users. Source: Adapted from Cleevely and Walsham (1980). USE OF BUSINESS AND RESIDENTIAL TELEPHONES 223 Table 11-4. Reasons Given for Most Frequent Telephone Contact by a Sample of Telephone Users from Three Mediutn-size Towns in Ethiopia, 1982 Times Number Auerage times mentioned of times mentioned per 100 Reasons for telephone contact mentioned per respondent respondents Conducting urgent business 717 0.72 22 Soliciting business 337 0.34 1o Placing or receiving orders 286 0.29 9 Handling administration 242 0.24 7 Giving or receiving information 237 0.24 7 Handling payments or finances 224 0.22 7 Arranging meetings 163 0.16 5 Ordering spares or repairs 138 0.14 4 Arranging transport 119 0.12 4 Negotiating 53 0.05 2 Other 736 0.74 23 Total 3,252 n.a. 100 n.a. Not applicable. Note: Users comprised both subscribers and public coin box and call office users. Of the three towns, Nazareth had 1,556 subscribers, Awassa had 547, and Shashemene had 500. In total 996 users were sampled: 586 from Nazareth. 222 from Awassa, and 188 from Shashemene. Source: Ethiopian Telecommunications Authority. Kenyan study, they most frequently identified urgent business (which in translation was generally interpreted to mean urgent business or so- cial matters) and soliciting business as the two principal reasons for telephone contact. These were then followed by the three reasons most frequently mentioned in the Kenyan survey: placing or receiving orders, handling administration, and giving or receiving information. There are sometimes, however, perceived problems in the strong re- lation between business telephones and business use. In Papua New Guinea, for example, a 1977 study found that business owners and managers felt that business and government telephones were used by employees too frequently for personal purposes, to the extent that em- ployers would sometimes fit locks on telephone dials to deter outgo- ing calls.'0 Such an action, of course, also denied the employer some benefits related to communication. Finally, the Kenyan survey threw some light on the nature of tele- phone use by the agricultural sector and reinforced the intuitive con- jecture that although agriculture is generally a less-intensive telephone Table 11-5. Percentage of Important Telephone Contacts between Sectors, for a Sample of Telephlone Subscribers, 'Waiters, and Coin Box IJsers in Two Regions of Kenya, 1980 Percent of Number contact of other SeCtOT TeCeiving call ~~outside own econonuic Sector receiving call Residential economic categories Sector making call 1 2 3 4 5 6 7 8 9 10 11 12 customers category contacted 1. Retail trade 14 5 13 39 8 8 7 85.7 6 2. Plantation and other agriculture 9 20 6 17 7 9 5 16 8 80.4 8 3. Transport and storage 5 26 12 15 15 4 5 16 74.4 7 4. Textiles, paper, and food processing 5 7 10 30 9 12 5 4 8 7 69.6 9 5. Wholesale trade 30 32 18 5 68.2 3 6. Utilities 33 6 9 6 15 21 66.7 5 7. Mining, metals, and chemicals 6 6 8 13 44 5 5 9 56.3 7 8. Construction 4 5 6 10 47 4 12 5 52.6 7 9. Banking and finance 6 6 48 6 6 25 52.1 5 10. Telecommunications, printing, and broadcast 4 4 4 13 50 9 7 7 50.0 7 11. Education and health ser- vices 12 5 50 12 10 19.8 4 12. Public administration and security 4 11 59 10 40.6 3 Note: This table should be read by rows. Values given are the percentage of time the calling sector mentioned the receiving sector as being a major recipi- ent of calls. Each respondent could mention lip to five sectors that were frequently called. Values of less than or equal to 3 percent were omitted to clarify the pattern. Source: Adapted from Cleevely and Walshanm (1980). USE OF BUSINESS AND RESIDENTIAL TELEPHONES 225 user, it needs, because of the diversity of its inputs and outputs, to communicate with a wider variety of other economic activities than do most other sectors. Table 10-9 showed that in the two regions sampled in Kenya the average calling rates for plantation and other agriculture were considerably below those for all other sectors. Such a finding is consistent with the input-output results observed in chapter 5, in which the agricultural sector purchased fewer communication services than did most other sectors. A low average calling rate, however, does not necessarily mean that the agricultural sector has little need to communicate rapidly with other sectors. In fact, of all the sectors surveyed in these two regions of Kenya, the agricultural sector initiated the greatest variety of tele- phone contacts with other sectors. From table 11-5 it can be seen that about 80 percent of all calls originating in the agricultural sector were directed to other sectors, with the most important being to food processing and textiles, education and health services, retail trade, chemicals and metal fabrication and manufacture, government, wholesale trade, and transport and storage. In fact, if one were look- ing for a sector whose communication linkages are mostly productive, it is noteworthy that agriculture was the only sector that had insignifi- cant contact with final consumers in the residential sector.'' Also of note from table 11-5 is that, other than agriculture, retail trade and transport and storage (both tertiary activities) were the most outward-communicating sectors, whereas at the other end of the spectrum both the public administration and security and the educa- tion and health services sectors (also tertiary activities) communi- cated with themselves and with each other well over 60 percent of the time. Residential Telephone Use Data available on the uses of residential telephones suggest that in general the largest proportion of calls made on such telephones per- tain to social or family uses. Table 11-6 shows that for samples of resi- dential subscribers from six countries, calls classified as social accounted for about 70 percent in three of them and around 50 per- cent in the remaining three. Of the developing countries shown, three had social calls in the 50 percent range, whereas one (Chile) ap- proached 70 percent. The lower proportion of social calls in Papua New Guinea could be related to the fact that a few business subscrib- ers might have been included in the sample and that in Papua New Table 11-6. Reasons for Using Residential Telephones in Several Countries (percentage of calls) Nepal, Nepal, Kathmandu, Papua United Thailand& Kathmandu, residential New Chile States and telephone Use Guinea, (urban) b Englandc (New York)d Bangkok Provinces Pokhara' applicantsg Social 50 69 74 67 51 40 52 34 Friends 26 39 - 36 26 18 - - Family 24 30 - 31 25 22 - - Business 31 20 17 16 42 12 24 45 Goods and services acquisition for hone - 7 5 4 7 46 - - Emergency services 17 3 4 - - - 14 8 r-z ~ Other 2 1 - 13 - 2 10 13 t__ - Not available. a. Papua New Guinea data exclude expatriates and may include some business subscribers. In the survey, the subscribers sampled were asked what they used their telephones for; there were on average 1.4 answers per respondent. The figures presented are percentages scaled down tO Suln 100. b. Figures represent the proportion of sampled telephone lines reported to be used most frequently for the given purpose. Figure given for "friends" in- cludes communication with kin. c. These results are for old customers in the British Post Office's 1971 residential research sample of current supply. Figures represent the proportion of telephone lines reported to be used most frequently for the given purpose. d. These results were observed when 90,000 subscribers lost their telephone service for twenty-three days because of a fire in the switching center. A sample of those residential subscribers was asked what types of calls they most missed being able to make. "Other" includes 10 percent given as "medical." e. These figures are based on a survey of 108 residential telephone subscribers and applicants in Bangkok metropolitan area and 100 in provincial cities and towns. "Family" includes communication with kin. Applicants used borrowed and public telephones and comprised 20 percent of the sample surveyed. Figures represent the proportion of sampled telephone lines reported to be used most frequently for the given purpose. f. These include calls both made and received. g. Waiters do not have telephones but use those of friends and other private subscribers. Source: Evans and Bryan (1977); Wellenius (1978), chap. 9; British Post Office personal communication; Srivisal and Tamnura (1980); Integrated Devel- opment Systems (1980); Wurtzel and Turner (1977), pp. 246-61. USE OF BUSINESS AND RESIDENTIAL TELEPHONES 227 Guinea (as in Nepal and Thailand) telephone density was so low that many friends and relatives of telephone subscribers did not have telephones.12 Nevertheless, residential telephones in all six countries were also used to an important extent for economic activities related to produc- tion and distribution. In the Chilean survey, 20 percent of the urban residential telephones were used most often to further work-related purposes, 7 percent to access goods and services for the household, and 3 percent to obtain services under emergency conditions (mainly to contact urgent medical assistance, firefighters, or police). In Papua New Guinea and overall in Thailand, approximately 31 and 28 per- cent, respectively, of the residential subscribers surveyed indicated business-related communication (excluding calls to suppliers of goods and services for the home) as the primary reason for using the home telephone, whereas in Nepal a high figure of 45 percent of the calls that persons waiting to receive a residential telephone made using the telephone of a neighbor or friend were reportedly related to busi- ness.13 Similarly, in urban areas of Sri Lanka, residential phones were used for business as well as for private purposes about 40 percent of the time among recent subscribers surveyed in Colombo and Kandy.'4 In provincial Thailand, each subcategory of business-related calls made from the home (calls to one's office, customers, and govern- ment services) had a considerably lower incidence than in Bangkok: only 12 percent of the families surveyed gave these as the most fre- quent uses of their residential telephone as against 42 percent in the metropolis; a far greater proportion of residential telephones in the provinces (46 percent) were mostly used to contact shops for home supplies (presumably a more viable and common practice in small places than in a large metropolis), bringing the total of residential telephones used primarily for purposes directly related to economic activity to 58 percent, even higher than in the Bangkok metropolitan area (49 percent). The study in Sri Lanka examined the frequency with which non- subscribers were able to borrow subscribers' phones and to leave mes- sages, an important aspect of telephone use in areas where density is extremely low. Although only one-third of businesses surveyed were willing to offer their telephones to borrowers, two-thirds of residential subscribers said that they would allow neighbors to use their tele- phone. Both groups were somewhat less willing to deliver messages left for nonsubscribers-one-quarter of businesses and two-fifths of residential subscribers said they would do this. The researchers con- cluded that in Sri Lanka, many telephone subscribers considered their 228 TELEPHONE ACCESS AND USE phones to be public property and appeared to use their phones very efficiently. In contrast, Pool and Stevens found that many telephones in wealthy households in Egyptian villages were underused. Social custom seems to play a larger role than economic factors in determin- ing the access that outsiders have to home phones.'" A survey of residential telephone subscribers in twelve small towns in Costa Rica did not ask the explicit purpose of telephone use but did ask if having a telephone in the home had changed how the re- spondents did their work. Of the 93 percent of the sample who were working, 79 percent said they had noticed a change, and 51 percent of these indicated that they used their home phone primarily for work purposes. When these respondents were asked how work had changed because of the telephone, 41 percent mentioned saving time, 16 per- cent getting information, 14 percent obtaining products, and 13 per- cent traveling less.'6 When the small-town residential subscribers in the Costa Rican sample were asked whether the installation of the telephone had caused a change in the town, 93 percent replied affirmatively. Of those, 44 percent stated that the telephone allowed more contact with family or made contact easier, 19 percent noted it reduced the need to travel or made travel easier and saved time or money, 13 percent noted the telephone induced more progress and development in the town, 9 percent stated it brought more information and news so the town was less isolated, 7 percent noted it allowed problems to be solved more rapidly, and 5 percent stated the telephone allowed more rapid help in emergencies.'7 One additional inference about how individuals use residential tele- phones for business can be made from the occupations of residential subscribers in the eighteen small towns in Syria, which were presented in table 10-12. Although, from the information available, it is not possible to calculate the extent to which residential telephones were used for business-related purposes, most persons in the occupations listed in table 10-12 (farmers, drivers, merchants, shop and industry owners, doctors, lawyers, and so forth) would certainly have occasion to use their telephones for reasons directly related to their em- ployment. Of course, the most frequent use of a residential telephone is not always regarded as the most important. For example, a 1971 survey of newly connected residential customers in Great Britain showed that they considered emergency uses to be the most important reason to have a telephone, whereas calls to family, kin, and friends were the most frequent use.'8 The families surveyed decided to apply for a USE OF BUSINESS AND RESIDENTIAL TELEPHONES 229 home telephone both because of the potential to make occasional but highly important calls and because of the ability to make more com- mon calls of lower individual value. A similar pattern was found elsewhere. In a 1973 survey of 1,500 households in and around Tokyo, an overwhelming 72 percent catego- rized the residential telephone as "very useful" in case of accidents.19 The second highest category to receive the "very useful" designation was work or profession, with 53 percent. Also, the residential sub- scribers sampled in the two countries with the lowest per capita in- come (Papua New Guinea and Nepal) tended to rank emergency use as much more prevalent than did residential subscribers in the other countries discussed in this section. In particular, applicants for resi- dential (as well as business) telephones in Papua New Guinea stated a concern for security, protection, and the possibility of an emergency as the main reason for wanting to have a telephone, even though, in fact, such calls were likely to be relatively infrequent; 41 percent of the existing subscribers nominated emergency and safety as major rea- sons for having a telephone, whereas only 25 percent gave these as reasons for actually using it.20 Telephones and Other Means of Communication The survey of residential telephone subscribers in Thailand also in- dicated the way in which subscribers perceived the importance of tele- phones relative to other forms of communication media. As table 11-7 shows, the telephone was overwhelmingly the preferred means of dealing with emergencies and setting up appointments, as well as communicating with kin and friends. The table also shows which other means of communication would have been preferred if the sub- scriber had not had a telephone, although over 40 percent still pre- ferred a telephone, albeit a public one. In Bangkok only 2 percent said they would not communicate at all.2- There were also notable differences between Bangkok and the prov- inces. For example, in the provinces the telephone was not preferred for setting up appointments as often as it was in Bangkok (although in both cases it was the means preferred most often); meetings were indicated somewhat more frequently in the provinces, possibly reflect- ing shorter travel distances and sometimes poorer-quality service. The public telephone was perceived to be more important in the provinces than in the metropolis as an alternative to a residential telephone, ac- companied also by a slightly greater preference for letters and tele- 230 TELEPHONE ACCESS AND USE Table 11-7. Preferences for Alternative Communication Media of 208 Residential Telephone Subscribers and Applicants in Thailand, 1980 Percent of subscribers Alternative sampled Item media Bangkok Provinces Preferred media to report an Telephone 86 86 emergency Other people 8 4 Other (fire alarm and so forth) 5 7 Unknown 1 3 Total 100 100 Preferred media to set up an Telephone 83 77 appointment Meeting 10 15 Letter 2 6 Messenger 1 2 Other - - Unknown 4 - Total 100 100 Preferred media to Telephone 45 42 communicate with kin and Visit 34 31 friends Letter 3 11 Other I I Unknown 17 15 Total 100 100 Preferred media to gather news Newspaper 41 38 Television 31 27 Telephone 13 15 Neighbor 1 5 Other 9 10 Unknown 5 5 Total 100 100 Other communication means Public telephone 43 49 preferred if the subscriber Travel to meet parry 16 16 did not have a telephone Letter 14 19 Telegram 7 11 Other 17i, 5 Unknown 3 - Total 100 100 - Not available. a. Neighbor's telephone, 6 percent; government radio, 5 percent; messenger, 2 percent; telex, 2 percent; and no communication, 2 percent. Source: Adapted from Srivisal and Tamura (1980). USE OF BUSINESS AND RESIDENTIAL TELEPHONES 231 grams and less preference for other means (including neighbors' telephones, which are in many instances not available in small com- munities with low telephone densities). In the Kenyan study of telephone use in two rural regions, both business and residential subscribers were asked wvhat means of com- munication they used to contact other economic sectors; they were al- lowed to list up to five methods. The results shown in table 11-8 reveal that the telephone and postal service accounted for almost 80 percent of all economic contacts, with the telephone being the main or only means (among those listed) in more than half of the contacts mentioned. The survey of residential telephone subscribers in the twelve Costa Rican towns noted earlier yielded a similar result. When the sample respondents were asked which means of communication they used most frequently to communicate with San Jose or any other city in the course of their work, 94 percent stated the telephone, with the other 6 percent noting telegraph, radio, messenger, or other.12 Finally, the survey of 474 males in nine small Egyptian villages showed that where access to telecommunications services was mini- mal and the long-distance telephone service was unreliable and con- gested, persons who communicated outside the village by means other than physical travel felt that telephones were important.23 Neverthe- less, even those who used one or more of the means of communica- tion listed in table 11-9 did so infrequently; the primary reason for this was that the persons they wished to contact could not be easily reached by existing postal or telecommunications facilities. Partly to compensate for this, village administrators used the telephone primar- Table 11-8. Means of Communication Used for Economic Contact by a Sample of Telephone Subscribers in Two Regions of Kenya, 1980 Percent of contacts Method using method, Only phone 16.6 Mainly phone 34.7 Phone and letter 17.7 Mainly letter 10.5 Only letter 0.1 Personal visit and other 20.4 a. Respondents were allowed to list up to five economic contacts. Source; Adapted from Cleevely and Walsham (1980). 2 3 2 TELEPHONE ACCESS AND USE Table 11-9. Means of Communication Used by 474 Male Respondents during Twelve Months in Nine Villages in Egypt Percent who Median messages had used listed sent or received Medium of communication by users communication medium during year Sent letters 16.0 5.0 Received letters 17.0 5.0 Talked on telephone 10.0 3.0 Sent telegrams 4.0 2.0 Received telegrams 3.0 2.5 Source: Adapted from Kamal, Dessouki, and Pool (1980). ily for leaving dictated messages. Similar results were obtained in the follow-up study, which covered 143 villages. Personal travel was the villagers' principal means of conveying information, and if adequate phone service had been available, 60 percent of this travel would have been unnecessary. The message system was haphazard at best, since messages were easily misunderstood in translation, and their delivery was uncertain and frequently delayed.24 Notes 1. World Bank data. 2. Wellenius (1971). The original analysis was made by Richard Meunier. 3. Jabif (1971). 4. National Council of Applied Economic Research, India (1978). 5. Economics Study Cell, Posts and Telegraphs Board, India. 6. Integrated Development Systems (1980). 7. When residential subscribers on the waiting list were asked the same question they ranked time savings as first, cost savings as second, and business promotion as third. As with the business subscribers on the waiting list, residential subscribers ranked public rela- tions fourth and social prestige last. 8. Cleevely and Walsham (1980). 9. Each respondent was allowed to mention up to five contacts. Respondents were asked to "please list the activities and locations of the most important businesses or farms or gov- ernment departments or individuals that you or this establishment deal with-and what is the main reason for your contact for each of them?" The 574 respondents in the sample gave 2,453 replies. 10. Evans and Bryan (1977). 11. Limited evidence of the importance of telecommunications to agriculture was also found in the survey of nine rural villages in Egypt (discussed briefly in chapter 10). In sev- eral of the villages examined, the cultivation of comniercial crops (rather than subsistence and traditional crops) was associated with fewer messages being carried among villages, on USE OF BUSINESS AND RESIDENTIAL TELEPHONES 233 the one hand, and much greater use of the telegraph and postal system, on the other. See Kamal, Dessouki, and Pool (1980). 12. In Papua New Guinea, almost 60 percent of the subscribers sampled stated that few or none of their friends or relatives in their town were telephone subscribers. The telephone density in 1977 in Papua New Guinea wvas 1.4, whereas in 1980, in Nepal it was about 0.12 and in Thailand, 1.1. 13. These figures are not strictly comparable and for Papua New Guinea do not neces- sarily indicate the relative frequency of each type of call. The Papua New Guinea data refer to multiple responses, at an average of 1.4 responses per subscriber. When the Papua New Guinea responses were prorated to sum to 100 percent, the business-related calls comprised 31 percent of the total. Business-related calls in Thailand were comprised of calls to shops, one's customers, one's workers, and government entities, in that order of importance. 14. Kojina, Hoken, and Saito (1984). 15. Kojina, Hoken, and Saito (1984) and Pool and Steven (1983). 16. Kilgour (1982), pp. 195-96. 17. Kilgour (1982) pp. 197-99. 18. British Post Office, Statistics and Business Research Department, personal com- nmunication. 19. Research Institute of Telecommunication and Economics (1974). 20. Of course, the emergency factor may be for receiving as well as for making calls. Al- though only 10 percent of the respondents in the Egyptian rural village surveyed reported using a telephone during the previous year (table 11-8), 30 percent stated that they would use a telephone for emergency purposes, and 9 percent claimed that at one time or another they had actually used a telephone to call a physician in an emergency. Kamal, Dessouki, and Pool (1980). 21. The figure of 2 percent who would not communicate seems to be too low since it im- plies an almost completely inelastic demand for communication, which has not been ob- served in other surveys. 22. Kilgour (1982) p. 201. 23. Kamal, Dessouki, and Pool (1980). Of the nine villages, three had small telephone exchanges (thirty to eighty-eight telephones), five had a few scattered direct lines, and one had no telephone access. Mvost of the telephones were in private hands. Telephone calls were generally viewed as time-consuming and difficult to make because call traffic was se- verely congested. Five of the villages had post offices, and three had telegraph offices. 24. Pool and Steven (1983). Chapter 12 Use of Public Call Office Telephones IN MANY DEVELOPING COUNTRIES the population living in towns, rural vil- lages, and low-income urban areas typically gain their first access to rapid two-way communication through public coin box telephones or public call offices (Pcos). Such service may be provided for reasons of facilitating general economic development, social equity, or national unity, even though the facilities sometimes do not generate enough revenue to cover their full financial costs, especially during the first few years of operation. Nevertheless, such public facilities are being provided with increasing frequency in the developing world, and in fact this provision has been encouraged by institutions such as the World Bank and the Inter-American Development Bank.' In reviewing the survey data and study results presented in this chapter, the generally poor quality of Pco service in developing coun- tries must be kept in mind. In many instances circuits are highly con- gested, telephone operators in urban areas give incoming Pco calls low priority, and transmission quality is poor, partly because of inadequate maintenance. Poor-quality service results in long waits for calls to be completed and ultimately in the cancellation of many calls. It also de- ters individuals from attempting to make any calls except those judged to be of high priority. Table 12-1 shows survey results for the town of Zway, Ethiopia, which reflect a typical situation. During a five-day sample period in February 1982, of the Pco calls urgent enough to be attempted, two-thirds were completed and one-third were canceled. Moreover, the interurban calling rate was higher for rcos than for most subscriber telephones. 234 Table 12-1. Interurban Telephone Calls, by Type of Subscriber, for the Manual-Exchange Service Town of Zway, Ethiopia, 1982 Interurban calls, Outgoing calls Total Incoming calls Total registered Attempts per Calls per calls per Number of Calls Calls call subscriber Calls subscriber subscriber Subscriber subscribers established canceled attempts per day received per day per day Government administration, public service, and political organizations 27 190 72 262 1.95 261 1.93 3.88 Business and commerce' 18 33 8 41 0.46 35 0.39 0.85 Registered as residential 32 18 6 24 0.17 16 0.11 0.28 Public call office 1 78 39 117 23.40 6 1.20 24.60 Total 78 319 125 444 n.a. 318 n.a. n.a. n.a. Not applicable. a. Cails were recorded for a period of five days for about twelve hours a day. b. Eight small hotels, nine retail stores, and one large-scale farm. Source: Ethiopian Telecommunications Authority. 236 TELEPHONE ACCESS AND USE Characteristics of Pco Tetephone Users Table 12-2 presents selected characteristics of rural or small-town users in five developing countries. In the 1978 Indian study of ten rural Pcos in five states, most Pco users outside large urban areas were individuals with relatively high levels of education, were either employers or self-employed rather than employees, and had incomes that were higher than those of most rural dwellers. Furthermore, a large proportion of the users were businesspersons, professionals, or civil servants, even though an aver- age of 70 percent of the population in the areas served by Pcos were engaged in farming. A feature of the 1981 Indian study, based on interviews of 174 Pco users in thirty-six villages from six districts in Andhra Pradesh State, is that users who gave agriculture as their principal occupation made up a relatively large 41 percent of those interviewed. Business and ser- vices (the definitions of which are somewhat unclear) made up 54 percent of the Pco users, which corresponds roughly to the 1978 fig- ure of 65 percent. The 1981 group was slightly less well educated than the 1978 group, whereas the income of the two samples appears to be distributed in roughly the same ratios. The interesting comparison with regard to income, however, is with the 1981 Indian survey of rural telephone subscribers discussed in chapter 10.2 Although only 35 percent of the subscribers had incomes of less than Rsl5,000 a year, almost 76 percent of the Pco users fell in that lower-income category. The study on Papua New Guinea presents a somewhat more re- stricted picture of the educational level of Pco users than the two Indian surveys. In Papua New Guinea, where 79 percent of the popu- lation had at most a primary education, this group accounted for only 5 percent of rural Pco callers and used Pcos infrequently. In contrast, persons with secondary, technical, vocational, or university education constituted only 3 percent of the population but comprised 75 per- cent of Pco users and tended to call frequently.3 Somewhat different results, however, were observed in Costa Rica. In a study of public telephones located in a cross-section of rural vil- lages, rco usage was fairly evenly distributed among economic and so- cial groups.4 The caller's occupation predictably tended to reflect the nature of the local economies (agricultural workers placed more than 80 percent of calls in an important agricultural area, businesspersons were the dominant callers in a river port, and so forth); users were not predominantly employers or government officials; no one small group dominated usage (about 80 percent of calls were placed by Table 12-2. Characteristics of Persons Using Rural Pco Telephones in Five Countries (percentage falling in category) India Costa Papua New Rica Guinea Mexico' Perub Characteristic (1978) (1981) (1975) (1976) (1978) (1985) Employment status Self-employed or employer 77.5 - 36.0 - - - Wage earner-employee 22.5 - 60.5 - - - Principal occupation Business 65.0 33.0 23.7 - 22.0 24.6 Professional and technician 7.5 4.0 23.2 - 22.0 23.8 Agriculture 7.5 41.0 32.7 - 15.0 9.7 Government service 12.5 - - - 20.0 - Administrative personnel - - 8.9 - - - Services - 21.0 - - - 33.3 Other 7.5 1.0 11.5 - 21.0 8.6 Income 0-¢300 a month - - 45.5 - - - 0300-¢600 a month - - 22.0 - - - ¢600+ a month - - 10.6 - - - No response - - 21.9 - - - O-Rs4,800 a year 20.0 - - - - - Rs4,800-Rs6,000 a year 7.5 - - - - - Rs6,000-RslO,000 a year 32.5 - - - - - RslO,000+ a year 40.0 - - - - - O-Rs5,000 a year - 20.1 - - - - Rs5,000-RslO,000 a year - 41.4 - - - - RslO,000-RslS,000 a year - 14.4 - - - - RsI5,000+ a year - 24.1 - - - - $262.53 average - - - - - 100.0 Educational level Primary or less 15.0 26.0 - 5.0 - 16.2 Middle 22.5 16.0 - - - - Matriculates (high school) 25.0 37.0 - - - 13.0 Secondary or technical - - - 50.0 - 35.7 Attend college - - - - - 17.6 Graduate 30.0 - - Postgraduate 2.5 - - College or university (degree) - 21.0 - 25.0 - 17.6 Still in school - - - 16.0 - - Other 5.0 - - 4.0 - - - Not available. Note: The category definitions are not completely comparable in all instances. a. Results of a survey of all Pco users for one month in the rural village of Metepec (state of Hidalgo), Mexico. Survey carried out by Centro de Investigaciones para el Desarrollo Rural in June/July 1980. b. Results of a two-year survey of seven small villages in rural Peru carried out by Florida State University with ENTEL, the Peruvian telecommunications authority. See Mayo and oth- ers (1987). c. The mean income level of income earners in the sample villages in Costa Rica was ¢372 a month. Source: See table 12-4, except for the Mexican and Peruvian studies. 237 238 TELEPHONE ACCESS AND USE users who called on average only two times a week or less, whereas heavy users, placing five or more calls a week, generated only 11 per- cent of the traffic); and the mean income of callers was close to that of the rural population at large. In a study of seven small villages in rural Peru, the typical user was male, thirty-three to thirty-five years of age, born outside the project zone but a resident for fifteen to seventeen years, well educated, a fre- quent traveler outside the project area, and either employed in a pro- fessional or technical position or the owner or manager of a business. Business owners and managers increased their share of total use of the telephone system from 14 to 34 percent over the two-year period of the study. One factor in this increase was that businesses learned to use the system more efficiently over time, by using calls as well as telegrams and letters to arrange appointments for future calls. Persons classified as employees or workers averaged 17.4 percent of total use, varying without a clear trend over the period, and use by housewives fell from 20.3 to 9.6 percent with no easily detectable explanation. Overall, the income of users was, on average, 14 percent higher than that of nonusers.5 With regard to the distance traveled to use public telephones, table 12-3 shows that in Papua New Guinea, Costa Rica, and Mexico, pub- lic telephones were used primarily by persons who lived or worked nearby. The average rural Pco user surveyed in Costa Rica traveled only about 0.3 kilometer to reach a Pco, although 5 percent of the callers traveled more than 2 kilometers, and in fact this latter group averaged 8.5 kilometers of travel per call.6 There was also some indica- tion that the distance traveled was not independent of the purpose of the call. Individuals making emergency calls, for example, traveled on average about 60 percent farther than the average caller. In Papua New Guinea, where telephone penetration was much lower than in Costa Rica, 81 percent of Pco users reported going to the telephone area specifically to place a call, and on average they traveled 6.9 kilometers.7 Seven percent of the users traveled more than 24 kilometers, and, in extreme cases, a couple of days' walk followed by some motorized transport were involved. In some areas it was not uncommon to make half-day boat trips just to make a telephone call. The results of the survey in the Mexican village are roughly consistent with those from Papua New Guinea except that in the Mexican sur- vey only 33 percent of the persons interviewed traveled to the village just to make a telephone call. In the 1978 Indian study, the average rural Pco served thirteen neighboring villages, the number varying between four and twenty-six USE OF PUBLIC CALL OFFICE TELEPHONES 239 Table 12-3. Distance Traveled to a PCo to Make a Telephone Call in Three Countries (percentage falling in category) Papua Costa Rica New Guinea Mexico Distance traveled (1975) (1976)a (1978)b 0-500 meters, 79 - - 0-2 kilometersd 95 - - 0-5 miles' 77 - 6-10 miles - 8 - 11-15 miles - 2 - 15+ miles - 7 - Within the village - - 68 Village to 5 kilometers - - 9 5-10 kilometers - - 11 10+ kilometers - - 12 - Not available. a. The Papua New Guinea results are for the 81 percent of the 916 respondents who made a special trip to make a call. b. Results of a survey of all Pco users for one month in the rural village of Metepec (state of Hidalgo), Mexico. Survey carried out by Centro de Investigaciones para el Desarrollo Rural in June/July 1980. c. From a sample of three villages (see appendix C). d. From a sample of eleven villages (see appendix C). e. Seventy-five percent of the natives of Papua New Guinea traveled less than 5 miles, compared with 90 percent of the expatriates. Source: Except for the Mexican study, see table 12-4. depending on the district concerned. Use was concentrated mainly in villages that had their own PCO, however, although occasionally users traveled much farther; the PCOS covered between 2.5 and 13.0 kilome- ters. The average distance traveled to make a call was tabulated for five rural PCos; the range was between 0.2 kilometer for two of the PCOs and 3.8 kilometers for one that served an area with a radius of 11.5 kilometers and contained seventeen villages. The average dis- tance traveled to place a call for all five Pcos was 1.1 kilometers.8 Finally, a study in Chile showed that 72 percent of all calls made from the rural PCOs surveyed were made by people who lived in the same village, 23 percent were made by users coming from other vil- lages (not necessarily for this sole purpose), and 3 percent were made by persons traveling between other places (mainly vehicles stopping by).9 Less information is available on urban rco users. One 1983 study in Senegal examined PCO users in both rural and urban environ- 240 TELEPHONE ACCESS AND USE ments.10 Interviews were solicited from 381 telephone callers and 305 telegram senders at eighteen sites throughout all eight regions of the country. The average user of public telecommunications facilities was a male between the age of thirty-one and fifty-six years, who lived within 2 kilometers of a post office and had slightly more than a pri- mary school education (six to eight years), which is well above the national average. Only 15 percent of users were female, and only 16 percent were engaged in agriculture or cattle raising. User occupa- tions differed markedly between the most urban Pco in a residential suburb of the capital, Dakar-Pikine, and a very rural Pco in Podor (see figure 12-1). The educational level was highest for persons using urban PCos, averaging almost nine years, but, surprisingly, was equally high in several of the most rural locations. It was hypothesized that government officials in these areas lacked their own telephones and were forced to rely on the Pco; in larger villages and towns, these offi- cials had their own telephones, and educational averages for Pcos were thus lower. In seven of the eighteen sites, no housewives (the vast ma- jority of women who used public telephones) used the PCO. These sites, which occurred in both rural and urban areas, tended to have the highest number of agricultural users. However, unlike all other sires, 30 percent of the Dakar-Pikine Pco users were housewives. This high figure was attributed to the important commercial role played by women in the markets of the urban capital. Purposes of Public Telephone Calls Some information is also available on the purposes of telephone calls made at town or village Pcos in developing countries. The sum- mary data presented in table 12-4 show considerable variation in the survey results among countries." Calls made to participate in activities directly related to the genera- tion or management of economic activity ranged from a high of more than 75 percent in the 1978 Indian survey to a low of between 25 and 30 percent in Kenya and Papua New Guinea, with an average among the nine surveys of about 50 percent."2 Mirroring this, calls to kin and friends ranged from roughly 20 percent in the 1978 Indian survey to nearly 60 percent in Costa Rica and Papua New Guinea, and over 70 percent in Kenya. The Kenyan results, however, reflected a tendency of business- USE OF PUBLIC CALL OFFICE TELEPHONES 241 Figure 12-1. Occupations of PCO Telephone Users in Llrban and a Rural Areas of Senegal, 1983 Dakar-Pikine Student 17% Jobless 7% Commerce 30% Artisan 7% -Government 3% 'Transport 3% Housewife 30% Podor Livestock 18% { |r n r n e n~~~Gveruet 9% Commerce 18% N . J,V, Artisan 9% Agriculture 45% Source: Christopher W. Nordlinger, "Users of Public Telecommunications Facilities and Their Benefits in a Developing Country: A Case Study of Senegal," in ITU, Information, Tel- ecommunications, and Development (Geneva, 1986). persons, who presumably value their time relatively highly, to use modes of communication other than coin box telephones. This was primarily because the quality of service was poor; to use a coin box telephone individuals generally had to queue for up to half an hour before booking the call, only to lose it when the operator called back and found the next person using the telephone. The problem was also Table 12-4. Reasons for Using Rural Pco Telephones in Several Countries (percent of calls) India Papua New (1978) Guinea (1976) Most Costa Most Remote Rural General recent Chile Korea Rica General recent Canada Mexico Kenya Use use call (1981) (1978) (1979) (1975) use call (1974) (1978) (1980) Call family or friends 5 23 32 28 43, 57b 75 60 35 50 76 Participate in economic activity 95 77 64 71 51 38 25 31 41 50Q 24 Business related - 72, 47 55 30 28 20, - - 3P - Government administration - - 7 - 16 9 - - - 7 Goods and services acquisition for the home (except emergencies) - - - 11 - - - - - - rz Emergency services - 5 10 5 5 1 5 - st 12t 41 Other - - 41 1 6 5 - gs I gh Total 100 100 100 100 100 100 100 100 100 100 100 - Not available. Note: The data presented in this table are not necessarily strictly comparable since the wording of survey questions, sampling procedures, and characteris- tics of the target populations varied somewhat among the countries listed. Overall, however, the results roughly indicate the actual situations. a. "Personal" calls, probably including calls to access goods and services for the honme. b. Includes calls related to both an emergency and health or a discussion of health problems. c. In addition to the categories listed, 14 percent of calls were made to collect farming information, 10 percent to conduct other business, 6 percent to conduct military affairs, and 3 percent to report national security and police matters. d. The Mexican data were from a survey of forty-six rural village rcos in five states. e. Includes calls to access goods and services for the home as well as to reach family and friends. f. Health purposes. g. Includes calls for emergency services. h. Includes 11.9 percent for weather information, which is imiiportant because transport is by small airplane since the region does not have road or rail access. Source: Appendix C. Figures in this table refer to the three-village survey; National Council of Applied Economic Research, India (1978); Economics Study Cell, Posts and Telegraphs Board, India (1981); Nicolai and Wellenius (1979); Chan-Kil (1979); Evans and Bryan (1977); Hudson and others (1979); Secretaria de Comunicaciones y Transportes Mexico (1979); and Cleevely and Walsham (1980). USE OF PUBLIC CALL OFFICE TELEPHONES 243 aggravated because local calls were untimed, and therefore once a connection was made users tended to talk at considerable length. In the Senegalese survey, the majority (53 percent) of calls made at the Dakar-Pikine Pco, the most urban of all surveyed, were made for financial reasons, while 20 percent were personal, but not urgent, and 27 percent were personal and urgent. No calls were considered to be related to government, perhaps because the respondents considered calls made to discuss government affairs with friends and relatives in government offices to be personal, based on their relationship to the person called.'3 In Papua New Guinea a high proportion of the calls initiated on public access telephones were received on business or government telephones. Seventy-five percent of the respondents indicated that a business or government telephone was the destination of the PCo call they had just made, and 41 percent of the PCo calls made to family or kin were made to business or government telephones. This partly re- flects the fact that in many instances friends or relatives in govern- ment were the caller's only contact who had access to a telephone. Such contacts thus delivered messages and information to other friends, relatives, or business acquaintances. Finally, as seen in table 12-4, in several of the countries surveyed, a relatively high 5 percent of the rural PCo calls were made expressly to obtain emergency services. Such calls would presumably have a high value to the caller. The emergency call figure of 10 percent in the 1981 Indian survey may include urgent calls for purposes other than obtaining emergency services. Not surprisingly, the income of persons making emergency calls followed a different pattern from that of persons making business or social calls. Of all emergency calls in the Indian survey, 61 percent were made by callers with annual incomes less than RslO,000 and 89 percent by callers with annual incomes less than Rs 15,000. These fragmentary results document the fact that a relatively high proportion of town or rural PCo calls relate directly to productive eco- nomic activity or to the provision of emergency services and advice. Given the probable high value of some of these activities, such calls probably account for most of the benefits derived from rural PCOS. Al- though the other major category of calls-those to kin and friends- presumably also affects the allocation of both human and physical resources, the most important effect of the calls is probably how they influence family and social change as well as the quality of life of rural dwellers. 244 TELEPHONE ACCESS AND USE Importance of Pco Calls and Alternatives for Communication Two complementary questions relate to the purposes for which Pco telephones are used: how do Pco users perceive the importance or ur- gency of their calls, and how would they communicate in the absence of a telephone? Table 12-5 presents information on the perceived ur- gency of calls made from rural Pcos in India and Papua New Guinea. In all three surveys more than 75 percent of calls were for communi- cation that was considered necessary immediately or on the same day. Hence, at least for these cases, the speed of the communication was highly important to public telephone users. Table 12-6 presents information from the surveys in Papua New Guinea, Costa Rica, Chile, and India (1981) on how communication patterns would change if the Pco at which the call was made were not available. In India and Chile, most of the Pco calls would have been replaced by either traveling to use the next nearest telephone or trav- eling to meet the other party. The postal service, telegraph, and mes- sengers would rarely have been chosen, apparently because they are slow and do not offer an immediate response.'4 Moreover, as was shown in table 12-4, more than 70 percent of all Pco calls in the Chilean survey and 64 percent in the 1981 Indian survey had to do with economic activities. In Costa Rica and Papua New Guinea, how- ever, where 40 percent or less of calls were related to direct economic activity, there may have been a larger potential for substituting letters Table 12-5. Perceived Urgency of Pco Telephone Calls in India and Papua Neuw Guinea (percentage falling in category) India Papua New Guinea Urgency (1978), (1981)b (1976) Immediate (emergency) 27.5 - 16 Same day 60.0 77.6 62 Within a few days 12.5 22.4 16 During the next week or so - - 6 - Not available. a. Data are for the last call made. b. Data are for the last call made. The two alternative responses were immediate urgency or could have waited. Source: Same as table 12-4. USE OF PUBLIC CALL OFFICE TELEPHONES 245 Table 12-6. Rural Pco Calls That Would Be Substituted by Alternative Means of Communication if Telephones Were Not Available in Five Countries (percent) Papua Costa New Guinea Rica Chile India Peru Alternative (1976) (1975) (1978) (1981) (1985) Travel to nearest other telephone - 3 22 5 - 21 69 Travel to meet other party 30 - 21 Letter 24 35 3 10 11 Telegram 18 18 3J 34 Messenger 10 - 2 - 32 Other - 26' - 7b 7, Would not communicate 18 18 48 3 11 - Not available. a. Includes 22 percent that would be replaced by obtaining information from radio broadcasts. See appendix C. b. Seven percent stated a combination of the listed alternatives. c. Radio is also an option in Peru. Source: See table 12-4 and Mayo and others (1987). or telegrams for calls."5 In Costa Rica, where information of value to farmers (wholesale market prices, weather forecasts, and so forth) is regularly broadcast by radio, the radio was considered to be an impor- tant alternative means of communication. In Peru, the choice of an alternative to telephone service changed significantly over time for users of the system; when telephone service was first introduced, 36 percent would have sent letters, but this de- clined to 11 percent two years later. In contrast, the number who would have sent a telegram increased from 14 to 34 percent, and the number who would have done nothing increased from 0 to 11 per- cent. This phenomenon was attributed to the increasing familiarity with all kinds of telecommunications that customers gained after the introduction of telephone service. The reason for the growth in the number of persons choosing to do nothing is more obscure; either their calls were so urgent that they felt nothing else would have been an adequate substitute, or their calls were not important enough for them to seek an alternative.'6 Table 12-6 shows that, except for India, in the absence of a PCo a substantial number of communications would not have occurred. The proportion of calls for which substitute communication would not have been undertaken was 18 percent for both Costa Rica and Papua 246 TELEPHONE ACCESS AND USE New Guinea and almost 50 percent for Chile. This suggests the possi- bility that the rural telephone generates communication and that al- though improved rural accessibility to telephone service results in some substitution among means of communication, it also increases the total volume of individual communications.'7 In the Senegalese study, the use of telegrams and telephones ap- peared to have limited substitutability. Table 12-7 shows that only 6 percent of telegraph users would have tried telephoning as an alterna- tive. The explanation lies in the difference between destinations. Two- thirds of all telegrams were sent for personal reasons, mostly to convey congratulations, condolences, and birth announcements, and telegrams were the chief means of sending messages abroad or to lo- cations without telephone service. Personal visits and telephone calls were generally too expensive or simply not possible under these cir- cumstances, so that posting a letter was often the only real option to sending a telegram.'" Constraints on Access to PCos Pco or coin box facilities in developing countries should be located where they are most accessible to the local population and where they are reasonably secure from vandalism. Hence, following the above ex- amination of who uses public telephones and for what purposes, it is useful to examine briefly the factors that most limit or constrain the access and use of such facilities. For purposes of this discussion, con- straints are grouped according to whether they relate to primarily physical, economic, or social and educational factors.19 Table 12-7. Alternative Means of Communication Chosen by Telephone and Telegraph Users in Senegal, 1983 (percent) Telephone Telegraph Alternative users users Do nothing 6 3 Use subscriber telephone 7 6 Go to another post office for the same communication 28 40 Post a letter 26 40 Send someone to the destination 8 2 Go to the destination in person 25 9 Source: ITU (1986). USE OF PUBLIC CALL OFFICE TELEPHONES 247 Physical Constraints The major constraint on public telephone use in developing coun- tries is clearly physical: the location of the telephone relative to that of the population. The uneven distribution of telephone service be- tween and within urban and rural areas has been noted throughout this book, which also documented that, in general, the farther a per- son is from a telephone, the less frequently he or she will use it, if at all. Related to this, public telephones tend to be used less if the qual- ity of voice transmission is variable or poor, and if the waiting lines or waiting times are long for booked long-distance calls.20 Other physical constraints include the placement of the telephone so that it is not readily or continuously available. In many rural towns in developing countries, the PCo is located in a post office, cafe, store, or other such place that is open only during daytime business hours. This obviously reduces some of the potential emergency call benefits of the telephone. In at least one South Asian and one North African country, rural rcos were used several hours each afternoon to deliver and receive the day's "telegram messages," thus precluding their regu- lar use for telephone calls. Also, in several areas in Korea, many rural Pcos were placed in the residences of village chiefs, who were often out of town. This substantially limited the public's ability to use the Pcos, and uncertainty about whether the telephone was available dis- couraged persons from neighboring villages from placing calls.2" Finally, an additional physical barrier to public telephone use is the difficulty of contacting individuals without subscriber telephones. Communicating with someone located at a place with only public telephones requires a means to relay the message or summon the party to the telephone. Such messenger facilities are often inade- quate. In the Korean villages surveyed, for example, incoming calls to rural PCos were held while the party was sought and brought to the telephone, and this process was considered to be too slow. It was fi- nally recommended that a public address system be installed to sum- mon people to the telephone. In Costa Rica and several other countries, small boys would hang around the rural PCo either to run and summon or to take a message to the person being called. In areas of some countries, however, the telephone message is simply passed on by word of mouth. Economic Constraints A second constraint is the cost of making the call. Part of the cost is incurred in getting to a telephone (opportunity cost of time and di- 2 48 TELEPHONE ACCESS AND USE rect expenditure) and, as such, is indirectly related to the physical ac- cess of the population to telephones.22 The other cost, of course, is the call charge, and since most calls made at rural or town Pcos are billed as long distance, the charges are in many instances significant for the average rural dweller. It can be argued, however, that the call charge is probably a factor holding down social calls to family, friends, and kin more than busi- ness or emergency calls, since, at least for subscriber telephones, the demand for business calls is generally thought to be more price inelas- tic than the demand for residential calls.23 Evidence supporting such a contention for calls made at Pcos in developing countries is, how- ever, limited. Social and Educational Constraints Social and educational barriers to the use of public telephones are widespread in developing countries. Among the more prominent bar- riers are insufficient experience with telephones and lack of adequate information about how to use them with confidence. In Papua New Guinea, for example, individuals needed to be taught what the tele- phone could do for them, where it was, and how to use it; this seemed to be important even in urban areas. Predictably, users tended to be mainly younger and better educated individuals, who had been ex- posed to ways of life in which the telephone plays an important role, and persons with white-collar occupations, where telephones are nor- mally used as part of work. In general, women tend to use Pcos significantly less often than men in most developing countries. In Egypt, no women were inter- viewed; in Senegal, the balance between male and female users was 85 to 15 percent; and in Peru, it was 68 to 32 percent. This reflects sociocultural restrictions as well as the lower average level of educa- tion and employment common among women in most countries. In Papua New Guinea, having to ask someone for assistance or per- mission to use a telephone inhibited use. For example, many of the radiotelephones serving remote rural areas were kept in the residences of government officials, schoolteachers, missionaries, or planters. Al- though everyone had a right to use these telephones or to send and receive messages over them, very few villagers did so; villagers would sometimes travel half a day to use a public telephone, although a ra- diotelephone station was located within 3 kilometers of their village. Apparently this was independent of the personality of the radio oper- ator: other images, social stigmas, procedural problems, and embar- USE OF PUBLIC CALL OFFICE TELEPHONES 249 rassments had to be overcome. This problem might be resolved if the radiotelephone were located in a central place and not in a mission or a school, perhaps somewhere that villagers helped to provide and that would be seen as "theirs."24 A related problem is the intervention of telephone operators. The involvement of an operator constrained telephone use in Papua New Guinea in two ways: users who only spoke a local tribal language had difficulty communicating with the operators establishing their call, when necessary, and, where this was not the case, the fact that the operator could listen in on the conversation inhibited use. This latter factor is a significant constraint throughout the developing world. Finally, the practice of metering calls and billing the individual re- sponsible for the telephone assumes that the person can control how others use the telephone. In societies in which considerable emphasis is placed on communal sharing, subscribers or vendors have difficulty restricting use by others or find doing so unacceptable. As a result, some choose not to obtain a telephone. Such problems have been ob- served in both Papua New Guinea and several countries in West Africa. Conclusions about Telephone Access and Use The information and analysis reviewed in this and the previous two chapters clearly show that telephones are an important means for fa- cilitating economic activity. This was found to be true for business, residential, public coin box, and PCo telephone service. In developing countries, on a national basis, within urban areas, and even within small towns and villages, the business community and, to a lesser ex- tent, government dominate access to telephone service.25 Even when a large proportion of telephone lines are connected to residences, the evidence suggests that, overall, telecommunications services are mainly used to facilitate and coordinate directly productive or distrib- utive activities of the economy. Telephones have a critical role in carrying out tertiary (or information) sector activities-government administration, finance, trade, commerce, services, transport-and are intensively used by managers and white-collar organizers in other sectors. Such groups tend to gain the first access to telephone service and tend to use it most frequently. Agriculture, at the other extreme, although it is one of the least-intensive users of telephones, appears to make contact with a relatively wider variety of economic activities than do most other sectors. 2 50 TELEPHONE ACCESS AND USE Telephones tend to be used primarily by individuals who function above the subsistence level, and they may be relatively widely used in regions or countries in which a significant proportion of the popula- tion has at least some formal education and is engaged at least to a minimal extent in market-related activities above the subsistence level. The limited empirical evidence on the use of business and residen- tial subscriber telephones suggests that it would be an oversimplifica- tion in developing countries to associate investment in business telephones wholly with "productive" uses and investment in residen- tial telephones solely with "unproductive" uses. Although in most in- stances business telephones are overwhelmingly used for work-related purposes, they also meet some of the personal and family needs of workers. Conversely, although residential telephones in most countries ex- amined are used more than 50 percent of the time to communicate with family members, kin, and friends, they are also used extensively to participate in the economy by performing work, gaining access to goods and services, and pursuing the organization and creation of economically relevant activities. They also allow the relatively higher- income and skilled people in rural areas to keep in touch with friends and family and to update information associated with creating and maintaining jobs. Finally, the use of both subscriber and Pco tele- phones for emergency purposes seems to be of universal importance in developing countries, perhaps even more so than in industrial countries. Although population groups functioning at a subsistence level do not make significant direct use of telephone facilities, little evidence supports the view that telephones are in any sense superfluous con- sumer goods for the high-income sectors of the population. Hence, opposing investment in telecommunications for distributional reasons is at best shortsighted and could, in the long run, retard the extent to which productivity could be increased through division of labor and could affect the level, distribution and spatial organization of eco- nomic activity. Assuring increased access to Pco and coin box telephones is, of course, an investment decision that can generally be financed and im- plemented rapidly by most telecommunications administrations in de- veloping countries.26 Assuring that potential high-benefit users receive first access to subscriber telephones and that the expanded Pco telephone program is adequately financed is at least partly a mat- ter of deriving proper pricing policies. Hence, in the following three USE OF PUBLIC CALL OFFICE TELEPHONES 251 chapters alternatives for pricing policy are reviewed, and a framework for telecommunications tariff policies is suggested. Notes 1. See, for example, Saunders and Warford (1979) and Saunders and Dickenson (1979). 2. Both the subscriber and the rco surveys are presented in Economics Study Cell (1981). 3. Individuals still in school were not included in the population percentage categories. Also in the survey in Papua New Guinea, 75 percent of the rco users were men, and users in the urban population over thirty-five years old were found to use public telephones rela- tively little. 4. A more complete description of the results of this study is presented in appendix C. 5. Mayo and others (1987). 6. Long trips, however, tended to have other objectives in addition to placing a call. 7. Telephone density in Papua New Guinea and Costa Rica was approximately 1.4 and 8.9, in 1976 and 1975, respectively. 8. National Council of Applied Economic Research, India (1978). 9. Nicolai and Wellenius (1979). 10. Nordlinger (1986). 11. One of the problems that arise in analyzing the various survey data on PC, use is the imprecise scope of categories used. However, the rough equivalencies used in table 12-4 are thought to represent reasonably well the situations as they were reported. 12. Answers relating to the purpose of the most recent call were considered to be more reliable than those requiring individuals to recollect general use over a longer period. 13. Nordlinger (1986). 14. Preferences for alternative means of making the calls varied considerably from one village or country to another, presumably reflecting, among other things, the relative costs of alternative means of communication and purposes of the calls. 15. Presumably, if another telephone had been available in the vicinity, it might have been a more common alternative mode of communication. In the United States when a switching center fire left 90,000 telephone customers without service for twenty-three days, relatively few residential subscribers increased their use of other modes of communication: 48 percent used emergency street phones set up by the telephone company, 33 percent (vir- tually everyone with a daily occupation) made calls from work, 10 percent wrote more let- ters, and less than 2 percent communicated by telegram. See Wurtzel and Turner (1977), pp. 246-61. 16. Mayo and others (1987). 17. In contrast, the Bangkok results indicated that only about 2 percent of families would not replace telephone calls made from the home by other means of communication should their residential telephone not be available. This suggests that the telephone may generate far less communication in large cities, which have fairly widespread alternative means of communication. 18. Nordlinger (1986). 19. These three categories are not necessarily mutually exclusive, and there are, of course, other relevant ways to categorize constraints on telephone usage. Bryan and Evans have, for example, suggested eleven constraints, some of them related to overcoming a rthreshold" to use and some to overcoming "barriers to increasing frequency of use." See Bryan and Evans (1979), pp. 172-87. 252 TELEPHONE ACCESS AND USE 20. In small towns and cities in many developing countries waiting several hours for a booked long-distance call to be connected is not uncommon. 21. Chan-Kil (1979). 22. Related to this, Bryan and Evans have noted that for relatively lower-income persons desiring to subscribe to telephone service, the lack of a suitable permanent structure for a home or shop can be a constraint to gaining a subscriber telephone. Bryan and Evans (1979) p. 174. 23. Littlechild (1979), p. 37; Meyer and others (1980), app. C; Alleman (1977), chap. 4; and Taylor (1980), chaps. 3 and 4. 24. Bryan and Evans (1979). 25. As shown in chapter 10, in small more-rural places with very limited telephone pene- tration, government or quasi-government telephone access sometimes exceeds that of the business community. 26. Some proponents of rural telephony argue that governments should subsidize rural telecommunications from general government revenues. Although in some instances rural facilities may need financial subsidies at least during their first few years of operation, it is not clear that these subsidies can or should come from general government tax revenues, where they must compete with sectors that do not earn revenue, such as education, health, nutrition, roads, and so forth. Part V Telecommunications Tariff Policy Chapter 13 An Introduction to Tariff Policy GENERAL AGREEMENT EXISTS in the literature on public enterprise tariff policy that, in the absence of good reasons to the contrary, prices charged should reflect the costs of providing services. There is, how- ever, some disagreement among accountants, economists, lawyers, pol- iticians, and so forth as to how those costs should be defined. Also, given the level and structure of telecommunications tariffs in both in- dustrial and developing countries, in many instances there are appar- ently "good reasons to the contrary"; prices charged for individual services generally do not reflect historical average, current average, or incremental costs. Objectives of Public Enterprise Tariff Policy This situation exists partly because telecommunications tariffs in practice attempt to satisfy different simultaneous objectives, which are often inconsistent with one another. These objectives can be grouped into three broad categories: those that relate to attaining financial goals, primarily aimed at assuring the financial viability of the tele- communications operating entity but also possibly at contributing to general government revenues; those that promote an efficient alloca- tion of a country's resources; and those that promote an equitable al- location of resources.' These objectives must be analyzed in light of a complex set of constraints facing policymakers; political and social ac- ceptability and the administrative feasibility of establishing any partic- ular tariff policy framework are among the primary considerations.2 255 256 TELECOMMUNICATIONS TARIFF POLICY Fitnancial Viabilit-y Subject to foreign exchange scarcity and, on occasion, unrealisti- cally low tariffs, most telecommunications operating entities in deve- loping countries can easily generate sufficient resources from users of the services to cover operation and maintenance costs as well as debt and interest payments. Likewise, they can usually generate a sufficient return on assets to attract needed capital and to cover a reasonable proportion of the future costs of expanding the system from internal cash generation. Subsidizing the sector from general government tax revenues is thus recognized as being inappropriate for several reasons. First, the sector as a whole does not require government financial subsidies.3 As outlined in chapter 1, in developing countries a large gap typically separates the supply and demand for telephone services: the number of potential subscribers on waiting lists sometimes ex- ceeds the number of telephones in service, and the average waiting time for a new connection can be as long as ten years. Likewise, business-hour traffic generated by telephone subscribers usually ex- ceeds the network's capacity. Given this excess demand, the market power (degree of monopoly) of the operating entity, and the fact that the short-run demand for telecommunications services is relatively price inelastic, particularly in urban areas, most telecommunications entities find it relatively easy to set their overall average charges at lev- els that cover their current costs and a reasonable portion (typically from 40 to 80 percent) of their expansion costs. In many instances they also generate substantial tax, interest, and dividend revenue for government. A second reason for not subsidizing telecommunications services relates to the inefficiencies involved in interfering with consumer choice by transferring, through general taxation, resources from the private sector to the public sector.4 It can be argued that the costs of the distortions in resource allocation resulting from such a subsidy are likely to exceed the benefits. A third argument against government financial subsidies relates to management. A case can be made that, if the management of tele- communications entities perceive that government subsidies will be forthcoming whenever costs exceed revenues, much of the incentive for cost-consciousness and efficient management is lost. Throughout the developing world, efficient management of public utilities is being undermined in this way (problems relating to management objectives and efficiency were outlined in chapter 3). INTRODUCTION TO TARIFF POLICY 257 Finally, from an administrative point of view, a stable and reliable source of revenue facilitates least-cost investment planning by tele- communications entities and supports their prudent financial man- agement. Telephone tariffs can provide such a reliable source of revenue; in many instances, government tax revenues prove to be less reliable. Numerous needs compete for tax revenues in developing countries, and financial subsidies for telephone service tend to be one of the easiest expenditures to cut during a national fiscal squeeze. Efficient Allocationi of Resources Price can be used to influence consumers so that the resources used to provide telecommunications services are not wasted. The price that consumers willingly pay for a good or service can signal at least the minimum value of that good or service to the consumer. If the price paid for telecommunications services exceeds the cost of production, this is a signal to expand output. If new output can only be sold at a price below the cost of producing it, this is a signal to re- duce production. To promote an efficient allocation of resources in operating a pub- lic utility sector dominated by what are essentially monopoly organiza- tions, as is the case of telecommunications in most developing countries, traditional welfare economics suggests that price should be set equal to the incremental or marginal cost of expanding output, or, if capacity is fully used, at a level sufficient to "clear the market."5 If, on average, the price of a telephone call is set equal to the incremen- tal cost of providing an additional telephone call, then consumers will indicate whether or not the value to them of an additional call is worth the cost. In effect, the responsibility for determining whether or not the telephone system (or some portion of it) should be expanded is therefore shifted from central planners or companv officials to the ultimate consumers of the service. Hence, if prices charged for differ- ent telecommunications services reflect the incremental costs of pro- viding those services, then signals are given about which services should be developed most quickly. In the short run, if supply cannot be expanded rapidly enough to meet demand, price can be increased to ration the supply of services (telephone lines or calls) to persons who value them the highest and, therefore, presumably derive the largest benefits from their use. Sev- eral problems are associated, however, with relying entirely on price to 2 58 TELECOMMUNICATIONS TARIFF POLICY promote efficient resource allocation. In particular, for monopoly tele- communications organizations, the tariff analyst must deal with con- ceptual problems of estimation, problems of demand forecasting and measurement, market distortions, the presence of externalities, and declining unit costs.6 These issues are discussed in chapter 15, which reviews some of the specific problems of estimating the marginal costs of providing telecommunications services. Equitable Allocation of Resources Most developing countries exhibit highly skewed income distribu- tions and large differences in regional income.7 Hence, some potential telephone users might not be able to pay telephone tariffs that cover the full cost of telephone service. This is particularly true in smaller towns or more remote low-traffic areas where the cost of providing service is generally higher than in larger urban areas, where incomes are low, and where the initial volume of traffic may not be large.8 As a result, providing selected telecommunications services at prices below cost may sometimes be desirable in order to spread the benefits of rapid communication widely and reduce the number of in- dividuals excluded from at least minimum access because of low in- comes. Even in higher-income areas such as North America, universal telephone service has been a primary goal of pricing policy for many years, and cross-subsidies between local and long-distance service and some urban and rural areas have been common." A certain degree of cross-subsidies are also the norm in most other countries in both the industrial and developing world.'° Traditional Approaches to Telecommunications Pricing Telecommunications services include several "outputs" that are usually charged for separately. Therefore, telecommunications tariff structures have typically included all or most of the following elements: a. A nonreimbursable fee for initial connection to the network b. A monthly, bimonthly, or quarterly rental, access, or subscrip- tion fee for terminal equipment and continued connection to the network, sometimes including a limited or unlimited number of local calls" INTRODUCTION TO TARIFF POLICY 259 c. Charges for local calls not included in the rental (increasingly, the length of local calls and time of day, as well as the number of calls, are being introduced into the tariffs) 12 d. Charges for long-distance and international calls, varying with duration, distance, time of day, and other factors, such as the use of operator assistance rather than subscriber dialing, person- to-person, and so forth e. Miscellaneous charges for business telephones, Pcos, private branch exchanges, data communication, telex, and so forth. In several developing countries, subscribers must also contribute capi- tal that gives them an equity share in the telecommunications entity. Sometimes this is treated as debt, and subscribers are reimbursed when service is terminated, with or without adjustments for inflation or interest. Such arrangements are discussed further in chapter 14. Where a regulatory body exists, the traditional approach to pricing is to establish an overall rate of return that yields an adequate flow of funds to cover the entity's operation and maintenance and a fair re- turn on capital. If, based on an examination of historical costs, the total revenue earned by an entity generates a reasonable rate of re- turn, when compared with the opportunity cost of capital, the regula- tory authorities can conclude that the overall level of charges is reasonable. A second concern of regulators is to ensure that a mo- nopoly telephone entity does not abuse its market power. Abuse in pricing practices can occur when an entity has excessively high prices or gives preference to some users either by charging other customers excessively or by cross-subsidizing its competitive services. A third concern of regulators is to provide incentives for the telecommunica- tions entity to extend access to telephone service. Within the frame- work of achieving an overall revenue requirement, both regulators and telephone operating entities have traditionally favored holding down monthly subscription charges (network access charges) in order to make basic service affordable and thus achieve the maximum penetra- tion feasible among users with lower incomes."3 Four key principles are traditionally taken into account in setting rates: (a) rate averaging across a company or a system, (b) value of service, (c) costs, and (d) usage. Rate averaging across a company means that all customers pay the same price for the same class of service throughout the company's operating territory. An example is when, within a particular exchange, all subscribers pay the same local exchange acces rate (monthly rental charge), regardless of their use 260 TELECOMMUNICATIONS TARIFF POLICY or distance from the switching center. The somewhat ambiguous value-of-service principle (defined as such by common usage) reflects the idea that a prospective buyer will pay a price that is related to the value derived from the service and that telephone services are more valuable to some classes of customers than to others. Thus, basic local flat-rate charges are higher at locations with more subscribers in the local calling area than in those with fewer, and rates for business subscribers are often higher than those for residential subscribers. Costs are taken into account by charging more for higher-cost ser- vices. Thus, operator-assisted long-distance telephone calls are priced higher than direct-dialed calls. Usage is a rating factor that can be as- sociated with both value and cost. For example, subscribers attach value to use and are willing to pay more for it, and increased use may, in turn, increase costs. Charges may be reduced for Sunday, weekend, or night service because such discounts contribute to improved net- work use at a low incremental cost to the system and thereby help re- duce peak-traffic demand. Type of User Telephone subscribers are sometimes categorized on the basis of the purpose for which their telephones are most likely to be used. Hence business and government users may be placed in one tariff cat- egory, whereas residential users are put into another.'4 Sometimes sep- arate categories are defined for more specific groups of nonresidential subscribers, such as businesses, industries, professionals, nonprofit or- ganizations, and government. There are two reasons for the basic business-residential dichotomy, one based on demand factors, the other on cost of supply. First, tele- phones are generally assumed to be more important to business firms than to private residences because they are, for competitive reasons, concerned with finding the least-cost means of communicating rap- idly. Business firms also tend to produce a much greater volume of communications. Therefore, the value of telephone service is assumed to be greater for business firms than for private residences, that is, the demand for business telephones is relatively more inelastic to price changes, and in any case such charges are tax deductible in many countries.15 If this is so, businesses can be charged somewhat higher fixed monthly fees for their subscription without greatly distorting the business demand for telephones, and the number of residential con- nections can be increased through lower monthly fees.'6 Hence, it is INTRODUCTION TO TARIFF POLICY 261 argued that new residential subscribers benefit because they have a telephone, and all existing subscribers, including business firms, bene- fit because more subscribers are connected to the network and can be called. The second reason for the business-residential dichotomy is that, although business and government calls are made primarily during peak business hours, many residential calls occur during nonbusiness hours when they do not burden the overall capacity of the system to handle traffic, if residential lines are intermixed with business lines in the same exchange area. Thus, since the marginal cost of an off-peak residential telephone call is in some instances relatively low, it is ar- gued that residential telephones should be encouraged through lower monthly fixed charges.'7 In industrial countries, a third argument in favor of lower residen- tial telephone rentals or connection charges is sometimes put forth. This argument suggests that for reasons of equity residential users (families) should be subsidized by business firms, which can better af- ford to pay more for telephones. This argument is essentially incor- rect, since business firms will pass on at least part of the increase in telephone charges to the final consumers of their goods and services, and it should not be given much weight without a detailed analysis of who ultimately bears the burden of higher charges to business firms. To the extent that upper- and middle-income groups have a greater proportion of residential telephones than do lower-income groups, and to the extent that business firms sell goods and services to lower- income groups who do not have residential telephones, it could be ar- gued that when business service subsidizes residential service, lower-income groups subsidize the residential telephones of groups with higher incomes. This may be a more powerful argument in favor of relatively higher charges for residential telephones in poorer coun- tries, where residential services tend to be in the hands of higher- income groups. An important alternative to subsidizing the access of residential users to telephone service across the board is to target sub- sidies toward either individuals or regions.'8 Finally, special interests-doctors, fire fighters, politicians, and so forth-frequently claim, usually without much supporting analysis, that social concerns qualify them for favorable tariff treatment. Inter- national organizations have even suggested charging different tariffs for different categories of users. For example, in the 1970s UNESCO pe- titioned telecommunications operating entities to set low concessional tariffs for journalists transmitting news. One argument against this 262 TELECOMMUNICATIONS TARIFF POLICY kind of subsidy is that such costs should be borne by governments, not telecommunications subscribers and users.'9 Although subsidies for special-interest groups may be neither a de- sirable nor a practical option, the rapid spread of new communica- tions technologies and services complicates the issue further. Very high capital investments are required to introduce new communica- tions technologies, and, where tariffs are related to costs, small and noncommercial users may not be able to use the new services, partic- ularly in a developing country.20 Concerns have been raised that these investments may only benefit the relatively few individuals and large businesses that have sufficient means to use the new technologies. Some authors argue that, over the long run, potential noncommercial users of new services (including such services as data banks and satel- lite distribution networks) may risk being marginalized through their inability to participate in increasingly important developments in communications technology.21 Size of the Local Service Area In many countries the size of the fixed monthly access fee that must be paid to retain telephone service varies according to the size of the local telephone service area, with higher fees being charged in larger service areas. In fact, distinguishing telephone tariffs by size of service area is recommended by CCItt.22 The rationale for such a pat- tern is once again based partly on the demand-oriented concept of value of service and partly on the cost of supplying service.23 From the value-of-service perspective a telephone may be more valuable to any one subscriber if more people can be called without incurring extra distance-related charges. Hence, subscribers and po- tential subscribers in larger local service areas are assumed to find tel- ephone service to be relatively more valuable than those in smaller areas. Also, it is argued that large settlements have more types of communication events for which local telephone service is an effec- tive medium, and telephones tend to be a lower-cost substitute for al- ternative means of communication (travel to order supplies, travel to arrange meetings, and use of messengers). From the point of view of telecommunications service entities, which have traditionally generated their highest profits and contribu- tions to fixed costs from long-distance service, any enlargement in the inexpensive local calling area would reduce long-distance revenues. These would normally be replaced by compensatory increases in monthly charges for local network access. INTRODUCTION TO TARIFF POLICY 263 A case for increasing monthly access charges as the local service area increases in size is sometimes made from the point of view of the cost of providing service, although empirical evidence supporting such a contention is limited. As the size of the service area increases, so do the requirements for switching capacity and connections be- tween exchanges. Also, expanding service in heavily built-up urban areas can be costly in terms of land, buildings, duct construction, and work on outside plant. As outlined in chapter 2, however, growth also reduces the average cost per subscriber added, because of economies of scale and improved network use as well as the opportunity for tech- nological innovation. Also, small service areas are generally located in remote rural areas, where low population density increases fixed in- vestment costs for each customer served. The net result of these op- posing trends varies from one situation to another.24 Distance Called Traditionally it has been accepted that charges for calls should in- crease with distance.25 Figure 13-1 outlines the typical relation be- tween distance and call charges and shows that although charges do increase with distance, long-distance charges tend to increase at a de- creasing rate.26 The principle that telephone charges should increase as the distance called increases is once again based on both cost of service and value of service. With regard to value of service, the number of alternatives to tele- communications for speedy and convenient communication tends to decline as distance increases, and the cost of such alternatives in- creases. Nevertheless, there is evidence that demand for telephone calls is more elastic as distance increases.27 On the cost side, a long-distance telephone call is generally more costly than one within a local exchange area simply because it in- cludes local service area costs (on both the sending and receiving ends) plus the costs of long-distance transmission and switching. In the early days of telecommunications, when long-distance transmis- sion was mainly through open wire lines, the associated costs were more nearly proportional to distance. With more modern technology, however, incremental (and average) long-distance costs tend to flatten out as distance increases. This is particularly so in international services using communication satellites, where cost is essentially in- dependent of distance. Costs are also a function of traffic volume. Shorter-distance calls over low-traffic routes can be much more costly on a per-call basis than longer-distance calls over high-traffic routes, 264 TELECOMMUNICATIONS TARIFF POLICY Figure 13-1. General Relation between Price and Cost for Local and Long-Distance Traffic Average cost and price per unit call Price Cost Distance particularly as microwave and fiber-optic transmission is implemented in an increasing number of situations. Even though long-distance tariffs usually increase at a decreasing rate with distance, tariffs typically overshoot substantially the cost of supplying high-volume long-distance service.28 As illustrated in figure 13-1, the tariff/distance curve traditionally levels out at a price that is well above cost, both for domestic long-distance and international service.29 This situation began to change during the 1980s in the United States, as competition in the provision of long-distance ser- vices increased. Significant reductions in both domestic and interna- tional long-distance charges have occurred in several other countries as well, including Canada, Japan, Mexico, and the United Kingdom. INTRODUCTION TO TARIFF POLICY 265 Time at Which the Call Is Made Since the use of telephones is not constant throughout a twenty- four-hour period, different charges for identical calls are frequently levied based solely on when the call is made.30 In some countries this applies both to local and long-distance calls, although more often only long-distance calls are metered. Although not universal among developing countries, such peak pricing of long-distance calls has be- come widespread in Western Europe, North America, Japan, and Australia, and peak pricing of local calls is becoming increasingly ac- cepted in many countries.3' The adoption of time-of-day call metering can have a moderate effect on the distribution of traffic at different times of day.32 Once again, arguments based both on cost and on demand factors support having higher call charges during daytime business and early evening hours than during nighttime and weekend periods. From the demand perspective, it is generally argued that business and govern- ment demand for telephone service is relatively inelastic during day- time business hours, since the ability to complete a call is more valuable during these peak hours than during evening and nighttime hours when most offices are closed. From a cost-of-service point of view, it is argued that consumers who make increased demands on the telephone system when it is operating at or near capacity should pay the incremental costs that must be incurred when the system needs to be expanded. Also, with peak pricing, there is some possibility of reducing the peak, shifting or smoothing the traffic load, and hence saving resources by reducing or postponing future expansion. Given this, calls that are made at off- peak times and that do not strain the system's capacity to handle traf- fic should then be charged only the incremental operation and maintenance costs they impose on the system and should not be re- quired to bear the costs of system expansion. The problems usually encountered with time-of-day pricing relate to technical metering difficulties and with improper choice of peak periods. For example, if a short peak period is selected, the call charge would have to be relatively high to recover capacity costs from the smaller number of calls. And, having a high call charge and a short peak period increases the likelihood that the peak will shift to adjacent lower call-charge periods. If the peak period is lengthy, however, subperiods exist when demand is low but charges are high, and there is thus no incentive to shift from high- to low-demand subperiods.33 266 TELECOMMUNICATIONS TARIFF POLICY Finally, to avoid confusing the subscriber too much, it has been suggested that no more than three price levels should be in effect dur- ing a twenty-four-hour period.34 Having three or four call-charge peri- ods creates the problem of educating and then periodically reminding telephone users which prices are in effect during which periods. Such an argument is, however, not totally convincing. If the size of the price differences among the periods were at all substantive, consumers would find it in their best interests to remember and use them to their best advantage. The costs of introducing time-of-day metering into different types of existing local exchanges vary considerably. The costs tend to be rel- atively high when metering is being added to existing step-by-step and rotary exchanges (where they still exist) but is considerably less costly when being added to crossbar facilities. With modern digital elec- tronic exchanges, the cost is very low. For example, the incremental capital cost required to time local calling on older nondigital ex- changes can range from as little as $2 to more than $50 per line, de- pending on the size of the local exchange, the type of switching equipment, and the extent of record-keeping required.35 During the mid-1970s the New York Telephone Company installed metering equipment in its large crossbar exchanges for $15 per line.36 Record- keeping and billing costs for time-of-day metering for local calls are estimated to range between $0.001 and $0.002 per call. In general, the larger the exchange and the more modern the equipment, the lower the metering cost. Notes 1. The objectives of policies governing public utility pricing may be categorized in many ways. Meyer and others list eight categories for telecommunications: universal service, static efficiency in resource allocation, equity for different kinds of users and services, financial self-sufficiency, preemption of uneconomic entry, consistency with expected technological change, administrative simplicity, and historical continuity. Meyer and others (1980) pp. 75-81. Baumol (1968), pp. 108-23, lists seven. 2. For a thorough examination of innovation and the relation between the theory and actual implementation of tariffs, see Mitchell and Vogelsang (1991), chap. 12. 3. This does not necessarily iniply that there should never be anv financial cross- subsidies within the sector. Also, different forms of subsidies, for example, per-unit subsi- dies, general deficit financing, and so forth, have different effects. 4. The same argument would apply under either private or public ownership of the tele- communications entity. 5. For a brief discussion of welfare economics and of marginal cost pricing in Britain and North America, respectively, see Littlechild (1979), chaps. 9 and 14; Mitchell and Vogelsang (1991) chaps. 3 and 4. INTRODUCTION TO TARIFF POLICY 267 6. Saunders and Warford (1977). 7. World Bank (1990) 8. In rural areas, fewer people can generally be reached without incurring toll charges than in urban areas; hence the value of telephone service in those areas might be less, thus justifying lower charges. The value of a telephone depends on many things, however, such as remoteness, emergency needs, cost of alternative means of communication, loneliness, and level of economic activity. 9. U.S. Congress. The Communications Act of 1934, as amended, 50 Stat. 189, sec. 1. 10. For examples from Western Europe, see Mitchell (1979). Additional information on tariff structure issues can be found in OECD (1990) and Mansell (1990). 11. Since the regulatory authorities in many countries have decided that competition in the supply of equipment on the customer's premises better serves the public interest, many subscribers can purchase telephone inStTUments from sources other than telephone compa- nies. Hence, the monthly charge can be separated into at least two parts: one part for a subscription to the network, which everyone pays, and a second part for rental of tele- phones, which subscribers who purchase their own equipment do not pay. 12. Historically, local calls have not been metered in North America. The current trend, however, is toward charging for individual local calls, resulting in the progressive introduc- tion of local measured service. In general, the metering of, and charging for, calls by num- ber and duration is referred to as usage-sensitive pricing in North America. See Schultz and Barnes (1984); Alleman and Emmerson (1989). Park and Mitchell (1987) argued that changing the technology for providing local service could reduce the cost of local usage suf- ficiently to make local measured service inefficient. 13. Kahn (1970), pp. 63-65. 14. This is generally the practice in North America and England, although less so in the remainder of Western Europe. Of the eighteen countries whose tariffs were surveyed in chapter 2, eight differentiated their connection fee and monthly rentals by category of subscriber. 15. The few existing empirical estimates of price elasticity of deniand for telephones tend to show that business demand for local and long-distance calls may be more price inelastic than residential demand. However, the evidence on connection charges and monthly rental or subscription fees is somewhat mixed. See Taylor (1980), pp. 168-74; Meyer and others (1980) armex C; Alleman (1977) chap. 4; Littlechild (1979) pp. 34-38; and Wilkinson (1983). 16. When the nionthly rental or subscription charge includes an unlimited number of local calls, the rental for business telephones may be higher, although the average price paid per call is lower, since businesses tend to use their telephones more intensely. 17. It may, however, be more adequate to reflect such time-of-day, traffic-related effects in the price for traffic (call charges) than to manipulate rentals or subscription charges on this account. Differentiating rental charges among classes of subscribers does nor create in- centives to shift calls at the margin from peak to off-peak hours. 18. The argument offered was that telecommunications should not charge the mass media on the basis of a so-called return on investment, since the exchange of media infor- mation is not commercially motivated but is directly related to the social and economic de- velopment of a country. Presumably, this argument was not meant to apply to the commercially oriented press of North America, Western Europe, Japan, and Australia. It also may not reflect the fact that in some developing countries, newspapers, television, and radio broadcasting are private business enterprises. 19. Helling (1981) pp. 27-29. 20. For instance, in 1986, one application planned for Brazil's domestic satellite was the national transmission of educational television programs for the Ministry of Education's FL]NTEVE network via a dedicated transponder However, switching FL NTEVE program distribu- 2 68 TELECOMM _'NICATIONS TARiFF POLICY tion from existing means (terrestrial microwave, at a late-night discount) to the satellite would have increased the network's transmission costs by a factor of ten. Under these cir- cumstances, use of the satellite for educational broadcasting was unrealistic without sub- stantial tariff concessions. See Nettleton and McAnany (1989). Even in industrial countries, small users may find new technology services too expensive to use. For example, a basic on-line data base in the United States, Mead Data Central's Lexis Financial Informa- tion Service, costs its subscribers $39 an hour for access, plus $6 to $35 per individual search, plus a monthly service charge of $50. Only a minority of very well-informed small businesses can be efficient enough in their searches to avoid incurring expenses higher than the profitabiliry derived from the information they retrieve. See Race (1990). 21. cCITT (1965). 22. In the United States, as in several other countries, pricing according to the size of service area has not been uniform. Although service areas are fairly consistent in size within states, local tariffs seem to be almost random among states. Meyer found that "in general nonuniformities in local telephone rates can be discovered almost at will, and it is difficult to explain many of the disparities by cost differences or density differences or other charac- teristics." See Meyer and others (1980) p. 86. 23. In the mid-1960s, a study in Texas showed that the average cost per main telephone decreased as rhe number of main telephones in the exchange area increased. At some size between 50,000 and 100,000 main telephones, however, unit costs leveled off and started to rise because of increasing loop lengths and more complex switching problems, which re- sulted in the need for more exchange and trunking equipment. See Bowers and Lovejoy (1965). See also the discussion of economies of scale in chapter 2 of this volume. 24. cInTT (1965). Mitchell and Vogelsang found that this is still generally true in the United States for interstate calls, even though competition and sharply reduced costs for transmission technology have made prices less sensitive to distance and have shrunk differ- ences between peak and off-peak rates. However, the introduction of optional tariffs means that in some cases, simplified, uniform per-minute rates that do not vary with distance are being offered during night and weekend hours, particularly for residential and small busi- ness users. For an example of this kind of tariff, see the discussion of AT&T's experimental 1983 Optional Calling Plan in chapter 8 of Mitchell and Vogelsang (1991). 25. In the United States, which has a diverse and decentralized regulatory situation, large anomalies exist in the relation between distance called and call charge. Generally, a 100-kilometer call among states is priced the same irrespective of the location of the calling and called points. However, a 100-kilometer call within one state may cost a subscriber more or less than a 100-kilometer call in a neighboring state and more or less than a 100- kilometer call between states. Another factor is the purely geographic basis used to deter- mine long-distance zones; customers who live near the boundaries of a particular area code often have to pay long-distance charges to call nearby locations in the neighboring area code. 26. No doubt other reasons contribute to this increasing inelasticity of demand at longer distances; for example, a relatively high proportion of long-distance calls are made by busi- ness and government users. See also Ministry of Supply and Services, Canada (1988). 27. Telex charges are often uniform throughout the country, reflecting what is sometimes a star-shaped network in which distance is a less meaningful factor of cost. 28. For a comparison of trunk call revenues and costs in Sweden, see Mitchell (1979) pp. 4-7. 29. Many countries in which the supply of telephone service is roughly equal to demand tend to have three distinct periods of peak use: between 10 a.m. and 11 a.m., around 3 p.m., and in the early evening. In many developing countries in which demand exceeds sup- ply, the daytime peak begins much earlier, ends later, and does not dip as much around mid-day. INTRODUCTION TO TARIFF POLICY 269 30. Mitchell (1979) and Mitchell and Vogelsang (1991). 31. In-depth analyses of the application of peak pricing in telecommunications have only become available in the past ten years, even for industrial countries; see Park and Mitchell (1987) and Sherman (1989). The importance of peak pricing has been long recognized in other sectors, however, notably electricity. See, for example, Boiteux (1949) and as trans- lated (1960). The classic article on peak-load pricing is Steiner (1957). 32. In some instances large multiexchange areas could also experience the complication of having different exchange areas experience call traffic peaks at different times. For a brief discussion of this point, see Selwyn and Borton (1980). 33. See Mitchell (1979). 34. Alleman (1977) pp. 118-27; Mitchell (1976), pp. 26-27; and Mitchell (1978). 35. Mitchell (1976 and 1978). 36. Microprocessor-based electronic devices, which collect and send telephone billing in- formation from different kinds of switching facilities to central offices, have lowered the cost of installing metering equipment, record-keeping, and billing for older switching equip- ment. See, for example, Rettig (1981). However, adding this kind of equipment to very old facilities may be shortsighted; doing so should be compared to the option of using the capi- tal expenditure to obtain all the advantages of newer switching equipment, which already includes this feature. Chapter 14 Tariff Policy for Economic Efficiency PRICES THAT ENHANCE ECONOMIC EFFICIENCY should be a major considera- tion in establishing a tariff policy for any country. This is particularly true in the developing world, where improving the standard of living is an important concern and resources often are very scarce. An eco- nomically efficient allocation of resources maximizes the value of out- put in an economy for any given income distribution. Since the consumption of resources is influenced by prices, economists have de- voted much attention over the years to specifying the set of prices that will be the most economically efficient for a given situation.' This chapter explores issues relating to economically efficient pricing of monopoly telecommunications services. Pricing As outlined in chapter 13, if an efficient allocation of resources is of primary concern, prices for monopoly telecommunications services should not be based on historical accounting costs nor on the some- what arbitrary value of service designations that may not explicitly re- flect significant differences in demand elasticities. The traditional analysis of historical accounting costs does not reflect current re- source choices, and, since the approach is backward looking, the level and mix of costs assigned to any particular service depend partly on when that service was instituted or expanded and on how large the changes in prices and interest rates were during the intervening period. 270 POLICY FOR ECONOMIC EFFICIENCY 271 A more appropriate approach (which applies when output can be expanded to meet demand) is to set prices on the basis of the forward-looking marginal or incremental costs of expanding output in order to maximize economic efficiency.2 These costs are used because they are tied to decisions to change the level of output. Thus, they can be avoided and are the only economic costs relevant to future consumption and investment decisions that affect the allocation of resources in an economy.3 Applying the concepts of marginal cost to the pricing of telecom- munications services is complex and encompasses several aspects of telecommunications tariffs. For example, the rationale behind estab- lishing fixed monthly fees for gaining access to the network is that costs associated with local loop infrastructure do not generally vary with the number of calls made or with the duration of a call.4 On the other hand, switching costs and long-distance transmission costs vary with the volume of traffic and are, therefore, reflected in usage charges-that is, charges per call and charges that vary according to duration of the call. A further example of the application of marginal cost pricing occurs when demand for use of the telecommunications network varies during the day. In this case, network use can be im- proved by providing discounted prices during off-peak hours, as dis- cussed in chapter 13.5 These discounted prices should be no less than the short-run marginal costs of providing off-peak service.6 Where networks will be expanded to meet demand, prices should be set no less than long-run marginal cost.7 Externalities Arising from Connections to the Network Marginal costs should, in principle, also include costs (or benefits) that are "external" to the market.8 In the field of telecommunications services, one of the most important externalities is that the value of telecommunications networks increases as the number of subscribers grows. A positive external effect is created each time a new subscriber is connected to a network, because the telephones of existing sub- scribers increase in value as the potential number of contacts in- creases.9 Theoretical problems arise when telephone installation immediately follows payment of the financial costs involved. Where no queues for service exist, the connection charge to an individual subscriber should, ideally, equal the cost of connection less the bene- fits (or disbenefits) accrued to persons or firms wanting (or not want- ing) to be in contact with that subscriber. This suggests that the 2 7 2 TELECOMMUNICATIONS TARIFF POLICY optimum connection price might in some instances be less than the connection cost and implies that a financial subsidy, which presum- ably should vary by class of subscriber (theoretically, by individual subscriber), may be necessary to allocate resources efficiently in these circumstances. However, the implementation of any subsidy scheme is never with- out cost. The particular external benefits discussed here accrue to other telephone users, and therefore, it may be argued, for reasons of equity, other telephone users should provide the subsidy. But there seems to be no completely satisfactory way of arranging this. A fixed monthly rental or subscription charge to finance the subsidy would not be appropriate, as it would produce exactly the same problems, and, for reasons discussed in the next section, a surcharge on call charges might also be inefficient. One option would be a subsidy from general tax revenue. However, such a subsidy scheme would be partic- ularly undesirable in most developing countries, because of a shortage of public funds, and because a government subsidy might have unfor- tunate consequences for the autonomy, management, and long-term planning of the telecommunications entity. Externalities Arising from Calls Similar theoretical problems are encountered in applying the rules of marginal cost pricing to calls, if it is accepted that a telephone call benefits not only the party who pays for the call but also the party called. In such a case, the optimal price of a call could be something less than marginal cost, and the justification for expanding the system should be determined not by just the revenues generated by callers but also by the benefits accruing jointly to both callers and callees. For example, if the value of a telephone call were the same for both parties, the price would be set equal to half the marginal cost charged to each party. Some judgment as to the size of the external effects is therefore necessary if marginal cost pricing of this joint product is to help determine the size of the investment.'0 The empirical problems of attempting to sort out the separate benefits of a joint product are immense; certainly the financial costs people are willing to pay to receive calls or the extent to which externalities are internalized by such things as mutual arrangements to return calls or reverse charges cannot be accurately identified. A further complication is that some calls made are not beneficial-they create external disbenefits rather than benefits. If, overall, external benefits outweigh external disbenefits, the case POLICY FOR ECONOMIC EFFICIENCY 273 for including externalities in telecommunications tariffs is strength- ened in the frequent instances in which competing services, such as public radio, transport, and the postal service, are subsidized at least partly because they generate external benefits. Other forms of com- munication have similar characteristics: the recipient of a letter does not pay; face-to-face communication is usually at the expense of the traveler, and so on." On balance, given the current state of knowledge about external benefits, the most practical solution is probably to ignore the externality problems discussed here; they are probably even less of an issue in developing countries, where the external disadvantages asso- ciated with unsatisfied demand dominate the situation. There is, how- ever, a clear need to research the issue of monopoly communications pricing so that judgments about second-best policies can be made more satisfactorily. Ramsey Pricing One of the most problematic results of marginal cost pricing is that where an industry is operating under conditions of declining average costs, as many analysts believe is the case in the telecommunications services sector, marginal cost pricing will not generate enough total revenues to cover costs (see chapter 2). As a result, either the opera- ting entity will go out of business, or a subsidy will be required. Subsi- dies, either from taxes on income or from sales, also have important disadvantages as a response to the problem. They distort economi- cally efficient pricing in other sectors of the economy and do not pro- vide an operating entity with suitable incentives to run its operations efficiently. Instead, many analysts have proposed marking prices above mar- ginal costs according to "ability to pay" or "what the traffic will bear" or "value of service." These concepts are comparable to marking prices above marginal costs inversely with the price elasticity of demand-an approach known as Ramsey pricing. The intuitive ra- tionale for this approach is that since prices above marginal costs pro- duce demand (and therefore output) below the optimum, output reductions are minimized if prices are marked up most where demand is most price inelastic, thereby minimizing the loss of economic efficiency.12 Ramsey pricing is, however, problematic. First, price elasticity of de- mand for telecommunications services is not entirely determined exogenously by the market. It is, in fact, very significantly affected by 274 TELECOMMUNICATIONS TARIFF POLICY government policy on sector competition and on authorizations for users to meet some of their own telecommunications requirements by means of radio communications. Therefore, price elasticity is not an independent factor guiding government tariff policy but is, in part, a result of government policy."3 Second, the categories of customers with very inelastic demand for telecommunications services are likely to include persons with no al- ternative, or only very expensive or less desirable alternatives, to the services offered by the operating entity. Charging much higher tariffs to the most vulnerable customers often cannot be supported when public policy fairness is considered.'4 In practice, the concept of charging more where demand is more price inelastic is usually applied to broad groups of customers, such as business subscribers. Price Rationing One of the most important elements of the economic inefficiency that exists in the telecommunications sectors of many developing countries is the inability to expand networks sufficiently to meet the demand for service (at prices that cover costs). Thus, in many coun- tries, a shortage of telephone lines is indicated by waiting times of months or years to obtain service; this is often accompanied by heavily congested network traffic during daytime business hours (see chapter 1). In this situation, administrative forms of rationing, such as giving priority to select groups, tend to be arbitrary and cumbersome and to invite management or administrative irregularities. Even an appar- ently nondiscriminatory form of rationing such as queueing (meeting demands in order of application) is almost always too rigid to allow an economy to function efficiently."5 Further, in a situation of unserved demand, setting the price of telecommunications services on the basis of calculated long-run marginal costs is not necessary. In- stead, price can serve as a rationing mechanism. Scarce telephone lines can thus be allocated to customers who will derive the highest value from their use by charging prices above cost to the point where the level of unmet demand is almost eliminated and additional funds are generated for the enterprise. To the extent that telephone lines are fungible (can be transferred from one user to another, usually within one exchange), the optimum way to set prices so as to ration the service is to increase the monthly line rental. This approach serves to ration the total supply of available lines, including both new and existing lines. An extension of this ap- POLICY FOR ECONOMIC EFFICIENCY 275 proach is to permit trading in "telephone rights" (that is, local net- work access) either officially, through a market exchange, or unofficially, in order to ration the scarce supply of local access lines in an economically efficient way. However, where monthly rental charges cannot be increased beyond a certain level, the scarce supply of new telephone lines is commonly rationed by charging high installation prices to new subscribers. Similarly, where network traffic is congested, increased peak-hour call charges are appropriate. A salient advantage of this approach, as noted in chapter 13, is that decisions about the importance and use of scarce telecommunications services relative to other goods and ser- vices are left in the hands of customers.16 One important variant of call charging in response to network congestion is to charge for un- successful call attempts. This would tend to reduce the common prac- tice of dialing repeatedly, which contributes to network congestion. Although industrial countries have generally chosen not to charge for unsuccessful call attempts, developing countries should perhaps con- sider this approach. From the viewpoint of economically efficient ra- tioning of scarce network capacity, such a charge should apply both if the call is blocked by the network and if the called line is busy. How- ever, charging for calls blocked by the network does not give the tele- phone company the best incentive to improve its call completion rate and may well be seen by customers as an unfair practice. Neverthe- less, even charging only for unsuccessful call attempts when the called line is busy would reduce network congestion. Call Metering The practice of charging for individual calls is a prerequisite for any form of efficient call traffic pricing. To keep tariff schedules adminis- tratively simple and easily comprehensible, however, call charges can never completely reflect the exact cost of making each call. Neverthe- less, groupings into general categories or along specific routes can roughly reflect costs and have been found to be administratively work- able and easily understood. Following from this, particularly in developing countries, special at- tention needs to be paid to the disproportionately large burden placed upon a network during peak hours by heavily used subscriber lines. Because they are busy a relatively high proportion of the time, these heavily used lines engender an unproductive circle of unsuccessful attempts to establish connections and worsening traffic congestion.17 In such instances, call metering and pricing for both local and long- 276 TELECOMMUNICATIONS TARIFF POLICY distance calls made during peak periods should be seriously considered. 18 In the short run, however, before metering equipment can be in- stalled, a temporary approach to this problem might be to use a mul- tiple tariff, in which the charge per time unit or per call for each telephone line increases during the billing period after a certain amount of use has occurred. The objective of such a policy would be to provide incentive to restrict the amount of traffic per subscriber line so that the line would not be continuously tied up, thus restrict- ing incoming calls.19 Automatic toll ticketing may also be desirable in some instances in developing countries. The cost of introducing it is low-about $10 to $20 per direct exchange line for larger crossbar switches that are al- ready installed and virtually nothing for new electronic digital switches-and the benefits can be very large. With automatic toll ticketing, the telecommunications operating entity can collect better traffic data to use in network planning and can expect to reduce sub- scriber complaints and related administrative problems. Given these benefits and the low equipment costs, one South Asian country in 1980 estimated that the automatic toll ticketing equipment that was projected to be installed in its (crossbar switched) network would have a payback period of much less than two years. Automatic toll ticketing can also have broader national benefits. It has the potential to make subscribers more conscious of their calling rates and habits and to help them pinpoint opportunities, as well as waste and ineffi- ciency, in their telephone use. Differentiating between Business and Residential Subscribers The practice in some countries of charging business subscribers a higher monthly rental fee than residential subscribers is also subject to special question in a developing country. As noted in the previous chapter, in situations in which telephone service is generally available on demand, the rationale for a residential-business dichotomy is based partly on the concept of value of service and its underlying premise that business demand for telephone service is more price in- elastic than residential demand. Hence, it is argued that increasing the price of business telephones may not significantly retard their use, whereas reducing the price of residential telephones can increase the penetration of telephones throughout the country and thus benefit subscribers and others.20 POLICY FOR ECONOMIC EFFICIENCY 277 In the typical developing country, however, in which demand greatly exceeds supply, any attempt to stimulate residential consumption through lower prices may not be justified when business demand is not being met.21 In particular, the highly skewed income distribution in most developing countries suggests that the final burden of tele- phone charges should be analyzed before accepting proposals to im- pose tariffs that are different for businesses than for residences.22 Long-Distance Telephone Tariffs The practice still common in developing countries of pricing high- traffic density, long-distance calls above their cost cannot, other things being equal, be defended on the grounds of economic effi- ciency when supply can be readily increased to keep up with demand. Although no definitive analysis similar to that in figure 13-1 has been undertaken for developing countries, the partial data that are avail- able closely match the patterns found in many industrial countries. Short-distance and local calls tend to be subsidized by high-density long-distance and international traffic, which, on average, generates revenues well above incremental and average costs. In fact, in many developing countries investment in the provision of international ser- vices typically yields incremental financial returns on the order of 100 percent annually. As outlined in chapter 1, although the sector as a whole typically obtains financial returns in real terms of between 15 and 30 percent on new investment, incremental financial returns on domestic long-distance services, although lower than for international services, are also usually much higher than the sector average.23 As an example of this, an analysis of the expected incremental financial re- turns on the 1981-85 telephone investment program in one large South Asian country suggested that investment in the larger urban areas was expected to yield an annual return of about 37 percent, and in smaller towns and rural areas it was forecast to yield 16 percent, whereas new investment in higher-traffic long-distance facilities was expected to yield almost 50 percent. The exact implications for economic efficiency of such tariff and investment policies are unknown for several reasons. First, reliable data are lacking on the elasticity of call traffic demand in various situ- ations in developing countries. Second, the elasticity of demand is not known for the many goods and services whose prices partly reflect the high charges that producers pay for long-distance telephone calls. Third, little is known about the calls that were not made, nor the in- 278 TELECOMMUNICATIONS TARIFF POLICY efficiencies incurred, because of inadequate service or high call prices. In general, however, a convincing argument can be made for the exis- tence of a global underconsumption of long-distance telephone and telex communication in developing countries, brought about partly by the practice of charging prices that are significantly higher than costs for long-distance calls and partly because the resulting revenues are not sufficiently devoted to investment in additional long-distance capacity. Arguments against overcharging for traffic for long-distance tele- phone, telex, data, and so forth include the overwhelming business orientation of their use and the likely distortions in resource use asso- ciated with using second-best alternatives for communication-often travel (rail, road, air) or postal services-that themselves may be fi- nancially subsidized by government. Furthermore, the implications of such a policy on regional development, national trade, and transport and energy consumption are unknown. Finally, any equity implica- tions of long-distance price-cost differences will vary greatly among countries and among regions within countries, depending on which groups have access to telecommunications services. In general, how- ever, residents of smaller towns and villages outside the main urban areas tend to make a disproportionate number of long-distance calls at the higher tariff rates. Given the above, telephone administrations in developing countries should closely examine the extent to which they overcharge on or underprovision long-distance networks and should think through some of the implications of so doing. In many instances, they might consider reducing long-distance tariffs over heavily used routes and perhaps making long-distance tariff structures less sensitive to dis- tance (see chapter 13 for a discussion of the sensitivity of costs to distance). Subscriber Financing In several developing countries, subscribers are required to contrib- ute capital to the telecommunications operating entity. This method, called subscriber financing, compels new subscribers to purchase bonds or shares of stock in the telecommunications enterprise in order to provide funds for network improvement or construction (it also serves to ration scarce supply). It works best where there is a large unsatisfied demand for service and no other acceptable source of sufficient capital is available. These schemes have been imple- mented in different forms in several countries. In Taiwan, it was an ir- POLICY FOR ECONOMIC EFFICIENCY 279 recoverable charge; in Japan and Thailand, a bond; in Argentina, Bolivia, and Brazil, a contribution to equity capital; and in Mexico, a contribution to either the equity or the debt capital of the telecom- munications authority. Subscriber financing was used with spectacular results in Japan, where it was a key ingredient in the growth of the Japanese telephone system from 550,000 lines in 1946 to 42 million lines in 1981.24 For periods of several years, such capital subscription or connection charges assisted the telephone systems in Brazil and Taiwan, among others, to expand at a rate of approximately 20 percent or more a year. Furthermore, during those years the rapid expansion of the sys- tem was attained without drawing on government capital.25 Indeed, such expansion was achieved in Brazil and Taiwan with substantial net financial contributions from the telecommunications authorities to government. An Investment Strategy to Complement Efficiency Pricing An analysis of tariff policy in developing countries in which tele- communications will remain as a government-regulated monopoly, and in which waiting lists and system traffic congestion are significant during daytime business hours, could in many countries yield a three- pronged strategy: set prices in the short run near a market-clearing level, speed up the overall rate of investment in the sector, and put an adequate amount of investment in the near term in public telephone facilities (Pcos and coin box telephones) both in the larger cities and in smaller cities, towns, and rural villages.26 When deciding to set prices in the short run at a level that comes close to clearing the market, a relevant question is, what does "close to clearing the market" mean? Roughly, one answer might be prices that bring about perhaps no more than a six- to eight-month waiting period for connections in most areas and a completion rate of approx- imately 40 to 60 percent for the first attempted call during busy hours. For the United States the normal call completion rate during busy hours is about 75 percent, and for Australia it is about 70 per- cent. In a large city in South Asia the figure is currently estimated to be 5 percent. The suggested six- to eight-month waiting period and the successful call rate of 40 to 60 percent must be considered only starting points for discussion.27 In some instances attaining an ac- ceptable successful calling rate during busy hours may be difficult be- cause telephones are used almost continuously, thereby generating 280 TELECOMMUNICATIONS TARIFF POLICY busy signals when a high proportion of the available telephones are si- multaneously trying to call one another. As outlined in chapter 13, a large unsatisfied demand for calls and connections at prices in excess of costs indicates that a rapid expan- sion of the sector is justified; the incremental benefits perceived by telephone subscribers and users exceed the incremental resource costs. As expansion takes place, the price may be gradually lowered so that demand and supply remain roughly in balance. Traditional eco- nomic welfare theory argues that the expansion should not be halted until the price has fallen to the level of marginal production costs, and supply equals demand, or, if as is more likely in this declining- cost sector, to a level at which a lower price would result in financial losses for the telecommunications entity. Not increasing the rate at which telecommunications services are expanded can cause another problem, in addition to the risk that communication-related inefficiencies would multiply throughout the national economy. In the absence of competition, the high profits, which keep rolling in because market-clearing prices are relatively high, have in several countries contributed to significant inadequacies and inefficiencies in the management of telecommunications entities. The financial incentive for efficient management is lost. Costs within the sector rise, and resources allocated to the sector may be increas- ingly misallocated-all the result of management inefficiencies. Limit- ing the natural growth of the telecommunications system, typically by government regulation of entry, prices, profits, investment, and man- agement control, would in these circumstances typically require fur- ther government action. Such action could include levying a special unit tax to avoid the excessive profit that would otherwise accrue to the telephone authority and to ration the decreed level of service so that higher-value uses are maintained. Given a short-run market-clearing price strategy for subscriber tel- ephones, increasing investment in Pco or coin box facilities through- out a country may be desirable for several reasons.28 First, Pco and coin box facilities are generally the least expensive way to provide wide telephone access to the most people. Second, the high market- clearing prices charged in the short run for subscribers' telephones may for a time prohibit their acquisition by smaller business and ag- ricultural establishments that occasionally have an urgent need for rapid two-way communication. As shown in chapter 12 public tele- phones provide a means through which these smaller commercial entities and government administrators in towns and rural areas can POLICY FOR ECONOMIC EFFICIENCY 281 satisfy their highest priority communication needs. Third, as re- flected in the evidence reviewed in chapters 1 and 7, public tele- phones in market centers and villages near highly traveled routes can increase the efficiency of a country's transport system by providing points at which drivers can receive new instructions from headquar- ters or report problems or breakdowns. Finally, evidence reviewed in chapter 12 and appendix C also suggests that such facilities provide a way for the general population to report or discuss emergency or health needs or simply to communicate with family or kin who have left the area. X-Efficiency Another aspect of economic efficiency, separate from the allocative efficiency discussed earlier, is known as X-efficiency, a term coined by Harvey Leibenstein.29 The concept incorporates the ideas of motiva- tion, incentive, and skill or, on the other hand, inertia, apathy, and ineptitude into the economic analysis of efficiency. At a practical level, the concept looks at how efficiency can be increased by working smarter or harder, by reducing costs, or by increasing quantity and quality of output. The basic X-efficiency hypothesis is that neither firms nor individu- als are always as productive as they could be, and, as a result, costs are not always kept to their minimum. Leibenstein linked the concept to competition, observing that competitive pressures engender efforts to reduce costs. All other things being equal, X-efficiency will be higher in a competitive than in a monopoly environment because each environment has different incentive characteristics.30 Many au- thors have found evidence to support the concept of X-efficiency.31 In some cases, the losses from X-efficiency can be larger than those from allocative inefficiency.32 Although the terminology of X-efficiency rarely appears in discus- sions of policy initiatives dealing with privatization and competition in the telecommunications sector (see chapters 3 and 16), the con- cept has clearly been a major motive in the process. For example, ac- cording to Sir Bryan Carsberg (director of the United Kingdom's telecommunications regulatory body, OFTEL), the main objective be- hind the privatization of British Telecom and other public enterprises in the United Kingdom during the 1980s was 282 TELECOMMUNICATIONS TARIFF POLICY to increase efficiency; to promote competition in the interests of creating incentives to become more efficient; to allow freer partici- pation in various markets; . . . and to foster efficiency by an appro- priate form of regulation in areas where competition cannot be effective.33 As an increasing number of countries around the world move to in- crease competition and privatization in the telecommunications sec- tor, X-efficiency can be seen as a significant source of increased potential productivity and improved performance. X-efficiency consid- erations have clearly become important in influencing decisions about the sector's organization and policy (as discussed in chapter 16), as well as in designing and selecting tariff regulation modalities for the sector (as discussed in chapter 15). Notes 1. Traditional normative welfare economics suggests that welfare will be maximized for natural monopolies if the price of each service is set equal to its marginal cost, if output is expanded to meet demand at those prices, and if several related assumptions are made. For a brief discussion of welfare economics and marginal cost pricing in telecommunications, see Littlechild (1979), chap. 9. 2. Kahn (1970), pp. 65-66. Kahn defines marginal cost as "the cost of producing one more unit; it can equally be envisaged as the cost that would be saved by producing one less unit." Although marginal cost and incremental cost are often used synonymously, marginal cost, strictly speaking, refers to the additional cost of supplying an infinitesimally small ad- dirional unit of output, while incremental cost sometimes refers to the average additional cost of a finite and possibly large change in output. 3. Lewis (1949 and 1968) For more on marginal costs and economic efficiency, see chap- ter 13. 4. The costs of subscriber network access consist of the subscribers' terminal equipment (if rented from the telephone operating entity), jack, drop wire, cable-pair to the local ex- change (switch), and line termination equipment at the switch. See Littlechild (1989), app. 1. 5. A discussion of varying charges according to when a call is made is found in chapter 13. Current peak-load pricing theory focuses on uncertainty and nonprice rationing; it is also concerned with devising better ways to handle the pricing of services so as to accom- modate the lumpy investments in capacity that occur as the sector expands. See Mitchell and Vogelsang (1991), p. 23. 6. Short-run marginal cost is simply the change in total variable costs caused by produc- ing an additional unit; Kahn (1970) pp. 71-72. To the extent that depreciation does not vary with use (that is, for a fiber-optic transmission system), depreciation costs are not in- cluded in short-run marginal cost. 7. Long-run marginal costs include capital costs. Setting prices equal to long-run mar- ginal costs is equivalent to setting prices so that the resulting net present value of an invest- ment program is zero. This assumes that the discount rate used to calculate net present POLICY FOR ECONOMIC EFFICIENCY 283 value is appropriate in that it reflects the opportunity cost of capital for projects of compa- rable risk. 8. Air pollution is an example outside telecommunications of a negative externality re- sulting from the production of some goods and services. 9. Of course, sometimes private contractual arrangements take this into account. A busi- ness firm may pay for the home telephones of key employees or children may pay for their mother's telephone so that they can keep in contact. For a more detailed discussion of externalities as they relate to telecommunications, see Squire (1973) and Littlechild (1979) chap. 12. 10. It might be argued that everything comes out in the wash, because of the identity of the number of calls received and made, and that price should equal marginal cost after all. This would allow the financial equivalent of marginal COSt to be recovered from telephone users. Nevertheless, if price is to be used to allocate resources efficiently, every decision to make a call should result from a conscious comparison of benefits and costs. A price based on the marginal (capacity plus operating) cost of the telephone enterprise alone may in some instances not allow this to be achieved. 11. It has also been argued that there can be no connection-related externalities unless there are call-related externalities, and, if the information content of a call is separated from the call-delivery of that information, then at least for business firms, the cost of the call will be reflected in the fees charged by the caller, which will be paid by the person re- ceiving the call; hence, there may be no market failure. See Jonscher (undated). 12. Hazelwood (1950-51 and 1968); Baumol and Bradford (1970); Littlechild (1979) pp. 128-89; and Culham (1987). 13. For example, if government policy favored introducing as much competition as possi- ble in the provision of telecommunications services, and if that policy had not been achieved in certain areas, it would seem perverse for government policy to support charging the remaining monopoly customers a markup higher than cost. 14. Writing about the suitability of Ramsey pricing concepts to pricing of telecommuni- cations services in the United Kingdom, Bryan Carsberg (director general of telecommuni- cations) wrote, "I am much more doubtful about the usefulness of the economic approach for determining the proper balance between local call charges and long-distance charges ... I am keen to ensure that the economic approach should not be used as an excuse by BT lBritish Telecom] to load the burden of any internal inefficiency onto those callers who have most need of telephone service." Culham (1987). 15. Especially since in developing countries waiting times to get new connections typi- cally are several years. If for political or other reasons the official telephone tariffs cannot be used to ration demand effectively, a secondary or black market might be encouraged to do so. In several developing countries telephone brokers use newspapers to advertise the availability of, or demand for, telephone lines. 16. Thus, high-value users who might otherwise be unknown to telecommunications or national planning or development authorities can increase their likelihood of gaining access to a telephone, telex, and so forth. 17. As noted previously, peak-period pricing in Europe has shifted traffic patterns. In several instances, in fact, sharply reduced off-peak rates stimulated calling at off-peak hours to such an extent that capacity was exhausted and overload conditions occurred. See Mitchell (1979). 18. One objection to call metering is that some telephone users pay more and use the telephone less. Nevertheless, it can be argued that pricing by number, time, and duration of the call tests the value of telephone use. Parties who use the service place a value on it at least equal to the price they are paying (calls or duration are no longer free), and to the ex- tent that busy signals are reduced by eliminating overly long or low-value calls, total satis- 284 TELECOMMUNICATIONS TARIFF POLICY faction from the system increases. Even if the average length of calls is relatively short before metering, metering could significantly compress the distribution of calls by length. 19. A variant approach has been tried in India. There, rather than immediately installing peak-period metering equipment for local calls, P&T instituted a tw4o-part call charge. The first 1,750 calls made (or pulses generated) by a subscriber during a quarter (billing period) were billed at the standard call charge of 3 paisa. All calls in excess of 1,750 were charged a 33 percent surcharge and billed at 4 paisa for each call. The objective was partly financial and partly to encourage heavy telephone users (who are generally business firms calling mostly during peak periods) to apply for additional telephones and to reduce the amount of time spent on each subscriber trunk dialing call; the network is highly congested, and all calls are bulk metered, with each pulse counting as one call. This solution is imaginative and could be effective if waiting lists were shorter, additional telephones were more readily available, or heavy users received priority telephone allocations. 20. In economic terms, such a tariff policy tends to maximize the economic measure of welfare-the sum of consumer and producer surplus. See Mitchell (1978) and Baumol and Bradford (1970). 21. The current practice in the United States and several other countries of charging more for the rental of touch-tone telephones, partly on the assumption that the demand for them is more price inelastic, may also be undesirable in developing countries, since touch- tone dialing permits more efficient use of the network. 22. There have been attempts to redistribute income directly through the monthly charge to rent a telephone. In Colombia, for example, both the monthly rental and the initial capi- tal contribution required of residential subscribers are related to the tax rate value of the dwelling in which the telephone is to be installed. 23. Two alternative explanations justify such a pricing and investment policy: (a) govern- ment has set revenue targets relatively high, and this has generated too few calls and too lit- tle investment, or (b) short-run financial constraints or other investment priorities have limited investment in long-haul facilities, and hence the high charges for long-distance calls are appropriate in the short run to ration the fixed capacity. Both explanations probably have some validity. 24. The Japanese financing program reached its peak between 1963 and 1972, when it generated 32 percent of the investment funds. The public accepted the program, which re- mained in effect until 1982. See Gellerman (1986). 25. From the point of view of the telecommunications operating entity, capital subscrip- tion charges only generate local currency; in most countries, considerable foreign exchange is also needed for the system expansion. 26. It is sonietimes argued that there are a few cases in which exceptions to the high connection fee should be granted to meet distributional or equity objectives of the coun- try's development policy. For example, education, health, and emergency services might be favored with lower connection fees and rentals, and, if it was thought desirable, differences in regional incomes or urban-rural incomes might be partially compensated by having higher connection fees for larger cities and lower fees for smaller cities and towns. This is, in fact, done in several countries that have instituted significant capital subscription re- quirements in addition to connection fees. It can be argued, however, that governments should subsidize these services directly and should not force other telephone subscribers to bear a disproportionate amount of the subsidy burden, particularly by charging lower rates for telephones used for public service activities. 27. For an interesting discussion of some of the economic issues involved in determining an appropriate grade of service, see Littlechild (1979), chap. 11. 28. If direct subsidies from government are not feasible, which is the situation in most developing countries, such facilities can usually be financed out of the financial surpluses generated from urban subscribers and the long-distance network, even if in the short run POLICY FOR ECONOMIC EFFICIENCY 285 some of them lose money. In an interesting simulation example, data for a Latin American country were used to simulate the possibility of generating extra revenue from urban ser- vices through price increases and using it to subsidize rural services. The results of the sim- ulation, which were partly predetermined by the assumptions made, showed that during 1978-87, the extra revenue from urban services more than canceled out the reduction in revenue from rural services and predicted that by 1987 rural traffic would be 30 percent higher than it would have been without the increased cross-subsidy. See Walsham (1979). 29. Although the terminology of X-efficiency was new at the time, the concept was not. See Leibenstein (1966 and 1973). 30. For example, if a monopoly firm fails to produce a required level of service in any given time period, the consequences to the firm will be minimal since its customers cannot easily go elsewhere. In a competitive environment, such mistakes or behavior send custom- ers elsewhere and lose business. 31. Scherer (1970) writes that "X-efficiency exists, and it is more apt to be reduced when competitive pressures are strong than when firms enjoy insulated market positions." Also see Frantz (1988). 32. Hollas and Hereen (1984). 33. Carsberg (1986). Chapter 15 Tariff Policy in Practice ALTHOUGH PRICING BASED ON MARGINAL COSTS is theoretically the ideal ap- proach from the viewpoint of allocative efficiency, several difficulties are associated with its practical implementation. Several areas in which such difficulties arise in the telecommunications sectors of de- veloping countries were discussed in chapters 13 and 14. These in- clude demand forecasting, income distribution, the need for telecom- munications operating entities to achieve adequate financial returns, and the presence of externalities.' This chapter examines additional problems involving the estimation of marginal costs and reviews sev- eral approaches to tariff analysis and regulation that are followed in practice.2 Estimating Marginal Cost Traditional economic welfare theory implies that marginal or incre- mental costs of expanding the system should be used at least as a benchmark to develop a tariff structure for telecommunications mo- nopoly organizations when demand can be met. Therefore, it is im- portant to have a forward-looking cost analysis for each telecommuni- cations product or service. Unfortunately, many conceptual as well as practical problems are involved in defining the incremental costs to the system of even just expanded telephone service. Problems are also associated with measuring those costs. For an incremental cost analysis, the output of the telephone au- thority (more telephones, private branch exchanges, local calls in big cities, local calls in small towns, long-distance calls over low-traffic 286 TARIFF POLICY IN PRACTICE 287 routes, long-distance calls over low-traffic routes, long-distance calls over high-traffic routes, data transmission, television transmission, telex, and so forth) is difficult to measure. Even when output can be adequately defined, difficulties of measurement arise because invest- ments might not be made in frequent, small increments. Investment streams for individual services or products can be "lumpy." This re- sults in conceptual and practical problems in determining the proper time horizon and the accompanying incremental quantity of output for the analysis. These issues are discussed briefly in following sections, but first the real costs, which are incurred by the economy as a whole, must be distinguished from the financial costs, which are paid by the telecom- munications entity itself. This involves the concept of shadow pricing. Economic Input Costs Market distortions frequently exist in developing countries. Hence useful estimates of the real economic costs of producing various goods and services are obtained by adjusting the financial costs and prices according to economic analysis. Competitive market prices gen- erally indicate the opportunity cost of inputs for the production of telecommunications services when those inputs are produced and supplied under near-competitive conditions or are freely imported. In some circumstances, however, it is possible to identify inputs whose prices do not reflect their real value to the economy; therefore they should be shadow priced.3 In particular, overvalued local currencies (resulting from artificial support such as foreign exchange controls and protective measures) and wages for unskilled workers determined by minimum wage laws or labor agreements (when unemployment and underemployment are widespread) tend to require the use of shadow values so that the re- source cost to the economy of using foreign exchange, unskilled local labor, or local production in the telecommunications sector can be properly estimated. As in the calculation of internal rates of return, which was discussed in chapter 8, transfer items such as import duties should be excluded from cost estimates. Such adjustments are espe- cially important for the telecommunications sector in developing countries, which is typically capital intensive and relies heavily on im- ported equipment and materials. Interest rates should also reflect opportunity costs of borrowed funds in the economy as a whole. This is not always reflected in the rates that enterprises actually pay, particularly when they are publicly 288 TELECOMMUNICATIONS TARIFF POLICY owned, as is often the case with telecommunications entities in deve- loping countries. Finally, income generated by the telecommunica- tions operating entity for the public sector may have a premium placed on it, relative to consumption, when public savings are inadequate.4 The foregoing adjustments allow a set of shadow prices, called effi- ciency prices, to be derived. A more detailed discussion of shadow pricing in the telecommunications sector is presented in appendix D. Capital Indivisibility and 17icremental System Costs Problems associated with setting price equal to marginal cost are particularly apparent in the presence of capital indivisibility, a condi- tion that is often encountered in telecommunications expansion pro- grams, particularly in smaller developing countries.5 For example, if the objective is to meet demand fully, switching and transmission ca- pacity might be installed to make up for local deficits in supply and to meet some portion of future demand. The initial costs of subscriber, exchange, and transmission facilities are relatively high in relation to operating and maintenance costs, and scale economies can usually be realized with some initial overprovisioning. Even where extending ca- pacity in fairly small increments is technologically possible, fluctua- tions in the availability of finance can mean that capacity must be extended in somewhat larger lumps. Strict marginal cost pricing in these circumstances may require periodic, large changes in price. If in such circumstances price were set equal to marginal system cost, price would equal short-run marginal cost when capacity is less than fully used, and, when demand increases so that existing capacity is fully used, price would be raised to ration the fixed capacity. This procedure would continue until subscribers, system users, or parties demanding new access or service reveal their willingness to pay a price for additional service equal to short-run marginal cost plus the an- nual equivalent of marginal capacity cost.6 At this point-where ex- isting capacity is fully used and consumers are paying a price that equals long-run marginal cost-investment in additional capacity can be justified. Once the investment has been carried out, however, price would need to be lowered again to the point where it reflects only short-run marginal costs; then the only real costs (or opportunity costs in terms of alternative benefits forgone) are for operation and maintenance. Price therefore plays two distinct roles: obtaining effi- cient use of resources invested in existing capacity and providing a signal to invest in additional system capacity. TARIFF POLICY IN PRACTICE 289 Alternative Approaches to Estimating Marginal Cost In view of these difficulties, several approaches have been developed to estimate or approximate marginal costs in the telecommunications services sector. In the simplest case-a telecommunications system in which investment roughly keeps up with demand and occurs simulta- neously in several parts of the network-the investment streams asso- ciated with one or more telecommunications service outputs may be relatively smooth. In such an instance a measure of marginal cost that has been used widely in electric power pricing-total long-run incre- mental cost (TLRIC)-seems to be an appropriate method.7 TLRIC is de- fined as TLRIC, = ,- Cl + rlk Q'+1 - Q~ Qk--l - Q where t = year or period for which the estimate is being calculated, C,+l = operation and maintenance costs in year or period t + 1, Q+= quantity of telecommunications service output in years or per- iod t + 1, 1 = the next major capital expenditure, r = the capital re- covery factor, or the annual payment that will repay a $1 loan over the useful life of the investment with compound interest (equal to the opportunity cost of capital) on the unpaid balance,8 and k = the year or period in which the very next major investment expenditure is completed. During the period t through k, the term rlk / (Qk- I - Qk) remains constant, reflecting the annual equivalent of marginal capacity cost for the next increment of investment. After the investment has oc- curred in year or period k, k is redesignated to be the next year or per- iod in which a large investment will be completed, so that the price reflecting TLRIC will be adjusted in a stepwise fashion over time. Other analytic methods have been developed to estimate incre- mental costs in the production of local exchange services. Four methods can be employed: econometric methods, engineering plan- ning models, optimization models, and engineering-economics proc- ess models.9 Each method has strengths and weaknesses. For example, the econometric method is based on historical data de- scribing expenses and output but cannot easily account for techno- logical change nor for new methods of production. Engineering planning models, in contrast, handle the costs of new technology well but require extensive data that may not be easily available to planners in developing countries. Optimization models can yield ac- curate information on specific design problems, but the relatively 290 TELECOMMUNICATIONS TARIFF POLICY narrow focus of this approach makes it difficult to use as a general model to obtain marginal costs of access and local use. The engi- neering-economics model avoids some of these problems and works with the major factors that determine incremental costs without the need for highly detailed data. A particular advantage of this kind of model is that it can be used to estimate the average incremental costs of supplying an increased level of service for a specified set of demand and other conditions.1L Tariff Analysis in Practice Tariff analysis is an extremely complex undertaking, due to the ex- istence of multiple objectives for tariff policy as well as to the prob- lems that affect the allocation and estimation of costs in concept and in practice. As a result, many approaches to analyzing tariffs for the telecommunications sector have been developed. Each approach is tailored to the specific needs of a particular country and its telecom- munications system, and none is clearly superior. Two approaches are described in this section. In the first example, TLRIC methodology was applied to the tariff structure of a telecommunications operating entity in a large Asian country, as shown in table 15-1.II The TLRic-derived marginal costs were compared with actual tariffs charged by the telecommunications administration in order to determine the extent to which tariffs actu- ally reflect costs of operation. Connection costs were estimated based on the cost of connecting a subscriber to a distribution point plus an allowance for inside wiring. Subscription charges were estimated by adding together the annual- ized investment cost, the operating and maintenance costs for an ex- change line, plus an allowance for local exchange equipment related to line termination. Local call charges were estimated from the annualized cost of traffic-sensitive local exchange investments plus operating and maintenance costs. These costs were spread across ex- pected peak local traffic. The comparison in table 15-1 shows that tariffs were significantly out of line with costs. Connection charges were generally too high, and subscription charges were far too low. Although charges for local calls on automatic lines appeared to be in line, no charge was made for local calls on manual lines. Trunk call charges were much too high. TARIFF POLICY IN PRACTICE 291 Table 15-1. TLRiC Analysis of Existing Tariffs in an Asian Country (U.S. dollars) Estimated Existing TarifJ component LRMO' tariff Connection charge Automatic 65.04 66.37-442.48 Manual 152.21 Mostly over 176.99 Monthly subscription Automatic 14.51 1.77-3.10 Manual 11.68 0.88-1.77 Local call charge (per pulse) Automatic 0.05 0.07 Manual 0.12 No charge Long distance charges (per three-minute call) Automatic (distance in kilometers) 25-100 0.07 1.99 100-300 0.10 2.99-2.39b 300-600 0.19 3.98-2.99b 600-1,000 0.24 3.98 Over 1,000 0.32 5.97 Manual (distance in kilometers) 25-100 0.17 0.99 100-300 0.22 1.19-1.49b 300-600 0.28 1.49-1.99' 600-1,000 0.32 1.99 Over 1,000 0.35 2.99 a. Estimated LRMC (long-run marginal cost) was used to illustrate an economic (as op- posed to a financial) incremental cost estimate. b. Exact comparisons with estimated LRMC prices were not possible since the existing tar- iff used different categories for distance. Table 15-2 shows the extent to which these existing tariffs satisfied five typical objectives of tariff policy. In particular, the economic effi- ciency of the tariff structure, as indicated by the relation of the tariffs to LRMC was contrasted with other objectives such as efficient network use and generation of tax revenue. The economic efficiency of this particular tariff structure was found to be bad to very bad, yet many other policy objectives were well met. The most extreme case appears in trunk, or long-distance, calls, where economic efficiency and assistance in rural development were very bad, due to high tariffs, even though the ability of this tariff to pro- Table 15-2. Existing Tariffs Evaluated in Terms of Policy Objectives Efficient Assistance Generation Economic network Service to rural of tax Charge efficiency' useb rationing, developmentd revenue' Connection charge Automatic x (+) n.a. vv x(+) v Manual x (+) n.a. vv x (+) v Monthly subsLriptioni Automatic x (-) n.a. x (-) v xx (-) Manual x ) n.a. x (-) v xx (-) Local call charge S. Automatic x v n.a. v x (-) Manual x x (-) n.a. v x (-) Trunk call charge Automatic xx (+) v n.a. xx (+) vv Manual xx (+) v n.a. xx () vv n.a. Not applicable. Note: Performance ratings for objectives are v, good; x, bad; (-), tariff too low; vv, very good; xx, very bad; (+), tariff too high. a. Economic efficiency is defined here as the convergence between long-run marginal cost and tariff revenues. b. Efficient network use refers to the degree to which peak-load pricing encourages off-peak use, suppresses excess peak demand, or both. c. Service rationing refers to the ability to ration scarce access of new subscribers to network connection. d. Assistance to rural development refers to the ability to keep prices low to increase access in poor and rural areas. e. Generation of tax revenue refers to the likelihood that sufficient revenues will be generated to enable transfers from the telecommunications authority to other government departments. TARIFF POLICY IN PRACTICE 293 mote efficient network use was good, and its ability to generate tax revenues was very good. Even so, network efficiency might be im- proved further by introducing lower, off-peak charges to take advan- tage of spare capacity, and additional revenue could be raised by increasing access charges, with little detrimental effect on economic efficiency and network use. The local pulse rate for automatic service was found to encourage efficient use to a certain extent but was too low, since the network continued to suffer congestion during busy hours. The manual local service, available at no charge, was clearly extremely inefficient, and these circuits were badly congested as a re- sult. As regards the goal of rationing access to service, the perfor- mance of the tariff was contradictory: high connection charges discouraged new applications, whereas low subscription charges en- couraged subscribers to hold onto existing lines. Both charges would have to be high in order to ration effectively. Assistance with rural de- velopment was facilitated through cross-subsidy of local charges but probably was counteracted completely by very high trunk charges, which are incurred more frequently by rural subscribers than by urban users. These tradeoffs highlight the complexity of determining the ap- propriate tariff structure in practice, where conflicting policy objec- tives must be balanced. A quite different approach is used by the Canadian telecommuni- cations regulatory authority (CRTC) in its cost inquiry phase III methodology.'2 This approach was developed specifically to identify the sources and recipients of cross-subsidies among broad categories of existing services and the existence, if any, of a cross-subsidy of competitive categories by monopoly categories. The approach uses historical accounting data instead of prospective long-run marginal costs. For a regulatory agency, such data have the particular advan- tage of being auditable; another advantage is that the phase III methodology is much less complex and less charged with conceptual problems than theoretically more appropriate forward-looking mar- ginal costs. The phase III methodology establishes eight broad categories for the allocation of costs and revenues (see appendix F for the defini- tions). In 1989, these categories were access (A); monopoly local (ML); monopoly toll (MT); competitive network (CN); competitive terminal-multiline and data (CT-MD); competitive terminal-other (CT-O); other (o); and common (C).13 Of particular importance is the creation of "access" (that is, sub- scriber network access) as a separate category.'4 As a result of an im- 294 TELECOMMUNICATIONS TARIFF POLICY mense amount of work over several years, standardized procedures were developed to allocate almost all revenues earned and costs in- curred in the provision of the carrier's services in a given year on the basis of causality. "Common" is a separate category used for costs that cannot be causally related to a particular service category. The resulting comparisons of revenues and costs or, alternatively, of net revenues per category permit the direction and magnitude of broad cross-subsidies to be assessed quickly. In this way, the results provide important information relevant to assessing possible changes in the balance or structure of tariffs for telecommunica- tions services. For example, the analysis shows whether competitive carrier services are cross-subsidized by monopoly services, an im- portant concern of regulators. Furthermore, the analvsis indicates the magnitude of the cross-subsidy from monopoly toll to sub- scriber access categories. To meet efficiency goals, the appropriate prices can be adjusted to reduce or eliminate cross-subsidies. Phase III methodology results for Bell Canada in 1987 are shown in fig- ures 15-1 and 15-2. Figure 15-1. Revenue/Cost Results for 1987, by Service Category, According to the Bell Canada Tariff Analysis Methodology Billions of Canadian dollars 7 6 Revenue 5 Costs 4 3 2 0 Access Monopoly Monopoly Comp. Comp. comp. Other Comnion Local Toll Network Terminal Terminal Multi-Data Other Serv,ice categories Source: Canadian Radio-Television and Telecommunications Commission. TARIFF POLICY IN PRACTICE 295 Figure 15-2. Revenue Surplus/Shortfall Results for 1987, b-y Service Category, According to the Bell Canada Tariff Analysis Methodology Billions of Canadian dollars 2,000 1,500 * 1,000 , 500 -500 -1,000 -1,500 -2,000 -2,500 Access Monopoly Monopolv Comp. Conmp. Comp. Other Common Local Terminal Network Terminal Terminal Multi-Data Other Service categories Source; Canadian Radio-Television and Telecommunications Commission. Tariff Regulation Until the early 1980s, telecommunications operating entities in most countries were either government departments or public enter- prises. Privately owned telecommunications entities mainly existed in North America and the Philippines. Under these circumstances, regu- lation used to plav a less prominent role in policy implementation in most countries than it does today, because governments exercised di- rect control over the telecommunications operating entities. Starting with the privatization of British Telecom in 1984, a global trend to- ward increased privatization has occurred in many countries. In some, the telecommunications operating entity was reestablished as a com- mercially oriented government-owned corporation (for example, in Fiji, France, the Netherlands, and Sri Lanka); in others, the operating entity was privatized (for example, in Argentina, Chile, Japan, Malay- sia, Mexico, and New Zealand). 296 TELECOMMUNICATIONS TARIFF POLICY Despite attempts to foster competition in the telecommunications sectors of many countries (often occurring as a complement to poli- cies of commercializing or privatizing the dominant telecommunica- tions service entity), significant elements of natural monopoly or market power continue to exist. In this situation, the question arises as to how governments should best regulate the sector in order to avoid monopoly abuses and optimize performance. This section, therefore, reviews some aspects of regulation as it applies to tariffs, al- though tariff regulation is only one option among many that can be used to control the behavior of a firm with market power.15 Rate-of-Return Regulation Key elements of traditional rate-of-return regulation were described in chapter 13. The approach imposes on the regulated firm a maxi- mum allowable rate of return on capital. According to this approach, the telecommunications sector in North America produced enormous benefits for the economy, as well as for the telephone system. The ratio of households with telephones to the number of all households increased from one in three in 1920 to nine in ten in 1970, and the total number of telephones rose by a factor of ten during this same period.16 Other evidence indicates that the productivity gains in the U.S. telecommunications industry between the 1950s and the 1980s were consistently higher than those in the U.S. economy as a whole."7 Despite these benefits, rate-of-return regulation has well-known drawbacks as well. These include weakened incentives for the regu- lated firm to maximize technical efficiency because of the cost-plus character of rate-of-return regulation, incentives (under some circum- stances) to use excessive amounts of capital relative to other inputs, incentives to engage in predatory, below-cost pricing in competitive markets, and the costly and time-consuming administrative proce- dures connected with the regulatory process.18 Price-Cap Regulation Because of the drawbacks of rate-of-return regulation, there has been widespread discussion of whether alternative regulations might improve economic performance. Most of this discussion has focused on price caps.'9 Under a price-cap approach, the regulated firm is required to en- sure that the aggregate price level of a basket of services will rise by TARIFF POLICY IN PRACTICE 297 no more than a certain amount. Typically, this amount is calcu- lated using the formula CPI - X, that is, the general rate of infla- tion, as defined by the consumer price index, minus the estimate X, which is the gain in productivity in the telecommunications sec- tor that is greater than the gain in the economy as a whole.20 The formula is then fixed for a specific period, for example, five years. Thus, price-cap regulation differs most significantly from tradi- tional rate-of-return regulation in that it sets limits on the rates that carriers may charge for their services, while eliminating almost all limits on the profitability of such services during the period in which the fixed price-cap formula is in effect. However, rate-of- return regulation and price-cap regulation are not totally different because both depend on forecasts of rates of return. Thus, even under the price-cap approach, the all-important estimate of the productivity gain X is partly determined by future rates of return calculated under various scenarios. The main advantage of the price-cap approach is that it forces a telecommunications enterprise to pass on a minimum assumed pro- ductivity gain of X to its customers in the form of reduced real prices and creates an incentive to increase productivity, which is reflected in profits that can be retained. By the same token, the enterprise is com- pelled to absorb any losses resulting from inefficiency. Additionally, the price-cap approach eliminates the inappropriate incentive that may exist under rate-of-return regulation to use excessive amounts of capital inputs, or to "gold plate." Finally, the approach is believed to be somewhat simpler to administer than rate-of-return regulation. Another advantage of price caps applies in some developing coun- tries, where telephone rates have the potential to be politically contro- versial. Once a price-cap tariff adjustment formula has been fixed for a predetermined period, tariff adjustments can be taken out of the po- litical arena to some extent and implemented automatically by the regulated firm. This gives the telecommunications entity greater secu- rity in its future revenues and allows it to make investments that will meet demand and generate profits. However, price-cap regulation does not solve all tariff regulation problems. The productivity gain factor can be difficult to determine, and carriers' profits are greatly affected by small variations in this fig- ure. Exogenous cost changes are also problematic, since the general- ized price indexes such as the current price index may not accurately measure all the cost changes occurring outside the firm's control.2" Finally, price-cap mechanisms can create an incentive for reducing 298 TELECOMMUNICATIONS TARIFF POLICY the quality of service and result in a cumulative deterioration in net- work infrastructure.22 In a sense, price-cap regulation shifts some of the focus of regulation from profits to quality of service. Price-cap regulation is currently the most prominent type of incen- tive regulation. It is designed to provide telecommunications enter- prises with an incentive to achieve efficiency gains and in this sense is linked to the discussion of X-efficiency in chapter 14. The method was first applied to British Telecom when it was privatized in 1984 and has subsequently been applied in the United States to AT&T, tO major local exchange carriers, and in many state jurisdictions, as well as in Mexico and other countries. Developing countries have not had widespread experience with the method. Because the track record of the price-cap approach is still incomplete, there is some concern that the incentives for improving efficiency will also provide incentives for reducing the quality of service. Further, if the price-cap method is not administered well, it can lead to profit levels that some would con- sider excessive. Notes 1. For a more general discussion of demand forecasting in telecommunications, see Littlechild (1979), chap. 3. For a discussion and review of the factors underlying telephone demand, see Taylor (1980) and Mitchell and Vogelsang (1991), chap. 4. 2. The objective of this book is not to discuss the premises underlying the portion of normative economic theory known as welfare economics. A good introduction to this sub- ject in the area of telecommunications can be found in Littlechild (1979), chap. 9. 3. Unless otherwise noted, all costs are assumed to be in constant prices with respect to a given base year, that is, net of inflation. 4. For a general discussion of shadow pricing, see Little and Mirrlees (1976) and Squire and van der Tak (1975). For further elaboration of the use of shadow pricing in public en- terprise pricing policy, see Munasinghe and Warford (1978). 5. This section relies heavily on Saunders, Warford, and Mann (1977). 6. In theory, the total cost of an additional lump of investment should be attributed to the very last unit of output. In practice, however, the output considered as incremental has to be averaged over a particular period of time. In fact, the investment can be justified when consumers show their willingness, over a given period, to assume the financial (and presum- ably the economic) burden of additional capacity investment measured over the same per- iod. See the definition of the capital recovery factor (r) in the Tl RIC estimation method. 7. Turvey and Anderson (1977). Where capital indivisibility is a major problem, as when inadequate data or price inflexibility do not permit the TLRic approach to be used satisfac- torily, a simple shorthand approach that has been widely used in World Bank public utility projects is the "average incremental cost" method, which is defined as the discounted pres- ent worth of system costs divided by the similar discounted value of incremental output. For a description, see Saunders, Warford, and Mann (1977). 8. r = 1(1 + i)" / (1 + i), - 1, where I is the investment cost, i is the appropriate inter- est rate, and n is the useful length of life of the investment. TARIFF POLICY IN PRACTICE 299 . 9. The material in this paragraph is drawn largely from parts of a paper by Mitchell (1989). 10. For a more detailed discussion of these methods, see H. Okazaki (1984); Roojsma (1985); Yoshita and Okazaki (1985); and Skoog (1980). 11. Tomlinson and Wellenius (1987). 12. The Canadian federal telecommunications regulatory agency is the CRTC (Canadian Radio-Television and Telecommunications Commission). For an explanation of the CRTC cost inquiry phase Ill methodology, see CRTC (1984); Bigham and Wall (1989); and Bigham (1990). 13. Terminology often differs between British and North American descriptions of iden- tical items in telecommunications systems; for example, subscriber apparatus (United King- dom) = terminal equipment or customer premises equipment (North America); subscriber local line (United Kingdom) = local loop (North America); exchange (United Kingdom) = switch or central office (North America); long-distance (both) = trunk (United King- dom) = toll or interexchange (North America). 14. Kahn and Shew (1987). 15. These include approval of interconnection arrangements, authorization to provide competitive services, authorizations for customers to provide their own telecommunications links, and nonsectoral antitrust legislation. 16. Selwyn and Lundquist (1989). 17. Federal-Provincial-Territorial Task Force on Telecommunications, Canada, (1988). 18. Johnson (1989). 19. See Littlechild (1983); Rohlfs and Shooshan (1988); Selwyn (1988); Director Gen- eral of Telecommunications, OFTEL (1988); Hartley and Culham (1988); and Beesley and Littlechild (1989). 20. In the United Kingdom, the equivalent of the consumer price index is known as the retail price index. In the United States, the Federal Communications Commission has pro- posed a price-cap formula based on the price index of the gross national product. See U.S. Federal Communications Commission (1989). 21. See Johnson (1989) and Bethesda Research Institute (1988). 22. Johnson (1989). Part VI Mobilizing Resources and Promoting Efficiency: Alternative Approaches Chapter 16 Restructuring the Telecommunications Sector FOR TELECOMMUNICATIONS, THE 1980s was a decade of revolution. The traditional modalities of the sector's organization were successfully challenged in industrial countries and gave way to structures that were vastly more complex and dynamic. A rapidly growing number of de- veloping countries undertook or began preparing to undertake a major overhaul of their telecommunications policies. A rapidly ex- panding body of literature addresses the issues and options for chang- ing the structure of telecommunications and draws lessons from the practical implementation of sectoral reforms.' This closing chapter provides a link between the approach taken in this book, which examines telecommunications from the standpoint of costs and benefits, and alternative approaches that emphasize pol- icy, structure, and regulation of the sector. To this end, we shall briefly examine several questions: How has the perception of the role of telecommunications in development been evolving in recent years? Why are structural changes taking place in the telecommunications sector, and what directions are they taking? What are the main issues that need to be addressed for these changes effectively to help over- come past constraints? Evolving Views on the Role of Telecommunications in Development Telecommunications is a huge business that moves at a breathtak- ing pace, commanding the attention of politicians, government au- 303 304 MOBILIZING RESOURCES thorities, transnational corporations, and the world's capital markets. The Economist, The New York Times, and other major periodicals reg- ularly carry editorial material and news on telecommunications. It is easy to forget that twenty years ago, only a few analysts perceived the importance of telecommunications and began to make the case that it should be a basic input in economic development. How did this turn- around occur? Until the early 1960s, telecommunications were often considered to be a service mainly consumed by the well-to-do for mostly trivial pur- poses. During the 1960s and 1970s, however, policymakers gradually began to recognize telecommunications systems as essential infra- structure for economic development. It was shown that telecommuni- cations services were used mainly in connection with a wide range of economic production and distribution activities, delivery of social ser- vices, and government administration. They also contributed to im- proving the quality of life and to achieving social, political, and security objectives. When they were available, telecommunications benefited a broad cross-section of the urban and rural population from many different income, educational, and occupational strata. These features generated high social and private returns (with appro- priate tariffs) from telecommunications investment and considerable capacity to mobilize resources. In the 1980s, information came to be regarded as a fundamental factor of production, along with capital and labor. The information sector accounted for one-third to half of GDP and of employment in OECD countries in the 1980s, and this figure is expected to reach 60 percent for the European Communities by the year 2000. Information also accounts for a substantial proportion of GDP in the newly indus- trialized economies and in the modern sectors of less-developed countries.2 In the 1980s, economic activity became increasingly inten- sive in information, and the globalization of capital flows, trade, man- ufacturing, and other activities advanced. This produced a strong demand for better, more varied, and less costly communications and information services. Growth in demand became intertwined with rapid changes in telecommunications technology fueled by advances in microelectronics, software, and optics. These changes have greatly reduced the cost of transmitting and processing information, altered the cost structures of telecommunications and many other industries, created new ways of meeting a wider range of communication needs at lower cost, reduced the dependence of users on established telecom- munications operating companies, and increasingly integrated infor- RESTRUCTURING THE SECTOR 305 mation and telecommunications technologies and services.3 These interrelated processes, illustrated in figure 16-1, are still well under way and show no signs of abating.4 In this context, telecommunications are now widely considered to be a strategic investment that is key to maintaining and developing competitive advantage at the level of the nation, the region, and the firm. Telecommunications constitute the core of, and provide the in- frastructure for, the information economy as a whole. From the stand- point of the user, telecommunications facilitate entry to the market, improve customer service, reduce costs, and increase productivity. They are an integral part of financial services, commodities markets, media, transportation, and tourism, and they provide vital links among manufacturers, wholesalers, and retailers. Moreover, industrial and commercial competitive advantage is now influenced not only by the availability of telecommunications facilities but also by the choice of network alternatives and the ability to reconfigure and manage net- works as corporate objectives change. Countries and firms that lack access to modern systems of telecommunications cannot effectively participate in the global economy. This applies to the least-developed countries of Africa and Asia as much as to the middle-income coun- tries in Latin America, East Asia, and Central and Eastern Europe, which aspire to join the ranks of industrial countries in the next de- cade or so. Changing Sector Structures As the views on how telecommunications relate to the economy have evolved, major changes have occurred in the organization of the sector. Telecommunications used to be regarded as a natural monop- oly and a relatively straightforward public utility. Economies of scale, political and military sensitivities, and large externalities made tele- communications a typical public service. Inidustrial Countries In most industrial countries, telecommunications services were pro- vided on a monopoly basis by government departments or state enter- prises (the main exceptions were Canada, Finland, and the United 306 MOBILIZING RESOURCES Figure 16-1. Driving Forces of Reform in the Telecommunications Sector Increased GlobalizationTehogia information <_ of economic Te inofglobal intensity activity _ .vaio Rapid growth/ and diversification / of user demands / Sector policv reforms more /between \ of global players / users and markets New Increased services Aggressive competition new business RESTRUCTURING THE SECTOR 307 Box 16-1. The Precursors: Liberalizatior and Privatization in the Ulnited Kingdooi The liberalization of the telecommunications sector and the eventual privatization of British Telecom was a flagship project for the Thatcher government's program of free market economic reform and one of the world's pioneer efforts to privatize a public telecommunications authority. In order to introduce competition, the markets for consumer equipment and cellular services were opened to new providers in 1981. In 1984, Telecom was privatized through the sale of 51 percent of shares in the do- mestic and foreign markets. A duopoly was set up in the area of basic fixed-link tele- communications between Telecom and a new operator, Mercury Communications, Ltd., initially a joint venture between Cable and Wireless, British Rail, and Barclays Bank. Also in 1984, the Office of Telecommunications (OFTEL) was created as an inde- pendent, nonpolitical, regulatory body, outside the Department of Trade and Indus- try (the ministry responsible for telecommunications policy and licensing). OFTEL was headed by a powerful director general for telecomniunications, with authority to en- force and amend rules, recommend licensing, and develop strategies. The regulator was intended to be strong enough to control the dominant operator (Telecom) and help Mercury develop as a competitor, but without intervening excessively in the market and thus undermining the benefits of competition. By 1992, the British telecommunications market had been thoroughly trans- formed into one of the most dynamic in the world. In particular, Telecom now has new drive and vigor and is very profitable. It has implemented a massive domestic investment program and successfully entered into new venues such as foreign invest- ments and information systems and services. More than 2 million telephone lines have been added, and over 90 percent of households now have telephone service. With some 2 million shareholders, Telecom ownership is widely distributed, includ- ing 96 percent of its employees. On the other hand, Telecom has experienced prob- lems with quality of service, and its high profitability has resulted in public clamor for lower tariffs. OFTEL, while remaining a relatively small organization with about 200 staff, has succeeded in playing a strong, proactive role in shaping British telecommunications policy. Price-cap regulation, pioneered by OFTEL and increasingly adopted in other sectors and countries, has reduced tariffs in real terms by about one-third. Policy has evolved flexibly in consonance with actual development of the competitive en- vironment. Effective competition with Telecom and Mercury has developed in im- portant markets, especially from cellular operators and a large number of cable television operators. Competition in basic network services has, however, been growing slowly under the duopoly system, and Telecom still controls 95 percent of voice telephony in Britain. Mercury remains quite small in size and scope and is competing mainly in digital local and long-distance services for businesses in large cities, still aided by significant pricing protection from OFTEL. A review of the duo- poly policy in 1990 resulted in a regulatory decision to open the core telecommuni- cations business to more competition. So far, this opening has not attracted new investors in new basic networks and services. Source: Drawn largely from King (1989); Laidlaw (1991); Carsberg (1991); and Gillick (1991). 308 MOBILIZING RESOURCES States). These monopolies generally succeeded in building and profit- ably operating national infrastructures, meeting the demand for basic telephone service, and introducing more advanced services. In the 1980s, however, major structural changes took place, driven by the twin forces of technological innovation and qualitative changes in de- mand. Deregulation and divestiture of the Bell system in the United States were followed by privatization and the introduction of competi- tion in the United Kingdom, Japan, and, more recently, Australia and New Zealand (see box 16-1). By the early 1990s, virtually all OECD countries were in the midst of, or had completed, restructuring their telecommunications sectors. Overall, these reforms increased the sys- tem's ability to satisfy user needs, greatly broadened user choices, in- creased productivity, and reduced prices. Developing Countries In developing countries, telecommunications services were initially run by foreign private companies and colonial government agencies. During the 1960s, most telecommunications operations were nation- alized and taken over by the public sector. Unlike the Prrs in the in- dustrial countries, however, these state monopolies generally fell short of meeting needs. There continued to be large unmet demand for tel- ephone connections, congested call traffic, poor quality and reliability of service, limited territorial coverage, and a virtual absence of mod- ern business services. In the 1970s, attempts to overcome these short- falls focused on obtaining a larger share of the limited public funds and external development credit and aid for telecommunications. Moreover, enterprises in creditworthy countries borrowed funds and, in some cases, imposed mandatory subscriber financing. These efforts resulted in sustained, rapid growth and modernization of telecommu- nications services. In Brazil and Costa Rica, for example, annual growth rates reached or exceeded 15 percent. Overall, however, the sector did not have access to enough capital, nor did it use its scarce resources efficiently. Deteriorating national economies and tight inter- national credit in the early 1980s further constrained investment and led to escalating shortages of supply and lower quality of service, even in countries that earlier had done fairly well. Starting in the mid-1980s, a growing number of governments in de- veloping countries realized that sectoral arrangements had to be over- hauled. This movement toward change was driven by the same factors underlying reforms in the industrial world, namely technological inno- vations and increased demand, amplified by five additional factors: RESTRUCTURING THE SECTOR 309 . State monopolies had reached the limit of their ability to acceler- ate the supply of telecommunications services. In particular, govern- ments realized that they could not provide the huge amounts of capital required to catch up with demand. For developing countries in Asia and the Pacific, Latin America, and Africa to catch up, as a group, with existing unmet demand for basic telephone service by the year 2000, they would have to invest about $25 billion to $30 billion a year, throughout the 1990s, or in real terms, about five times the av- erage level of investment achieved in the 1980s. To this figure at least $10 billion a year should be added for upgrading facilities in Central and Eastern Europe and in the former U.S.S.R., and this does not in- clude the unknown, but rapidly growing, amount needed to develop more advanced business services and facilities. Although reorganizing the enterprise and improving management allow telecommunications entities in some developing countries to use their scarce resources more efficiently, in the context of inadequate government policies, these internal changes alone have limited potential. . In recent years, many developing countries have begun to adopt market-oriented economic strategies, including measures to liberalize trade, promote competition, deregulate financial and capital markets, reduce restrictions on foreign investment, and restructure public en- terprises. In order for these broad economic reforms to be effectively implemented, adequate telecommunications infrastructures urgently needed to be developed. The constituency for telecommunications de- velopment quickly grew to embrace a crowd of powerful and vocal users, including multinational companies, domestic and foreign inves- tors, industrialists, traders, and bankers, all demanding the services they needed to succeed under the new economic strategies. Telecom- munications became a central theme in multilateral and bilateral ne- gotiations on trade in services. Suddenly, governments felt the need to become involved in telecommunications.5 The new economic strat- egies also provided a context in which the new models of sectoral or- ganization became politically acceptable. In several countries (such as Argentina), the state-owned telecommunications operator was the first target of efforts to restructure state enterprises. . Popularly elected governments found that public dissatisfaction with service and, in many countries, extensive corruption of tele- phone company personnel resulted in widespread public support for major reform initiatives. . Parallel changes in the telecommunications sectors of industrial countries raised international awareness of a wide range of sectoral policy issues and options and demonstrated the viability and increas- ing political desirability of alternatives to state monopoly. 3 10 MOBILIZING RESOURCES .Telecommunications operating companies in industrial countries, repositioning themselves in their own changing domestic and regional markets, aggressively pursued new business opportunities in develop- ing countries. Also, foreign banks sought to shift their exposure in highly indebted developing countries from nonperforming loans to new investment opportunities. Telecommunications investments in developing countries appeared to be particularly promising. Main Directions of Change Although fairly universal policy issues and options face govern- ments attempting to reform the telecommunications sector, their rela- tive importance, the sectoral solutions adopted, and especially the strategies to implement them are highly country-specific.6 Yet, all tele- communications reforms so far involve some degree of change along each of four directions: (1) commercializing and separating opera- tions from government; (2) containing monopolies, diversifying sup- ply of services, and developing competition; (3) increasing private sector participation; and (4) shifting government responsibility from ownership and management to policy and regulation. This is true for developing as well as for industrial countries. Commercializing operations. Telecommunications operations are being reorganized along the lines of commercial companies. There is widespread agreement that regardless of who owns them, telecommu- nications operating entities perform best when they are run as profit- driven businesses. Achieving this goal with state-owned entities involves either transforming them into companies or otherwise creat- ing for them conditions that approximate the freedoms, incentives, and discipline of a commercial enterprise. The continuum along which state operating entities are separated from governments begins with the reorganization of government departments into state enter- prises; state enterprises are then transformed into state-owned compa- nies, which are transferred, in turn, to mixed state/private ownership and then from mixed to totally private ownership. Whatever the start- ing point, moving farther along this continuum tends to free the en- tity from rules that may be appropriate for government administration but not for a dynamic, high-technology, capital-intensive business. All reforms are moving in this direction, although only some have com- pletely privatized state operations. At the same time, improvements are also being undertaken in internal organization and management, such as reorganizing the enterprise into cost and profit centers, sub- contracting functions that can be performed efficiently by other RESTRUCTURING THE SECTOR 311 organizations,7 establishing or improving commercial accounting and management information systems, and emphasizing customer service, cost awareness, financial discipline, and staff performance. Diversifying supply and developing competition. Allowing new enter- prises to supply services and networks can attract fresh sources of capital and management to the telecommunications sector, develop rivalry among service providers regarding performance and price, and generate cost benchmarks to guide pricing of monopoly suppliers. A single monopoly operating enterprise, whether state-owned or private, is increasingly unable to meet equally well the large, varied, and rap- idly changing demands of all types of users. Diverse alternatives are available. They include (a) monopolies divided by regions; (b) joint ventures in which the established operator and other enterprises pro- vide specific new services or facilities (for example, very small aper- ture terminal, packet-switched data, cellular); (c) build-operate- transfer and related arrangements with foreign operating companies, equipment manufacturers, and investors for creating parts of the pub- lic network; (d) selected specialized networks licensed to meet the needs of major communication-intensive sectors of the economy (such as banking, tourism, mining) and to attract resources from them for investment; (e) independent public telephone companies li- censed in unattended areas (such as new industrial estates, residential developments); (f) licensed extensions of the public telephone net- work (for example, cellular, rural subscriber radio); and (g) articu- lated rules that govern voluntary commercial relations between dedicated and public telecommunications networks. In the context of broad economic liberalization, an essential ele- ment of reforming the telecommunications sector is a policy on com- petition. Competition, or a credible threat of competition, is likely to spur established operating enterprises to focus attention on custom- ers, improve service, accelerate network expansion, reduce costs, and lower prices. Competition also widens user choices and accelerates the introduction of new services and facilities. Elements of competi- tion can be effectively introduced in the early stages of reform and then extended by stages to affect most or all segments of the market. Modalities include competitively awarded concessions to a monopoly or duopoly, regulated entry, and unrestricted competition.8 Techno- logical changes are making competition possible in a widening range of market segments. Competition in the provision and maintenance of terminal equipment belonging to the customer and value added services is beneficial in virtually all countries. The introduction of some competition in long-distance networks makes sense because the 3 1 2 MOBILIZING RESOURCES increased volume of traffic is reducing the importance of economies of scale. At present, large economies of scale still exist in local net- works, which makes competition in the provision of wired local ser- vices viable only in exceptional situations, such as highly developed urban business districts. New radio technologies (for example, cellu- lar, personal communication networks, mobile satellite), although still more costly than wired telephones, already offer competitive alterna- tives for business and high-income residences under certain circum- stances, such as when conventional telephone lines are very scarce or perform poorly and when rapid installation is highly valued (for ex- ample, when existing networks must be upgraded quickly, as in East- ern Europe, or when disaster relief is crucial). Increasing pritate sector participation. A larger role is being ac- corded the private sector in many countries. Given the right policy environment, private enterprise and investors have responded vigor- ously to business opportunities in telecommunications. As private sector participation increases, the telecommunications sector can be expected to attract new sources of capital, management, and technol- ogy and also to contribute to the development of the private sector overall. More and more developing countries are privatizing their state telecommunications enterprises. This has mainly involved selling the telecommunications provider to a consortium of domestic and foreign investors and foreign operating companies. Privatization by itself does not, however, guarantee improvement of the sector and is not always feasible. Many other avenues are available for private participation in tele- communications. Existing enterprises may divest or outsource con- struction, maintenance, transportation, routine design, billing and collection, directory services, operator assistance, and other functions traditionally undertaken internally. State-owned enterprises can be re- organized under company law, and shares can be sold gradually to the public even if the state retains a controlling interest. Bonds and other debt instruments can be floated in domestic and, sometimes, foreign markets. Franchising, leasing, revenue-sharing, build-operate-transfer, and other arrangements offer options of growing interest. Private fi- nancial entities can be established to invest in profitable new tele- communications ventures; for example, facilities could be operated by other entities, such as the existing or a new telecommunications com- pany, under lease, revenue-sharing, or some other arrangement. Sub- scriber financing schemes can raise a large proportion of the local funds needed for expansion when funds are in severely short supply, and they also can promote the participation of users in company own- RESTRUCTURING THE SECTOR 313 ership and debt. Most of the options to diversify supply and introduce competition are suited to private enterprise. Separating and developing regulation. Policy and regulatory func- tions are being separated from operations. As long as telecommunica- tions operations remain in the public sector, the political system provides, however imperfectly, for reconciling diverse objectives such as commercial efficiency and broader national and regional develop- ment. As operations move away from direct government control and the number of participants in the telecommunications business in- creases, however, this arrangement breaks down and the functions of policy, regulation, and operation must be developed separately. At this point, the nature and extent of additional sectoral reforms that can be undertaken are constrained by the development of institutions ca- pable of effectively formulating and regulating the implementation of policy. In particular, regulation is essential when public sector monop- olies are privatized, and an effective regulatory capacity must be in place at the time of sale. Regulatory frameworks and institutional ar- rangements must be developed early to monitor operator perfor- mance, to prevent abuses of market power, and to promote competi- tion. Alternatively, temporary arrangements could be made to oversee the transition from public to private ownership and to deal with regu- latory matters that require immediate attention while a more perma- nent capability is being built. The Beginnings of Telecommunications Reform in Developing Countries By 1992, major reforms of the telecommunications sector had been completed or were well under way in at least fifteen developing coun- tries, and a comparable number were in preparation. The pace and scope of structural change have varied considerably among regions. The reform movement got to an early start in Latin America. Privatization of state telecommunications enterprises was completed in Chile (1987), Argentina (1990), Mexico (1990; see box 16-2), and Venezuela (1991), and in 1992 reforms were being prepared in Pan- ama, Peru, and Uruguay and being considered in Bolivia, Nicaragua, Paraguay, and other countries. Following the collapse of the communist regimes in Central and Eastern Europe and in the former U.S.S.R., rapid progress was made in outlining broad strategies for reforming the sector in this region as well (see box 16-3). By 1992 all of the new governments were developing Box 16-2. Privatization in a Large Developing Country: The Case of Mexico At the time of the devastating 1985 earthquake in Mexico City, telecommunications in Mexico had already reached a state of near crisis. In 1989, the government responded by announcing a comprehensive plan to modernize the sector. This plan sought to establish a new regulatory framework that would promote efficiency, competition, and private in- vestment; to privatize Telefonos de Mexico (TELMEX, the mainly state-owned monopoly telephone company) while regulating price and quality of service; and to introduce competition in local and long-distance telephone services as well as in cellular and other new services. These far-reaching objectives were largely achieved by 1991. New telecommunications regulations enacted in 1990 established a comprehensive, modern sectoral framework in which policy formulation, licensing, and regulatory func- tions were exercised by the government and networks and services were provided largely by the private sector in an in- creasingly competitive marketplace. The telephone tax was abolished, and tariffs were increased on average and re- balanced to approximate international practice. The Secre- taria de Comunicaciones y Transportes (scT), the ministry responsible for telecommunications, was reorganized to be- come the telecommunications regulatory agency. Operating responsibilities were transferred from the SCT to Telecomuni- caciones de M6xico (TELECOM), a new autonomous state en- terprise that became the domestic and international satellite carrier. TELMEX was reorganized along the lines of a modern commercial company, labor contracts were renegotiated to allow management greater flexibility and facilitate techno- logical innovation, and capital structure was changed to pre- pare for privatization. A controlling interest of 20.4 percent was sold in 1990 to a consortium of private domestic and foreign investors in association with U.S. and European tel- ecommunications operators. The remainder of state-owned TELMEX shares were sold to employees and the public at large, in domestic and foreign capital markets, in successive tran- ches in 1991 and 1992. The new TELMEX has a monopoly over all basic services and networks for six years, after which competition will be allowed. During the reserved period, TELMEX'S concession es- tablishes stringent requirements for growth and quality of service, including quantified obligations to extend service to rural areas. Tariffs are subject to declining caps as well as limits to increases in residential charges. The market for customer premises equipment, private networks, and value added services has been liberalized, and a growing range of network choices is available to businesses. In each of the country's nine cellular regions, one TELMEX subsidiary and one independent operator compete on a duopoly basis. The foundation for competition in long-distance services after 1996 is being laid in several ways. The nine independent cel- lular operators have formed an association to coordinate their relations with TELMEX and the government, and three of the companies have linked themselves via a small digital mi- crowave network and tandem exchange. TELMEX has emerged from the reform as a very successful, viable private company. It has met or slightly surpassed its concession targets in nearly all areas. Growth in telephone lines exceeded 12 percent for 1991, quality and reliability of service have improved in most parts of the country (and a major network improvement program is under way in the metropolitan area of the capital), a new digital overlay net- work has been completed for large businesses using voice and advanced data and other services, substantial productiv- ity gains have been realized, and prices have been further re- aligned toward costs while remaining within the mandated aggregate limits. Profits have risen dramatically, and over 60 percent of the 1991-96 investment program will be financed from internally generated funds. Share prices have increased more than four times in value. Regulation has made slower progress. The SCT'S response to several regulatory challenges has been good so far, but its new organizational structure was only approved in 1992, and there is an acute shortage of managers and senior profes- sional staff able to deal with the rapidly increasing amount of regulatory issues. The agency may be overwhelmed by the technical demands of monitoring TELMEX'S performance and dealing with regulatory issues of the North American Free Trade Agreement being negotiated. Source: Based on Casasus (1991) and on internal World Bank sources. 315 3 1 6 MOBILIZING RESOURCES Box 16-3. Catching up with the Industrial World: Central and Eastern Europe The telecommunications sector was much neglected by the former commu- nist regimes in Central and Eastern Europe (CEE). The current technology lags fifteen to twenty years behind world practice, and there are on average only about 11 telephone lines per 100 inhabitants compared with 30 to 50 lines per 100 inhabitants in Western Europe. Furthermore, service is poor quality and unreliable. Today, CEE countries face three challenges as they develop their telecommunications networks: they must greatly accelerate the rate of growth and increase the scope of services; they must begin im- plementing the latest digital technologies; and they must adapt their insti- tutions to a newly competitive environment. In response to these challenges, most CEE countries aim to double or tri- ple investment and to achieve an annual growth rate of 10 or 12 percent to reach an average density of about 30 telephone lines per 100 inhabitants by the year 2000. The offices administering posts, telegraphs, and telephones (Prrs) are being gradually commercialized and have been, or will soon be, separated from regulatory and policymaking functions in the ministries. Most countries plan to follow the framework for introducing competition and structural reform set out by the 1987 European Commission's Green Paper on Telecommunications.' Existing domestic manufacturers of equip- ment are being strengthened by joint ventures with foreign companies that try to give them a significant role in the growth efforts. Implementation of these strategies has barely begun. A rapid explosion in demand, resulting from efforts to develop market economies, has put increasing pressure on countries to liberalize faster, but the legacy of so- cialist institutions and business processes has made this difficult to put into practice.2 In particular, regulatory structures and policies are still very underdeveloped.' Financing is another stumbling block; estimates of what the region will need over the next decade range between $3.5 billion and $6 billion a year, amounting to some 2 percent of GNP, and foreign exchange, in particular, will be difficult to obtain in sufficient quantities.4 Overall, these goals are probably just barely feasible, if operators use every means available to them to reduce costs, increase revenues, attract outside debt and equity, and control financial planning. In addition, an integrated approach covering policy, regulation, investment financing, and enterprise organization and management is essential for each coun- try, so that all the efforts to change occur in harmony. Finally, signifi- cant help from the West is required, both in financing the changes and in providing technical and management expertise, to bring reforms to fruition. Notes 1. As Europe draws toward closer integration under the single European Community market of 1992, policy consistency in the area of telecommu- nications is being strongly influenced by the 1987 EC Green Paper on Tele- communications and by subsequent EC legislation. Common goals are (1) an open Ec-wide market for terminal equipment, (2) an open market for value added services and liberalization of data services, while accepting na- tional differences in basic network and voice telephony, (3) separation of regulation and operations, (4) development of consistent technical stan- dards to promote a single terminal market, (5) mutual recognition of serv- ice licensing to promote a single services market, and (6) definition of Open Network Provision to give new service providers fair access to the fa- cilities of the network infrastructure and basic services. 2. East Germany is, of course, an exception. Technically, however, it faced perhaps one of the worst situations; one-quarter of its switches were installed before 1934, none were installed after 1965, and many were fully depreciated two or three times over. Unification of the Federal Republic of Germany and the German Democratic Republic was completed in October 1990. Unification of the two PTTs occurred by the introduction of West German legislation (itself recently reformed in 1989-90 along the lines of the 1987 Ec Green Paper) into the East German PTT. The government of Germany made the political decision to bring East German regions up to the general level of West German telecommunications by the year 2000, and a massive investment program was immediately begun to achieve this goal. The total cost is estimated to be around DM60 billion, which will un- doubtedly exert a significant drag on other investment. 3. This is the case even in Hungary, where reforms of the sector are most advanced. Reforms have been under way since 1989, when the Hun- garian Telephone Company (Matav) was separated from the former Prr (Magyar Posta). A ten-year investment program aims to give Hungary a density of 28 lines per 100 inhabitants by 2000, requiring investment of $4.5 billion; a separate three-year program aims to install business services and a digital overlay network at a cost of $1 billion. Less attention has been given to matters of regulation. 4. Even without foreign exchange, much can be done. Between 1973 and 1983, the Bulgarian rPT implemented a concerted effort to expand and upgrade its network, based largely (about 80 percent) on internally generated funds. Municipalities, cooperatives, and state-owned enterprises also helped construct local networks, which were then transferred to PTT ownership. In spite of institutional and structural limitations, Bulgaria achieved a density of 17 lines per 100 inhabitants, nearly twice the re- gional average. The program was derailed by problems with technology transfer and funding cutbacks after 1983. The total cost was estimated at Leva 1.6 billion. Source: Based largely on four papers: Muller and Nyevrikel; Neumann and Schnoring; Nulty; and Ungerer, all in Wellenius and Stern (1991). See also Whitlock and Nyevrikel (1992). 317 3 1 8 MOBILIZING RESOURCES the details of implementation, and Hungary, Poland, and Ukraine, among others, were building state-of-the-art business networks through joint ventures (with, for example, foreign operating companies and through other investment and management modalities). Reform has been slower and more limited in Asia: there have been corporatization (1987) and subsequent partial privatization (1990) in Malaysia, corporatization and liberalization of nonbasic services in In- donesia (1990), decentralization of operations in India (1985, par- tial) and China (1988, extensive), and reorganization of telecommu- nications departments to become state enterprises in Sri Lanka and Fiji (1990; see box 16-4). The pace appeared to be picking up in 1992, however, with Pakistan preparing for privatization, Thailand embarking on a series of major build-operate-transfer ventures, and the Indian government exploring options to restructure the telecom- munications operating entity and possibly open the market to new service providers. In relative terms, the least-developed countries have the most to gain from sectoral reforms, yet efforts to overcome telecommunica- tions shortages in Sub-Saharan Africa have so far been confined largely to improving the performance of state entities (see box 16-5). Although some countries have explored new modalities, including joint ventures with Prrs and the introduction of some competition, most governments still hesitate to consider broader reforms and pri- vatization. This mainly reflects their concern with limited resources, especially skills to prepare and implement such programs. Slow prog- ress is compounded by small markets, extreme paucity of existing fa- cilities, disproportionate size of social needs, and sometimes un- friendly economic policies, which make investment in these countries less attractive to foreign operators and investors than opportunities in other regions. Nevertheless, telecommunications privatization is now under way in Cote d'lvoire and Guinea, other governments have ex- pressed interest in their experience, small private ventures (for exam- ple, cellular) led by local entrepreneurs have been successful in a few countries, and regional organizations are starting to examine broad aspects of telecommunications restructuring. Issues and Lessons The initial experiences that developing countries have had in re- forming their telecommunications sectors are generally encouraging. Where such changes have occurred, telephone service has expanded and improved at a faster pace, productivity has increased, new services RESTRUCTURING THE SECTOR 319 Box 16-4. Restructuring without Privatization: Sri Lanka In the mid-1980s, the telecommunications sector in Sri Lanka was so weak that it constituted a considerable drag on overall economic development. There were only 0.7 telephone lines per 100 inhabitants, and the quality of service was very poor. The monopoly operating entity, the Department of Telecommunications (SLDT), a government department within the Ministry of Posts and Telecommu- nications, was a civil service organization suffering from inefficiency, noncom- petitive remunerations, and low staff morale. In response to general dissatisfac- tion with this state of affairs, in 1985 the government appointed a Presidential Commission of Inquiry to consider reorganizing the SLDT and the telecommuni- cations sector overall. The Commission recommended a radical restructuring that would transform the SLOT into an independent, autonomous, government-owned, but commer- cially oriented company and the creation of an independent regulatory agency, the National Telecommunications Commission. A foreign operating partner would contribute capital and expertise once the new company was established. In 1986, the Telecommunications Board of Sri Lanka was appointed to imple- ment these goals. At the same time, an act of parliament was drafted to provide the legal framework for these changes, several licenses were issued to new opera- tors in cellular and radio paging services, and the market for subscriber terminal equipment was opened to competition. However, opposition from organized labor and SLOT management, mounting political opposition to privatization of state enterprises generally, and growing civil unrest in parts of the country even- tually brought this process to a halt. The Board was dissolved in March 1989. In 1990, the government adopted a scaled-back version of the original plan. The operating functions of SLDT were transferred to a new, government-owned corporation, Sri Lanka Telecom (SLT), which is autonomous and commercially oriented, but under overall government control. A new telecommunications act was passed by parlianient in 1991, which confirmed the establishment of SLT and created an independent regulator, the Telecomniunications Authority, headed by a director general, answering to the minister of posts and telecommunications. Addressing the concerns of the opposition was the key to implementing this limited restructuring. In response to SLDT management's concern for loss of sen- iority and imposition of outside control, new management was only hired in certain critical areas, such as finance and marketing. Organized labor feared lay- offs, loss of pensions, and loss of the social status conferred by being in the civil service; therefore, the transition was made flexible, and employees were given the option of maintaining their civil service status in the new SLT (only 17 percent of personnel chose to keep their civil service status, and 80 percent chose to accept the new working conditions, which required greater effort and efficiency, but which rewarded these efforts with higher reniuneration). Finally, a massive training effort was undertaken, with the assistance of the ITU and the United Nations Development Programme to introduce comnsercially oriented work practices into the new company. Source: Based largely on Watson (1991) and Mendis (1989). 320 MOBILIZING RESOURCES Box 16-5. The Beginnings ot Reform in Africa Telecommunications development faces its greatest challenge in Sub-Saharan Africa. Densities for the region are among the lowest in the world, and most rural areas have no access to services at all. Investment is sporadic, reflecting the overall acute short- age of investment capital for the region, and project implenientation and subsequent maintenance are complicated by an equally acute shortage of skilled personnel. An- other significant limitation is the prevalence of the traditional structure of posts, tel- egraphs, and telephones (rT7). In the past four or five years, however, some partial reforms have begun to show positive results. Benin. Modest reforms were begun in Benin in the mid-1980s, in concert with a $70 million investment program funded by a group of five multilateral agencies. The investment plan required the Office of Posts and Telecommunications (OPT) of Benin to implement an internal institutional reform program, which included (a) re- form of the financial functions, including bill collections, the postal savings bank, and separation of the OPT treasury from the state treasury, (b) modernization of ad- ministrative functions such as marketing, human resources, and procurement, and (c) internal preparation for the institutional separation of posts and telecommunica- tions at the end of five years (1988 to 1993). At the same time, the government of Benin was encouraged to give greater responsibility and autonomy to the OPT by re- placing its direct involvement in day-to-day management with a system of contrac- tual agreements between the state and the enterprise based on multiyear development plans. The results have been encouraging. As of early 1992, the internal reforms were working well, especially in the area of management, and the investment program was on schedule. Financial reforms have also been successful. The rate of uncollected bills fell from more than 50 percent to reach about 10 percent, and OfT'S treasury was separated from that of the state. However, many significant problems remain: the government continues to interfere in daily management, and entrenched civil service structures have limited the changes in employment patterns. Guinea. A more ambitious reform effort is under way in Guinea as part of a gov- ernment effort to privatize state-owned companies, which began in 1985. In the case of the telecommunications sector, an interministerial committee was created in 1989 to supervise institutional reform, and extensive outside assistance in the form of technical and financial consultants is being used. It has been determined that the rTT should be separated into two parts: a niixed-capital, commercially oriented mo- nopoly telecommunications provider, SOTELGLI, and a governnsent-owned Office des Postes Guineennes (orc.). This separation is to be finalized on March 1, 1993. In the meantime, three preparatory processes are under way: (a) a search for private sector partners for the new SOTELOUi enterprise, (b) operational and technical assist- ance aimed at developing structures for the new entities, and (c) engineering studies to support the investment program that will accompany the restructuring. Madagascar. Reform of Madagascar's telecommunications sector is also progress- ing at a brisk pace with the assistance of outside consultants. A current plan calls RESTRUCTURING THE SECTOR 321 for (a) separating operational functions from the ministry, which will retain only reg- ulatory responsibilities, (b) unifying all telecommunications operators (domestic ser- vices are currently provided by a department within the ministry, and international services are provided via the Societe des Telecommunications Internationales de Madagascar, STIMAD) into a single, mixed-capital, commercially oriented monopoly, (c) drawing up a social plan to guide the new operator's personnel planning and em- ployee relations, and (d) implenienting an investnment program to improve the net- work and introduce new services. Implementation is currently set for the end of 1993, but a great deal of preparatory work still lies ahead. Niger. Reform plans in Niger are still being discussed, but the government is well aware that the current structure is not viable. Among the ideas being consid- ered are (a) integrating the satellite services provider STIN into the Office of Posts and Telecommunications, (b) separating Posts and Telecommunications into two offices, giving each enterprise responsibility for achieving certain objectives, along with the autonomy needed to do so, (c) decentralizing managerial responsi- bility within each organization, (d) creating a Human Resources Department within the telecommunications operating entiry together with a scheme for devel- oping human resources, (e) improving internal management, such as procure- ment practices, maintenance, and financial management, (f) creating a separate management department for the capital city, Niamey, and its immediate sur- roundings, where the majority of telephones in the country are located. Rwanda. A comprehensive restructuring program for the PTT in Rwanda was drawn up by outside consultants between 1989 and 1991. This plan called for sepa- rating posts and telecommunications into RWANDATEL, a mixed-capital company, and the Office National des Postes (ONP), a state-owned company. Both would be auton- omous in management and administration, have a commercial orientation, and be granted a thirty-year monopoly. In addition, the policymaking and regulatory func- tions would be reserved for the ministry in charge, while all operational responsibili- ties would be invested in RSANDArEL. The law creating the oNr was passed in March 1991. The law establishing RWAN[DATEL is still under study but is due to be adopted in July 1992; it calls for a first stage, under which state control of capital would be 99 percent, progressing to a second stage, in which the participation of national and foreign capital would rise to 49 percent. The law creating a director general for com- munications, with regulatory authority over the two companies, will be promulgated after RWAN1DATEL is created. The restructuring effort is taking place in concert with a large-scale investment program funded by three multilateral lenders. Source: Drawn from a series of papers presented at a seminar on telecommunica- tions restructuring, sponsored by the World Bank, the ITL, and ACCT, held in Tunisia in May 1992. The papers include Jacques Hababib Sy, "African Nations and Access to Telecommunications Services: Political Economy and Legal Issues"; H. Vignon, "La restructuration des tel6communications au Benin"; Mamadou Pathe Barry and Mohamed Sylla, "Le cas de la Guinee"; Mamiharilala Kasolojaona, "Colloque sur la restructuration du secteur de telecommunications Malagasy"; Anonymous, "Restruc- turation du secteur des tel&communications dans Niger"; and Habyalimana Malien, "Colloque sur la restructuration du secteur des telecommunications de Rwanda." 322 MOBILIZING RESOURCES have become available, and in some cases, international capital mar- kets have been tapped effectively. Sectoral reform is not, however, without dangerous pitfalls. Although it is still too early to assess the extent to which these reforms have overcome past constraints, several key areas of concern have already emerged. Privatization. Privatization of state telecommunications enterprises is not universally feasible, nor does it necessarily improve perfor- mance. In particular, the timing and method of privatization in a given country are determined largely by relatively narrow and some- what unpredictable windows of political opportunity and by broader developments in national economic strategy. In each case, govern- ments must clearly identify the conditions necessary for privatization to be successful and the extent to which these conditions are likely to be met. A growing body of global experience with privatization shows the risks involved (to governments as well as to investors and opera- tors) and highlights ways in which these risks can be reduced. International market considerations are also increasingly important. In the near future, several telecommunications enterprises are ex- pected to be sold in both industrial and developing countries. Timing and preparation will increasingly influence the extent to which privat- ization succeeds.9 Commercialization of operations, organizational and financial restructuring of enterprises, renegotiation of labor con- tracts, and improvement of available information on the enterprise are actions that can make a particular offering more attractive. Privatization, moreover, is not a one-shot deal, but rather a com- plex process of introducing private capital and know-how into tele- communications operations. The process has various aspects: (a) separating operations from government and nontelecommunications activities (for example, posts, manufacturing); (b) restructuring the telecommunications operator as an independent state enterprise that is financially self-sufficient and financially autonomous from the gov- ernment; (c) reorganizing the enterprise internally so that it runs as a business; (d) restructuring the telecommunications enterprise under private company law; (e) devising a privatization strategy (including decisions on controlling interest, employee stock ownership, tranching of stock sales, and residual state ownership) and changing the compa- ny's capital structure to enable implementation of this strategy; and (f) carrying out the sale. These facets may all be dealt with over a rel- atively short period of time (Argentina, Mexico, and Venezuela each did so in less than two years), or they may develop in stages over RESTRUCTURING THE SECTOR 323 longer periods (as occurred in Chile and Malaysia). Some stages (for example, internal reorganization) may be left to the new owners. Designing and implementing this complex process properly require strong and visible high-level political commitment, clear allocation of authority and resources to manage the process, and expert (including foreign) assistance on policy, regulatory, legal, and financial matters. Early in the reform process, the government must clarify its position on tradeoffs among the conflicting interests that inevitably arise from privatization, such as conflicts among existing operators, organized labor, prospective buyers, potential competitors, investment bankers, the treasury, equipment suppliers, large users, and the public at large. 10 Regulation and competition. The single most troubling issue in re- cent reforms has been the slow progress made in developing regula- tory capabilities. All major reforms have been predicated on the expectation that effective public regulation of the privatized monopo- lies-especially with respect to prices, service obligations, intercon- nection, competitive behavior, and access to the public domain (including use of the radio spectrum)-can be implemented fairly quickly. Yet building regulatory institutions in countries with little or no regulatory tradition in any sector is proving to be an arduous and slow task. Whereas some developing countries have carried out satis- factory privatization in a short period of time (Argentina, Mexico, and Venezuela did so in a little more than a year), and at least one country (Chile) completed privatization some five years ago, so far not one has a properly functioning telecommunications regulatory system (see box 16-6). The largest privatized companies are operating with little or no competition and in a regulatory vacuum in which critical regulatory responsibilities regarding licensing, pricing, techni- cal and accounting standards, and performance monitoring, to name just a few, are not being properly discharged. In a market dominated by one operator and lacking effective and proactive regulation, com- petitors are unlikely to emerge and become firmly established, and nu- merous forms of anticompetitive behavior may become entrenched. The search for better regulatory solutions merits high priority in the design of sectoral reforms. Like privatization, development of reg- ulation is not a one-shot affair, and it is intertwined with the develop- ment of the telecommunications market. Lack of an initial constitu- ency for regulation might be partly overcome by establishing a council or advisory board that represents users and other key interests and Box 16-6. Regulation and Competition: Chile Five Years After Chile began a slow but ultimately very successful process of reforming its telecom- munications sector by adopting in 1975 a national development strategy oriented to- ward free-market economics. Two mainly state-owned companies provided most telecommunications services as monopolies: Compania de Telefonos de Chile (crc) had over 90 percent of local telephone lines, and Empresa Nacional de Telecomuni- caciones (ENTEL) was responsible for most long-distance and all international facili- ties. The Ministry of Transport and Communications (MTC) had overall responsibil- ity for the sector, while the National Development Corporation exercised state ownership of CTC and ENTEL, and the Ministry of Economy approved tariffs. In 1977, the Sub Secretaria de Teleconiunicaciones (sLBTEL) was creared within MTc as the specialized regulatory agency for telecommunications, responsible mainly for techni- cal and administrative matters. A national telecommunications policy was promul- gated in 1978, which set the basic principles that guided subsequent liberalization and privatization of the sector. A new telecommunications law in 1982, and an addi- tional law in 1987, established the legal framework for reforms. In 1986, the government began to sell some of its shares in ENTEL to company em- ployees and the public at large through the Santiago stock exchange. In 1987, the government sold 35 percent of CTC shares with a controlling interest to Australian investor Alan Bond. That same year, the remaining ENTEL shares were sold to various domestic and foreign investors, and 20 percent were sold to Spain's Telef6nica. In 1990, Bond sold his share of CTC (now 50 percent. after additional investments in accordance with the terms of the initial sale) to Telef6nica, which then became a major force in Chilean telecommunications. Concurrently with privatization, the market was made more competitive. The reforms have greatly revitalized the telecommunications sector. Investment has been substantial. Telephone lines, which had been growing at a paltry 5 percent a year in the 1970s and 1980s, expanded more than 20 percent in the early 1990s; by the end of 1991, CTc had 0.9 million lines in service, 35 percent of them installed in the previous two years. Networks have been quickly modernized; over 70 percent of telephone lines are digital, and extensive optical fiber cable and satellite networks are being developed. Many new services have been introduced, and some market seg- ments have become very competitive. The supply of customer premises equipment has been fully liberalized, cellular telephone service is offered by three companies in the main cities, and data and private networks are provided by several new carriers. CTC successfully floated $100 million in new shares in the U.S. market in 1990. Re- flecting improvements and good financial performance, the value of ENTEL and CTC shares roughly doubled in one year. Not all is well, however. CTC and ESTEL still have too much market power, and strong competition in the core telephone business is slow in getting started. SUBTEL is not strong enough nor sufficiently well staffed to serve as an institutional locus for the regulatory system. Excessive involvement in litigation has delayed solutions to regulatory problems. Two court battles have remained unresolved during the past three years. First, CTC'S planned expansion of long-distance facilities is being resisted by ENTEI., and CTC, in turn, opposes ENTEL gaining increased direct access to end users; second, the government is seeking to force Telef6nica to divest either its CTC or its ENTEL holdings in order to avoid possible monopoly consolidation. No practical means have been developed to safeguard social service obligations, especially in rural areas; cross-subsidies are largely ruled out by mandatory cost-based pricing rules, and attempts at direct government subsidy have had limited success. Source: Drawn largely from Melo (1991) and Ramajo (1991). RESTRUCTURING THE SECTOR 325 sets an agenda for the regulator. Restructuring the sector so that strong competitors are present in key market segments from the out- set would generate demand among dominant operators for regulatory action and shift the regulatory burden to the technical issues of interconnection and standards. However, since there are limits to how much competition a particular market can sustain, another possibility would be to divide the regulatory function into discrete tasks, some of which could be subcontracted." The quality and progress of the telecommunications sector will be tied to the level of development of the country as a whole. Arrange- ments that are possible and necessary in the more advanced develop- ing economies, which participate vigorously in competitive global markets, may be neither affordable nor necessary in less competitive ones. In many cases, optimal regulatory arrangements may not be possible, and compromise solutions will be necessary.'2 Continued techtnological change. Rapid technological changes are likely to bring greater opportunities for lower-cost and reliable expan- sion of networks to developing countries, giving them the chance to leapfrog intermediate stages of network development. However, these changes also make designing, procuring, and managing new networks much more complex. For example, wireless technology for personal communication has emerged as a strong challenger to the fixed net- work; the cost of optical fiber systems continues to fall even as capac- ity increases; the new synchronous format for transmission systems permits flexible and inexpensive access to data streams; and faster computer technology is increasing the call-processing capacity of ex- changes significantly. These developments are redefining the optimal network structure and reducing costs.3 Broadband ISDN (Integrated Services Digital Network) has also arrived, permitting integration of video, data, and voice, although standards have yet to be established. The eventual result of this proliferation of new systems may be either a multitude of competing technologies or a limited number of proven technologies that emerge through market dominance and de facto standardization. The growing number of network alternatives has in- creased the importance of strategic planning and the need for revised solutions to designing and planning networks. The challenge is to re- duce overall cost while ensuring the development of an integrated network that is robust, manageable, and consistent with broader de- velopment goals. In this context, the prevailing approach to network planning, which is based on long-term master plans followed by piece- meal projects, is likely to produce a suboptimal outcome. 326 MOBILIZING RESOURCES Technological innovation also raises the question of whether plans based largely on today's dominant technologies are capable of evolv- ing and accommodating major systemic changes. Some innovations will lower and eventually remove entry barriers that currently block new competitors in certain areas. For example, radio technologies such as new analog and digital cellular telephony, personal communi- cation systems, and satellite mobile services, which are all reaching the market, will accelerate the demise of the wired local telephone network as a natural monopoly. Other developments, however, will have precisely the opposite effect. For example, new generations of optical fiber cables are capable of providing ever-cheaper bandwidth at a negligible marginal cost of transmission, reviving arguments for a natural monopoly in parts of the system where they no longer apply. The potential impact of technology on regulatory strategy is espe- cially intriguing. Policymakers should use continuing technological changes to lessen dependence on regulation because these changes will enable a larger portion of telecommunications operations to be submitted to the disciplines of the market. This raises the possibility that good regulation, if and when it is finally in place, may no longer be important. If that were the case, setting up temporary, self- extinguishing arrangements to oversee the initial years following re- form would be preferable to developing self-perpetuating regulatory bureaucracies. Despite further technological changes, however, es- tablished service providers will likely retain considerable market power, which will continue to need regulation. Furthermore, some analysts believe that despite growing competition, successive waves of liberalization will tend to increase the need for regulation in the future. Contvergetice of telecommuniicationis and inforniatics. Telecommuni- cations are becoming an integral part of a much broader family of in- formation technologies and applications. Distinct economic activities are increasingly being structured as different end applications of a common business, namely the organization and management of tele- communications and information resources. Examples include bank- ing, stock trading, broadcasting, publishing, library and information services, and national statistical services, as well as traditional tele- communications and computing. As the supply of telecommunica- tions facilities grows and diversifies, the structural focus shifts from the conventional dichotomy between networks and users to a new, less well-defined sector, which requires organizing and managing facilities that are available from many sources and suit particular business RESTRUCTURING THE SECTOR 327 needs. In this context, the boundaries among businesses become blurred, and, in particular, the strict identity of a telecommunications company is no longer clear.'4 Policymakers are just beginning to look into the implications of these trends for economic development. The effects that at least four areas have on telecommunications policy and regulation require at- tention. First, the structure of the developing telecommunications supply is likely to influence the pace and direction of the overall in- formatization of the economy. An initial step is to explore the possi- bility of organizing the telecommunications sector in a way that allows a wide range of suppliers (including users and companies that resell to others) to develop facilities and services. This approach enhances the potential for innovation but makes it more difficult to establish industry standards and thereby risks slowing the development of criti- cal new services and limiting interconnectivity among user groups. Second, the choice of technologies to modernize and expand the main telecommunications network influences the development of the sector and the applications of information technology in the future. In particular, generalized networks such as ISDN may provide solutions that are cost-effective for certain categories of users but not for oth- ers, delay the development of facilities-based competition, and simply cost much more than other solutions. However, some observers be- lieve that this kind of common framework, once established, would provide a more viable platform for multiple uses than diversified net- work structures, which could be especially valuable in developing countries, and could in the long run eventually reduce costs. Third, the question arises of the role that dominant telecommunications companies play in providing nontraditional services. In particular, some analysts emphasize the economies that can be achieved when these companies enter the market for value added services; for exam- ple, they can add intelligence and processing capabilities to their ex- isting networks and make these facilities accessible to users. Others, however, feel that new services should be implemented by alternative providers, as part of an overall effort to build up an increasingly com- petitive environment. And fourth, the development of advanced tele- communications services is likely to have important externalities. The pace of informatization of an economy will be affected by policy deci- sions as to whether telecommunications companies should merely re- spond to demand or whether they should also play a proactive role in, for example, building up a critical mass of data terminals among small and medium businesses to facilitate the development of infor- mation services. 328 MOBILIZING RESOURCES Closing Remarks: Looking Forward About 80 percent of all developing countries have yet to address structural change in the telecommunications sector. Change is, how- ever, inevitable. Not only has the traditional model based on state monopoly lost favor in an increasingly procompetitive, pro-private en- terprise global environment, but its technological and economic un- derpinnings have also shifted substantially. The question is not whether to reform, but rather how and when. Governments that fail to guide change in an orderly manner risk having crucial policy deci- sions made for them by more assertive players, and such decisions may not be in the country's best long-term interest. These govern- ments are also likely to find it increasingly difficult to sort out the new opportunities and pressures that both domestic and foreign sources are placing on their telecommunications sector. Holding the fort is, in this environment, a lost cause. There is, however, no universal blueprint for reform. Although a wealth of relevant knowledge can be gained from the experience of other countries, particular models cannot be moved readily from one country to another. The timing and pace of reform, the solutions adopted, and the implementation strategies followed are all highly de- pendent on the conditions in a particular country, and they are intri- cately interwoven with each country's broad economic strategy and political processes."5 As more developing countries revise the structure of their sector, more varied solutions are likely to appear. So far, the most ambitious reforms have been driven by high-level government decisions to privatize state enterprises, and the rest-market structure, policy on competition, regulatory development-is left to emerge as best it can. This approach is probably not well suited to all developing countries. An important measure of success will be the ability of future reforms to attract investment capital creatively. To a large extent, the design of sectoral solutions may well revolve around this one issue. Future reforms will also be measured by the extent to which they secure a better balance among the three pillars of reform, namely private par- ticipation, competition, and regulation. Regulatory arrangements that limit reliance on new institutions will become increasingly desirable, and unorthodox solutions are likely to be tried. The interdependence between competition and regulation will be played up to increase reli- ance on markets whenever possible as well as to generate demand for effective regulation. RESTRUCTURING THE SECTOR 329 In closing, it is worth suggesting that pragmatism is a valuable in- gredient in whatever approach is eventually taken. Choices among in- creasingly numerous solutions can and should be made only after assessing realistically their potential contribution to the sector's growth and performance, rather than by citing doctrine. In particular, the decision to allow private ownership of telecommunications opera- tions can be decided for each country based on its access to capital, management, and technology, regardless of whether one believes that private is inherently better than public. Likewise, a policy governing competition should be a means to achieve specific objectives for ser- vice and the network, rather than a mere reflection of the assumption that market forces know best. Telecommunications companies in many developing countries were, after all, initially private and subsequently nationalized to forward practical development goals, 'not just to satisfy nationalistic fervor. And markets are not infallible; they have also given us industrial pollution, crowded airports, and the massive collapse of financial institutions. Notes 1. Two books based on World Bank seminars on telecommunications reform in develop- ing countries (Kuala Lumpur, 1987, and Washington, D.C., 1991) present the main policy issues and options as well as country experiences in preparing and carrying out reforms. See Wellenius and others (1989) and Wellenius and Stern (1991). World Bank (1992) summa- rizes the main features and trends of telecommunications in the developing world as viewed by a number of Bank staff and managers through successive rounds of discussion. This chapter draws heavily on these publications, and some parts are taken verbatim from the last reference. A short study commissioned by the International Finance Corporation gives examples of various modalities of growing private sector participation in telecommunications in develop- ing countries. See Ambrose, Hennemeyer, and Chapon (1990). An advisory group on telecommunications policy, established by the ITU secretary general in 1988 and chaired by the late Poul Hansen, produced an excellent report that was en- dorsed by the ITU'S plenipotentiary conference in 1989. Chapter 5 of that report gives rec- ommendations that apply to most developing countries and a checklist of policy, regulatory, and legal issues that any comprehensive reform process should address. See ITU (1989). A bibliography and the summary findings of two ITU surveys are presented in ITU (1991). 2. Loosely speaking, the information sector comprises all activities that involve the pro- duction, processing, and distribution of information and knowledge, as distinct from physi- cal goods. It includes activities that primarily comprise the handling of information, such as banking and government, as well as the information components of other activities, such as accounting in a factory and management of a farm. The information sector thus includes activities traditionally counted under the primary, secondary, and tertiary sectors. The in- 330 MOBILIZING RESOURCES formation sector has been quantified by a number of researchers in the United States, Eu- rope, and Japan since the 1950s. Data for developing countries are more limired. However, several studies in the Asia and Pacific region in the early 1980s, using data from the late 1970s, give some indication of the information sector as a proportion of CDP: Singapore, 25 percent; Indonesia, 19 percent; Malaysia, 14 percent. 3. Nulty (1991). 4. This figure was originally published in Wellenius (1989). It is shown here as modified in ITU (1991). 5. Wellenius (1990). 6. A checklist of aspects that need to be addressed chrough sector policies, regulation, and legislation is given in ITL (1989), p. 37. 7. It has been proposed that subcontracting can also be an effective tool among different organizational units of the same operating company. See Bruce (1991). 8. The following are examples: international competitive bidding for a ten-year license to provide cellular services in a given region (more than one or two are seldom possible in terms of market size and radio spectrum capacity); competitive supply of subscriber terminal equipment (also called customer premises equipment; for example, telephone sets, private branch exchanges) subject to technical standards and type approval to ensure network com- patibility; and unrestricted competition in the provision of services such as shared data pro- cessing, information, electronic mail, packet-switched data, and store-and-forward facsimile and telex. 9. A growing number of major telecommunications operating companies from industrial countries are entering ventures in developing countries as part of their strategies to globalize their business. Some are already involved in several developing countries, and the trend is likely to continue. However, there are limits to the pace at which these companies can divert internal human and financial resources to foreign ventures, and as the number of opportunities grows, firms can be increasingly selective. 10. For example, sale price can be enhanced by giving the new owners extended monop- oly privileges; however, reducing service costs and promoting responsiveness and innovation require increased competition and regulation. 11. Novel options may include placing regulatory authority with an existing government department but contracting out critical functions. Some (untested) possibilities include re- taining internationally reputable audit firms to monitor compliance with franchise and other obligations (or requiring the main operating companies to retain such firms to report periodically); contracting out recurrent procedures and conflict resolution to local manage- ment or legal consultants, probably with foreign associates for specialized assistance; adopt- ing technical standards and type approvals from another country; and retaining the spectrum management agency of another country to set up a local branch backed by the agency's established norms, practices, processing hardware and software, and expertise. 12. As an extreme example, entrusting monopoly services to an experienced foreign op- erator or investor may quickly remove communication bottlenecks in critical productive sec- tors, even if the absence of competition and effective regulation keeps prices high and compromises the pace of future innovation. In some countries, having to pay high prices for communication, which accounts for only a small proportion of the total costs incurred by most businesses, may be less of a handicap to users than not having the services at all. 13. For example, large switching nodes fed by remote interface electronics close to the customer are becoming more cost-effective than current solutions based on smaller switches collocated with demand centers and connected to customers through cables. Such networks can have several forms, including fiber-to-the-curb and disaggregation of the switch itself. 14. To a large extent, this relates to the convergence of what used to be quite different technologies. For example, with digitalization of telecommunications systems, differences RESTRUCTURING THE SECTOR 331 among voice, data, text, graphics, and video can only be made at the user end, although they all are bit-streams for purposes of processing and transmission. Major telecommunica- tions equipment now essentially consists of specialized computers and software, and its op- eration and maintenance are akin to those of a data processing center; the similarity among network management centers for telecommunications. power, and railway systems is striking. 15. Attempts to interpret events in the telecommunications sector in the context of the political process are starting to appear. One of the first published papers focusing on deve- loping countries is Cowhey (1991). The practical importance of the subject can hardly be overstated, since it is essential to understand, for example, what and when reforms are pos- sible in a particular country, how reforms are likely to mobilize support and opposition, and what welfare distributional effects they would have. Appendixes Appendix A Applications of Telecommunications for the Delivery of Social Services Heather E. Hudson THIS BOOK HAS EXAMINED the role of telecommunications in economic development, with examples of applications for production, trade, and management. Telecommunications can also play an important role in facilitating the delivery of social services, particularly to rural and iso- lated populations. This appendix provides an overview of applications of telecommunications for health care and education, with examples drawn from numerous projects in developing countries. The past quarter century has been marked by dramatic technologi- cal developments in computers and telecommunications and the growing importance of information in all aspects of human life. Ac- cess to information, and to the facilities that produce, store, and transmit information, is now considered vital to development, so that the classifications "information rich" and "information poor" may mean more than distinctions based on GNP or other traditional indica- tors of development.' Reductions in infant mortality as well as in contagious diseases and other preventable health problems are priorities for health care throughout the developing world. Development planners also realize that their populations must become better educated as well as health- ier to participate in an increasingly globalized economy. Yet develop- ing countries have a chronic shortage of teachers and health care workers. Telecommunications and information technologies can help overcome these shortages by providing training and consultation to practitioners as well as instruction to learners, both children in school and adults in the community or on the job. 335 336 APPENDIXES Changing Technologies New communications technologies including video recorders, satel- lite receivers, microcomputers, and facsimile machines are often con- sidered the tools and toys of the industrial world. Nevertheless, advances in communications technology now make it possible to ex- tend reliable communications to any village or encampment, no mat- ter how isolated. Satellite and radio technology can be combined to reach virtually any location, whether in the desert or the jungle or on a remote island. These telecommunications links can transmit voice and data, and often video as well, so that the power of new technolo- gies can be harnessed for development. Examples of these technolo- gies and services are discussed below. Communication satellites. Communication satellites now make it possible to extend basic communication and broadcasting services throughout a country or region, including the most remote islands and villages. Satellites can be used to provide basic telephony as well as radio and television networks with multiple sound tracks for local languages, where appropriate. The advantage of a satellite system is that the earth stations can be installed wherever they are needed, no matter how remote the location. In contrast, terrestrial technologies typically extend out from urban areas as funds are found to build more repeaters or string more wire and cable. Satellite services are considered to be highly robust, since an outage at an earth station af- fects only that location, whereas a cut cable or damaged repeater af- fects the entire network past that point. In addition, satellite earth stations can be located in or near the communities they serve, whereas microwave repeaters are often located in inaccessible loca- tions such as mountain tops, and cable may be buried in the ground or under water. Small low-cost earth stations such as those used for rural telephony with domestic satellites and the VISTA terminals used with INTELSAT sat- ellites can bring voice and data communications to isolated regions. These earth stations may be installed in any community or project site without being connected to the national network by expensive terrestrial links. They may serve the surrounding territory through line-of-site radio links.2 Satellites can also be used to transmit educational programming such as rural development information to villages, instruction to stu- dents from primary to university level, and professional training to TELECOMMUNICATIONS FOR SOCIAL SERVICES 337 teachers, nurses, and other development personnel. Such services need not be one way: teleconferencing via satellite includes simple audio networks that serve as an improved version of the multiparty high-frequency radio services available in many developing regions and enhanced systems that offer computer conferencing, audio- graphics (graphic communication over narrowband channels), freeze- frame video, or motion video using new bandwidth compression techniques. New applications of radio technology. Advances in radio technology such as cellular radio and rural radio subscriber systems offer afford- able means of reaching less-isolated rural customers. These technolo- gies make it possible to serve rural communities without laying cable or stringing copper wire. For example, digital microwave radio links provide a local loop via radio instead of traditional copper wire.3 Cel- lular radio systems may also be used to reach isolated communities without installing wires. In some cases, these technologies may be combined with satellite earth stations, so that the earth station be- comes a hub with radio links to surrounding communities. The appli- cation of solar power to eliminate reliance on diesel generators and community power supplies also makes these technologies increasingly attractive for developing countries. Electronic mail. Communication via computer is a means of ex- changing information immediately. Microcomputer users throughout the world may now interact using various electronic mail networks. Messages may be sent from one computer to another through a host computer equipped with communications and message-processing software including so-called mail boxes for subscribers. These services are cheaper than voice communications and overcome the differences in time zone that hinder real-time communications. Users may dial into local nodes of packet-switched networks to reduce transmission costs. Specialized electronic mail networks have been established for users in developing countries.4 Computer conferencing. Another application of computer commu- nications is computer conferencing, that is, the interaction of many users through a central host computer. Each conference member may share ideas with the others and respond to their comments. Partici- pants may log on at their convenience, thus avoiding the need to ac- commodate different schedules and time zones. 338 APPENDIXES Facsimile. A technology with widespread development applications is the facsimile machine, which enables any type of hard copy, includ- ing print, graphics, and handwritten messages, to be transmitted over a telephone line. Facsimile boards may now be installed in personal computers to allow a message created on the computer to be sent di- rectly to a remote facsimile machine. Audio con ferencing. A thin route service with considerable promise for development applications is audio teleconferencing. Several sites can be linked together through a bridge at a switching point or through a common frequency assigned on a satellite audio channel. This service can be used to hold meetings of, for example, project personnel, to tutor students in rural areas, or to train staff in the field. CD-ROM (Compact Disk, Read-Only Memory). Information in the form of data bases, full texts of journals, video images, and other graphics may now be stored on compact discs and retrieved with a rel- atively inexpensive reader attached to a microcomputer. The advan- tages of CD-ROM include vast storage potential, low cost, durability, and ease of use. In addition, CD-ROM can be used on a stand-alone basis, without the need for on-line access to data bases. Of course, the discs must be frequently updated to keep information current.5 Access to on-line data bases. Although data bases in CD-ROM format are proliferating rapidly, much specialized information may be more readily accessible through on-line data bases. Users with a computer terminal and modem can dial into a data base and search for infor- mation using key words or phrases. Relevant information is then dis- played as a citation, abstract, or sometimes a complete text of documents. Users can then select the relevant information and download it into their own computer or print it out. Data broadcasting. The flow of information within the developing world has been hampered by the cost of distribution and by the lack of access to telecommunications facilities in rural areas. VSATS (very small aperture terminals) now allow information from news services to be disseminated to virtually any location. News services transmit copy by satellite from a hub earth station that may be shared with other data, voice, and video customers. These micro earth stations may be powered using photovoltaics or portable generators.6 TELECOMMUNICATIONS FOR SOCIAL SERVICES 339 VSATS for interactive data communications. Microcomputers or termi- nals linked to mainframes via interactive VSAT technology can be used to collect and update information from the field. A VSAT network called NICNET operated by the Indian government's National Infor- matics Centre now links 160 locations and will be expanded in the next stage to more than 500.7 Similar systems may be used for access- ing centrally stored patient records from any clinic, collecting epi- demiological data, and downloading information from centralized data bases. Desktop publishing. Enhanced graphics capabilities of microcompu- ters now make it possible to produce newsletters and other printed material without typesetting. These features may be particularly valu- able in countries where newspapers, texts, and development materials in local languages are scarce and costly to produce. Development agencies can now produce their own materials in-house. Storefront desktop publishers may also spring up, as they have in the United States, in association with copy shops. This approach enables many small users to share desktop publishing equipment and software. These technologies and services may be combined to enable devel- opment workers to access and share information electronically in the following ways: Electronic messaging. Facsimile transmission and electronic mail may be particularly viable alternatives to sending hard copies of corre- spondence and documents through the mail in areas where service is often slow or unreliable. These technologies can also be used to link staff in the field with one another and with headquarters. Electronic rneetings. Managers, development experts, and project staff may now stay in touch electronically rather than through face-to- face meetings. Audio conferencing allows individuals at several sites to participate in the same meeting, while computer conferencing al- lows group members to interact at their convenience by reading and contributing to discussions stored on a host computer. These elec- tronic meetings do not offer the richness of face-to-face interaction, but they may be an important supplement to meetings where trans- portation costs severely strain limited travel budgets. Access to data bases. Computer terminals or microcomputers with modems linked to the telecommunications network can provide ac- 340 APPENDIXES cess to data bases anywhere in the world. Agricultural researchers, for example, may consult the data bases of the Food and Agriculture Or- ganization of the United Nations in Rome. Health researchers may search the data base of the National Library of Medicine in Bethesda, Maryland. Others may use specialized development data bases such as those for agriculture and energy in India and for development project management in Malaysia.8 Costs may be reduced even further if these searches can be local- ized. Data bases may be downloaded onto computers within the coun- try on tapes, floppy disks, or laser disks, with updates transmitted at regular intervals using telecommunications. For example, the same VSAT technology used to transmit information from new services may also be used to download updates. The search then becomes local and saves the cost of connect time. Dissemination of information. Information may be transmitted from the field and from regional centers to desktop publishing locations via telecommunications networks. For example, development workers and reporters in the field could send reports by facsimile, which would then be edited and published in urban newsletters. Posters and news sheets could be sent by facsimile to rural communities, and newslet- ters could be sent either directly to the communities or to regional centers, where they would be duplicated and dispatched to schools, clinics, and government offices in their territory. Information ob- tained from sources such as news services, data bases, and telecon- ferences could also be disseminated to development workers through- out the country or region via facsimile. Applications for Health Care Delivery Primary health care is being provided to an increasing degree in de- veloping countries by paraprofessionals drawn from the local popula- tion. These paraprofessionals may be community health workers or health aides with a few weeks or months of training, medical ex- tenders or assistants with about one year of training, or practical field nurses. The duties of primary health care workers vary widely by country but generally focus on preventive, promotional, and simple curative care. This includes early diagnosis and treatment of common illnesses, maternal and child health care, midwifery, family planning, treatment of injuries, and referral of patients to higher-level facilities, TELECOMMUNICATIONS FOR SOCIAL SERVICES 341 if available. Primary health care workers may also organize immuniza- tion and mass treatment programs; provide guidance and education on nutrition, family planning, and hygiene; monitor epidemics, water quality, and sanitation; and collect demographic and health infor- mation.9 The second level of health care varies according to the location and complexity of the health care system. It normally refers to services available at larger health centers and small district or regional hospi- tals. Such services may be dispensed by nurses, paramedics, or general practitioners. The third level of care is normally the most sophisti- cated technologically and has the greatest mix of health personnel and services. It usually includes medical specialists based in major hospitals in larger urban centers. In some countries, highly specialized hospitals treating specific health problems, such as leprosy or mental illness, constitute a fourth tier.'° The health sector in developing countries is thus characterized by both breadth and specialization. It is composed of many services that require practitioners with widely varying levels of expertise, such as physicians with specialist training, general practitioners, registered nurses, other nursing personnel, nurses' aides, medical technicians, paramedics, midwives, and various kinds of community health workers or health aides. In most developing countries the shortage of trained medical work- ers is critical. Hence, to make the most efficient use of the limited number of trained personnel, individuals with the most training tend to be concentrated in a few locations. Because resources are generally inadequate, individuals living in towns, villages, and rural areas out- side major urban centers have limited access to organized health care; the same is often true for lower-income inhabitants of urban fringe and squatter areas." Health care in developing countries must be dis- pensed by individuals with less training and less backup than their counterparts in industrial countries. Partly as a result of this and partly as a result of inadequate infrastructure, the health care sector in developing countries encounters major administrative, quality con- trol, and logistical problems. Comrnunications and the Health Care System The benefits of a rapid and reliable system of communication to help with hierarchical, spatial, and logistical problems should be self- evident. The relatively large number of lesser-trained health workers 342 APPENDIXES has a critical need to communicate up the hierarchy of expertise; rapid and reliable communication is necessary for the system to func- tion as a system. With rapid communication, the limited higher-level expertise can be focused more effectively on the highest-priority prob- lems, thus increasing the scope and quality of health services available per trained doctor or nurse. Likewise, lower-level care can be adequate when backup and support are available, and the general population's perception of the quality of that care improves. In Australia, Japan, and the industrial countries of Western Europe and North America, the value of telecommunications in health care has long been understood. Examples include telephone access to all medical personnel, computerized patient records that are accessible on-line, mobile telecommunications links with ambulances and other emergency vehicles, and applications for training and consultation. In developing countries, where resource constraints are more binding, telecommunications are also used to support health care, although at a less capital-intensive level. Nevertheless, the functions fulfilled are much the same. In particular, the application of telecommunications to primary health care services (known in some contexts as telemedi- cine) offers a key to maintaining the system: building morale and maintaining confidence, providing emergency assistance, allowing consultation, facilitating administration and logistics, maintaining su- pervision and quality assurance, and supplying education and train- ing. The following are examples of how telecommunications can support these functions. Emergency Assistance Emergency communications offer the most dramatic evidence of why communications are a critical component of health care delivery because getting help quickly can save lives. Telecommunications play an important role in health services in many developing countries. Studies in Costa Rica, Egypt, India, and Papua New Guinea showed that about 5 percent of rural calls were made for emergencies and medical reasons.'- These calls saved lives and reduced suffering, which are highly significant indirect benefits. During natural disas- ters, such as earthquakes and floods, and epidemics, communication systems are used to secure assistance and to coordinate the logistics of emergency relief. Two-way radios are used in many developing countries to coordi- nate disaster relief activities. Portable satellite stations may be TELECOMMUNICATIONS FOR SOCIAL SERVICES 343 brought in on trucks or small planes to provide reliable emergency communications. Portable satellite terminals were used following the 1980 Mount St. Helens eruptions in the United States and the 1985 earthquake in Mexico City.'3 In the South Pacific, the World Health Organization has used the experimental PEACESAT satellite network to summon medical teams during outbreaks of cholera and dengue fever and to coordinate emergency assistance after typhoons and earth- quakes.14 Mobile services such as cellular telephones can replace regu- lar telephone service when cables are damaged. Cellular phones were used to coordinate disaster relief activities following the 1989 San Francisco earthquake. Emergency location beacons transmitting signals to a satellite en- able rescuers to find ships lost at sea or downed aircraft. Telecommu- nications systems can also help prevent disasters. For example, following several severe cyclones that caused great loss of life in Ban- gladesh in the mid-1970s, the government implemented a cyclone early warning telephone system consisting of single telephone instal- lations in several coastal areas previously without access to tele- communications. Consultation Many developing countries now use paraprofessionals to deliver basic health services, particularly in rural areas. These workers receive basic training in the treatment and prevention of common health problems but need supervision and assistance in diagnosing and treat- ing uncommon diseases and coping with serious health problems. Telecommunications links between village clinics and regional hospi- tals or health centers can be used to provide consultation and supervision. Rapid communication allows paraprofessionals to maintain regular contact with a medical doctor or nurse. More than 100 Alaskan vil- lages are equipped with earth stations that are used for the dedicated medical network, long-distance telephone service, and television re- ception. The medical network is a shared audio conferencing system that includes health aides (who typically have only six weeks of train- ing) and sometimes provides in-service training. Each day during a scheduled "doctor call," a physician at the regional hospital contacts each health aide to discuss current cases, to provide advice on diag- nosis and treatment of patients, and to authorize evacuations if necessary."' The Alaskan satellite network also provides twenty-four- 344 APPENDIXES hour emergency monitoring; by pressing a button on a telephone handset, a village health aide can activate an alarm at the nearest re- gional hospital. 16 In Guyana, rural health workers called "medex" use a two-way radio network to communicate with headquarters in Georgetown to check on the delivery of drugs and supplies and to receive advice on major health problems. They may also request emergency evacuations and follow up on patients referred to the hospital. The medex, who have received about one year of specialized training and are therefore more self-sufficient than their Alaskan counterparts, only use their radios for advice on difficult or complicated cases.'7 The Georgetown train- ing staff offer refresher sessions and "grand rounds" over the radio. At night, chatting over the radio helps medex reduce their sense of isola- tion and boosts morale.'8 Administration and Logistics Another important use of a two-way communication system is to administer rural health services, including ordering drugs, checking on delayed shipments of supplies, transferring patient records, coordi- nating staff travel, and evacuating patients. In Guyana, for example, the dedicated radiotelephone network was used most frequently for administration. Medex with two-way communication facilities received drugs more quickly and kept a more complete supply than medex without them.'9 Also, administrative problems that took weeks or months to resolve by mail or in person were resolved within hours by radiotelephone. Medical services in some countries use small planes to take physi- cians to field sites rather than bring patients to central hospitals. This approach, known as the "Flying Doctor," began in Australia in 1928. People living on remote ranches or stations used two-way, high- frequency radios to contact regional flying doctor bases. Doctors flew in to provide assistance, often piloting the small bush planes them- selves. Fifty years later the "mantle of safety," the area served by the flying doctors, covered more than 80 percent of Australia.20 Today, the flying doctor service is replacing the use of high-frequency radios with telephone service provided either terrestrially (typically via digi- tal microwave links) or via satellite. Similar health communication networks are also found in other parts of the developing world. Flying doctor services in several East African countries (including Kenya, Malawi, and Tanzania) use two- way radio networks to link nurses at rural clinics with headquarters TELECOMMUNICATIONS FOR SOCIAL SERVICES 345 and to coordinate the aircraft used to transport doctors to the clinics and to evacuate seriously ill patients. Two-way radio is a vital support for this service, with radiotelephone communications linking the headquarters with all field hospitals and clinics and with the airplanes themselves through a common frequency installed in the planes' radios.2 Supervision and Quality Assurance The deprofessionalization of health care and the decentralization of services exacerbate the problems of providing a relatively consistent quality of health care service in all locations.22 Hence, supervising medical performance and monitoring the use of services offered be- come critically important.23 Periodic visits by senior personnel are the most effective form of su- pervision, although in many countries difficult terrain or shortages of vehicles, fuel, and expert staff mean that such visits are infrequent. Where available, the telecommunications system has been used to supplement personal contact by scheduling regular periods for tele- phone contact with each health worker to monitor or check on proce- dures and, on occasion, to prompt the health worker about specific functions and duties. Such frequent and interactive contacts with health workers may help maintain the quality of service dispensed by individual workers and reduce regional disparities in quality of service. Another facet of health care administration in developing countries is that the management expertise necessary to organize and operate a complicated system of health services for rural and semiurban popula- tions is in very short supply.24 Rapid means of communication allow the limited number of managers to work more effectively and over a wider area as well as to manage available resources rather than only react to problems as they emerge. Morale and Confidence An important factor in the acceptance of primary health care by rural populations is the knowledge that a doctor or nurse is on call and ready to back up the local health worker's advice to a patient. Rural health aides have reported that they feel more confident in their diagnoses and treatment of patients and that patients are, in turn, more willing to accept treatment that is backed up by a tele- phone consultation with a physician.25 A second important factor involves the morale of field workers dis- 346 APPENDIXES pensing primary health care. The ability to maintain daily contact with health workers in other villages and other sectors of the system reduces feelings of isolation and boosts the sense of security and the morale of health workers. In Guyana, for example, health workers stated that being able to contact one another by radiotelephone greatly improved their morale.26 In Lesotho, rural field nurses were adamant that health staff who were assigned to more isolated regions should be provided a means of communicating with one another.27 Applications of New Technologies In many industrial countries, several new technologies are being used to augment the functions of the basic telecommunications net- work for health service uses. A carryover from two-way radio technol- ogy is the conference circuit, which has been highly useful for in-service, continuing education. In Alaska, the public health service medical network consists of five dedicated audio channels. Four are assigned to the regional service units, so that two hospitals and their associated villages share a channel. The fifth is dedicated for commu- nication between hospitals to ensure privacy at the secondary and ter- tiary levels. The system can be configured for special courses and seminars for all health aides, for all nurses involved in coronary care, or for other uses simply by having the target group select the assigned frequency at the scheduled time.28 A simple, low-cost add-on to individual telephones is the speaker phone; both the patient and the health worker, or groups such as ex- pectant mothers, can listen in and participate in a dialogue. Hard copy can be transmitted (for example, between regional hospitals and administrative headquarters) by low-cost facsimile or by microcom- puters equipped with modems.29 Computerized record-keeping systems are also used by many agen- cies in several countries, including the public health service in Alaska and the southwestern United States, to store and update patient data. These systems allow patient records to be accessed and updated at different locations, which is particularly advantageous both for migra- tory populations who receive treatment at various clinics and for pa- tients transferred within the system. Computerized records can also be sorted and grouped according to any variable so that it is possible to monitor, for example, heart patients with a history of rheumatic fever and to provide up-to-date lists that itinerant health workers can use to treat children requiring vaccinations or women requiring peri- odic checkups. TEILECOMMUNICATIONS FOR SOCIAL SERVICES 347 Another narrowband technology, slowscan video, transmits still pic- tures over a telephone line. Slowscan has been used to transmit X rays and pictures of dermatological lesions as well as for teaching. Tele- phone channels can also be used to transmit electrocardiograms, heart sounds, and ultrasound. Such relatively low-cost add-ons to the voice network are generally most useful for communication between the secondary and tertiary levels of the health care system, for exam- ple, between the general practitioner or nurse at the regional hospital and the specialist at a major medical center. Some use has been made of full-motion video for consultation, but this has not been cost- effective, although lower-cost digital compressed video systems are now available that use a fraction of the bandwidth used by conven- tional analog video systems. Recently, a nonprofit organization called SatelLife, based in Cambridge, Massachusetts, and the former Soviet Union, organized a telecommunications system called Healthnet. Designed to operate in- dependently from existing telephony networks, its goal is to increase the flow of information about recent medical research from industrial countries to developing countries, as well as to provide an electronic mail network that serves subscribers in developing countries who could not communicate otherwise because telephone service is poor or nonexistent. By means of inexpensive radio equipment (based on ham radio technology) transmitting signals to a small, low-orbit, store-and-forward satellite, users can query colleagues, hold confer- ences via electronic mail, order and receive medical literature from sources such as the New En glanid journal of Medicine (available at no cost through a subsidization agreement), and request data base searches at low, subsidized rates. As of 1991, demonstration sites had been set up in Kenya, Tanzania, Uganda, Zambia, and Zimbabwe. Applications for Education and Training Telecommunications and information technologies can also be used to train health care workers, teachers, and other field staff, as well as to educate students in the classroom and adults at home or in the workplace. The vast majority of such applications are found in in- dustrial countries, although some are beginning to appear in develop- ing countries. In the area of vocational technical training, most telecommunications-based programs serve relatively high-level stu- dents, such as engineers interested in keeping abreast of changes in 348 APPENDIXES their field or businesspersons seeking additional training in subjects such as marketing or accounting that can be taught at a distance. In the area of distance education, as in vocational training, telecommu- nications media are nearly always used in combination with other, more traditional media, such as books and papers sent through the mail, audio and video cassettes, and occasional small meetings of stu- dents with a tutor or instructor. Training and Contirnuinrg Education Field staff usually have little opportunity to take refresher courses or discuss new techniques once they return to the field after training. Most health agencies in developing countries provide only limited continuing education through infrequent and costly (in time and travel) meetings or seminars. In a few countries, however, telecommu- nications systems are used to offer continuing education, either di- rectly as a medium for interactive instruction or as support for health workers studying on their own. Audio conferencing systems have been especially useful for instruction. Two-way radiotelephone systems gen- erally have a few channels that are shared by many locations, each using the channel in turn or at scheduled intervals. Although such systems lack privacy, they allow rural health workers at all stations to listen, participate, and learn. Satellite systems can also be used for conferencing by having one audio channel that is shared by all sites. In Peru, the Rural Communication Services Project linked seven rural communities, three via satellite and four via very high-frequency radio and then via satellite, to the national network. More than 650 audio teleconferences concerning agriculture, education, and health were carried out during the project.30 In Alaska, health aides listened to consultations between their colleagues and physicians using an ex- perimental satellite network. This conferencing feature was later in- cluded in a much expanded network.3" In Guyana, a physician using a two-way radio network presented a case in a grand-rounds format once a week and tested health workers on diagnosis and treatment.32 Television programs that offer continuing education to health staff and public health education both to patients in hospitals and to the general public are increasingly available in the United States. Chan- nels offering educational programs and entertainment are received via satellite and distributed via internal cable systems to patients' rooms. Continuing education channels for health care professionals are also distributed nationally via satellite.33 TELECOMMUNICATIONS FOR SOCIAL SERVICES 349 Distanice Education Satellite technology can reach distant learners at their workplace, at school, or at home, no matter how isolated. In China, the Televi- sion University distributes programs via INTELSAT and terrestrial facili- ties to more than I million students at their work sites. Indian educators use INSAT to transmit adult education programs and supple- mental materials for classroom use to villages equipped with low-cost antennas. In North America, the National Technological University transmits technical courses to participating companies so that em- ployees can continue their education without leaving the work site. The Learn Alaska network distributes educational programs to village schools where teachers tape them on video cassette recorders for classroom use.34 Using telecommunications for distance learning may reduce the rate of student dropout and staff turnover at remote locations. The University of the South Pacific operates a satellite-based audio conferencing network linking its main campus in Suva, Fiji, with its agricultural college in Western Samoa and with extension centers in nine Pacific island nations. The system is used to administer exten- sion service activities and courses, offer tutorials for students taking correspondence courses, and extend outreach services to the people of the region through, for example, consultation, in-service training, and seminars offered by the United Nations and other development agencies. The benefits of this experimental network have been signifi- cant. The savings in travel time and costs have been at least ten times the cost of using the network. Dropout rates of correspondence stu- dents in courses with effective satellite tutorials have also been reduced.35 In the Caribbean, the University of the West Indies has established a teleconferencing network called UWIDITE to link its campuses in Barbados, Jamaica, and Trinidad with extension centers throughout the region.36 It uses a combination of satellite and terrestrial audio links. Pilot Projects to Encourage Applications of Satellites During the 1980s two ambitious initiatives were undertaken to en- courage the use of satellite technology for development. The Rural Satellite Program, sponsored by the U.S. Agency for International 350 APPENDIXES Development (AID), was designed to help countries use satellites to help solve development problems. Pilot projects were carried out in Indonesia, Peru, and the West Indies. INTELSAT'S Project SHARE pro- vided free access to INTELSAT satellites for health and education dem- onstrations and projects. AID Rural Satellite Program By the late 1970s, AID had identified major obstacles to using satel- lites for development. These included the small number of visible and ongoing projects that could serve as models, the lack of planning ca- pacity and relevant experience in developing countries with access to satellites, and the lack of awareness among development agencies and telecommunications lenders of the role satellites could play in rural development.37 In 1979, AID undertook an exploratory program to help the developing world test the use of satellite communications for development. The experimental aspect of the program was not in the technology, although certain hardware innovations were introduced; rather, it was in the applications of that technology for development. The goal of the program was to help developing countries that al- ready had access to a satellite to use it for development rather than plan new satellite capacity more appropriately designed for develop- ment use. The program concentrated on interactive narrowband tele- communications rather than on broadcast community television because of the experience with audio conferencing in Alaska and the South Pacific, the low cost of narrowband equipment, and the simple production techniques required. In Indonesia, AID supported a project that used Indonesia's PALAPA satellite to enhance higher education by linking thirteen new universi- ties in the Indonesian archipelago. These Eastern Islands Universities suffered from a severe shortage of specialized faculty, particularly in basic sciences and agriculture. The satellite audio conferencing system enabled a professor at one institution to teach students at several lo- cations. The network was also used to train faculty, handle adminis- tration, and enable administrators to meet electronically between infrequent face-to-face meetings. Sites were also equipped with fac- simile machines and electronic blackboards that could transmit mate- rial written or drawn on graphic tablets. At most locations, the sites accessed existing Perumtel earth stations through telephone lines. The Peru Rural Communications Services Project was developed and administered by ENTEL Peru (Empresa Nacional de Telecomunica- TELECOMMUNICATIONS FOR SOCIAL SERVICES 351 ciones del Peru) with support from AID. The goal of the project was to use satellite communications via INTELSAT to provide basic telephone service and teleconferencing to support development activities in an isolated region of Peru. The project provided public telephone service and audio conferencing facilities to seven communities in the Depart- ment of San Martin, a high-altitude jungle area east of the Andes. Satellite earth stations were installed in three communities, and four smaller towns were linked via very high-frequency radio to one of the earth stations. The teleconferencing activities were developed in cooperation with Peruvian agriculture, health, and education ministries and incorpo- rated a wide variety of administrative, training, diffusion, and promo- tional strategies. A total of 658 audio teleconferences were sponsored by the ministries and ENTEL during 1984 and 1985, involving almost 12,000 participant hours.38 The health sector was the most active and successful user, perhaps because it had a stable organizational struc- ture and a pressing need for communication to support its operations. ENTEL itself became a major user of teleconferencing for training pur- poses. At the end of the two-year period, ENTEL transferred responsi- bility for developing teleconferencing to its commercial sector, with plans to promote teleconferencing among government agencies and private business. The third satellite project supported by the AID Rural Satellite Pro- gram was the University of the West Indies Distance Teaching Experi- ment (UWIDITE). This activity built on the experience gained in a previous experiment called Project Satellite, which used NASA'S ATs-3 and ATs-6 satellites to link the Jamaica and Barbados campuses with the St. Lucia extension center in 1978. UWIDITE began with five dedi- cated teleconferencing rooms, one at each of the main campuses (Barbados, Jamaica, and Trinidad) and one at each of two extension centers (Dominica and St. Lucia). Each room was equipped with audio conferencing equipment and slowscan television; telewriters and microcomputers were added.39 The main applications of the network were for extension studies (courses for credit toward a university degree), extramural studies (special nondegree courses), and extension services to support agri- cultural development and information distribution. A typical weekly schedule included in-service classes for teachers, meetings of the UWIDITE coordinators, class sessions for Challenge Examinations (which allow students to take their first year of university in their home countries), continuing medical education classes, and medical consultations.40 352 APPENDIXES INTELSAT announced Project SHARE (Satellites for Health and Rural Education) in August 1964 as part of its twentieth anniversary cele- bration. INTELSAT offered to provide free satellite time for health and education purposes; users had to provide their own earth station facil- ities and support for the projects. Project SHARE was initially limited to sixteen months and was twice extended through the end of 1987. Its activities involved sixty-five countries on five continents. Project SHARE applications were divided almost equally between video events or video conferences that were one-time or limited events and those that were recurring and longer term. Of the twenty com- pleted projects summarized in the final evaluation report, eleven in- volved video conferences and four used audio conferencing. Most of the video conferences were one-time events, whereas the audio con- ferences were generally ongoing. Three projects involved computer communications. One project, China's Television University, used television for distance education. Several projects, including those by the University of the South Pa- cific, China's Television University, the World Health Organization, and the Pan African News Agency, found satellite to be a cost- effective means of meeting their goals. The University of the South Pacific, based in Suva, Fiji, took advantage of Project SHARE as a tran- sitional step to achieving operational service using the INTELSAT Pacific Ocean satellite and existing INTELSAT earth stations in eight countries. The network is used for tutorials and administration and now pro- vides facsimile, data, and slowscan video in addition to audio teleconferencing.4' China used Project SHARE as the first step in implementing a na- tional Television University. The goal was to deliver university instruc- tion to students at the workplace, because the shortage of places for qualified students was critical. Courses were initially transmitted over the terrestrial network used for broadcast television, but much of the country was not covered by this system. Using INTELSAT, educational programs were transmitted to fifty-three locations. Following the pilot project, China purchased two transponders and leased a third from INTELSAT. By mid-1988, some 5,000 TVRRO terminals built in China had been installed, and the Television University had an estimated stu- dent body of more than I million. INTELSAT's board of governors has authorized a follow-up to Project SHARE known as Project Access. More market driven, Project Access is designed to stimulate service to rural and remote areas while empha- sizing the potential for commercial applications. It also provides free TELECOMMUNICATIONS FOR SOCIAL SERVICES 353 use of spare space segment capacity for educational, health, or closely related social services. Project Access is intended for areas where communications are limited or where access to new communications can be developed.42 Applications for Education in Industrial Countries During the 1980s, industrial countries increasingly turned to tele- communications and information technologies for education and training. Their use of satellites for distance education provides valu- able models for developing countries, as does their use of media for community education and development. The following examples are particularly applicable to developing countries that are attempting to increase access to education and practical training in order to enhance the economic opportunities and well-being of their popu- lations. Satellites for Distance Education Canada, the United States, and Australia have pioneered the use of satellites for making education available, accessible, and affordable for populations living in remote or isolated communities. Canada. Canada has used its own domestic satellite system since 1973 to deliver broadcasting and telephony services to its vast iso- lated northern areas. In the mid-1970s, it undertook a series of exper- iments on the Communications Technology Satellite or Hermes, a joint project of the Canadian Department of Communications and the United States' NASA. Several of these experiments were continued on Canada's own ANIK B satellite, and the most successful are now op- erational services. For example, TV Ontario uses an ANIK satellite to deliver educational programs previously available only in the most densely populated areas to remote communities throughout Ontario and directly to schools, libraries, cable television systems, master an- tenna systems, and low-power television transmitters. In British Columbia, the Knowledge Network delivers educational programs for credit to isolated communities, where they are retransmitted over cable systems or viewed at community centers and colleges. An audio conferencing link enables students to interact with instructors located in the VancouverNVictoria area.43 354 APPENDIXES United States. In the United States, educational television pro- grams are delivered to schools and workplaces via satellite. In rural areas, small high schools cannot offer a wide array of vocational and college preparatory classes. A satellite network called TI-IN, based in Texas, offers specialized courses in foreign languages, mathematics, and science, enrichment courses, and staff development programs. Participants interact with teachers using toll-free telephone lines. At the university level, the National Technological University offers grad- uate courses in engineering and computer science via satellite to em- ployees at their workplaces throughout the country. Employers support the program because it enables their professional staff to im- prove their knowledge without having to attend graduate programs at distant universities. Continuing education programs are also distributed to homes via satellite. One example is the Learning Channel, an outgrowth of the Appalachian Community Services Project begun more than a decade ago to provide educational opportunities to small communities in the region's isolated hill country. Now the Learning Channel dissemi- nates adult education channels via satellite to local cable television systems throughout the United States. Australia. The School of the Air originated in Australia to supple- ment the correspondence courses sent to children in the outback. Using two-way radios to contact students on remote farms or ranches, the School of the Air offered some instruction and enrichment but, most important, provided a link with other children and "real teach- ers." Distance education has expanded in Australia with the creation of AUSSAT, the domestic satellite system. A pilot project demonstrated that educational programs could be delivered to rural and remote areas via satellite and received using small satellite earth stations. Sev- eral states now operate distance education networks to reach both children and adults. Applications for Education in the Community Industrial countries have developed applications of telecommunica- tions and information technology at the community level that may be adapted to developing countries. Two examples discussed here are telecottages in Scandinavia and community radio in Canada. Telecottages. Telecottages are rural sites equipped with computer and telecommunications facilities that are used by community resi- TELECOMMUNICATIONS FOR SOCIAL SERVICES 355 dents. The official name of telecottages is Community Teleservice Centres (CTSCS). Their aim is to provide small communities with ac- cess to telecommunications, data processing, and computer-assisted services. A telecottage may be defined as "a centre where IT lin- formation technology] apparatus is placed at the disposal of the citi- zens of a specific local community within a marginal geographical location, so that communal use may be made of the facilities avail- able. The purpose of the CTSC is to counteract some geographically determined disadvantages which affect the local community, whether they are of an economic, educational, or cultural nature, or whethet those disadvantages concern employment, services, or other infra- structure facilities."44 Approximately twenty-five telecottages have been established in Denmark, Finland, Norway, and Sweden. Telecottages form part of a strategy of economic diversification that aims to reduce dependency on agriculture in some areas and on resource industries such as for- estry and mining in others. Attraction of small and medium-size busi- nesses may be hampered by geographical and sociocultural distance from markets and decision centers, lack of easy access to information, lack of appropriate services including training programs, and other factors. Locally available and affordable access to telecommunications and information technologies is seen as a means of overcoming these barriers. Telecottages are located in schools, libraries, local government buildings, and other converted or shared buildings such as coffee shops. They may contain an office, classroom, meeting room, and work facilities as well as a small kitchen. Equipment in a telecottage may include personal computers, printers, telephones, a facsimile ma- chine, and possibly video equipment. Telecottage services could in- clude word processing and desktop publishing, bookkeeping for small businesses, and training and continuing education both in the use of the technologies and in other topics available on computer disks or videotapes or via teleconferencing. An important feature of the telecottage is the facilitator, who takes care of the equipment but also serves as a computer consultant, trainer, and development worker. This role is similar to that of the animator in Canada. Communit-y radio in Canada. Canada's native peoples are scat- tered in hundreds of villages across the vast northern regions. Most villages are isolated from urban centers. In the far north, the only access is by boat or plane. In these villages, people generally speak 356 APPENDIXES their own languages, although children are now educated in English or French. Community radio began in northern Canada in about 1970, as an experiment in giving native people the opportunity to communicate in their own languages. The Canadian Department of Communica- tions funded a pilot project in the early 1970s using a community de- velopment approach. The Canadian Broadcasting Corporation sup- ported community radio by facilitating access to its own radio transmitters. Pilot projects with federal and provincial support ex- plored a range of community media models, including local access, television transmitters, portable video, and community radio. A strong influence was the National Film Board of Canada's Challenge for Change Program, which used film as a participatory community development tool. The French concept of animator was applied to community radio so that residents learned how to use the medium to inform the community but also to collect information and opinions and thus give the community a more powerful voice in its own development. Community radio stations in Canada are nonprofit, controlled by the community, and responsible for programming and production. Some produce all their own programming, typically a few hours a day in the morning, at noon, and in the evening. Others also broadcast network programs from the Canadian Broadcasting Corporation and other networks. Community radio stations are found across the coun- try, particularly in small towns and remote communities. In 1983, the federal government recognized the importance of na- tive broadcasting in its Northern Broadcasting Policy, which estab- lished the Northern Native Broadcast Access Programme. This program has provided more than Can$40 million to enable thirteen communications societies to establish production facilities, train broadcasters, and produce and distribute programming in native lan- guages and oriented toward native themes and concerns. Many of the organizations distribute their programs via satellite. Native broadcasters have combined the local and development- oriented approach of community radio with the distribution power of a satellite so that programs produced by native Canadians can be dis- tributed via satellite to community radio stations that retransmit them to their own communities. In some regions, native organizations also produce native language television programs that are carried via satellite to supplement the national network programming distributed from southern cities. The Wawatay Television Network produces pro- grams in Cree and Ojibway Indian languages that are transmitted to TELECOMMUNICATIONS FOR SOCIAL SERVICES 357 thirty remote communities in northern Ontario on the ANIK C satel- lite, using an audio subcarrier of rv Ontario, the provincial govern- ment's education channel. The Inuit Broadcasting Corporation now broadcasts five hours of Inukititut (Eskimo language) programming each week using time on the Northern Television Service channels of the Canadian Broadcasting Corporation. Programs are uplinked from satellite stations in the Northwest Territories and Arctic Quebec. The Need for Integrated Planning The examples in this appendix demonstrate that telecommunica- tions technologies can be used to improve delivery of health care, ed- ucation, and other social services. For these applications to occur, however, telecommunications planners must be educated about them and social service planners and practitioners must be made aware of the potential of the technology. In the glitter of all this new technol- ogy, we must not lose sight of the lessons learned from decades of work with instructional media. As Wilbur Schramm noted fifteen years ago, motivated students can learn from any media-whether large or small-including satellite-delivered television, audio cassettes, and workbooks.45 Educators need to work with technologists to deter- mine which mix of technologies is most appropriate for a specific application. Demonstrations and pilot projects can be effective means of testing new technologies and extending awareness. Careful planning is re- quired, however, to move beyond the experimental stage. Again, both telecommunications and social service professionals must be involved throughout the process. As the evaluators of the rural satellite project in Peru pointed out, "the transfer of sophisticated telecommunica- tions technology is not, and probably will not be for the foreseeable future, a straightforward exercise. Even the most thorough plans re- quire revision, endless patience, and dedication if they are to be im- plemented successfully. New and innovative communication strategies require extensive promotion as well as innovative management struc- tures, especially when changes in standard operating procedures are involved."46 In many cases, the most difficult hurdle for telecommunications applications in the social sectors is their high cost. Particularly in the area of satellite applications, tariffs are commonly much higher than nonprofit organizations can afford to pay on an ongoing basis, even if the necessary equipment can be obtained from outside donors. For 358 APPENDIXES example, the Learn Alaska network, which was set up by the suddenly oil-rich state government of Alaska when the price of oil was high, suffered significant cutbacks as the price of oil fell. In fact, most so- cially oriented satellite projects depend on special tariff subsidies, and when subsidies are terminated, the projects often cannot continue without the full support of the government, as is the case with the Chinese Television University. Even where simple voice telephone ap- plications are called for, social service entities in some developing countries find it difficult to pay related costs. For example, in Guate- mala, the Ministry of Health does not have sufficient funds to pay the monthly telephone bill of its headquarters in Guatemala City. Yet the rewards are well worth the effort. Telecommunications ad- ministrations and national planning agencies must remember that the telecommunications network is more than a financial asset and a source of revenue. It is a vital strategic resource for development. By meeting the challenges posed by new and converging technologies, developing countries can maximize the benefits of telecommunica- tions as a strategic resource for national and regional development. Notes 1. ITU (1984) and Hills (1990). 2. Hudson (1990). 3. Parker and others (1989). 4. International Development Research Centre (1989). 5. National Research Council (1990). 6. Parker (1987). 7. Blair (1988). 8. Stover (1984). 9. World Bank (1980), p. 57, and Evans, Hall, and Warford (1981). 10. In Latin America, instead of resembling a pyramid, the structure of health staffing 'resembles a sand clock with a slip neck of technicians and two receptacles representing a cadre of auxiliaries and a cadre of professionals." See Golladay (1980), p. 10. 11. Dorozynski (1975) and Golladay (1980), pp. 5-11. 12. Hudson (1984). 13. Goldschmidt, Hudson, and Lynn (1980). 14. Hudson and others (1979). 15. Hudson and Parker (1973). 16. When feasible, arrangements can be made to evacuate critically ill patients. In such situations, it is not clear whether rapid two.way communication results in an overall in- crease or decrease in hospital admissions. Some difficult and expensive evacuations may not be necessary, whereas others, which might be marginal, may be easier to arrange. See Hudson and Parker, (1973), p. 7. 17. Goldschmidt, Forsythe, and Hudson (1980). 18. Hudson (1984). 19. Hudson (1984). TELECOMMUNICATIONS FOR SOCIAL SERVICES 359 20. Page (1977). 21. African Medical and Research Foundation (1979). 22. Rockoff (1975). 23. Golladay (1980), p. 41. 24. Evans, Hall, and Warford (1981), p. 1126. 25. Hudson and Parker (1973). Information conveyed directly to the village concerning patients who are away for medical treatment helps allay the fears of family and friends. 26. Goldschmidt, Forsythe, and Hudson (1980). 27. African Medical and Research Foundation (1979). 28. Hudson (1990). 29. Hudson (1980). 30. Mayo and others (1987). 31. Hudson and Parker (1973). 32. Goldschmidt, Forsythe, and Hudson (1980). 33. Hudson (1990). 34. Hudson (1990). For more in:depth discussion of technology in distance education, see Tiene and Futagami (1987); Perraton (1986); Jenkins (1988); and Lockheed, Middleton, and Nettleton (1991). 35. Pierce and Jequier (1983). 36. Block (1985). 37. Block and others (1984). 38. Mayo and others (1987), p. iii. 39. Block and others (1984), p. 29. 40. Block and others (1984), p. 29. 41. INTELSAT (1988b). 42. INTELSAT (1988a), p. 29. 43. Hudson (1990). 44. Qvortup (1989). 45. Wilbur Schramm (1976). 46. Mayo and others (1987), p. iv. Appendix B Posts and Telecommunications as an Integrated Communications Sector Douglas Goldschmidt IN THE MAIN TEXT of this book the postal system was mentioned several times: chapter 3 discussed some of the advantages of organizational and management autonomy for postal and telecommunications ad- ministrations, chapter 5 reviewed joint postal and telecommunications consumption patterns by sector for several countries using input- output analysis, and chapters 11 and 12 outlined survey results on the substitutability of postal services for telecommunications services. Traditionally, the postal and telecommunications sectors have for many purposes been viewed as one, both organizationally and analyti- cally. The primary reason for this, in spite of their vastly different re- quirements of capital and labor inputs, is that both services facilitate the same general function: two-way communication on a spatial basis. Hence, in most countries telecommunications historically have been treated as an extension of the publicly owned postal monopoly.' Both services have also traditionally been tied closely to transport infrastructure and facilities, and, as discussed in chapter 7, transport acts as both a complement to and a substitute for other communica- tions services. In spite of this, the tendency has been to treat the transport sector as separate from the communication sector, particu- larly as modern communication has expanded beyond the physical movement of objects. Nevertheless, the functions performed by the transport, postal, and telecommunications sectors are increasingly merging and overlapping, and the interaction of these sectors may not 360 AN INTEGRATED COMMUNICATIONS SECTOR 361 be efficiently facilitated in the future within the traditional institu- tional and regulatory framework that exists in many countries. Substitutability of Telecommunications and Postal Communication The telephone user surveys presented in chapters 11 and 12 gave ample evidence of the substitutability of telephones and posts, and also travel. In Thailand a sample of residential telephone subscribers ranked writing a letter as the third most preferred means by which to set up appointments and communicate with kin and friends. The tele- phone ranked first, and a personal meeting or visit ranked second. Likewise, in both Costa Rica and Papua New Guinea, when rural public call office (PCo) users were asked how they would communicate if the PCO were not available, writing a letter was the first alternative in Costa Rica and the second (behind travel) in Papua New Guinea.2 For a sample of telephone subscribers in two rural regions of Kenya, the telephone and postal service together accounted for almost 80 percent of all economic contacts that subscribers made; mainly tele- phone or only telephone accounted for about 51 percent of the con- tacts, mainly letter or only letter accounted for about 11 percent, and an equal combination of telephone and letter accounted for 18 percent. The respondents in these surveys might not be typical of the gen- eral population, however, since they were by definition telephone sub- scribers or Pco users. The only information reviewed on postal use from a more general sample population in a developing country came from a survey of nine rural villages in Egypt. There, of almost 500 male inhabitants polled, about 20 percent had sent or received a let- ter during the preceding year compared with about 10 percent who had talked on a telephone.3 In the past two decades, in most developing and industrial coun- tries, telephone use grew at a faster rate than postal communication. For example, figure B- 1 presents the situation in the United King- dom, where the number of letters sent each year has declined since 1970, while the volume of telephone calls has increased dramatically. Such changes do not result entirely from the substitution of tele- phone calls for letter writing. Although part of the change reflects a substitution of telephone calls for letters, the rapid increase in calls is probably more closely linked to the major expansion of telephone access in Great Britain.4 The overall demand for telecommunications 362 APPENDIXES Figure B-i. Inland Letter and Telephone Traffic in the United Kingdom, 1950-76 Number per year (millions) 20,000 - 15,000 - Letters __~- 10,000 , /Telephone calls 5,000 - 0 I 11111111 1 1 1 1950 1955 1960 1965 1970 1976 Source: Smith (1978). Reproduced with permission of the Controller of Her Britannic Majesty's Stationery Office. also increased as quality of service improved and as new services, particularly relating to data and message communications, were introduced. Postal and telecommunications services are not perfect substitutes even for the transfer of information. Although information is moved spatially by both means, the postal service is not as efficient in facili- tating two-way interaction.5 For certain types of messages, speed is highly valued. For example, in the Papua New Guinea and both of the Indian studies reviewed in chapter 12, more than 75 percent of all calls made by the Pco users sampled were perceived as being necessary immediately or as having to take place during the day on which the telephone calls were made. AN INTEGRATED COMMUNICATIONS SECTOR 363 Electronic Message Transfer Several chapters of this book cited instances in which postal and telecommunications functions merged. In the study of villages in rural Egypt, village administrators used telephones primarily to dictate mes- sages, which were then hand carried to the intended recipient. Also, the general public sometimes used telephones to transmit telegraph messages, which were then written down and delivered; in Egypt such messages are called phonograms. Likewise, in Myanmar several pro- vincial public telephone offices were closed each afternoon while (telegram) messages were dictated over the telephone lines for tran- scription and hand delivery at the other end; similar telegraphy systems have been observed in parts of the Philippines and in other developing countries. These, of course, are rudimentary examples of the ultimate convergence of the information transfer function of posts and telecommunications: electronic mail. Electronic mail systems are being introduced throughout the indus- trial and developing world. These systems are similar to a traditional postal system in that documents are sorted and transmitted to spe- cific addresses. The difference is that the documents (messages) are transmitted electronically over telecommunications networks rather than through transport networks, and they need not be printed on paper. One form of electronic mail is the transmission of messages using telecommunication/computer interfaces. These systems use a variety of computers ranging from inexpensive home or small business micro- computers to the largest mainframe systems. Essentially, the elec- tronic mail system is a software package that provides computer sorting, mailbox (storage), and transmission functions and an inter- connect, which is accessible from as many other computers or termi- nals as desired.6 Such small (or sometimes extensive), mostly privately owned and operated computer-telecommunications systems are be- coming common in North America and parts of Western Europe within commercial enterprises of all types and sizes, as well as among universities and research institutions.7 In the public sector, the United States Postal Service, among other postal carriers in industrial countries, has begun to implement a na- tional electronic message transfer system that integrates conventional systems of postal collection and distribution with computer, facsimile, and telecommunications facilities.8 Increasingly, large-volume users will be able to send tapes or magnetic cards to local postal centers where the messages will be read and routed electronically; others will 364 APPENDIXES send bundles of nonenveloped message pages, which will then be fac- simile encoded and transmitted. Ultimately, individuals with hard copy messages may be able to access the system through coin- operated facsimile machines (electronic mailboxes), which will be lo- cated in post offices or other public places (large stores, shopping centers, bus and railway terminals, and so forth). Individuals with ter- minals in their home or small business may be able to submit mes- sages directly over local telecommunications links.9 As a first step, the U.S. Postal Service inaugurated in January 1981 regular international electronic postal service (Intelpost) between the United States and London, and other Western European countries will be added to the system. In January 1982 an Electronic Computer Oriented Mail ser- vice (E-COM) was inaugurated that allows large commercial mailers to send computer-generated messages to twenty-five specially equipped postal facilities, where the messages are printed, stuffed into enve- lopes, and delivered as regular first-class mail.', Given the rapid advancement of technology, in the foreseeable fu- ture, a major function of the current postal system in all countries- the transmission of bills, invoices, and funds-may be largely replaced by electronic banking services using telecommunication/computer in- terfaces. Not only will funds be automatically debited from accounts and detailed accounts maintained for many transactions, but custom- ers will be able to use simple data entry devices, such as push-button telephone sets, for banking functions. In fact, if falling data processing and transmission costs bring the cost of an electronic message below the cost of a first-class letter, electronic mail could substantially displace the manually carried first- class letter.'' In particular, electronic message transfer could replace the first-class mail generated by business and industry. The system could transmit market information, product specifications, technical details, purchase orders, contract copies, and so forth. This could di- rectly reduce market response and purchase lead times, which would permit companies to reduce inventory levels, while providing an equal or improved level of customer service. ' If electronic mail systems are widely adopted, major savings in labor, paper, vehicles, and energy could also be realized.'3 The Postal System in Developing Countries It is likely that both public and privately owned electronic mail sys- tems of varying degrees of complexity will soon be relatively wide- AN INTEGRATED COMMUNICATIONS SECTOR 365 spread in many developing countries. Already telecopiers, data terminals, teletvpewriters, facsimile machines, and so forth are found in the larger cities of many developing nations. The widespread devel- opment of domestic computer/communications networks is now under way; in fact, packet-switching systems being planned or imple- mented by many telecommunications authorities in developing coun- tries will directly facilitate the development of electronic message systems. However, such systems are mostly for communications within and among the larger urban and commercial areas.14 In the absence of reliable telephone systems in many smaller towns and rural places, it is unlikely that electronic mail will be a significant factor in less ur- banized places during the next decade. The absence of modern electronic mail services in rural areas will not, however, be a roadblock to development. In rural areas, where transport and other communication services are often highly inade- quate, the traditional post office can continue to perform several functions, including the transport and movement of bulky items. In fact, some of the characteristics that make the postal service dif- ficult to manage-the need for an elaborate transport system and the large volume of labor required-offer a special opportunity. If properly managed, the postal systems in developing countries can provide a valuable scheduled movement of correspondence that does not re- quire immediate response as well as of newspapers, supplies, freight, and even people. The postal system can also be an attractive vehicle for implementing and supporting a variety of rural development pro- grams. Personnel operating a postal system have in some countries provided banking services in areas without banks, disseminated infor- mation on behalf of government and other public and private enter- prises, maintained PCOS, and provided various social services by virtue of their physical placement in small towns and villages. No other communications media or civil service maintains such an extensive pool of labor in remote locations. The following sections cite brief ex- amples of the postal system's rural potential to support agriculture, health, and nutrition extension activities and business; education and literacy; transport systems; rural banking; and the generation of employment. Rural Agriculture, Extension Activities, and Commerce The need to disseminate information remains crucial to rural devel- opment; without adequate systems to facilitate the movement of in- 366 APPENDIXES formation, maintaining many of the traditional development activities is difficult and costly.'5 In the absence of a reliable postal service and at least minimal telephone access, rural development or extension agencies and private commercial enterprises tend either to use courier services or to operate their own relatively high-cost transport and communication systems. Agriculture The growth in recent years of agricultural extension programs in developing countries has increased the need to disseminate various types of printed information and administrative correspondence regu- larly to small towns, villages, and rural areas. Historically, this was also the case in industrial countries. For example, in the United States the postal system had a significant effect on the agricultural sector in the early 1900s. One poll taken in 1912 indicated that farm- ers were very responsive to published agricultural information. Of over 3,500 farmers polled, 40 percent stated that agricultural periodicals delivered through the rural free delivery postal service were most help- ful. An extension agent with the U.S. Department of Agriculture noted at the time that, "So far as this survey is an indication, the ag- ricultural press would seem to be the most efficient of our agricultural extension agencies in reaching the farmer.'16 Health and Nutritioni The existence of postal delivery to rural towns can also help dis- seminate information on health, family planning, and nutrition pro- grams. A letter carrier who often handles few letters and parcels, but who travels long distances, could easily distribute posters, pamphlets, drugs, and medical supplies to a local health worker as well as other materials to government agencies at very low marginal cost. In effect, the size of a government's rural health extension program could in some instances be increased by making the postal service an inte- grated part of the effort.'7 The fact that postal carriers are generally not used in such a way and are in many instances underemployed probably reflects bureaucratic structures that resist interagency coop- eration, a general disinterest in or ignorance of the possibilities of- fered by the post, or a distrust of the postal service resulting from experiences with traditionally poor or nonexistent delivery services in rural areas and small towns. AN INTEGRATED COMMUNICATIONS SECTOR 367 Commerce and Trade As with the telephone, many commercial or business-related bene- fits of the post are so obvious as to be transparent. Presumably, the ordering of supplies, remittance of bills and payments, and shipment of parcels and produce can all be facilitated by a reliable postal ser- vice. The post also provides a conduit of information for rural dwell- ers, which can be essential for the conduct of trade and business. In 1900 the postmaster general of the United States noted, with regard to the then experimental rural free delivery service, that Rural delivery has now been sufficiently tried to measure its effects. The immediate results are clearly apparent. It stimulates social and business correspondence, and so swells the post receipts. Its intro- duction is invariably followed by a large increase in the circulation of the press and of periodical literature. The farm is thus brought into direct daily contact with the currents and movements of the business world. A more accurate knowledge of ruling markets and varying prices is diffused, and the producer, with his quicker com- munication and larger information, is placed on a surer footing. The value of farms, as has been shown in many cases, is enhanced. Good roads become indispensable, and their improvement is the essential condition of the service. The material and measurable benefits are signal and unmistakable."8 More recently, a retrospective study of rural America documented that the introduction of rural free delivery service affected the way in which farmers conducted business and ultimately led them to employ more precise and careful business practices as the level of business and finan- cial transactions they conducted through the post increased.19 A Caveat The postal system, nevertheless, should not be given too much credit; the problems of disentangling the joint contributions of di- verse inputs are immense. For example, the value of farms at the turn of the century in the United States likely increased at least as much from the improvement in the road system as from the introduction of the rural postal delivery system. Also, the introduction of the parcel post in the United States appears to have had a somewhat negative effect on stores in small towns. The introduction of the parcel post al- lowed large mail-order companies to market a diverse selection of 368 APPENDIXES merchandise at prices sometimes lower than those charged bv small local stores. Although many local stores attempted to compete by di- versifying their merchandise, the overall effect of rural parcel post seems to have been to injure local retailing.20 Such competition was probably beneficial for the country as a whole, but it would be diffi- cult to determine the net effects on a particular village or small town of losing its store or shop to such competition. In any case, comparing turn-of-the-century North America with the current situation in many developing countries is partly invalid because so many changes have taken place in communication tech- nologies. Today there is a much wider variety of communications media for conveying market and other information to rural areas. Hence, attempts to facilitate the flow of information among rural areas, markets, and urban centers in developing countries now require both users and suppliers to make many more decisions about the cost and effectiveness of posts, telecommunications, transport, broadcast- ing, and various other communication media, as well as about the value of time. Education and Literacy In the United States in the early 1900s, as in many other countries, postal deliveries were viewed as one way to promote an informed and literate electorate.2' Early accounts of the rural free delivery system emphasized that the post improved the level of education of both adults and children.22 This was partly reflected in the large increase in the number of newspapers delivered to rural areas after the introduc- tion of the system. By 1911 more than I billion newspapers and mag- azines a year wvere delivered on rural routes-a volume greater than that of all other rural mail combined. By 1919 this number had doubled.23 In developing countries, the postal system has also helped promote literacy through the transmission of information among correspon- dents and, more important, through newspapers, periodicals, and some books. Literacy campaigns that do not include a variety of printed literature have been difficult to sustain. It has been argued that the audiovisual media sometimes seen in rural development proj- ects, although useful, cannot provide the diversity or depth of infor- mation provided by the print media, and, even more important, printed material, through its convenience, gives learners more control AN INTEGRATED COMMUNICATIONS SECTOR 369 over their learning; they are able to work at their own pace and to choose when and where to inform themselves.24 An efficiently functioning postal system can also facilitate the im- plementation of correspondence-type education programs. In develop- ing countries, in particular, the demand for school places generally exceeds the capacity of national economies to supply them, and this has led to a search for alternative education methods that reach larger numbers of people at low cost. One alternative becoming in- creasingly accepted is distance teaching; usually a form of correspon- dence school relying on radio broadcasts, printed literature, and in selected instances periodic group discussions and telephone conversa- tions.25 Administrators of distance teaching programs in both indus- trial and developing countries generally prefer to rely on the postal system to deliver instructional material, to return exams, to deliver as- signments, and even on occasion to answer individual student questions.26 When a reliable postal system is not in place, however, such educa- tional programs can be impeded, which incurs additional costs; staff of the International Extension College, in London, an organization that promotes university and other forms of adult education in deve- loping countries, have encountered many such cases. In Nigeria, for example, the postal service is considered unreliable enough that the local extension university must maintain its own private courier sys- tem.27 In Mauritius, all course materials for the Mauritius College of the Air are delivered by courier service.28 Similarly, private courier sys- tems are needed to supplement the post both in Indonesia, to assure the reliable and timely conveyance of audiovisual and printed teacher training materials to rural areas, and within the University of the South Pacific, to disseminate material among campuses located on different islands.29 Postal Services and Transport Postal systems by definition depend on transport facilities. In par- ticular, mail delivery to rural areas and to towns within those areas in many countries is unreliable in the absence of adequate roads, naviga- ble waterways, or small airstrips. Although the mail can be delivered by messengers traveling by bicycle, on foot, or by animal, such deliv- ery inevitably is slow and haphazard. In many countries the development of the postal system has been 370 APPENDIXES closely correlated with an improvement of travel conditions and the upgrading of roads.30 One of the conditions attached to the establish- ment of the rural free delivery service in the United States was that local residents had to maintain their roads; the post office was delib- erately used to promote the construction and maintenance of roads. An assistant postmaster general noted in 1912 that, The department LU.S. Post Officel very much desires that post- masters, rural carriers, and substitute carriers shall not only consti- tute themselves apostles of good roads and spread the propaganda, but that they shall, by their works, arouse interest and emulation in others. Many postmasters and rural carriers have been instrumental in forming good road clubs and associations, the results of which have been vast improvement in the condition of highways, and, in several notable incidents, the appropriation of enormous sums of money for the rebuilding and improvement of entire county high- way systems.31 This emphasis on upgrading postal roads is apparent through the years in various reports of U.S. postmaster generals.32 Following a general improvement in roads throughout the United States, the practice of maintaining postal roads was abandoned. How- ever, the case is suggestive.33 Given improved road conditions, the postal service can also act as a form of local freight and general delivery service as it does, for ex- ample, in rural Sweden. There, a service introduced in the early 1970s by the Swedish post office allows rural farming communities to con- tract with the post office for the house-to-house delivery of basic ne- cessities. The Swedish system allows the customer to order goods by telephone or letter from any shop in either the locality where the rural postman begins his route or on the route he follows while mak- ing his rounds. The shop packages the ordered goods and either drops them at the post office or hands them to the postman while he is on his rounds. The parcels are delivered as mail, and the postman may also collect payment for the goods. The Swedish government consid- ers that this service makes rural isolation more bearable and, more important, decreases the amount of travel that rural dwellers must undertake.34 In an age in which the postal service is normally relegated to the delivery of mail and strictly defined types of freight parcels, it is easy to overlook the fact that the post can also transport people. In Scot- land the post office transports both goods and persons in rural areas.35 Some years ago the Scottish post noted that its vans were AN INTEGRATED COMMUNICATIONS SECTOR 371 underused-postal mail volume, as traditionally defined, was insuffi- cient to use existing capacity fully. To correct this, the post office ex- panded the list of goods that could be carried by allowing vans to transport bread, meat, and other foodstuffs; newspapers; and the ma- terials and final products of cottage industries. The incremental costs of such an action were small. The Scottish post office now has few rules about what can be carried; such decisions are left to the discre- tion of the local staff. Hence, some offices transport parcels that do not conform to postal regulations elsewhere. In the absence of reliable local transport in many of the remote areas of Scotland, the post of- fice also uses small eleven-seat buses rather than vans on certain rural routes to carry mail, some cargo, and passengers.36 The Postal Service and Rural Banking Many rural areas in developing countries are isolated from banking services of any kind. Urban banks sometimes hesitate to enter rural areas and smaller towns because of the high cost of maintaining of- fices, the small per capita savings potential, and the small size of credit requirements (relative to urban areas). The lack of small town or rural banks does not, however, imply that rural inhabitants do not require banking services. Many farmers and businesspersons periodi- cally have surplus cash, which can be profitably placed in savings ac- counts; require facilities to transfer funds to creditors; and require regular sources of credit. Farmers in developing countries, in particu- lar, are often at the mercy of private lenders who offer capital at usu- rious rates, when such capital is available at all.37 As a result, postal banking systems can, acting in concert with other government agen- cies or private banks, provide credit for rural development needs at competitive market interest rates. The post office has been important in establishing banking facili- ties in several countries. The British, French, German, several other Western European, and Japanese post offices traditionally have of- fered savings account facilities for small savers, and this practice has been followed in many former colonies. The post can provide such additional rural services largely because investment and operating costs are smaller than those of other postal functions; the labor- intensive nature of the postal service can be exploited to provide a va- riety of services requiring face-to-face contact. When the postal service maintains counter and postal box services in a community, the 372 APPENDIXES addition of various banking services does not involve substantial in- cremental costs. Hence, postal savings facilities have been established in smaller towns and rural areas at a much lower cost than would be incurred for similar services at conventional commercial banks. The Universal Postal Union has contended that such facilities, because of their closeness with the people, actually encourage habits that lead to savings and thrift.38 Rural post offices are also sometimes used as a means to transfer funds, either through giro (checking) accounts or through the sale of money orders; as a credit facility; and as a means for rural inhabitants to pay their taxes.39 The Universal Postal Union also notes that postal giro accounts provide governments with funds for capital or operating expenses, since they tend to be stable over time.40 The proceeds from such ac- counts are sometimes used to finance national development pro- grams, as in India where approximately 10 percent of the national development plan was at one time financed from postal savings.41 The funds from these accounts are also sometimes used to finance further postal expansion.42 Typical examples of postal savings banks are found in Egypt and India. In Egypt, where three out of five citizens live in rural villages, the Egyptian Postal Savings Bank has encouraged small savings ac- counts through the use of "thrift accounts" (accounts specifically de- signed for small savers).43 In India in 1970, the Indian Post Office Savings Bank had Rsl9,160 million44 in deposits, with 24 million de- positors. This bank has no limit on the number of withdrawals-a policy that attracts small deposits. In the past twenty years, however, the growth of postal financial ser- vices in many countries has stopped. In 1979, only about one-third to half of developing countries offered giro or postal savings services, and the number of services had not grown since 1970.45 Although the average deposits in postal giro and savings accounts have increased since 1970, the growth in the assets of postal savings banks has been slower than the growth in the gross domestic products of countries of- fering the service, indicating a relative decrease in real terms in these assets. Overall, savings banks and giro accounts together represent only about 13 percent of the national savings in developing countries where they are offered, compared with 40 percent in industrial coun- tries where they are offered.46 The Universal Postal Union suggests that one of the reasons for the slow development of postal banking is the restrictive policy of postal and finance ministry authorities concerning tariffs, interest rates, and the use of the funds collected.47 In developing countries, AN INTEGRATED COMMUNICATIONS SECTOR 373 some governments have borrowed money from the postal savings bank at nominal or no interest without recording the interest savings as revenues to be credited to the post office-an "accounting error" that vastly underestimates the post office's contribution to government revenues. These governments frequently make no provisions to repay the "borrowed" funds.48 Although such practices generate savings in interest paid by the national government, they tend to limit the amount of funds available for rural credit, to limit the extent to which competitive interest rates can be paid to small depositors, and to cast doubt on the overall integrity of the postal system. Competition and Monopoly in Posts and Telecommunications In the United States in the early years of rural free delivery service, when alternative forms of freight carriage were not widely available, the postal system was an important source of rural freight transport.49 Today, however, almost all freight delivery services in the United States and many other industrial countries are privately owned, and in recent years numerous private message and parcel handling services have appeared, often in direct competition with government postal services. Such change indicates that as economic development takes place and as the volume of message, parcel, and freight traffic increases, private carriers, which may have stronger incentives to control costs and manage their operations efficiently, sometimes provide cheaper and better service than government monopolies. In New Zealand, for example, a network of commercial delivery services that carries farm supplies, commodities (including food), and the mail has evolved to serve rural areas.50 One observer noted that "these private transport operators are able to carry out the delivery of mail in conjunction with the delivery of other goods at a much lower cost than would be the case if the Post Office carried out the work."5" The post office in several developing countries has also by necessity adopted such tech- niques. For example, in Ethiopia, Kenya, and Tanzania, among others, the post office contracts private transporters to carrv mail and postal parcels to remote places along with cargo and passengers. Part of the reason private carriers can sometimes effectively under- cut the government postal service is that in many instances the objec- tives of private carriers are more limited and better focused. Private carriers can concentrate on achieving a low-cost and reliable move- 374 APPENDIXES ment of information, parcels, goods, freight, and possibly people. In providing such service they have a greater flexibility to choose the most efficient mix of cargo and inputs in terms of both labor and technology. Government monopolies, however, tend to be affected by a wide variety of policies made by officials who know little about the practical aspects of the business and who generally try to achieve mul- tiple objectives. Partly because of this, they have less incentive to pro- vide good service, minimize costs, pay competitive wages, and generate increasing revenues; many postal authorities can rely on rev- enue subsidies from government whenever they have financial short- falls for whatever reason. Also, government postal monopolies rarely keep their accounts on a commercial basis. Hence, no one knows how much the service or portions of the service actually cost. Also, governments themselves sometimes use the fact that the postal service is labor intensive and generates employment as a reason to prevent the introduction of new cost-reducing technology. They sometimes contend that automatic letter sorting and routing equip- ment, mobile post offices, modern computers, and electronic mail, data, and funds transfer would reduce employment.52 Although this argument may have some short-run merit in countries with surplus labor, the objective of generating employment has been frequently car- ried to an extreme to the detriment of efficient and innovative postal service.53 The possibility that an efficient, well-managed postal system could generate much more employment indirectly in business, com- merce, and agriculture than it could possibly generate directly tends to be overlooked. For example, recently in the Cote d'Ivoire, the reli- ability of the postal service was improved by reorganizing the mail transport system within the post office and by making minor improve- ments in sorting. As a result, the volume of letters handled in Abidjan increased from 100,000 to more than 200,000 a day in only two years. Technological innovation is, of course, another factor contributing to the likelihood that increased competition will emerge in the postal sector in developing countries. As outlined earlier in this appendix, as technology and consumer demand have evolved, so have the means for providing numerous transfer functions. The development of new telecommunications technology has not only provided opportunities for more efficient transmission of messages, it has also fostered the emergence of new business firms, institutions, and organizations to provide various types of postal-related services, often in direct compe- tition with the government post office. AN INTEGRATED COMMUNICATIONS SECTOR 375 If there is a lesson in all of this, it might be that it is becoming less appropriate to condone the postal system as a static government mo- nopoly. Rather, it might better be thought of in a functional context as a service that can be supplied with varying degrees of coverage and efficiency by different institutional and market structures. Evidence reviewed in the main text shows that for increased productivity and labor specialization to occur, rural as well as urban areas increasingly require efficient and rapid movement of information and goods. In some developing countries the postal monopoly has served a purpose and can continue to improve; in other situations mixed public or pri- vate systems that transfer data and other forms of messages electron- ically and carry the post, other goods, and perhaps even people might be more effective. The introduction of public and private electronic mail and data systems and the established private parcel and freight delivery in North America, Western Europe, Japan, Australia, and many developing countries show that the actual mix of competition and monopoly can be determined mostly by market demand, the mix of technological and service opportunities, and the potential for cre- ating additional incentives for efficient management. Other things being equal, since competition stimulates manage- ment and service efficiency, competition might be allowed to develop for the supply of postal and other communication functions wherever feasible-particularly in the provision of transport (including trans- port of postal and freight materials and of people) and some telecom- munications services (electronic message, data, and funds transfer). Of course, some government regulation would still be necessary to as- sure widespread access to postal, telecommunications, and other ser- vices. However, the possibility that increased competition might foster the greater overall development of such functions, might further the innovative integration of communication and transport alternatives for specific needs, and could serve as a benchmark for judging effi- ciency in the remaining government-administered operations deserves increased attention. If, however, public ownership of and monopoly in the primary postal and telecommunications functions are maintained, and if the two services are separated organizationally and operated commercially for the management and staffing reasons outlined in chapter 3, the two entities might be allowed to compete in some of the more capital- intensive communication areas (information transfer). Interaction and cooperation would, however, probably best serve them both by providing more labor-intensive rural services such as letter and parcel 376 APPENDIXES delivery, rural banking and funds transfer, the operation of Pcos, and so forth. A key for the efficient management of each entity would be to keep separate commercially oriented accounts for each individual service so that its financial position could be monitored, and in- formed management decisions could be made. Notes Note: The auchor acknowledges helpful comments from Jeremy J. Warford. 1. One major exceprion to this was the private development of telegraphy and telephony in the United States. 2. However, writing a letter was not an important alternative to telephone communica- tion in the Chilean and the 1981 Indian surveys. 3. In rural Egypt, however, the long-distance telephone links were extremely congested, and inhabitants of the nine villages had very limited access to both the postal system and telephones; only five of the villages had post offices, three had small telephone exchanges, five had a few scattered private telephone lines, and one had no access to a telephone. 4. Between 1970 and 1976 the number of telephones in Great Britain increased from 14 to 21 million. Telephone density increased from 25 to 39 telephones per 100 persons. 5. Analysis of a sample of about 500 respondents from rural areas in Lincolnshire, En- gland, showed that for social contacts, "contacts by post form the basis for a network of dis- tant friends and relatives who are infrequently visited, whereas visits and phone calls appear to be interdependent modes of social contact, especially over short and middle distances." See Clark and Unwin (1981), p. 55. Results consistent with these are described in Thorngren (1977), pp. 374-85. 6. For an overview of international private electronic mail systems and other communi- cations networks, see Elbert (1990). 7. Rapid technological advances have greatly extended the spread of facsimile message communications systems. Commercially available analog facsimile machines, using standard telephone lines, are available for less than $1,000 and can transmit one page every 20 sec- onds. Even faster machines, using specially conditioned telephone lines, are able to transmit digitally much higher volumes of information. Although these machines vary in cost, they offer the opportunity to expand electronic mail systems within an enormous array of com- mercial and government enterprises. 8. National Research Council (1981). 9. Lee and Meyburg (1981). 10. It has been contended, however, that the U.S. Postal Service will over the longer term be unable to compete in the electronic message transfer market in the United States partly because of its inexperience in competing with private business firms and partly be- cause (a) it is behind private industry in technology development, (b) it has performed too little market research, and (c) it would be required to provide a standard set of services in competition with a set of diverse and unstandardized offerings by other firms. Sorokin (1980), p. 122. 11. Pool (1979), pp. 187-88. 12. Langley and Pisharodi (1981). 13. Lee and Meyburg (1981). 14. In 1982, the government of Singapore, in a move somewhat counter to that dis- cussed in chapter 3, merged the Telecommunications Authority of Singapore with the Post AN INTEGRATED COMMUNICATIONS SECTOR 377 Office. One of the principal reasons for this merger was that within the highly urbanized Singapore area most postal mail is used to transfer nmessages and thus could be eligible for electronic mail service; packages or parcels represent a relatively small portion of total mail handled. Anocher reason unique to the urban Singapore environment was that telecommu- nications staff outnumber postal staff. 15. Much has been written on the general need for information for rural development. For case studies on rural uses of information, see McAnanv (1980). pp. 3-18; Development Communication Report (1988); Parker and others (1989): and Hudson (1984). 16. Fuller (1964), p. 301. 17. Postal and telecommunications services are also complementary in the administra- tion of health care prograns; see appendix A. 18. U.S. Government (1990), p. 5. 19. Fuller (1964), p. 312. 20. Fuller (1964), p. 312. 21. Such objectives are still partly reflected in the preferential postal rates given to maga- zines, newspapers, and brochures in manv countries. 22. U.S. Government (1902), p. 16, and U.S. Governnient (1901), p. 25. 23. Fuller (1964), p. 294. 24. Bates (1982). 25. Telephone teaching is also an increasingly viable option for some forms of distance- related education in developing countries. The major advantage is that it provides immedi- ate two-way interaction betxveen teachers and students. Stich teaching, of course, also relies extensively on the postal service. See Bates (1982) and app. I of that book. For a descrip- tion of the history and practice of telephone teaching in the United States, see Rao (1977), pp. 473-86. 26. For an informative discussion of distance learning programs in, among others, Brazil, England, Israel, Kenya, Korea, Malawi, Mauritius, Nicaragua, the former Soviet Union, and the United States, see Perraton (1982). See also Jenkins (1988); Development Commuinica- tion Report (1988); Asian Development Bank (1987); and Lockheed, Middleton, and Nettleton (1991). 27. Personal interview by Douglas Goldschmidt with Tony Dodds, International Exten- sion College, London, Eng., 1979. 28. Dodds (1982). 29. An issue that has been raised to defend rural telecommunications investment, and indirectly to attack rural postal spending, concerns the incidence of rural literacy in deve- loping countries. Where literacy is lowv, the importance of the post for carrying information may be questioned. Unfortunatelv, few developing countries have reliable data on this issue. Although the levels of household (as opposed to individual) literacy may in many instances be higher than is generally documented, few developing countries have reliable data on ei- ther household literacy or village literacy (for example, the existence of people in the village who will read and write for others). 30. The post has also frequently been used to subsidize rural transport facilities through postal carriage contracts. Even today, air service to rural villages in Alaska is highly depen- dent on U.S. government postal contracts, which historically were Lised as much to support continued rural transport as to ensure niail carriage. 31. P. V. Degraw, Fourth Assistant Postmaster General. Quoted in U.S. House of Repre- sentatives (1912). 32. See U.S. Congress (1971), pp. 29-31. 33. One small exaniple of the effect of the rural free delivery systenm on rLural transporta- tion was noted by the U.S. Congress in 1914. The Post Office found that in the northern regions of the country the inhabitants didn't like to go out after major snow storms until the road had been traveled. The mail carrier was generally the first person out and, hence, 378 APPENDIXES provided the opening of the road, which then elicited other travel, a minor, but interesting, externality. See U.S. Congress (1913) 34. Union Postale (1974). 35. Carpenter (1974). 36. Similar delivery systems run by private transporters are discussed in the last section of this appendix. 37. Donald (1976), chaps. 8-12. 38. Universal Postal Union (1974), p. 13. 39. Universal Postal Union (1974), p. 13. 40. Universal Postal Union (1974), p. 13. 41. A. M. Narula (1972). 42. Simonelli (1967), p. 31. 43. Chaffar (1974). 44. Approximately $2.3 million at early 1982 exchange rates. 45. Universal Postal Union (1980), pp. 11-12. 46. Universal Postal Union (1980), p. 13. 47. Universal Postal Union (1980). 48. World Bank data. 49. Fuller (1964). 50. These privately owned New Zealand delivery services correspond closely in function to the publicly owned Scottish post office and rural Swedish post office services discussed earlier. 51. Williams (1972). 52. A similar argument is sometimes used to oppose telecommunications investment; telecommunications systems are felt to be too capital intensive. 53. The government's objective of generating employment usually is accompanied by the necessiry of holding down wages of postal employees. This makes it difficult for the postal system to hire and retain well.qualified staff, which, in turn, directly contributes to poor ef- ficiency and management of the system. Appendix C A Survey of Rural Public Call Offices in Costa Rica IN 1970 A PROGRAM was initiated by Instituto Costarricense de Electricidad (ICE), Costa Rica's power and telecommunications utility, to install public call office (Pco) telephones on a concessionaire basis in small towns and villages without telephone access. By early 1974, eighty-two small towns had been provided with at least one Pco. How- ever, an unexpectedly rapid increase in costs and a decrease in finan- cial operating surpluses from urban and long-distance telecommunica- tions services reduced the resources available to continue the program. In the face of these financial constraints, ICE, in conjunction with the University of Costa Rica and with input from World Bank staff, initiated a study of rural Pco telephone demand, use, and bene- fits. The initial objective of the study was to provide input needed to clarify priorities for expanding this rural service further. When the study was being designed, unresolved conceptual prob- lems about how to identify and measure telephone benefits as well as data and resource constraints arose. Therefore, the analysis of bene- fits was confined primarily to estimating the short-run price elasticity of demand for Pco telephone use and to identifying some of the char- acteristics of Pco users, the purposes for which Pco telephones were used, the patterns of use, and the substitutability of Pco telephones for other modes of communication. For the study, a rural rco tele- phone possessed the following characteristics: a. It was administered by an ICE concessionaire, who was usually the owner of a business (a shop, bar, cafe, small hotel, restau- 379 380 APPENDIXES rant, and so forth). In a few instances the concessionaire was an employee of the local municipality, a local telegraph operator, or a member of the rural police b. It was a public access telephone, that is, the telephone did not belong to any particular subscriber; it belonged to ICE and could be used by anyone who could pay the call charge c. No private telephone subscribers were located in the village where the Pco telephone was located or within approximately 10 kilometers of the village. Most of the rural telephones were connected to the national network through single- or multichannel ultra or very high-frequency radio equipment. Two types of data were gathered for the study. First, secondary data on the economic, demographic, and telephone use characteristics of the villages were obtained from three sources: a housing and popula- tion census and an agricultural census, both carried out in 1973 by the Costa Rican Department of Statistics and Census, and Pco call traffic information collected and tabulated by ICE. The data covered ninety-two rural Pco telephones in eighty-two villages. Relatively de- tailed census information was available for sixty-four of the villages; somewhat incomplete data were available for the rest. Call traffic data were compiled by ICE both before and after a tariff increase, which raised the call charge for each metered pulse from 12 to 15 centimos on January 1, 1975. To supplement the secondary data, primary data were collected by (a) telephone concessionaires who interviewed telephone users imme- diately following the completion of calls in eleven villages for seven- teen weeks during about ten months in 1974 and early 1975 and (b) two University of Costa Rica students who interviewed callers follow- ing calls made at rco telephones in three villages (Puerto Cortes, Santa Clara, and Llano Grande) for one week in October 1974. Three aspects of the study are reported here: a. Estimation of the short-run price elasticity of demand for Pco telephone calls based on the observed volume of call traffic be- fore and after the January 1975 tariff increase b. Statistical analysis of the variations in telephone use among sample villages to identify some of the economic and demo- graphic determinants of use at the village level c. Examination of the survey data to determine who used rural Pcos (what are the characteristics of users) and for what purposes. A SURVEY IN COSTA RICA 381 Estimation of Price Elasticity of Demand for Calls To derive a simple estimate of the price elasticity of the demand for PCO calls, call traffic before the January 1975 tariff increase (Q1) was measured as the number of metered pulses generated by a village PCO telephone from September through December 1974. Call traffic after the January 1, 1975, tariff increase (Q,) was measured as the number of metered pulses from January through April 1975. Given unit-pulse prices (Pi and P,) before and after the price increase (12 centimos and 15 centimos per pulse, respectively), an estimate of the price elas- ticity of demand for telephone traffic for each PCO was calculated using the standard definition for price elasticity: Price elasticity = - Q 2 Q+Q, / P] - Pi 11/2(Q, + Q2,) I/L/2(P, + P,) For the ninety-two PCO telephones located in the eighty-two rural vil- lages, the estimated price elasticity per village averaged -0.29. Price elasticity estimates calculated for this relatively short period-four months before and after the price change-have the ad- vantage that few other factors changed significantly in the villages in that short period of time. A disadvantage, however, is that the elastic- ity estimates could have been biased in several ways. On the one hand, PCO users might not have had sufficient time to adapt their be- havior to the higher price. For example, they could have continued some of their old calling and travel habits without fully perceiving or adjusting to the new price. In this case the price elasticity would have been underestimated. On the other hand, although call traffic did on average initially fall following the price increase, this could have been a relatively short-term reaction, which might have been followed after several months by a return to previous levels. In this case the esti- mated price elasticity would have been an overestimate of the price elasticity of demand that would be relevant over a longer period. Bias could also have resulted from the presence of a secular trend in telephone use. For example, in the villages where the telephone had been installed less than six months before the price change, call traffic increased relatively rapidly as villagers became accustomed to using the new facility. This upward trend partly offset the decrease in call traffic resulting from the price change; the average estimated price elasticity of demand for the sixty-five villages that had a PCO in service for six months or more was -0.51. 382 APPENDIXES Seasonal variations in local economic activity also affected the esti- mates. After removing villages that had acquired a Pco within six months of the January 1, 1975, price change, the remaining villages that had positive price elasticity estimates were in tourist areas where the period from January to March was the busiest season or were rela- tively far from the capital city of San Jos6 and were partly involved in tourism. The average estimate of the price elasticity of demand for vil- lages in the tourist area was +0.29, whereas faraway villages also partly involved in tourism had an average price elasticity of +0.44.' Excluding from the sample the villages involved in seasonal tourism and those far from San Jose and partly influenced by seasonal tour- ism (twenty-six of the eighty-two villages in the sample), the average estimated price elasticity of demand for the remaining fifty-six villages was -0.57. Table C-i groups the remaining villages by the primary demographic characteristics or type of economic activity in their area. As can be seen, however, the estimated price elasticity among these villages still varied widely. The concessionaire survey of eleven villages, all of which had access to a Pco telephone for six months or more, showed that the average duration of calls decreased from 3.90 minutes between September and December 1974 before the price changed to 3.42 minutes between January and April 1975 after the price changed. This 12.3 percent re- duction of the average duration, when compared with the 25 percent increase in call charge, gives roughly an average price elasticity of de- Table C-i. Average Estimated Price Elasticity of Demand for Rural Pco Telephone Traffic (Metered Pulses) for Selected Groups of Villages in Costa Rica, 1974-75 Avterage estimated price elasticitv Type of area of demand Frontier or border -1.23 Island -1.19 Coffee and sugarcane -1.03 Banana and cocoa -0.45 Cereal and cattle -0.43 Overall average -0.57 Note: Excluding villages with less than six months of telephone access and those located in tourist and distant tourist-related areas, all of which were experiencing major growth in telephone traffic or xvere subject to predictable seasonal increases in traffic. Source: Traffic data supplied by ICE. A SURVEY IN COSTA RICA 383 mand for call duration of -0.49.' This estimate is not far from both the average call traffic price elasticity for rco telephones in service more than six months (-0.51) and the average for older and nontourism-related Pco telephones (-0.57). This implies that rco users on average may have reacted to the increase in call charges by reducing the duration of their calls rather than the number of calls actually placed. Table C-2 shows that the effect of the price change on call dura- tion varied considerably with the primary purpose of the call. Calls related to agriculture and industrial activities remained virtually un- changed, whereas those related to professional services, emergency services, and inquiries about or discussions of health decreased about 20 percent in average duration. The negative price elasticity of de- mand for the duration of calls varied by purpose of call between -0.02 and -0.84. The price elasticity of demand for call duration was positive for calls pertaining to work opportunities. Table C-2. Average Duration of Outgoing Calls and Price Elasticities, bIy Primary Purpose of Call, for Eleveni Concessionaire-Survey Villages in Costa Rica, 1974-75 Duration (minutes) Price September elasticity of to januarN demand for December to call Primary purpose of call 1974 April 1975 duration Family, kin, and friends 3.97 3.56 -0.41 Inquiry about health" 3.35 2.75 -0.72 Commerce 3.88 3.40 -0.45 Agriculture and industry 3.83 3.81 -0.02 Government administration 4.24 3.95 -0.27 Professional services 3.97 3.16 -0.82 Work opportunities' 3.12 3.34 0.28 Sports, religion, study 3.67 3.13 -0.59 Emergency services 3.61 2.85 -0.84 Unknown 3.61 3.14 -0.77 Average 3.90 3.42 -0.49 Note: On average, the concessionaire survey was taken once every two to three weeks dur- ing roughly seventeen survey weeks. All of the village surveys overlapped at least during the period between September 1974 and April 1975. a. Encompasses calls to family or friends inquiring about health matters or discussing someone's health. b. Calls to inform about, or obtain information on, work opportunities. Source: World Bank data. 384 APPENDIXES Determinants of Rural Pco Telephone Use In an initial cross-sectional analysis of the determinants of Pco tele- phone use in rural Costa Rica, an attempt was made to specify and test an explanatory model of the form: Q = f(P, Y x . .X,) where Q = quantity of telephone use per capita, P = price for using the telephone, Y = average yearly income of the caller, the caller's household, or the village of the caller, and X = selected other eco- nomic and demographic variables that could be associated with varia- tions in use. As can be seen, two of the most important explanatory variables were assumed to be price and income. Given the data available, how- ever, it was not possible to identify completely acceptable proxies for these two variables nor for the quantity of use. The problem of data existed partly because telephone calls are not a homogeneous good. To examine empirically the relation between the price paid for a good and the quantity of that good consumed at different prices, price Imlust be varied for a specific uniform good. However, Pco telephone calls were of different durations and made over different distances, at different times of the day and week, and in some cases with varying technical quality. Each telephone call was es- sentially a different good for which consumers would presumably be willing to pay a different price. The information on call charges was also inadequate to construct a totally acceptable proxy for price. In rural Costa Rica, as in many countries, calls are charged on the basis of duration, distance, and time of day. The price of a daytime call was 12 centavos per pulse be- fore January 1, 1975, and 15 centavos per pulse afterward. Call dis- tances were divided into three ranges: short distance, with a pulse frequency of one per twenty seconds; medium distance, with a pulse frequency of one per ten seconds; and longer distance, with a pulse frequency of one per five seconds. When the census data were col- lected (before January 1975), calls of short, medium, and longer dis- tance cost 36 centimos, 72 centimos, and 144 centimiios per minute, respectively. Hence, although the posted price paid by users was uniform (the price per pulse was constant), the price paid varied only with the nonhomogeneity of the calls (with time of day, call duration, and dis- tance), and the price variable was the average call charge paid in the village, not the more relevant price paid by each caller. As a result, an A SURVEY IN COSTA RICA 385 adequate proxy for the price variable could not be satisfactorily de- rived from the available data. Deriving a suitable income proxy was also problematic. For many villages the closest proxy was a piece of census information for each village called "average wage per month." This variable only reflected wages of wage earners, however, and in some villages the proportion of recorded wage earners to total population was relatively small. It was not clear whether or not their earnings varied among villages in a roughly similar manner as average village incomes. Given these and other problems in deriving adequate proxies, it was decided that the best feasible course of action was simply to search empirically for variables that were associated with differences in traf- fic among villages. The two search techniques used were stepwise re- gression analysis and factor analysis. Statistical Regression AnalNsis For the regression exercise, three variables were specified to reflect telephone use (Q): a. The average monthly number of pulses per capita (or, equiva- lently, the average call revenue per capita) generated by a village b. The average weekly number of calls per capita initiated by a village c. The average weekly minutes of call time per capita generated by a village. The pulse was tabulated by ICE for the sixty-four rural villages for which relatively derailed census information was available. The num- ber of calls and the duration of calls were obtained from the conces- sionaire's survey of users and were therefore available for only eleven villages. None of these three variables was fully satisfactory as a measure of the quantity of telephone service that each village consumed during an average week or month. The pulses variable provided information affected by the number of calls, the duration of each call, and the dis- tance of each call-the relative importance of the three factors being unknown. The number of calls variable and the minutes variable also provided mixed information on the duration, number, and distance of calls. Finally, none of the variables provided information on the qual- ity of the transmission. Given these shortcomings, the three quantity- of-use variables were interpreted loosely as general proxies for village telephone use. 386 APPENDIXES Twenty-four other economic and demographic variables were avail- able or could be tabulated from census data to reflect differences in population characteristics and in economic activity among the sixty- four villages. These are listed in table C-3. All economic, demo- graphic, and traffic data examined in this exercise refer to the period before the tariff increase. In an initial analysis of these data, a stepwise multiple linear regres- sion computer algorithm (least squares estimation) was used to select Table C-3. Economic and Demographic Variables Available for the Cross-Village Examination of Rural Public Telephone Use in Costa Rica, 1973-74 Variable number Description 1. Locally owned farm area out of total farm area in the village zone 2. Cultivated area as a percent of total farm area in the village zone 3. Quantity of fertilizer used per fertilized hectare in the village zone 4. Number of locally owned farms as a percent of total farms in the village zone 5. Percent of milk cattle out of total cattle in the village zone 6. Index of expenditures on machinery in the village zone 7. Average area of each farm in the village zone (hectares) 8. Percent of houses in the village within a 2-kilometer radius of telephone 9. Number of houses within a 2-kilometer radius of telephone 10. Percent of population 0 to 6 years old 11. Percent of population 7 to 14 years old 12. Percent of population 15 to 19 years old 13. Percent of population 20 to 24 years old 14. Percent of population 25 years and older 15. Illiterate population as a percent of population 10 years and older 16. Percent of employers and self-employed out of the economically active population in the village zone 17. Average wage per month of wzage earner in the village zone (colones) 18. Average number of rooms per house in the village zone 19. Percent of houses with septic tanks in the village zone 20. Percent of houses with radios 21. Number of shops, offices, health clinics, bars, and so forth in the village where the rural telephone is located 22. Students as a percent of the population 7 to 24 years old 23. Distance from the village to San Jose (kilometers) 24. Total population in the village zone a. The village zone includes the village and surrounding area in which the population is assumed to have convenient access to the village. In Costa Rica this was generally an area within a 10-kilometer radius of the telephone. Source: Department of Statistics and Census, Costa Rica. A SURVEY IN COSTA RICA 387 a small set of statistically significant regressors (independent vari- ables), which together explained the largest proportion of the varia- tion among the sixty-four villages in the values of the dependent variables (telephone use). This involved regressing, in turn, the three variables reflecting telephone use on the twenty-four economic and demographic variables. Three equations that emerged from these stepwise regressions are shown in table C-4. In these equations, Q, = average monthly pulses per capita Q, = average weekly number of calls per capita Q3 = average weekly minutes of calls per capita X73 = distance between the village and San Jos6 (kilometers) X17 = average monthly wage for wage earners X70 = percent of houses with radios X9 = percent of houses in the village within 2 kilometers of the telephone X22 = percent of the population between 7 and 24 years old who are students. These three regression equations revealed a consistent direction of association between the independent variables (regressors) and the telephone use variables.3 All of the regressors were statistically signifi- cant at the 0.05 level except for the wage variable (X,7) in equation 1, the distance variable (X23) in equation 2, and the radios variable (X20) in equation 3. The direction of association between the regressors and the telephone use variables was roughly consistent with what was expected in the following ways. First, if differences in the wage variable (X17) roughly reflected differences in relative income among the villages, the positive relation between the three telephone use variables and the wage variable showed that more traffic per ca- pita tended to be generated by higher-income villages than by lower- income villages. Second, on the one hand, the positive relation between the tele- phone pulses per capita variable (Q,) and the distance to San Jose (X73) was in line with the fact that in general most calls were made to San Jose (see table C-17) and that, for a given call, the frequency of pulses was by definition a function of the distance of the call. On the other hand, a positive relation was also found between distance to San Jose and both the number of calls (Q) and the duration of calls (Q3). This is consistent with the likelihood that the greater the dis- tance over which communication took place, the more the telephone was the least-cost way to communicate; the more distant villages had fewer feasible alternatives for communicating rapidly with the major 388 APPENDIXES Table C-4. Regression Equations in Which Selected Economic and Demographic Variables Are Regressed on Proxies for Village Telephone Traffic Equation Sample Dependent Constant number size variable term X23 1 64 Q, 5.652 0.039 (2.86) 2 11 Q. 0.729 0.005 (1.89) 3 1I QI -4.130 0.024 (2.29) - Not available. Note: Values of t-statistics based on students' distribution of t are presented in parenthe- ses below the regression coefficients. Source: World Bank data. economic, financial, and government center in the country, or the al- ternatives were more expensive. Third, the negative relation between telephone use and the propor- tion of houses with radios (X,7) was more difficult to explain, al- though telephones and radios were, to some extent, substitutes. The prevalence of radios as a means of receiving certain types of informa- tion (weather forecasts, agricultural reports, market prices, and politi- cal events) may have decreased the need for telephone use. The interviews of individual users in the three-village sample partly sup- ported this possibility (see table C-13). Equation 1 is the only equation that examined differences among the sixty-four villages. The inclusion in this equation of independent variables representing the percentage of village households within 2 kilometers of the telephone (X9) and the percentage of village popula- tion between 7 and 24 years old who were students (X,,) significantly improved the explanatory power of the model. The positive associa- tion between the population concentration variable (X9) and the tele- phone use variable (Ql) suggested that the more accessible the telephone, the more frequently it was used.4 To the extent that the student variable (X72) reflected the general educational level of the village, its positive association with telephone use (Q,) perhaps indi- cated that use increased with the general educational level of the pop- ulation. A related explanation for the significance of this variable is that students had a wider geographical community of interest or a greater familiarity with the telephone than did most villagers.5 A SURVEY IN COSTA RICA 389 Adjusted coefficient of X X20 X9 X1 determination 0.008 -0.168 0.068 0.002 0.451 (1.82) (-2.96) (2.36) (2.47) 0.004 -0.026 - - 0.648 (2.41) (-2.24) 0.022 -0.060 - - 0.655 (3.27) (-1.20) Factor Analysis Among the reservations concerning this stepwise multiple linear re- gression exercise is that the final selection of three or five regressors was influenced by the size and pattern of the collinearity among the original data. As a result a further exercise was undertaken to identify regressors that explicitly represent the underlying patterns of associa- tion among the original twenty-four economic and demographic vari- ables. Factor analysis was used to help identify those underlying patterns. Factor analysis is a statistical technique that aids in summarizing the general patterns of association among variables on the basis of their intercorrelations. Factor analysis can help disentangle complex relations in a set of data and present the data summarized in major and distinct regularities. Although the algebra of factor analysis is somewhat complicated, the technique is generally to extend into n di- mensions an examination of the extent to which vectors representing measurable characteristics tend to cluster when examined for several observations. Highly related characteristics tend to cluster together, whereas unrelated characteristics can be set at right angles to one an- other. Algebraically, these clusters are set forth as vectors or linear combinations of the original variables that explain the maximum pos- sible amount of variation present, given that the variables are com- bined linearly. After redundancies in the original sets of data are identified and eliminated, proxies for the remaining variables or char- acteristics can be introduced into least squares regression analysis.6 Summary results of a factor analysis of the twenty-four economic and demographic variables are presented in table C-5. The columns of figures are the factor loadings, that is, the coefficients or weights of Table C-5. The First Five Factors from a Factor Analysis of Twenty-four Economic and Demographic Variables Compiled for Sixty-four Villages in Costa Rica, 1973-74 Variable Factor Factor Factor Factor Factor number Description 1 2 3 4 5 1. Percent locally owned farm area out of total farm area in the village zone' -0.006 -0.017 -0.150 0.381 0.420 2. Cultivated area as a percent of total farm area in the village zone 0.299 -0.112 -0.348 0.055 0.041 3. Quantity of fertilizer used per fertilized hectare in the village zone 0.534 -0.031 0.122 0.062 0.156 4. Number of locally owned farms as a percent of total farms in the village zone -0.221 -0.243 0.244 0.187 0.509 5. Percent of milk cattle out of total cattle in the village zone 0.228 -0.399 -0.191 0.215 -0.148 6. Index of expenditures on machinery in the village zone 0.513 -0.194 -0.041 0.079 0.169 7. Average area of each farm in the village zone (hectares) -0.265 0.643 -0.108 -0.100 -0.063 8. Percent of houses in the village within a 2-kilonmeter radius of a telephone -0.106 0.455 -0.048 -0.158 0.360 9. Number of houses within a 2-kilometer radius of a telephone 0.814 0.141 0.127 -0.079 0.091 10. Percent of population 0 to 6 years old -0.287 0.307 0.429 0.594 -0.333 11. Percent of population 7 to 14 years old 0.025 -0.216 0.761 -0.326 0.002 12. Percent of population 15 to 19 years old 0.035 -0.186 -0.312 -0.074 0.141 13. Percent of population 20 to 24 years old 0.111 0.381 -0.631 0.428 0.239 14. Percent of population 25 years and older 0.216 -0.285 -0.483 -0.618 0.123 15. Illiterate population as a percent of population 10 years and older -0.611 0.278 0.031 0.132 0.164 16. Percent of employers and self-employed out of the economically active population in the village zone -0.423 -0.150 0.197 -0.173 0.261 17. Average wage per month of wage earner in the village zone (colones) 0.403 0.519 -0.070 -0.336 -0.285 18. Average number of rooms per house in the village zone 0.345 -0.255 -0.028 0.230 -0.433 19. Percent of houses with septic tanks in the village zone 0.388 0.447 -0.044 0.028 -0.468 20. Percent of houses with radios 0.245 -0.374 -0.002 -0.040 0.080 21. Number of shops, offices, health clinics, bars, and so forth in the village where the rural telephone is located 0.676 0.198 0.350 0.016 0.311 22. Students as a percent of the population 7 to 24 years old -0.004 0.024 0.207 -0.343 -0.142 23. Distance from the village to San Jose (kilometers) -0.131 0.522 0.089 -0.293 0.323 24. Total population in the village zone 0.705 0.098 0.340 0.315 0.261 Percent of total variance explained 22.4 16.0 12.6 10.9 10.4 Cumulative percent of total variance explained 22.4 38.3 50.9 61.8 72.3 a. The village zone includes the village and surrounding area in which the population is assumed to have convenient access to the village. In Costa Rica this was generally an area within a 10-kilometer radius of the telephone. Source: World Bank data. 392 APPENDIXES each variable in a linear combination with all variables in the analy- sis. The factor loadings represent the degree of association between the individual variables and the vector made up of a weighted combi- nation of all the variables, which is sometimes referred to as a factor, characteristic vector, or eigenvector. Each factor represents an inde- pendent dimension of the total variation of all variables in the analy- sis. The absolute value of a factor loading may be interpreted as an ordinal measure of the degree to which each variable is involved in each factor or cluster of variables. As shown in table C-5, the first factor explains approximately 22 percent of the total variation among all twenty-four variables. The size of the factor loadings on the first factor shows that population (variable 24), number of houses within 2 kilometers of a telephone (9), number of shops (21), illiterate population (15), fertilizer use (3), and machinery expenditures (6) were highly interrelated and to- gether reflected a dimension in which highly ranked villages tended to be larger, to have a better educated population, and to make relatively greater expenditures on machinery and fertilizer. Factor 1 might then be designated as reflecting a general dimension of village size, educa- tion, and technology use. The second factor explains an additional 16 percent of the varia- tion among the twenty-four variables, bringing the cumulative varia- tion explained to 38 percent. Given the relatively high loadings of variables 7, 8, 17, 19, and 23, this vector seems to represent a general dimension in which highly scored villages tended to be high-income, densely clustered, large-farm villages located far from San Jose. The third and fourth factors are dominated by variables describing the age distribution of the population. Villages ranked highly on the third vector had a relatively high proportion of population less than 15 years old, whereas villages ranked highly on the fourth vector had a relatively higher proportion of their population in the age group less than 6 years old or between 20 and 24 years old. The fifth and last factor reviewed here adds an additional 10 per- cent to the explained variation, bringing the total variation explained by the five factors to approximately 72 percent. Villages ranked highly according to this vector tended to have a relatively large proportion of locally owned farms and farm areas that had small houses and no sep- tic tanks. Given that these five factors represent five underlying patterns of association among the twenty-four variables examined, highly loaded variables from these factors can be selected as proxies for each factor. A SURVEY IN COSTA RICA 393 The following five variables have the largest absolute loading on each of the five vectors and, as a result, were chosen to represent the gen- eral dimension reflected by each vector. X9 = Number of houses within a 2-kilometer radius of a tele- phone. This variable represents the general dimension of vil- lage size, education, technology use (factor 1) X7 = Average area in hectares of each farm in the village zone. Villages in which farm areas were larger tended to be higher- income, densely clustered, large-farm villages relatively far- ther from San Jose (factor 2) X,, = Percent of population between 7 and 14 years old (factor 3) x14 = Percent of population 25 years and older (factor 4) X4 = Locally owned farms as a percent of the total number of farms in the village zone. Villages in which the percentage of locally owned farms and locally owned farm areas was rel- atively large tended to have small houses with no septic tanks (factor 5). These five variables were regressed, in turn, on the three telephone use variables previously examined, with Q, expressed as average monthly revenue (pulses times 0.12 colones) per capita rather than as pulses. The results of the three regressions (linear with least squares estimation) are presented in table C-6. Although the direction of the association between X., X7, Xll, and X4 and the Pco call traffic variables is consistent among the three re- gressions, only X9 and X7 are statistically significant at the 0.05 level in any of the regressions. Given this, it can be tentatively concluded that villages that exhibited greater amounts of rco telephone use per capita tended to be larger, higher-income farming villages relatively far from San Jose, in which the farms were larger and the population was relatively better educated and engaged in more technically progressive agriculture. Results of the Determinants Analysis In summary, the cross-sectional analysis of the determinants of PCo telephone use in Costa Rican villages yielded interesting but not nec- essarily surprising results. If it is assumed that the per capita tele- phone use variables that were examined correspond roughly to crude indexes of benefits that villages derive from telephones, it might then be concluded that the rural villages in Costa Rica that tended to re- 394 APPENDIXES Table C-6. Regression Equations in Which Economic and Demographic Variables Selected through Factor Analvsis Are Regressed otl Proxies for Village Telephone Traffic Equation Sample Dependent Constant number size variable term X9 4 64 QE 2220.40 7.11 (1.95) 5 1I Q. -3.72 0.002 (0.001) 6 11 Q, -10.37 0.007 (0.005) Note: Standard errors are presented in parentheses below the regression coefficients. Source: World Bank data. ceive the greatest benefits from rural Pco telephone service had on av- erage some of the following characteristics:7 a. They were relatively large. Several variables relating to village size were highly weighted on the factor represented by X9 b. The village population was relatively better educated and en- gaged in more progressive agricultural techniques. Several vari- ables relating to both of these characteristics were highly loaded on the factor represented by X9, and the student variable (X22) was significant in equation 1 c. The village was better off economically. The wage variable was statistically significant in equations 2 and 3 and was highly loaded on the factor represented by XI. In addition, X7 itself was the farm size variable, which, other things being equal, could partly reflect the stock of wealth per family d. The telephone was relatively convenient to use for a greater pro- portion of the population. The percentage of households within 2 kilometers of the telephone was statistically significant in equation I and highly loaded on the second factor represented by X7 e. They were relatively far from San Jos6. This suggests that as dis- tance as well as the cost of some other forms of rapid communi- cation increased, telephones increasingly became the least-cost solution for communicating rapidly in rural Costa Rica. The dis- tance from San Jose variable was significant in equation 1 and was highly loaded on the vector represented by X7. A SURVEY IN COSTA RICA 395 Adjusted coefficient of X, X' Ix4 XI determination 13.20 -136.98 -77.37 37.59 0.210 (8.82) (135.50) (109.51) (27.08) 0.019 -0.001 0.055 0.017 0.880 (0.003) (0.001) (0.038) (0.010) 0.073 -0.107 0.151 0.067 0.561 (0.026) (0.330) (0.320) (0.088) Characteristics of Pco Users The determinants exercise provided some insights into the types of villages that might receive the greatest immediate benefits from rural telephone investment in Costa Rica.8 This and the following sections review some of the characteristics of the persons using Pco telephones in rural areas. The basic questions examined are who uses the tele- phones and for what purposes? To address these questions two sources of data were examined. One was the responses obtained in the eleven-village concessionaire survey from which the number and duration of calls were tabulated. The sec- ond source was the October 1974 survey of telephone users in three villages: Puerto Cortes, Santa Clara, and Llano Grande. Selected eco- nomic and demographic characteristics of the three villages are shown in table C-7. Survey information on the characteristics of rural PCO telephone users is presented below. Employment Characteristics In the three user-survey villages approximately 97 percent of the callers who answered the question on employment status claimed to be employed. As can be seen from table C-8, more than 60 percent of those persons were wage earners, rather than self-employed or employ- ers of others. With regard to their occupations, the largest percentage overall (32.7 percent) consisted of agricultural workers followed closely by businesspersons, professionals, and technicians. 396 APPENDIXES Table C-7. Selected Characteristics of Three Villages in Costa Rica in Which Surveys of Public Telephone Use Were Made, 1974 Characteristic Puerto Cortes, Santa Clara Llano Grande Locationi Southern Costa Rica Northern Costa Rica Central valley Main economic River port; banana Cattle breeding; General agriculture; activity plantation cereal growing; vegetables; flowers; technical school dairy products Percent of houses within 2 kilonmeters of the telephone 91 39 35 a. Puerto Cortes is the largest of the three villages. Souirce: World Bank data. Table C-8. Employment Characteristics of Callers from a User Survey of Three Rural Villages in Costa Rica, 1974 Emplomment characteristics Puerto Santa Llano of callers Cortes Clara Grande Total Employment status of the caller Wage earner 199 73 131 403 (52.0) (65.8) (74.9) (60.5) Self-employed 157 30 36 223 (41.0) (27. 0) (20.6) (33.5) Employer 17 0 0 17 (4.5) (0.0) (0.0) (2.5) Principal occupation of the caller Professionals, technicians, and similar workers 98 37 1 136 (25.9) (42.0) (0.8) (23.2) Higher public officials 14 0 0 14 Adniinistrative personnel and (3.7) (0.0) (0.0) (2.4) similar workers 45 7 0 52 (11.9) (7.9) (0.0) (8.9) Businesspersons and similar occupations 120 5 14 139 (31.7) (5.7) (11.6) (23.7) Workers in the service sector 3 6 5 14 (0.8) (6.8) (4.1) (2.4) Agricultural workers 65 29 98 192 (17.2) (32.9) (81.0) (32.7) Nonagricultural laborers 33 4 3 40 (8.7) (4.5) (2.5) (6.3) Note: For each question the number of affirmative answers is presented first and the per- centage of answers falling in that category is presented below in parentheses. There are 674 total observations: 383 in Puerto Cortes, 114 in Santa Clara, and 177 in Llano Grande. Soturce: World Bank data. A SURVEY IN COSTA RICA 397 The occupations of the callers predictably varied according to the economic characteristics of the village in which they lived. In Puerto Cortes, a relatively large river port, businesspersons constituted the greatest proportion of callers (31.7 percent). In Santa Clara, profes- sionals and technicians, to some extent from the local technical school, constituted the most frequent callers, whereas in Llano Grande, an agricultural area, agricultural workers made 81 percent of all calls. Inicotmie Level The average income of income earners in the three user-survey vil- lages was approximately 0372.23 ($43.79) a month in 1973.9 The mean income of persons making Pco telephone calls who answered the income question was ¢377.87 ($44.45) a month. This suggests that at least in these three villages rural telephone access was not dominated by either high- or low-income users. Table C-9 shows the distribution of calls according to the income of the caller for the three villages. The median income of the callers who answered the in- come question was somewhat less than the mean income. Of course, the median income for the villages as a whole might also have been less than the mean. Information on income and calls broken down for each of the three villages is presented in table C-10. Again, as shown in the more ag- gregate data, relatively lower-income villagers participated heavily in Pco telephone use.', Table C-9. Sumnmary Distribution of Calls, ba Income Class of Caller, for Three Rural Villages in Costa Rica, 1974 Incotime group Number (colones per momtih) of calls Percentage No response 148 21.9 0-75 23 3.4 75-150 95 14.1 150-225 87 12.9 225-300 102 15.1 300-375 37 5.5 375-450 23 3.4 450-525 58 8.6 525-600 30 4.5 600-800 26 3.9 800+ 45 6.7 Total 674 100.0 Souirce World Bank data. 398 APPENDIXES Table C-10. Distribution of Calls and Income of Caller, by Village, Jor Three Rural Villages in Costa Rica, 1974 (colones per month) Average income of Percentage Income income of total Village group earners calls Puerto Cortes 0-300 180.69 35.1 300-600 431.31 32.2 600+ 1,063.33 12.9 No response" - 19.8 Santa Clara 0-300 188.00 17-8 300-600 425.88 11.0 600+ 774.79 19.1 No response - 52.1 Llano Grande 0-300 182.39 79.2 300-600 463.17 4.8 600+ 1,154.00 0.8 No response - 15.2 - Not available. a. No response includes no answers and those who claimed no income because they did not wvork. The latter group comprised about 3 percent of the total. Source: World Bank data. Distance Traveled to Make Calls Both the eleven-village concessionaire survey and the three-village user survey provided information on the distances Pco telephone users traveled to make calls. The user survey revealed that overall 78.6 per- cent of all calls were made by individuals who traveled at most 500 meters to make the call. The number within 500 meters ranged from 70.8 percent of the callers in Puerto Cortes (the largest community of the three) to 99.1 percent in Santa Clara. The concessionaire survey showed that on average 94.8 percent of the calls were made by indi- viduals who lived within approximately 2 kilometers of the telephone. These results reaffirm the telephone convenience factor observed in the statistical examination of the determinants of telephone use. Frequency of Calls Table C-II, which summarizes the frequency of calls made during the sample week, suggests that Pco telephone use in the three user- survey villages was not dominated by any one small group of users. A SURVEY IN COSTA RICA 399 Table C-l I. Percentage of Calls Made by the Same Caller for Three Rural Villages in Costa Rica. 1974 Number of calls made during the sample week Puerto Santa Llano by the same person Cortes Clara Grande Average 1-2 78.8 82.1 76.1 78.8 3-4 8.4 13.4 12.7 10.3 5-10 11.3 4.5 7.0 9.1 11-15 1.0 0.0 1.4 0.9 15+ 0.5 0.0 2.8 0.9 Total 100.0 100.0 100.0 100.0 Source: World Bank data. Almost 79 percent of the callers made only one or two calls during the week, and more than 89 percent made four calls or less. Table C-12, however, which compares the mean monthly income of callers with the frequency of calls, indicates that users who made more than two calls a week tended to have at least average or above-average incomes.'I Hence, although the benefits of the rural telephones were not generally restricted to any particular income group or small group of Pco users, the 21 percent who did use the telephone three or more times in the week tended, on average, not to be among the lowest- income callers. Table C-12. Average Monthly Income of Callers According to the Number of Calls Made during the Sample Week for Three Rural Villages in Costa Rica, 1974 (colones) Number of calls made during the sample week Puerto Santa Llano by the same person Cortes Clara Grande Average 1-2 350.90 393.13 159.12 321.45 3-4 429.79 520.80 170.75 369.89 5-10 511.13 404.00 172.50 439.82 11-15 303.67,' 15+ 300.00 a. Sample size is too small for the estimate to be reliable. Source: World Bank data. 400 APPENDIXES Table C-13. Alternative Communication Preferences for Callers in Three Rural Villages in Costa Rica, 1974 Alternative method of communication if Pco Puerto Santa Llano telephone were not available Cortes Clara Grande Total Would nor have communicated 64 53 3 120 (16.8) (46.5) (1.7) (17.9) Telegram 107 15 0 122 (28.1) (13.2) (0.0) (18.2) Radio 132 0 16 148 (34.6) (0.0) (9.1) (22.1) Post office 57 45 131 233 (15.0) (39.5) (74.9) (34.8) Other telephone 5 1 14 20 (1.3) (0.9) (8.0) (3.0) Other 16 0 11 27 (4.2) (0.0) (6.3) (4.0) Note: For each question the number of affirmative answers is presented first and the per- centage of answers falling in that category is presented below in parentheses. There are 674 total observations: 383 in Puerto Cortes, 114 in Santa Clara, and 177 in Llano Grande. Source: World Bank data. Preferences for Altertnative Forms of Communication Callers in the three user-survey villages were also asked the means of communication they would have used if the Pco telephone had not been available in the village. The results of the responses are pre- sented in table C-13. Although responses varied widely among vil- lages, the first preference of almost 35 percent of the respondents was the postal service. The next most frequently listed alternative was the radio, primarily because it was the dominant alternative in Puerto Cortes. Third overall was a telegram. On average almost 18 percent of the callers would not have communicated at all if the telephone had not been available; the range was a large 2 to 46 percent. Overall, the wide variation in choice of alternative communication suggests that preferences, or more likely feasible alternatives, tended to be location specific. Purpose and Characteristics of Calls Table C-14 summarizes information from the three user-survey vil- lages on the primary purposes for which telephone calls were made. A SURVEY IN COSTA RICA 401 Table C-14. Primary Purpose of Pco Calls Made in Three Rural Villages in Costa Rica, 1974 Puerto Santa Llano Primary purpose of call Cortes Clara Grande Total Family, kin, and friends 180 84 82 346 (47.4) (73.7) (46.3) (51.5) Inquiry about health, health- related topics, or someone's health 30 0 6 36 (7.9) (0.0) (3.4) (5.4) Commerce 72 2 26 100 (18.9) (1.7) (14.7) (14.9) Agriculture and industry 16 0 28 44 (4.2) (0.0) (15.8) (6.6) Government administration 33 16 10 59 (8.7) (14.0) (5.6) (8.8) Professional services 18 3 3 24 (4.7) (2.6) (1.7) (3.6) Informing of, or obtaining information on, work opportunities 10 4 3 17 (2.6) (3.5) (1.7) (2.5) Sports, religion, study, and other 20 4 16 40 (5.3) (3.5) (9.0) (6.0) Emergency services I 1 3 5 (0.3) (0.9) (1.7) (0.7) Note: For each question the number of affirmative answers is presented first and the per- centage of answers falling in that category is presented below in parentheses. There are 674 total observations: 383 in Puerto Cortes, 114 in Santa Clara, and 177 in Llano Grande. Source: World Bank data. Table C- 15 presents similar information from the eleven conces- sionaire-survey villages, which are grouped according to prominent economic or demographic characteristics. These two tables show that although the relative importance of the reasons for calling did vary a few percentage points from village to village, overall the reasons given were roughly consistent. Both surveys showed that the most frequent purpose for making a rural PCO telephone call was communicating with family or friends, for whatever reason. Calls related to trade and commerce ranked second in frequency, and government administration calls were third in the three-village survey, while sports, religion, and study were third in the eleven-village survey. Table C-15. Percentage of Calls Generated in Eleven Concessionaire-Survey Villages in Costa Rica, Grouped by Characteristics of the Area, by l'urpose of Call, 1974-75 Frontier; Coffee; Banana; Cereal; Primarily Primary purpose of call Distant border sugarcane cocoa cattle tourist Total Family, kin, and friends,, 67.8 58.3 45.6 54.8 58.1 74.7 59.7 Inquiry about health, health-related topics, or someone's health 4.3 3.2 12.1 5.7 6.0 5.8 5.9 2 Commerce 9.3 20.5 10.6 10.6 14.1 11.5 12.6 Agriculture and industry 5.2 1.4 5.9 3.3 3.0 1.2 3.1 Government administration 1.8 3.0 3.9 2.7 6.1 1.0 3.7 Professional services 0.6 0.1 2.6 1.1 1.6 0.6 1.1 Sports, religion, study, and other 1.0 0.5 11.1 4.0 6.8 0.8 4.7 Emergency services 0.1 0.3 0.1 0.2 0.5 0.2 0.3 No reply 7.9 12.4 8.2 13.4 4.0 4.3 8.6 a. Includes calls to discuss work opportunities. Source: World Bank data. A SURVEY IN COSTA RICA 403 The fact that economic and demographic factors affect the types of calls made from a village can be seen from the results of the conces- sionaire survey, which are presented in table C- 15. Calls to family and friends were highest in the tourist and the more distant areas that were also associated partly with tourism-almost 75 and 68 percent, respectively, of all calls made from these areas. A partial explanation for this could be that individuals in distant areas felt relatively iso- lated and thus relied more heavily on the telephone as a means of personal contact. Also, since the business of distant tourist areas is people, a higher percentage of tourist-generated personal calls might be expected from such areas. A relatively large proportion of commer- cial calls were made from border areas, perhaps because these areas are often associated with trade. Agricultural or industrial calls were made with approximately the same frequency from the three crop- producing areas and the distant areas; the share of such calls was rela- tively smaller in border and tourist areas. These results essentially reinforce what was observed earlier in the examination of the employment characteristics of callers in the three- village user sample. The share of agricultural or industrial calls tended to be average or higher in crop-producing areas, commercial calls were highest in trade-oriented border areas, and calls to family or friends tended to dominate tourist areas. Similarly, calls about religion, sports, government administration, and professional services ac- counted for higher shares in the villages surrounding San Jose. Hence, overall, the reasons for rural Pco telephone use predictably tended to reflect the economic and demographic characteristics of the area. Survey information on the distances that callers traveled to the Pco telephone by purpose of the call showed that although the distances traveled were on average about the same, individuals tended to travel longer distances to make certain types of calls. Findings from the three-village user survey showed that on average individuals traveled relatively longer distances to make calls about agriculture or industry, health, and emergency services. Similar results for calls about agricul- ture or industry and emergency services were found in the eleven- village concessionaire survey (see table C-16); callers making agricul- tural or industrial and emergency services calls traveled 81 and 58 percent farther, respectively, than the average distance traveled. Both surveys indicated that individuals calling family or friends traveled about average distances. Agricultural or industrial callers probably travel relatively farther because farmers or farm workers are likely to be located farther from the center of the villages, where Pco telephones 404 APPENDIXES Table C-16. Mean Distance That Callers Traveled to Make Calls, by Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 (kilometers) Callers residing more than 2 kilometers from the All Primary purpose of call telephone callers Emergency services 9.00 0.41 Inquiry about health 8.16 0.36 Family, kin, and friends: 8.73 0.27 Commerce 7.56 0.26 Agriculture, industry 9.39 0.47 Government administration 6.99 0.12 Sports, religion, study, and other 7.14 0.30 Professional services 7.29 0.41 No answer 9.28 0.12 Total 8.46 0.26 a. Includes calls to discuss work opportunities. Source: World Bank data. Table C-17. Percentage of Calls to San Jose, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 Primary purpose Florencia Llano Penas of call TC. Guapiles Grande Miramar Orosi Blancas Family, kin, and friends' 49 70 43 45 53 67 Inquiry about health 40 71 37 42 49 64 Commerce 8 71 53 44 48 73 Agriculture, industry - 73 35 39 45 59 Government administration 100 69 28 39 50 62 Professional services 50 72 38 36 25 80 Sports, religion, study, and other 37 65 45 47 55 69 Emergency services - 40 - 49 - 38 No answer 37 67 40 41 53 66 Total percentage of calls going to San Jose 38 69 44 44 50 68 - Not available. a. Includes calls to discuss work opportunities. Source: World Bank data. A SURVEY IN COSTA RICA 405 tend to be located. This is, of course, opposite the situation for calls about government administration. Finally, callers seeking emergency services may travel relatively longer distances because such calls are viewed as being more important, and thus individuals are willing to travel farther to make them. Information on the most frequent destination of rural calls, by the primary purpose of the call, is provided in table C-17. On average, 60 percent of all calls made in the eleven concessionaire-sample villages went to San Jose. The two most frequent types of calls to San Jose were related to commerce and to professional services. This, of course, is consistent with the fact that San Jose is the business and government center of Costa Rica. The least-frequent reason for mak- ing calls to San Jose was for emergency services. This might indicate that many emergencies must be dealt with quickly, and therefore a neighboring town or provincial center is a more appropriate place to seek assistance. Finally, information was also collected on how the primary purpose of calls varied with the duration of calls as well as with the day of the week and the time of day the calls were made. Table C-18 shows the average duration of calls made in each of the eleven concessionaire villages. The shortest duration was on average for calls seeking emer- gency services, when time is of the essence. At the other extreme were Total Poasocitos percentage Playas Puerto Puerto Santa San of calls to Jaco Cortes Jimrnnez Clara Vito San Jose 60 66 50 47 67 60 48 78 45 45 83 58 69 76 50 57 74 65 55 80 49 54 83 59 55 57 23 45 68 55 54 73 36 62 74 55 80 100 100 50 50 50 46 79 60 53 77 64 49 63 51 44 67 61 59 67 50 48 69 60 406 APPENDIXES Table C-18. Mean Duration of Calls, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 (minutes) Primary purpose of call Call duration Family, kin, and friends' 3.70 Inquiry about health 3.09 Commerce 3.62 Agriculture and industry 3.79 Government administration 4.00 Professional services 3.39 Sports, religion, study, and other 3.19 Emergency services 2.88 No answer 3.48 a. Includes calls to discuss work opportunities. Source: World Bank data. calls relating to government administration. This would be consistent with situations in which public sector employees are reimbursed for their official calls and therefore lose some of the incentive to econo- mize on telephone call charges. Table C-19 shows how the purposes for which calls were made var- ied by day of the week. The major changes in daily calling patterns seem to be that calls related to commerce, agriculture, government administration, and professional services tended to fall off on the weekend, whereas the proportion of calls to family and friends made on the weekends tended to increase. This is as would be expected when a five- or six-day workweek is followed by a one- or two-day weekend holiday. Table C-20 shows types of calls made, by the time of day. Once again, calls about commerce, agriculture, government administration, and professional services varied together and tended to be made dur- ing the working day (6 a.m.-6 p.m.), with the largest number occur- ring during the morning (6 a.m.-12 noon). For no category were more calls made at night (6 p.m.-6 a.m.), when reduced call charges were supposed to be in effect.'2 Calls to family and friends and those relating to inquiries about health and religion, study, and sports did not fall off as much as did business-related calls after 6 p.m., but they did decline. Table C-21, which summarizes calling time information from the three user-survey villages, shows roughly the same general patterns, although differences among the time slots were not as pro- nounced as they were in the eleven concessionaire-survey villages. The Table C-19. Percentage Distribution of Calls per Day, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 Primary purpose of call Monday Tuesday Wednesday Thursday Friday Saturday Sunday Family, kin, and friends" 55.8 56.9 58.0 56.4 57.8 63.6 69.2 Inquiry about health 5.6 5.7 5.8 6.1 5.4 5.7 7.2 Commerce 15.0 14.5 13.6 13.5 13.1 11.6 7.0 4b- Agriculture and industry 3.8 3.8 2.9 3.5 3.3 2.5 2.0 2-j Government administration 4.8 5.0 4.3 4.6 4.4 2.2 0.8 Professional services 1.4 1.6 1.4 1.3 1.3 0.7 0.1 Sports, religion, study, and other 4.9 4.1 4.7 4.0 5.3 4.9 5.0 Emergency services 0.3 0.3 0.8 0.1 0.1 0.2 0.3 No answer 8.4 8.2 8.6 9.9 9.0 8.1 8.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 a. Includes calls to discuss work opportunities. Source: World Bank data. 408 APPENDIXES Table C-20. Percentage Distribution of the Times of Day during Which Calls Are Made, by Primary Purpose of Call, for Eleven Concessionaire-Survey Villages in Costa Rica, 1974-75 Morning Afternoon (6 a.m.-12 (12 noon-6 Night Primary purpose of call noon) p.m.) (6 p.m.-6 a.m.) Family, kin, and friends 35.6 36.8 27.6 Inquiry about health 39.4 34.9 25.7 Commerce 49.4 39.7 10.9 Agriculture and industry 47.3 35.9 16.8 Government administration 52.3 40.4 7.4 Professional services 50.8 40.5 8.7 Sports, religion, study, and other 36.6 38.2 25.3 Emergency services 46.7 30.4 22.8 No answer 36.2 40.0 23.8 a. Includes calls to discuss work opportunities. Source: World Bank data. Table C-21. Summary Characteristics of Calls for Three Rural Villages in Costa Rica, 1974 Puerto Santa Llano Call characteristics Cortes Clara Grande Total Call placed between 6 a.m.-12 noon 143 28 68 239 (37.5) (24.6) (38.4) (35.6) 12 noon-6 p.m. 140 32 54 226 (36.7) (28.1) (30.5) (33.6) 6 p.m.-6 a.m. 98 54 55 207 (25.7) (47.4) (31.1) (30.8) Outgoing call 352 73 114 539 (91.9) (64.0) (64.4) (80.0) Outgoing collect call 61 21 3 85 (17.3) (28.8) (2.6) (15.8) Call made on own behalf 359 113 160 632 (93.7) (99.1) (90.4) (93.8) Trip to town or Pco telephone location made only to place a call 346 113 172 631 (91.0) (99.1) (97.7) (94.0) Note: Numbers in parentheses are percentage of total calls. Source: World Bank data. A SURVEY IN COSTA RICA 409 one exception to the general pattern was in Santa Clara, perhaps be- cause staff and students of the technical school were in session during much of the daytime hours. Table C-21 also shows that outgoing calls comprised approximately 80 percent of all calls made in the three user-survey villages. For the eleven concessionaire-survey villages outgoing calls comprised 76 per- cent of all calls, perhaps because it is inconvenient to receive incom- ing calls at a village PCO. In the three-village sample, almost 16 percent of the outgoing calls were collect calls, and approximately 94 percent of the trips made by callers to the establishment in which the telephone was located were made specifically to use the telephone. In these villages at least, outgoing calls tended to be planned in advance. Factor Analysis of User Characteristics and Purposes of Use In the above sections, information on characteristics of persons using rural PCO telephones and the purposes of those calls was com- piled and reviewed on a two-dimensional basis. An attempt was also made to examine the relations within the user/usage survey data more thoroughly to understand the size and direction of the collinearity among the individual variables. The exercise involved a factor analysis of thirty-eight economic, demographic, and telephone use variables collected in the user survey of Puerto Cortes, Santa Clara, and Llano Grande. The results of this factor analysis were for the most part disappoint- ing in that they generally revealed only obvious or trivial relations. For example, if more calls were made by self-employed persons and businesspersons, fewer calls were made by wage earners. One of the few interesting results that emerged was that on the fourth vector in the three-village factor analysis and the third vector in the individual Puerto Cortes analysis the highly loaded variables suggested that call- ers who lived farther than the average distance from the telephone tended to take more time to reach the telephone and complete the call and therefore tended to make the trip not only for calling, but for other purposes as well. Furthermore, since they made these trips for multiple purposes, they tended to allocate a relatively small portion of the cost of the trip to the call.'3 Also the second vector in the three-village analysis and vectors one or two in each of the three individual village analyses indicated that 410 APPENDIXES call charges tended to be lower for outgoing calls. The meaning of this result is unclear, but the relation was consistent for all three vil- lages. It could indicate that call charges on incoming collect calls tended to be higher than average since it takes additional time to find and bring the person being called to the Pco, or that when someone takes the unusual step of calling a village Pco collect, the reasons for the call are sufficiently complex that the relevant information takes longer to convey. Summary of the Costa Rican Rural Pco Telephone Study In evaluating the priorities for investment in a rural telephone pro- gram, it is useful to understand the type and amount of benefits asso- ciated with alternative configurations of investment and the segments of the population and of the economy that are likely to benefit the most. This examination of the rural Pco telephone program in Costa Rica as it existed in 1974 and 1975 provided information relating to both of these questions. Evidence suggested that benefits to individuals using Pco tele- phones tended to be greatest in rural villages that possessed one or more of the following (sometimes collinear) characteristics:14 a. Village income per capita was higher than average b. The village had a relatively large population c. The village was located relatively far from the major economic, social, and government center of San Jos6 d. The educational level of the population was above average e. The population tended to be clustered more closely around the site where the telephone was located. Within villages telephone use appeared to be relatively widespread with no one small group dominating the others. There also appeared to be no major income bias among the callers; overall both higher- and lower-income villagers used the Pco. The purposes for which calls were made tended to be correlated with the economic base and demographic characteristics of the vil- lage. By far the most prominent purpose for which calls were made, however, was to maintain contact with family or friends, for whatever reasons. Hence, in addition to facilitating economic and government administrative activity, the rural public telephone program also im- proved the general quality of life of rural residents by allowing them to stay in touch and exchange information with family and friends. A SURVEY IN COSTA RICA 411 A smaller but significant proportion of calls were to discuss health matters with family and friends and for emergency purposes, and, on average, people tended to travel farther to make such calls. This sug- gests that the telephone's spatial zone of influence varied somewhat by the type of call. Finally, the most common substitute for the PCo telephone was the postal service. PCo telephones did, however, evidently generate new ac- tivity and communication. Overall, approximately 18 percent of the callers in the three user-survey villages stated that if the telephone had not been available, they would not have communicated with the person they called. Notes Note: The analysis and write-up presented in this appendix were done by Robert J. Saunders. Antonio Canas of icE directed the collection of telephone use data provided by ICE. Marco Tristan of the University of Costa Rica directed the tabulation of the census data and the collection of the survey data. Inputs to the initial study design and analysis were made by RI J. Saunders, S. C. Littlechild, M. Tristan, and J. J. Warford. Research assistance or comments from M. Gellerson, K. Challa, J. Lu, and B. Wellenius are also acknowledged. 1. Telephone traffic records for the previous twenty-four months for villages involved at least partly in tourism provided evidence that the positive price elasticities observed for sev- eral of these villages were likely the result of predictable seasonal variations resulting from tourism. 2. Such a calculation assumes that the distribution of calls according to destination and time of day remained unchanged. 3. The signs of the regressor or independent variable coefficients are the same for all three equations. 4. This conclusion is consistent with survey results from other countries reviewed in chapter 12. 5. Both of these explanations are also supported by the survey results outlined in chapter 12. 6. For a more detailed discussion of factor analysis, see Harmon (1967) and Anderson (1958). 7. The factors examined were interrelated so that it is not totally valid to speak of the ef- fects of any one of them in isolation. 8. This statement assumes that the volume of Pco use per capita is directly proportional to the size of benefits derived, which may not always be a valid assumption. 9. National Census of 1973 and Cespedes (1971). 10. The survey data on income for Santa Clara could be biased downward. The inter- viewers reported that the no-response rate was high partly because some of the technical school employees were reluctant to reveal their incomes. 11. It is possible that concessionaires making multiple telephone calls themselves partly brought about this result. 12. Although reduced call charges were supposed to be in effect betxveen 6 p.m. and 6 a.m., several concessionaires charged the same tariff throughout the twenty-four-hour per- 412 APPENDIXES iod and therefore made larger profits on night calls. The fact that most concessionaires closed their business at night also affected the number of night calls, since the pcos could not be used. 13. Example 21 in chapter 9 outlines an attempt to estimate some travel and time costs associated with rco telephone calls in Puerto Cortes. 14. A summary of these results was first published in Saunders and Warford (1977). Appendix D Shadow Pricing for Telecommunications THE BORDER PRICING APPROACH to shadow pricing, outlined by Little and Mirrlees and elaborated by Squire and van der Tak, is used here as a basis for explaining the relevance of shadow pricing for estimating telecommunications costs.' With the border pricing approach, the numeraire or unit of account is defined to be uncommitted public in- come at border prices, that is, basically a foreign exchange numeraire expressed in units of the local currency converted at the official rate of exchange. Therefore, directly imported inputs, such as cables or ex- change equipment, whose foreign exchange costs are known, are al- ready in border prices (converted at the official exchange rate). Locally purchased inputs, however, which are expressed in domestic market prices, need to be converted to border prices by an appropri- ate conversion factor.2 The ways in which commodities normally traded internationally (tradables) and those typically produced locally (nontradables) may be analyzed are summarized below. For tradables, which are those items for which there is a world sup- ply of imports and a global demand for exports, the cost, insurance, and freight border price for imports and the border price for exports may be used, with a suitable adjustment made for the marketing mar- gin. The free trade assumption is not required to justify the use of border prices, since domestic price distortions can in effect be ad- justed for by netting out the effect of all taxes, duties, subsidies, and so forth. In the telecommunications sector the most important tradables are capital goods, such as switching and transmission equip- ment, and subscriber apparatus and cables. 413 414 APPEN DIXES For nontradable commodities, primarily labor and land, the associ- ated border-priced marginal social cost (Msc) is the relevant resource cost, unless the input is supplied through the decreased consumption of other users, in which case the border-priced marginal social benefit (MSB) of this forgone consumption would be a more appropriate mea- sure of social cost. More generally, if both effects are present, a weighted average of MSC and MSB should be used, with the MSB tending to dominate in the short-run supply-constrained situation and the MSC being more important in the longer run when output may be ex- panded. The MSC may be determined by successively decomposing the nontradable into its constituent inputs, through several rounds, until all that remains are the ultimate tradable inputs, which can be valued at border prices, and nontraded primary factors such as labor and land, which may also be shadow priced as discussed below. The MSB for intermediate and final consumption should be evaluated in terms of the forgone social profit and consumer surplus, respectively, plus the income transfer effects of any changes in price. In practice, the MSC is often used because of data constraints. In the telecommunications sector the most important nontradable input is labor. The procedure for estimating the shadow wage rate (SWR) can be illustrated by considering the use of unskilled labor in a labor-surplus country, for example, workers employed in constructing buildings or laying cable. The forgone output of workers used in the telecommunications project is the dominant component of the SWR. Complications arise because the original income earned may not re- flect the marginal product of labor, and, further, for every new job cre- ated, more than one worker may give up his old employment. Seasonal activities, such as harvesting, must also be allowed for. In theory, if the laborer has to work harder in his new job than in his old one, then the disutility of forgone leisure must also be included in the SWR, but in practice this component is ignored. Overhead costs in- curred (for example, transport) should also be considered. The fore- going can be summarized by the following basic equation: SWR = a nm + c u where m and u are the forgone marginal output and overhead costs of labor in domestic prices, and a and c are corresponding factors for converting these values into border prices. These effects on consumption patterns also need to be considered. Suppose a worker receives a wage W. in his new job and that the in- come forgone is W0, both in domestic prices. (WO may not necessarily SHADOW PRICING 415 be equal to the marginal output forgone, m.) Assuming, as is likely for low-income workers (and also in the interests of simplicity), that all the increase in income (Wn - W0) is consumed, this increase in con- sumption will result in a resource cost to the economy of b (W'- WO), where b is the border-priced MSC of increasing consumption (in domestic prices) by one unit. The increased consumption also pro- vides a benefit given by w (W " - W0), where u! represents the MSB (in border prices) of increasing domestic-priced, private sector consump- tion by one unit. Therefore,3 SWR = a ni + c u + (b - w)(Wl, - WO). In the case of land inputs, the appropriate shadow value placed on this primary factor depends on location. In most cases, it is assumed that the market price of urban land is a good indicator of its eco- nomic value in domestic prices, and the application of an appropriate conversion factor, for example, the standard conversion factor, to this domestic price will yield the border-priced cost of urban land inputs. Rural land that has an alternative use in agriculture may be valued at its opportunity cost, that is, the net benefit of forgone agricultural output. In the case of telecommunications, this can usually be as- sumed to be negligible. The shadow price of capital may be reflected in the discount rate or accounting rate of interest (ARI), which is defined as the rate of de- cline in the value of the numeraire over time. An appropriate dis- count rate would be the opportunity cost of capital (occ), which refers to the rate of return to the economy as a whole resulting from an additional unit of public investment and may be used as a proxy for the ARI in the pure efficiency price model. A simple formula for ARI, which includes consumption effects, is ARI = OCC IS + (1 - s)U' / bl where s is the fraction of the yield from the original investment that will be saved and reinvested. Use of such an explicit shadow pricing framework would allow cor- rections for major distortions in the market mechanism. The shadow prices of inputs may then be incorporated into a given formula to yield the basic marginal costs in border prices. However, these border- priced marginal costs have to be divided by an appropriate conversion factor to arrive at the domestic-priced marginal costs, which may be used as a basis for determining the domestic price of telecommunica- tions services, as in several of the examples cited in chapters 8 and 9. 416 APPENDIXES Ideally, this conversion factor should depend on the alternative use of income by each type of consumer, but in practice some "average" proxy like b or the standard conversion factor is often used. Notes 1. Little and Mirrlees (1976) and Squire and van der Tak (1975). 2. For inputs that are not important enough to merit individual attention or that lack sufficient data, a standard conversion factor may be used. It is equal to the official ex- change rate divided by the more familiar shadow exchange rate. For a more detailed discus- sion of alternative interpretations of the shadow exchange rate, see Squire and van der Tak (1975). 3. The consumption term (b - w) disappears if, at the margin, (a) society is indifferent to the distribution of income (or corisumption), so that everyone's consumption has equiv- alent value; and (b) private consumption is considered to be as socially valuable as uncom- mitted public savings, that is, the numeraire. Appendix E World Bank Loans and Credits Table E-1. Loans and Credits b-v the WVorld Bank and the International Development Association for Telecommunications Projects, 1962-89 (millions of U.S. dollars) Loan or Total Source of credit project Fiscal year and country funding amount cost 1962 Ethiopia Bank 2.9 6.2 1963 India IL.A 42.0 122.0 1964 Costa Rica Bank 9.9 12.6 El Salvador Bank 9.5 13.6 Subtotal 19.4 26.2 1965 India ILDA 33.0 228.0 1966 Venezuela Bank 37.0 100.0 Ethiopia Bank 4.8 10.8 Subtotal 41.8 110.8 1967 Jamaica Bank 11.2 18.3 East African Community Bank 13.0 26.7 Colombia Bank 16.0 27.6 Subtotal 40.2 72.6 1968 China, Rep. of Bank 17.0 50.0 Singapore Bank 3.0 9.5 Papua New Guinea Bank 7.0 15.4 Subtotal 27.0 74.9 417 418 APPENDIXES 1969 Malaysia Bank 4.4 49.0 Upper Volta IDA 0.8 1.2 Pakistan IDA 16.0 42.0 Ethiopia Bank 4.5 25.4 India Bank/TDA 55.0 361.0 Subtotal 80.7 478.6 1970 Costa Rica Bank 6.5 9.5 Nepal IDA 1.7 4.2 Singapore Bank 11.0 37.1 Yugoslavia Bank 40.0 95.0 Pakistan IDA 15.0 35.3 East African Community Bank 10.4 28.3 Subtotal 84.6 209.4 1971 Indonesia IDA 12.8 22.1 Iran Bank 36.0 149.2 India IDA 78.0 290.7 Colombia Bank 15.0 39.1 Malaysia Bank 18.7 94.0 Venezuela Bank 35.0 294.5 Subtotal 195.5 889.6 1972 Iraq Bank 27.5 39.7 Guatemala Bank 16.0 21.1 Costa Rica Bank 17.5 32.2 El Salvador Bank 9.5 12.7 Fiji Bank 2.2 5.7 Mali IDA 3.6 4.3 Subtotal 76.3 115.7 1973 Papua New Guinea Bank 10.0 17.2 Bangladesh IDA 7.3 12.1 Thailand Bank 37.0 102.8 Senegal Bank 6.3 8.9 Nepal IDA 5.5 7.9 East African Community Bank 32.5 53.3 India IDA 80.0 534.1 Subtotal 178.6 736.3 1974 Iran Bank 82.0 194.0 Upper Volta IDA 4.5 5.6 Ethiopia IDA 21.4 37.1 Cote d'Ivoire Bank 25.0 53.6 Costa Rica Bank 23.5 54.2 Trinidad and Tobago Bank 18.0 30.5 Bangladesh IDA 20.0 87.0 Subtotal 194.4 462.0 WORLD BANK LOANS AND CREDITS 419 1975 Pakistan IDA 36.0 67.5 Colombia Bank 15.0 52.1 Guatemala Bank 26.0 45.9 Egypt IDA 30.0 173.4 Ethiopia IDA 16.0 60.6 Ghana Bank 23.0 29.5 Burma IDA 21.0 30.9 Zambia Bank 32.0 78.2 Subtotal 199.0 538.1 1976 Fiji Bank 5.0 14.0 Syria Bank 28.0 145.6 Thailand Bank 26.0 146.1 Niger IDA 5.2 6.5 Subtotal 64.2 312.2 1977 India Bank 80.0 415.0 Colombia Bank 60.0 167.7 Subtotal 140.0 582.7 1978 Lebanon Bank 14.5 33.7 Egypt IDA 53.0 210.0 Costa Rica Bank 10.6 94.1 El Salvador Bank 23.0 51.6 India Bank 120.0 818.5 Subtotal 221.1 1,207.9 1979 Nepal IDA 14.5 25.2 Thailand Bank 90.0 307.2 Kenya Bank 20.0 63.5 Subtotal 124.5 395.9 1980 Burma IDA 35.0 93.0 Sri Lanka IDA 30.0 36.3 Subtotal 65.0 129.3 1981 Burundi IDA 7.7 9.1 Rwanda IDA 7.5 17.5 Oman Bank 22.0 97.2 Colombia Bank 44.0 110.0 Inidia IDA 314.0 1,619.4 Subtotal 395.2 1,853.2 420 APPENDIXES 1982 Uruguay Bank 40.0 204.8 Tanzania IDA 27.0 47.0 Cameroon Bank 7.5 11.1 Pakistan IDA 40.0 266.1 Mali IDA 13.5 25.3 Egypt Bank 64.0 141.3 Upper Volta IDA 17.0 40.6 Thailand Bank 142.1 492.2 Kenya Bank 44.7 117.9 Subtotal 395.8 1,346.3 1983 Bangladesh IDA 35.0 61.8 Uganda IDA 22.0 26.0 Subtotal 57.0 87.8 1984 Algeria Bank 128.0 312.4 Guatemala Bank 30.0 208.0 Thailand Bank 8.5 Subtotal 166.5 520.4 1985 Ethiopia IDA 40.0 151.8 Kenya Bank 32'6 86.1 Nepal ILIA 22.0 65.3 Oman Bank 23.0 227.8 Philippines Bank 4.0 4.8 Subtotal 121.6 535.8 1986 Cote d'lvoire Bank 24.5 116.7 Laos IDA 3.9 7.8 Senegal IDA 22.0 156.9 Subtotal 50.4 281.4 1987 Burundi ICDA 4.8 25.4 Hungary Bank 70.0 833.1 India Bank 345.0 2,050.0 Indonesia Bank 14.5 17.9 Morocco Bank 125.0 674.5 Pakistan Bank 100.0 817.2 Tanzania IDA 23.0 60.0 Subtotal 682.3 4,478.1 1988 Jordan Bank 36.0 338.1 WORLD BANK LOANS AND CREDITS 421 1989 Benin IDA 16.0 65.3 Ecuador Bank 45.0 330.0 Fiji Bank 8.1 47.9 Ghana IDA 19.0 173.0 Togo IDA 16.0 44.6 Uganda IDA 52.3 58.8 Western Samoa IDA 4.6 16.3 Subtotal 161.0 735.9 Total" 3,933.0 16,875.4 a. Does not include four small loans totaling $24.25 million made before 1962. Source: World Bank data. Appendix F The Canadian Radio-Television and Telecommunications Commission Costing Methodology: Cost Allocation Categories THE EIGHT BROAD CATEGORIES established by the Canadian federal tele- communications regulatory agency, the CRTC, in 1989 as part of the Cost Inquiry Phase III are presented here in detail. Although the re- quirements for an appropriate approach to the situation in each country are necessarily different, this case illustrates how costs can be allocated in order to analyze a telecommunications tariff (see chapter 15 for a complete discussion of costing methodologies). Access (A) The facilities associated with Access and their associated costs are composed of three separate components: the subscriber premises equipment, the loop, and the serving central office equipment. Subscriber Premises Equipment Costs related to the equipment and connections provided on a mo- nopoly basis pursuant to "Attachment of Subscriber-Provided Termi- 422 CRTC COSTING 423 nal Equipment," TELECOM Decision CRTC 82-14, November 23, 1982, are included in the Access category. In some cases, costs associated with terminal equipment giving the subscriber access to a service that can only be provided by the carrier are included in this category. A specific example is the private branch exchange line card required by a centralized system for reporting emergencies. All other subscriber premises equipment and connection costs are included under the categories of Competitive Network or Competitive Terminal. Unless otherwise specified and affirmed by the carriers, the boundary between the Access and Competitive Terminal categories is the demarcation point on the customer's premises. Loop The loop includes all costs associated with (a) the provision by the company, for any purpose, of a facility connecting the customer's premises with the nearest serving central office and (b) the termina- tion and protection of that facility at both the customer's premises and the serving central office. These interconnecting facilities normally consist of outside plant but also include related equipment required to derive the intercon- necting capability, such as subscriber line carrier systems and line concentrator equipment. The interconnection may also be provided by a radio system, in which case the loop includes all costs associated with the radio base station equipment and facilities connecting the serving central office. Serving Central Office Equipment Costs associated with all central office switching equipment not sensitive to traffic but needed to give customers access to the public switched telephone network are included under Access. Costs associated with equipment located in a central office re- quired specifically to derive the loop facility, such as subscriber line carriers and line concentrators, are included under Access, while those associated with equipment for specific services, such as circuit conditioning and signaling equipment, are included as appropriate under the categories of Monopoly Local, Monopoly Tool, or Compet- itive Network. 424 APPENDIXES Monopoly Local (ML) The Monopoly Local service category includes costs related to the provision, operation, and maintenance of the local switching equip- ment and interoffice transmission facilities that are required to estab- lish and maintain communication services within the local calling area and that generally are not provided by another supplier. Monopoly Toll (MT) The Monopoly Toll service category includes costs related to the provision, operation, and maintenance of the switching equipment and transmission facilities that are required to establish and maintain communication services between local calling areas and that generally are not provided by another supplier. Competitive Network (CN) The Competitive Network category includes costs related to the provision, operation, and maintenance of the facilities that are re- quired to establish and maintain communication services and that are, or can be, provided by another supplier, with the exception of the facilities included under Access. Costs associated with terminal equipment (equipment that is spe- cific to competitive network service and that is provided by the com- pany as an integral part of the service) are included in this category. Competitive Terminal-Multiline and Data (CT-MD) As specified in Decision CRTC 86-5, costs and revenues associated with the Competitive Terminal category are subdivided into two dis- tinct service categories: Competitive Terminal-Multiline and Data; and Competitive Terminal-Other. The CT(MD) category includes costs associated with key telephone systems, private branch exchange systems, and all telephone sets be- hind key and private branch exchange systems, including those be- hind Centrex systems subject to the final classification of proprietary sets and consoles. Also included are costs associated with all data ter- CRTC COSTING 425 minal equipment at the customer's premises that is not integral to the operation of the channel provided by the carrier. This category also includes costs associated with inside wiring on the customer's side of the service demarcation point that is associated with multiline and data terminals. Costs associated with terminal equipment that may be located on a customer's premises but that is provided by the company as an integral part of a network service shall be included as appropri- ate under the categories of Competitive Network, Monopoly Toll, or Monopoly Local. Competitive Terminal-Other (CT-O) Costs associated with terminal equipment other than those identi- fied above are included in the Competitive Terminal-Other cate- gory. This category consists primarily of costs associated with single-line telephone sets. Other (o) The Other service category includes the causal costs associated with activities and services that, in general, do not relate directly to the provision of telecommunications services and are not included in other service categories. Examples include the following: * Building space rented to others . Subscription charges for tariffs . Arrangements for cable television lessees . Billing service arrangements for CNCP * Communications seminars . Data processing for BCE Common (c) The Common category includes costs that cannot be causally re- lated to a particular category of service and that, with the support of empirical evidence, are shown to be fixed. References The word "processed" describes works that are reproduced from typescript by mimeograph, xerography, or similar means; such works might not be cataloged or commonly available through libraries or might be subject to restricted circulation. African Medical and Research Foundation. "AMREF Reports: Progress in Develop- ment of Various AMREF Projects for the Period January-June 1979." Nairobi, Kenya, 1979. Processed. Alleman, J. H. The Pricing of Local Telephone Services. Special Publication no. 77- 14. Washington, D.C.: U.S. Department of Commerce, Office of Telecommu- nications, April 1977. Alleman, James H., and Richard D. Emmerson, eds. Perspectives on the Telephone Industry: The Challenge for the Future. New York: Harper and Row Publishers, Ballinger Division, 1989. Ambrose, W., P. R. Hennemeyer, and J. P. Chapon. "Privatizing Telecommunica- tions Systems: Business Opportunities in Developing Countries." Discussion Paper no. 10. International Finance Corporation, Washington, D.C., 1990. American Telephone and Telegraph Corporation. The World's Telephones. Morris Plains, N.J.: AT&T Long Lines Overseas Department, August 1980, June 1981, and July 1982. . The World's Telephones: January 1987-1988. Whippany, N.J., 1989. T The World's Telephones: January 1989. Whippany, N.J., 1990. Analysys, Ltd. "Final Report of Phase 2." Prepared for CEC DGXIIIF ORA Program. Cambridge, Eng., 1990. Processed. ____ . A Study of the Economic Implications of Stimulating Applications of IT6eT in Rural Areas, 2 vols. CEC DGXIIIF ORA Program. Cambridge, Eng., 1989. Processed. Anderson, T. W. An Introduction to Multivariate Statistical Analysis. New York: John Wiley and Sons, 1958. Anonymous. "Restructuration du secteur des telecommunications dans Niger." Paper presented at a seminar on telecommunications restructuring sponsored by the World Bank, the ITU, and ACCT, Tunisia, May 1992. Arrow, Kenneth J. "The Economics of Information." In Michael L. Dertouzos and Joel Moses, eds., The Computer Age: A Twenty-Year View. Cambridge, Mass.: M.I.T Press. 1980. 426 REFERENCES 427 . "Political and Economic Evaluation of Social Effects and Externalities." In Julius Margolis, ed., The Analysis of Public Output. New York: National Bu- reau of Economic Research, Columbia University Press, 1970. Ashby, Jacqueline, Steve Klees, Douglas Pachico, and Stuart Wells. "Alternative Strategies in the Information/Education Projects." In Emile G. McAnany, ed., Communications in the Rural Third World. New York: Praeger Publishers, 1980. Asian Development Bank. Distance Education in Asia and the Pacific, 2 vols. Ma- nila, 1987. Baer, Walter S. "Defining Productivity Gains: Telecommunications and Produc- tivity." Telecommunications Policy (December 1981), pp. 329-30. Baeza, S., J. Bunster, and 0. Schenone. "Evaluaci6n econ6mica de la inversi6n en telfonos p6blicos de larga distancia en zonas rurales." Documento de Trabajo no. 37. Instituto de Economia, Universidad Cat6lica de Chile, November 1975. Processed. Balson, David, Robert Drysdale, and Bob Stanley. Computer Based Conferencing Systems for Developing Countries. Report of a workshop held in Ottawa, Can- ada, October 26-30, 1981. Ottawa: International Development Research Centre, 1982. Barry, Mamadou Pathe, and Mohamed Sylla. "Le cas de la Guinee." Paper pre- sented at a seminar on telecommunications restructuring sponsored by the World Bank, the ITU, and ACCT, Tunisia, May 1992. Bates, Tony. "Options for Delivery Media." In Hilary Perraton, ed., Alternative Routes to Formal Education: Distance Teaching for School Equivalency. Balti- more, Md.: Johns Hopkins University Press, 1982. Baumol, William J. "Reasonable Rules for Rate Regulation: Plausible Policies for an Imperfect World." In A. Phillips and 0. Williamson, eds., Prices: Issues in Theory, Practice, and Public Policy. Philadelphia: University of Pennsylvania Press, 1968. Baumol, William J., and David F. Bradford. "Optimal Departures from Marginal Cost Pricing." American Economic Review, vol. 60 Uune 1970), pp. 265-83. Bebee, E. L., and E. T W. Gilling. "Telecommunications and Economic Develop- ment: A Model for Planning and Policy Making." Telecommunications Journal, vol. 43, no. 8 (August 1976), pp. 537-43. Beesley, M. E. "The Value of Time Spent Travelling: Some New Evidence." Economica, vol. 32 (May 1965), pp. 174-85. Beesley, M. E., and S. C. Littlechild. "The Regulation of Privatized Monopolies in the United Kingdom." Rand Journal of Economics, vol. 20, no. 3 (Autumn 1989). Beier, G., A. Churchill, M. Cohen, and B. Renaud. "The Task for the Cities of the Developed Countries." World Bank Staff Working Paper no. 209. Washing- ton, D.C., July 1975. Bell Canada. Cost Inquiry Phase III Manual. Hull, Quebec, 1989. Bethesda Research Institute. "Regulatory Systems for the 1990s: A Comparative Analysis of Price Caps and Rate of Return Regulation." Bethesda, Md., 1988. Processed. 428 Bigham, Fred G. "Integrating Telecommunications Costing by Class of Service with Incentive Regulation." Paper presented at NARUC Advanced Regulatory Studies Program, Williamsburg, Va., February 1990. Processed. Bigham, Fred G., and G. W. Wall. "A Canadian Approach to the Determination of Broad Categories of Revenues and Costs." Paper presented at the confer- ence on telecommunications costing in a dynamic environment, San Diego, Calif., 1989. Processed. Blair, Michael L. "VSAT Systems in Developing Countries." Proceedings of the Pa- cific Telecommunications Conference. Honolulu, Hawaii, February 1988. Block, Clifford. "Satellite Linkages and Rural Development." In Heather E. Hudson, ed., New Directions in Satellite Communications: Challenges for North and South. Norwood, Mass.: Artech House, 1985. Block, Clifford, Douglas Goldschmidt, Anwar Hafid, Gerald C. Lalor, and Angel Velasquez. "Satellite Telecommunications for Rural Development: The AID Rural Satellite Program and Its Projects in Indonesia, Peru, and the Carib- bean." Proceedings of the Pacific telecommunications conference, Honolulu, Hawaii, January 1984. Booz Allen and Hamilton. "A Microeconomic Study of the Benefits of Improved Telephone Service in Selected Areas of the Philippines, 1984." London, Eng. Processed. . "Study of the Economic Benefits of New Telecommunications Services in Costa Rica, 1986." London, Eng. Processed. Boiteux, M. "La tarification des demandes en pointe." Revue Generale d'Electricite, vol. 58 (1949), pp. 321-40. Translated as "Peak Load Pricing." Journal of Business, vol. 33 (April 1960), pp. 157-79. Bower, L. B. "Telecommunications Market Demand and Investment Require- ments." TelecommunicationsJournal, vol. 39 (November 3, 1972), pp. 172-81. Bowers, D. A., and W. F. Lovejoy. "Disequilibrium and Increasing Costs: A Study of Local Telephone Service." Land Economics (February 1965), pp. 36-38. Bradshaw, Ted K. "Rural Development and Telecommunications Potential and Policy." Working Paper no. 524. Berkeley: Institute of Urban and Regional De- velopment, University of California, October 1990. Bruce, Robert R. "Franchising and Subcontracting for Services and Facilities: New Options for Attracting New Sources of Investment." Roundtable on East- ern European telecommunications, Badacsonytomaj, Hungary, May 1991. Bruce, Robert R., Jeffrey P. Cunard, and Mark Director. The Telecom Mosaic: As- sembling the New International Structure. London: International Institute for Communications, 1988. Bryan, Elizabeth B., and David A. Evans. "Access to Calls." Background papers for a seminar on social research and telecommunications planning, TELECOM, Australia, August 1979. Processed. Burkhead, Jesse, and Jerry Miner. Public Expenditure. Chicago: Aldine Publishing Company, 1971. Carpenter, Trevor C. "The Rural Postal Services of Scotland." Union Postale, vol. 5 (1974), pp. 66-68. REFERENCES 429 Carsberg, Sir Bryan. "The Liberalization of Telecommunications: Experience in the United Kingdom." Roundtable on Eastern European telecommunications, Badacsonytomaj, Hungary, May 1991. . "Regulating Private Monopolies and Promoting Competition." Long Range Planning, vol. 19, no. 6 (1986). Casasus, Carlos. "Privatization of Telecommunications: The Case of Mexico." In Bjorn Wellenius and Peter Stern, eds., Imnplementing Reforms in the Telecommu- nications Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Castilla, Adolfo, Maria Cruz Alonso, and Carlos Tirado. Telecomunicaciones y desarrollo en Espana e lberoamerica. Madrid: FUNDESCO, 1989. ccirr. Blue Book, vol. 2. Geneva: ITU, 1965. ____ . GAS-5 Handbook: Economic Studies at the National Level in the Field of Telecommunications. Geneva: ITU, 1968. . GAS-5 Handbook: Economic Studies at the National Level in the Field of Telecommunications. Geneva: ITU, 1976. ____ . National Telephone Networks for the Automatic Service, Third Plenary As- sembly, Geneva, 1964. Geneva: ITU, 1965. . "Optimum Allocation and Use of Scarce Resources in Order to Meet Telecommunications Needs in Urban or Rural Areas of a Country." C>AS-5 Eco- nomic Studies (1981-84), no. 8. Geneva: ITU, 1984a. _ . "Socio-Economic Aspects of Telecommunications Development in Iso- lated and/or Underprivileged Areas of Countries." C.AS-5 Economic Studies (1981-1984), no. 4. Geneva: iTu, 1984b. Center for Science and Technology for Development. ATAS Bulletin: New Informa- tion Technologies and Development, issue 3. New York: United Nations, June 1986. Cespedes, V. H. Costa Rica: Income Distribution and Consumption of Some Foods. San Jose: University of Costa Rica, 1971. Chaffar, Y. A. "Some Aspects of Postal Administrations' Role in Financing Devel- opment Plans." In "Proceedings of the 1973 Postal Executives' Seminar, Tokyo, March 1-20, 1973." Universal Postal Union, Tokyo, 1974. Processed. Chandler, A. D. The Visible Hand: The Managerial Revolution in American Busi- ness. Cambridge, Mass.: Harvard University Press, 1977. Chan-Kil, Chung. "Evaluation Study on Rural Telephone Expansion Project under IBRD Loan." Korea Rural Economics Institute, Seoul, December 1979. Processed. Chapuis, Robert J. 100 Years of Telephone Switching. Part 1, Manual and Electro- mechanical Switching, Studies in Telecommunications vol. 1. Amsterdam: North-Holland, 1982. . "Telephony Is a Heavy Industry." Telecommunications Journal, vol. 42-X (1975). Chapuis, Robert J., and A. Joel. Electronic Computers and Telephone Switching: A Book of Technological HistorY. Studies in Telecommunication no. 13. New York: Elsevier Science Publishers/North-Holland, 1990. 430 Charles, Jeff. "Approaches to Teleconferencing Justification: Towards a General Model." Telecommunications Policy (December 1981), p. 297. Christaller, Walter. Central Places in Southern Germany. Translated by Carlisle W. Baskin. Englewood Cliffs, N.J.: Prentice-Hall, 1966. Chu, Godwin C., Alfian Chote Srivisal, and Boonlert Supadhiloke. "Rural Tele- phone in Indonesia and Thailand." Telecommunications Policy (June 1985), pp. 159-69. Clark, David, and Kathryn Unwin. "Telecommunications and Travel: Potential Impact in Rural Areas." Regional Studies, vol. 15 (1981), pp. 47-56. Cleevely, D. D. "Modelling the Role of Telecommunications in Regions of a De- veloping Country." Working Paper. Department of Engineering, Division of Control and Management Systems, University of Cambridge, Cambridge, Eng., February 1979. Processed. Cleevely, D. D., and G. Walsham. "Interim Report on Modelling the Role of Telecommunications within Regions of Kenya." Department of Electrical Engi- neering, University of Cambridge, Eng., May 1980. Processed. Clippinger, J. H. "Can Communication Development Benefit the Third World?" Telecommunications Policy (September 1977). Cole, Barry, ed. After the Break-Up: Assessing the New Post-AT&r Divestiture Era. New York: Columbia University Press, 1991. Communications Studies and Planning International. "The Impact of Telecom- munications on the Performance of a Sample of Business Enterprises in Kenya." Research report commissioned by the ITU. New York, August 1981. Processed. _____. "A Study of the Economic Benefits of Improved Telephone Service in the Philippines." Information, Telecommunications, and Development. Geneva: ITU, 1986. Communications Studies Group. "The Effectiveness of Person-to-Person Tele- communications Systems." Long Range Research Report no. 003/ITF. Post Of- fice Telecommunications, London, May 1975. Processed. Contreras, Eduardo. "Brazil and Guatemala: Communications, Rural Modernity, and Structural Constraints." In Emile G. McAnany, ed., Communications in the Rural Third World. New York: Praeger, Publishers, 1980. Courville, Leon, Alain De Fontenay, and Rodney Dobell, eds. Economic Analysis of Telecommunications: Theory and Applications. New York: North-Holland, 1983. Cowhey, P. "The Political Economy of Telecommunications Reform." In Bjorn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunica- tions Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Crandall, Robert W. After the Breakup: IJ.S. Telecommunications in a More Com- petitive Era. Washington, D.C.: Brookings Institution, 1991. Crandall, Robert W., and Kenneth Flamm, eds. Changing the Rules: Technological Change, International Competition, and Regulation in Communications. Wash- ington, D.C.: Brookings Institution, 1989. REFERENCES 431 Crockett, Bruce. "The Changing Role of Satellites in Telecommunications." Tele- communications (June 1989), pp. 37-45. CRTC (Canadian Radio-Television and Telecommunications Commission). "Re- port of the Inquiry Officer with Respect to the Inquiry into Telecommunica- tions Carriers' Costing and Accounting Procedures: Phase 111-Costing of Existing Services." 1984. Processed. Culham, P. G. "A Method for Determining the Optimal Balance of Prices for Telephone Services." OFTEL Working Paper no. 1. OFTEL, London, March 1987. Cyert, R. M., and J. G. March. A Behavioral Theory of the Firm. Englewood Cliffs, N.J.: Prentice-Hall, 1963. Demac, Donna, and Joseph Pelton, eds. Telecommunications for Development: Ex- ploring New Strategies-An International Forum. Washington, D.C.: INTELSAT, 1986. Department of Trade and Industry, United Kingdom. "Competition and Choice: Telecommunications Policy for the 1990s-A Consultative Document." HMSO, London, 1990. Deutches Institut fur Wirtschaftforschung. Economic Evaluation of the Impact of Telecommunication Investment in the Communities. Berlin, 1984. Development Communication Report, vol. 63, no. 4 (1988). Director General of Telecommunications, OFTEL. "The Regulation of British Telecom's Prices-A Consultative Document." London, 1988. Processed. . "Responses to OFTEL'S Consultative Document on the Future Regulation of British Telecom's Prices." OFTEL Working Paper no. 3. London. Dodds, Tony. "The Mauritius College of the Air." In Hilary Perraton, ed., Alter- native Routes to Formal Education: Distance Teaching for School Equivalency. Baltimore, Md.: Johns Hopkins University Press, 1982. Donald, Gordon. Credit for Small Farmers in Developing Countries. Boulder, Colo.: Westview, 1976. Dorozynski, Alexander. Doctors and Healers. Ottawa: International Development Research Centre, 1975. Drewer, T M. "Econometric Models of International Communications." Report no. 28. British Post Office, Statistics and Business Research Department, Lon- don, 1973. Processed. Economics Study Cell, Posts and Telegraphs Board, Ministry of Communica- tions, India. "India's Rural Telephone Network." New Delhi, May 1981. Processed. Elbert, Bruce R. International Telecommunication Management. Boston, Mass.: Artech House, 1990. Engvall, Lars. "A Socio-Economic Study on Usage of Telephone Services." In ITU, Information, Telecommunications, and Development. Geneva, 1986. Estabrooks, Maurice F., and Rodolphe H. Lamarche, eds. Telecommunications: A Strategic Perspective on Regional, Economic, and Business Development. Selected papers from a conference in Ottawa, Canada. Moncton, Canada: The Cana- dian Institute for Research on Regional Development, 1986. 432 Evans, D. A., and E. B. Bryan. "Access to Calls. A Perspective on the Economics of Telecommunications in Papua New Guinea." Study sponsored by the Papua New Guinea Department of Posts and Telegraphs and the World Bank. The Implementation and Management Group Party, Ltd., Sydney, Australia, De- cember 1977. Processed. Evans, John R., Karen Lashman Hall, and Jeremy Warford. "Shattuck Lecture- Health Care in the Developing World: Problems of Scarcity and Choice." New England Journal of Medicine, vol. 305 (November 1981), pp. 1117-27. Fargo, Dan S. "International Telecom Spending Slows Down." Telephonm (Febru- ary 22, 1982), pp. 78-85. Favout, R. "Improving Urban Highway Transportation through Electronic Route Guidance." In D. Brand, ed., Urban Transportation Innovation. New York: American Society of Civil Engineers, 1970. Federal-Provincial-Territorial Task Force on Telecommunications, Ministry of Supply and Services, Canada. "The Effect of Changing Technology on the Structure of Costs for the Provision of Public Long-Distance Telephone Ser- vice." Consulting Report no. 3. (December) 1988a. Processed. ____. "The Effect of Toll Competition on Prices, Costs, and Productivity of the Telephone Industry in the United States." Prepared by David Chessler and Associates, 1988b. Processed. Flamm, Kenneth. "Technological Advance and Costs: Computers versus Com- munications." In Robert W. Crandall and Kenneth Flamm, eds., Changing the Rules: Technological Change, International Competition, and Regulation in Com- munications. Washington, D.C.: Brookings Institution, 1989. Fordyce, S. W. "NASAR Experience in Telecommunications as a Substitute for Transportation." U.S. National Aeronautics and Space Administration Head- quarters, Washington, D.C., April 1974. Processed. Fox, R. W. B. "A Review of the Short-term Forecasting Model for Total Trunk Calls." Report no. 27. British Post Office, Statistics and Business Research De- partment, London, 1973. Processed. Frantz, R. X-Efficiency: Theory, Evidence, and Applications. Boston, Mass.: Kluwer Academic Publishers, 1988. Fuller, Wayne. RFD: The Changing Face of Rural America. Bloomington: Indiana University Press, 1964. Gellerman, Robert F. "Subscriber Financing of Telecommunications Investment." Telecommunications Policy, vol. 10, no. I (March 1986). Gellerman, Robert F., and S. Ling. "Linking Electricity with Telephone Demand Forecast: A Technical Note." IEEE Transactions on Communications, vol. coM-24, no. 3 (March 1976), pp. 322-24. General Motors Research Laboratories and Delco Radio Division. "A Design for an Experimental Route Guidance System." Prepared for the Bureau of Public Roads, Federal Highway Administration, Contract no. FM-11-6626. 1962. Processed. Giaoutzi, M., and P. Nijkamp, eds. Informatics and Regional Development. Aldershot, Eng.: Avebury, 1988. REFERENCES 433 Gille, Laurent. "Growth and Telecommunications." In ITU, Information, Telecom- munications, and Development. Geneva, 1986. Gillespie, Andrew E., John B. Goddard, Mark E. Hepworth, and Howard Williams. "Information, Communications Technology, and Regional Develop- ment: An Information Economy Perspective." Science, Technology, Industry Re- view, OECD, no. 5 (April 1989), pp. 85-111. Gillick, David. "Telecommunications Policy in the U.K.: Myths and Realities." Telecommunications Policy. vol. 15, no. 1 (February 1991), pp. 3-9. Globerman, Steven. "Economic Factors in Telecommunications Policy and Regu- lation." Paper presented at IRPP conference on competition and technological change: the impact on telecommunications policy and regulation in Canada, Toronto, 1984. Processed. Goddard, J., and R. Pye. "Telecommunication and Office Location." In R. C. Smith, ed., "Impacts of Telecommunications on Planning and Transport." Re- search Report no. 24. Department of Transport and Department of the Envi- ronment, London, July 1978. Processed. Goldschmidt, Douglas. "Financing Telecommunications for Rural Development." Telecommunications PoliLy (September 1984), pp. 181-203. _ "Telephone Communications, Collective Supply, and Public Goods: A Case Study of the Alaskan Telephone System." Ph.D. diss., University of Penn- sylvania, Philadelphia, 1978. Goldschmidt, Douglas, Victor Forsythe, and Heather Hudson. "An Evaluation of the Medex/Guyana Two-Way Radio Pilot Project." Academy for Educational Development, Washington, D.C., December 1980. Processed. Goldschmidt, Douglas, Heather E. Hudson, and Wilma Lynn. "Two-Way Radio for Rural Health Care: An Overview." U.S. Agency for International Develop- ment, Office of Education, Development Support Bureau, Washington, D.C., May 1980. Processed. Goldschmidt, Douglas, Karen Tietjen, and Willard D. Shaw. Design and Installa- tion of Rural Telecommunications Networks: Lessons from Three Projects- Technology. Washington, D.C.: Academy for Educational Development, AID Rural Satellite Program, 1987. Golladay, Frederick. "Health Problems and Policies in the Developing Countries." World Bank Staff Working Paper no. 412. Washington, D.C., August 1980. Gordon, D. A., and H. C. Wood. "How Drivers Locate Unfamiliar Ad- dresses-An Experiment on Route Finding." Public Roads (une 1970), pp. 44-47. Gorelik, M. A., and 1. B. Efimova. "The Economic Efficiency of Development of Long Distance Telephone Communication." Vestnik Svvazi, Moscow, no. 5 (1977), pp. 28-30. Gorelik, M. A., 1. B. Efimova, and E. Kareseva. "How to Determine the Eco- nomic Efficiency of Long Distance Telephone Communication." Vestnik Svwazi, Moscow, no. 7 (1975), pp. 26-28. Gorelik, M. A., and E. Kareseva. "Standards and Assessment of the Economic 434 Efficiency of Long Distance Telephone Communications." Vestnik Svyazi, Moscow, no. 8 (1975), pp. 30-31. Green, James H. The Dow Jones-Irwin Handbook of Telecommunications. Home- wood, Ill.: Dow Jones-Irwin, 1986. Greenhut, M. L. Microeconomnics and the Space Economy. Chicago: Scott Foresman Company, 1963. _. A Theory of the Firm in Economic Space. New York: Appleton-Century- Crofts, 1970. Guiscard Ferraz, J. L. "The Infrastructural Role of the Telecommunications Sec- tor and Its Impact on Economic Development." In ITU, World Telecommunica- tions Forum 1987, Part IV. Geneva, 1987. Hagget, P. Locational Analysis in Human Geography. London: Edward Arnold, 1965. Halina, J. W. "Communications and the Economy: A North American Perspec- tive.' International Social Science Journal, vol. 32, no. 2 (1980), pp. 266-67. Hall, Peter. "Moving Information: A Tale of Four Technologies." In J. Brotchie, M. Batty, and Peter Hall, eds., New Technologies and Spatial Systems. London: Unwin Hyman, 1990. . "Transportation." Urban Studies, vol. 6, no. 3 (1969), pp. 408-35. Hansen, Suella, David Cleevely, Simon Wadsworth, Hilary Bailey, and Oliver Bakewell. "Telecommunications in Rural Europe: Economic Implications." Tele- communications Policy (June 1990), pp. 207-22. Hardy, Andrew Peter. The Role of the Telephone in Economic Development. Stanford, Calif.: Institute for Communications Research, Stanford University, January 1980. Harkness, R. C. "Communications Substitutes for Intra-Urban Travel." Joint ASCE-ASME Transportation Engineering Meeting, Reprint no. 1453. Seattle, Wash., 1971. Processed. Harmon, Harry H. Modern Factor Analysis, 2d ed. Chicago: University of Chi- cago Press, 1967. Hartley, Nicholas M., and Peter Culham. "Telecommunications Prices under Mo- nopoly and Competition." Oxford Review of Economic Policy, vol. 4 (summer), no. 2 (1988), pp. 1- 19. Hazelwood, A. "Optimum Pricing as Applied to Telephone Service." Review of Economic Studies, vol. 18 (1950-51), pp. 67-78. Reprinted with amendments as "Telephone Service." In Ralph Turvey, ed., Public Enterprise: Selected Read- ings. Harmondsworth, Eng.: Penguin Books, 1968. Helling, A. G. Henrik. "Tariffs for National and International Telecommunica- tions Services." Asia Pacific Telecommunity, Bangkok, March 10, 1981. Processed. Heymann, H. J. "The Relationship between Telecommunications and the Na- tional Economy: A Pragmatic Approach to Its Quantitative Determination Using a Macro-Economic Cross-Sectional Analysis." In ITU, World Telecommu- nications Forum 1987, Part IV. Geneva, 1987. REFERENCES 435 Hills, Alex. "Feasibility of Mobile Satellite Service in Alaska." Telecommunications Policy (December 1988), pp. 369-78. Hills, Jill. "The Telecommunications Rich and Poor." Third World Quarterly, vol. 12, no. 2 (April 1990), pp. 71-90. Hobday, Mike. "Telecommunications-A 'Leading Edge' in the Accumulation of Digital Technology? Evidence from the Case of Brazil." Information Technology for Development, Oxford University Press, vol. 1, no. 1 (1986). Hoffman, Kurt, and Michael G. Hobday. "The Third World and U.S. Telecom- munications Policy." In Paula R. Newberg, ed., New Directions in Telecommuni- cations Policy, vol. 2. Durham, N.C.: Duke University Press, 1989. Hollas, D. F., and R. S. Hereen. "An Estimation of the Deadweight and X-Efficiency Losses in the Municipal Electric Industry." Journal of Economics and Business, vol. 34 (1984), pp. 269-81. Hudson, Heather E. Communication Satellites: Their Development and Impact. New York: Free Press, 1990. _____. "Telemedicine: Some Findings from the North American Experience." Academy for Educational Development, Washington, D.C., June 1980. Processed. ______. When Telephones Reach the \Village: The Role of Telecommunications in Rural Development. Norwood, N.J.: Ablex Publishers, 1984. Hudson, Heather E., Douglas Goldschmidt, Edwin B. Parker, and Andrew Hardy. "The Role of Telecommunications in Socioeconomic Development." Report prepared for the ITu by Keewatin Communications. Washington, D.C., May 1979. Processed. Hudson, Heather E., Andrew P. Hardy, and Edwin B. Parker. "Impact of Tele- phone and Satellite Earth Station Installations on GDP." Telecommunications Policy, vol. 6, no. 4 (December 1982), pp. 300-07. Hudson, Heather E., and Edwin B. Parker. "Medical Communication in Alaska by Satellite." New England Journal of Medicine, vol. 289, nos. 1351-56 (De- cember 20, 1973). Hughson, Terri L., and Paul S. Goodman. "Telecommuting: Corporate Practices and Benefits." National Productivity Review. vol. 5 (Autumn 1986), pp. 315-24. Huntly, L. R. "Some Ideas Regarding Economics of Telecommunications Engi- neering." Electrical Communication, vol. 42, no. 1 (1967), pp. 6-21. Implementation and Management Group Party, Ltd. "A Study of Remote Area Telecommunications in the Northern Territory," vol. 2. Report prepared for TELECOM Australia. Sydney, May 1980. Processed. Integrated Development Systems. "Economics of Telecommunications in Nepal." Prepared for Nepal Communications Corporation. Kathmandu, August 1980. Processed. INTELSAT. Annual Report. Washington, D.C., 1988. . Project SHARE: A Final Report and Evaluation. Washington, D.C., 1988. International Development Research Centre. Sharing Knowledge for Development: IRDC'S Information Strategp for Africa. Ottawa, 1989. 436 International Energy Agency. Energy Balances of Paris: OECD Countries, 1987- 1988. Paris: OECD. ILO. "Employment Income and Equality: A Strategy for Increasing Productive Em- ployment in Kenya." Geneva, 1972. Processed. _____. "Studies in Development: A Program of Development, Equity, and Grants for the Philippines," vol. 1, March Report. Geneva, 1972. Processed. Isard, Walter. General Theory: Social, Political, Economic, and Regional. Cambridge, Mass.: M.l.T. Press, 1969. ITU. Benefits of Telecommunications to the Transportation Sector of Developing Countries: A Case Study in the People's Democratic Republic of Yemen. Geneva, March 1988a. . The Changing Telecommunications Environment: Policy Considerations for the Members of the ITU. Report of the Advisory Group on Telecommunications Policy. Geneva, February 1989. _ . Contribution of Telecommunications to the Earnings/Savings of Foreign Ex- change in Developing Countries: Case Studies of 20 Kenyan Firms. Geneva, 1988b. . "Development of Manpower Resources for Telecommunications Admin- istrations in Africa." Paper prepared for the third African telecommunications conference, Monrovia, Liberia, December 8-19, 1980a. Processed. _.____* Information, Telecommunications, and Development. Geneva, 1986. . The Missing Link: Report of the Independent Commission for World Wide Telecommunications Development. Geneva, 1984. . "Restructuring of Telecommunications in Developing Countries: An Empirical Investigation with ITU's Role in Perspective." Geneva, May 1991. . Socio-Economic Benefits of Improved Telecommunications in Developing Countries: Results of a Research Study in Vanuatu. Geneva, 1988c. . "Some Aspects of Multi-National Telecommunication Training in Deve- loping Areas." Paper prepared for the third African telecommunications confer- ence, Monrovia, Liberia, December 8-19, 1980b. . World Telecommunications Forum 1987, Part IV. Symposium on eco- nomic and financial issues: the role of telecommunications in the infrastruc- ture and its impact on economic growth, Geneva, 1987. . Yearbook of Public Telecommunication Statistics. 17th ed. Geneva, 1990. Ivanek, Ferdo, Timothy Nulty, and Nikola Holcer. The Impact of Technological Progress on Telecommunications Equipment Manufacturing in NICS. World Bank Technical Paper no. 145. Washington, D.C.: World Bank, 1991. Jabif, R. "Analysis y modelaci6n de la demanda telef6nica comercial." Departa- mento de Electricidad, Universidad de Chile, Santiago, 1971. Processed. Jenkins, Janet, ed. Commonwealth Co-operation in Open Learning: Background Pa- pers. London: Commonwealth Secretariat, 1988. Jipp, A. "Wealth of Nations and Telephone Density." Telecommunications Journal (July 1963), pp. 199-201. Johansen, Robert, Kathleen Hansell, and David Green. "Growth in Telecon- REFERENCES 437 ferencing: Looking beyond the Rhetoric of Readiness." Telecommunications Pol- icy (December 1981), p. 289. Johansen, Robert, J. Vallee, and Kathleen Spangler. Electronic Meetings. Reading, Mass.: Addison-Wesley, 1979. Johnson, Leland L. "Price Caps in Telecommunications Regulatory Reform." Rand Note no. N-2894-MF/RC. Rand Corporation, Santa Monica, Calif., 1989. Jonscher, Charles. "Benefits, Costs, and Optimal Prices in Telecommunications Services." Draft working paper. Harvard University, Cambridge, Mass., un- dated. Processed. _______. "Economic Benefits of Telecommunications Investments: A Synthesis of Quantitative Evidence." In ITU, World Telecommunications Forum 1987, Part IV Geneva, 1987. ______. "Productivity and Growth of the Information Economy." In Meheroo Jussawalla and Helene Ebenfield, eds., Communication and Information Eco- nomics: New Perspectives. New York: Elsevier Science Publishers, 1984. ____. "A Theory of Economic Organization." Ph.D. diss., Harvard University, Cambridge, Mass., 1980. Jussawalla, Meherro. "The Information Economy and Its Importance for the De- velopment of Pacific Region Countries." In ITL, Information, Telecommunica- tions, and Development. Geneva, 1986. Jussawalla, Meheroo, and Helene Ebenfield, eds. Communication and Information Economics: New Perspectives. New York: Elsevier Science Publishers, 1984. Jussawalla, Meheroo, Donald Lamberton, and Neil Karunaratne, eds. The Cost of Thinking: Information Economies of Ten Pacific Countries. Norwood N.J.: Ablex Publishing, 1988. Kahn, Alfred E. The Economics of Regulation: Principles and Institutions, vol. 1. New York: John Wiley and Sons, 1970. Kahn, Alfred E., and W. Shew. "Current Issues in Telecommunications Regula- tion: Pricing." Yale Journal on Regulation (Spring 1987). Kamal, Ahmed, Ali E. Hillal Dessouki, and Ithiel de Sola Pool. "Communication System in Rural Egypt." Communication Needs for Rural Development Re- search Report no. 11. Cairo University, 1980. Processed. Karunaratne, Neil D. "An Input-Output Approach to the Measurement of the Information Economy." In Meheroo Jussawalla, Donald M. Lamberton, and Neil D. Karunaratne, eds., The Cost of Thinking: Information Economies of Ten Pacific Countries. Norwood, N.J.: Ablex Publishing, 1988. Kasolojaona, Mamiharilala. "Colloque sur la restructuration du secteur de telecommunications Malagasy." Paper presented at a seminar on telecommuni- cations restructuring sponsored by the World Bank, the ITU, and ACCT, Tunisia, May 1992. Katsoulis, M. "Energy Impacts of Passenger Transportation." Bell Canada, Montreal, March 1974. Processed. ______. "Travel/Telecommunications Substitution-Its Potential for Energy Con- servation in Canada." Bell Canada, Montreal, February 1976. Processed. 438 Kaul, S. N. "Benefits of Rural Telecommunications in Developing Countries." Paper presented at the ITU/OECD expert meeting on telecommunications in de- veloping countries, Paris, December 18-20, 1978. Conference Working Paper no. 8, Document no. CD/R (78) 39.8. OECD, Paris, 1978. Processed. _ . "Cost Return Analysis and Economies of Scale in Telephone Network Expansion, India." Economics Study Cell, P&T Board, New Delhi, March 1980. Processed. . "Planning of Telecommunications Services for Development." NCAER Margin, vol. 2, no. 3 (April 1979), pp. 71-80. Kavanagh, N. J., and R. Smith. "Measurement of Benefits of Trout Fishing: Pre- liminary Results of a Study at Grafham Water." Journal of Leisure Research, vol. 1 (1969), pp. 316-32. Khadem, Ramin. "An Economic Assessment of Mobile Satellite and Cellular Radio Systems." In ITU, World Telecommunications Forum 1987, Part IV. Geneva, 1987. Kilgour, Mary Cameron. "The Telephone in the Organization of Space for Devel- opment." Ph.D. diss., Political Economy and Government, Harvard University, Cambridge, Mass., 1982. King, John A. C. "The Privatization of Telecommunications in the United King- dom." In Bj6rn Wellenius, Peter A. Stern, Timothy Nulty, and Richard D. Stern, eds., Restructuring and Managing the Telecommunications Sector. Wash- ington, D.C.: World Bank, 1989. Kinsman, Francis. The Telecommuters. New York: John Wiley and Sons, 1987. Kiss, Ferenc. "Productivity Gains in Bell Canada." In Leon Courville, Alain De Fontenay, and Rodney Dobell, eds., Economic Analysis of Telecommunications: Theory and Applications. Amsterdam: North-Holland, 1983. Klein, A. "Trafico telef6nico interurbano y una tipologia socioeconomica." Facultad de Ciencias Fisicas y Matematicas, Universidad de Chile, Santiago, 1971. Processed. Kochen, Manfred. "Opportunity Costs in Computer Conferencing during and for Economic Development." In Meheroo Jussawalla and D. M. Lamberton, eds., Communication Economics and Development. Elmsford, N.Y: Pergamon Press, 1982. Kojina, Mitsuhiro, Junichiro Hoken, and Masaru Saito. "Report: The Use of Tele- phones in Sri Lanka." Telecommunications Policy (December 1984), pp. 335-38. Kollen, J. H., and J. Garwood. "Travel/Communications Tradeoffs: The Potential for Substitution among Business Travellers." Headquarters Business Planning Group, Bell Canada, Montreal, April 1975. Processed. Kottis, G. "The Service Sector in Greece: A Regional Perspective." FAST Paper no. 99. Commission of the European Communities, April 1986. Krechmer, Ken. "The Hidden Costs of ISDN." Telecommunications (October 1989), pp. 25-34. Laidlaw, Bruce. "The Evolution of Telecommunications Policy in the United Kingdom." In Bj6rn Wellenius and Peter Stern, eds., Implementing Reforms in REFERENCES 439 the Telecommunications Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Lamberton, Donald. "Information Economics: 'Threatened Wreckage' or New Paradigm?" CIRCIT Working Paper no. 1990/91. Processed. ______. "The Theoretical Implications of Measuring the Communication Sec- tor." In Meheroo Jussawalla and D. M. Lamberton, eds., Communication Eco- nomics and Development. Pergamon Policy Studies on International Develop- ment. Honolulu: East-West Center, 1982. Lange, Siegfried, and Helmet Rempp. "Qualitative and Quantitative Aspects of the Information Sector." Karlsruhe Institut fur Systemtechnik und Innovations- forschung, Karlsruhe, Federal Republic of Germany, 1977. Processed. Langley, C. John, Jr., and Rammohan Pisharodi. "Discussion of 'Resource Impli- cations of Electronic Message Transfer in Letter-Post Industry,' by Alfred M. Lee and Arnim H. Meyburg." Transportation Research Record, vol. 812 (1981), pp. 63-64. Lathey, C. E. "Telecommunications Substitutability for Travel: An Energy Con- servation Potential." Report no. 75-58. Office of Telecommunications, U.S. Department of Commerce, Washington, D.C., 1975. Processed. Lee, Alfred M., and Arnim H. Meyburg. "Resource Implications of Electronic Message Transfer in Letter-Post Industry." Transportation Research Record, vol. 812 (1981), pp. 59-63. Leibenstein, Harvey. "Allocative Efficiency vs. 'X-Efficiency."' American Economic Review, vol. 56 (June 1966), pp. 392-415. . "Competition and X-Efficiency." Journal of Political Economy, vol. 81 (May/June 1973). Levy, Mildred B., and Walter Wadycki. "Lifetime versus One-Year Migration in Venezuela." Journal of Regional Science, vol. 12 (December 1972), pp. 407-14. Lewis, W. A. "Fixed Costs." In Allen L. Unwin, ed., Overhead Costs. 1949. Re- printed in D. Munby, ed., Transport. Harmondsworth, Eng.: Penguin Modern Economics, 1968. Linn, Johannes F. Cities in the Developing World: Policies for Their Equitable and Efficient Growth. New York: Oxford University Press, 1983. Little, 1. M. D., and J. A. Mirrlees. Project Appraisal and Planning for Developing Countries. New York: Basic Books, 1976. Littlechild, S. C. "The Effects of Postal Responsibility and Private Ownership on the Structure of Telephone Tariffs: An International Comparison." University of Birmingham, Birmingham, Eng., June 1980. Processed. . Elements of Telecommunications Economics. Stevenage, Eng.: Peter Peregrinus, Ltd., 1979. . "Regulation of British Telecommunications Profitability." Department of Industry, London, 1983. Processed. Lockheed, Marlaine E., John Middleton, and Greta Nettleton, eds. "Education and Technology: Sustainable and Effective Use". PHREE Background paper 91/ 32. World Bank, Washington, D.C., 1991. 440 L6nnstr6m, Sven, Folke Marklund, and Ingemar Moo. A Telephone Development Project. 3d ed. Stockholm: L. M. Ericsson, 1975. Luhan, P. "The Yearbook of Common Carrier Telecommunication Statistics, Pub- lished by the ITU: An Indispensable Tool for Gauging the Development of World Telecommunications." Telecommunications Journal, vol. 56 (November 1989), pp. 245-50. Machlup, Fritz. The Production and Distribution of Knowledge in the United States. Princeton, N.J.: Princeton University Press, 1962. Mack, M., and G. Lee. "TELECOM Australia's Experience with Photovoltaic Sys- tems in the Australian Outback." Telecommunications Journal, vol. 56, no. 8 (1989), pp. 513-20. Malgavkar, P. D., and V. K. Chebbi. "The Impact of Telecommunications Facili- ties on Rural Development in India." In Dan J. Wedemeyer and M. R. Ogden, eds., Telecommunications and Pacific Development: Alternatives for the Next Dec- ade. New York: Elsevier Science Publishers and North-Holland and PTC, 1988. Malien, Habyalimana. "Colloque sur la restructuration du secteur des telecommunications de Rwanda." Paper presented at a seminar on telecommu- nications restructuring sponsored by the World Bank, the ITU, and ACCT, Tuni- sia, May 1992. Mansell, Robin. "The Role of Information and Telecommunication Technologies in Regional Development." Science. Technology, Industry Review, OECD, no. 3 (April 1988), pp. 135-73. . Telecommunications Rate Restructuring: Issues, Patterns, and Policy Prob- lems for OECD Countries. Paris: OECD and iccp Research Publications, 1990. March, J. G., and H. A. Simon. Organizations. New York: John Wiley and Sons, 1958. Martin, James. Future Developments in Telecommunications, 3d ed. Englewood Cliffs, N.J.: Prentice-Hall, 1991. Martin, R. L., and R. Rowthorn, eds. The Geography of De-lndustrialization. Lon- don: MacMillan, 1986. Mayer, Martin. "The Telephone and the Uses of Time." In Ithiel de Sola Pool, ed., The Social Impact of the Telephone. Cambridge, Mass.: M.l.T Press, 1977. Mayo, John K., Gary R. Heald, Steven J. Klees, and Martha Cruz de Yanes. Peru Rural Communication Services Project: Final Evaluation Report. Tallahassee: Center for International Studies, Learning Systems Institute, Florida State University, 1987. Mazumdar, Dipak. "The Urban Informal Sector." World Bank Staff Working Paper no. 211. Washington, D.C., July 1975, and World Development, vol. 4, no. 8 (1976). McAnany, Emile G. "The Role of Information in Communicating with the Rural Poor: Some Reflections." In Emile G. McAnany, ed., Communications in the Rural Third World. New York: Praeger Publishers, 1980. -______ ed. Communications in the Rural Third World. New York: Praeger Publish- ers, 1980. REFERENCES 441 McKean, Roland N. "The Use of Shadow Prices." In Samuel B. Chase, Jr., ed., Problems in Public Expenditure Analysis. Washington, D.C.: Brookings Institu- tion, 1968. Medinikov, D. "What Is the Effect of Dispatcher Communication in Farming?" Vestnik Svyazi, Moscow, no. 4 (1975), p. 35. Melo, jose Ricardo. "Liberalization and Privatization in Chile." In Bj6rn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunica- tions Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Melody, William H. "Telecommunications Implications for the Structure of De- velopment." In ITU, World Telecomniunication Forum 1987, Part AV Geneva, October 1987. Mendis, Vernon L. B. "Phased Privatization with Proposed Foreign Participation: The Sri Lanka Experience." In Bj6rn Wellenius, Peter Stern, Timothy Nulty, and Richard Stern, eds., Restructuring and Managing the Telecommunications Sector. Washington, D.C.: World Bank, 1989. Meyer, John R., Robert W. Wilson, M. Alan Baughcum, Ellen Burton, and Louis Caouette. The Econonmics of Competition in the Telecommunications Industry. Cambridge, Mass.: Oelgeschlager, Gunn, and Hain, 1980. Meyers, Robert A., ed. Encyclopedia of Telecommunications. San Diego, Calif.: Ac- ademic Press, 1989. Miernyk, William H. The Elements of Input-Output Analysis. New York: Random House, 1965. Ministry of Labour and Housing and Ministry of Physical Planning of Local Gov- ernment, Sweden. "Planning Sweden-Regional Development Planning and Management of Land and Water Resources." Allmanna Forlaget, Stockholm, 1973. Processed. Ministry of Supply and Services, Canada. "Elasticity of Demand for Long Dis- tance Telephone Service." CRTC Consulting Report no. 5. Prepared by Steven Globerman Associates, Ltd., 1988. Processed. Mitchell, Bridger M. "Incremental Capital Costs of Telephone Access and Local Use." Rand paper no. R-3764-ICTF. Rand Corporation, Santa Monica, Calif., 1989. _____. "Local Telephone Costs and the Design of Rate Structures." In Leon Courville, Alain De Fontenay, and Rodney Dobell, eds., Economic Analysis of Telecommunications TheorY and Application. New York: Elsevier Science Pub- lishers, 1983. . Optimal Pricing of Local Telephone Service. R-1962-MF. Santa Monica, Calif.: Rand Corporation, November 1976. . "Optimal Pricing of Local Telephone Service." American Economic Re- view, vol. 68, no. 4 (September 1978), pp. 517-37. . "Pricing Policies in Selected European Telephone Systems." In H. Dordick, ed., Proceedings of the Sixth Annual Telecommunications Policy Research Conference. Lexington, Mass.: Lexington Books, 1979. 442 . "Pricing Subscriber Access to the Telephone Network." In A. Baughcum and G. R. Faulhaber, eds., Telecommunications Access and Public Policy. Paper no. P-6815. Santa Monica, Calif.: Rand Corporation, 1984. Mitchell, Bridger M., W. G. Manning, and J. P. Acton. Peak-Load Pricing: Euro- pean Lessons for U.S. Energy Policy. Cambridge, Mass.: Ballinger Publishing Company, 1978. Mitchell, Bridger M., and Ingo Vogelsang. Telecommunications Pricing: Theory and Practice. Cambridge, Eng.: Cambridge University Press, 1991. Mody, Ashoka. "Information Industries in the Newly Industrializing Countries." In Robert W. Crandall and Kenneth Flamm, eds., Changing the Rules: Techno- logical Change, International Conmpetition, and Regulation in Communications. Washington, D.C.: Brookings Institution, 1989. Moon, A. E., and others. "Technology Assessment of Telecommunications/ Transportation Interactions," vol. 3, "Contributions of Telecommunications to Improved Transportation System Efficiency." Stanford Research Institute, Menlo Park, Calif., May 1977. Processed. Moss, Mitchell L. "Urban Development in a Global Economy." In Maurice F. Estabrooks and Rodolphe H. Lamarche, eds., Telecommunications: A Strategic Perspective on Regional, Economic, and Business Development. Moncton, Can- ada: Canadian Institute for Research on Regional Development, 1986. ______ ed. Telecommunications and Productivity. Reading, Mass.: Addison- Wesley, 1981. Muller, Jurgen, and Emilia Nyevrikel. "Closing the Capacity and Technology Gaps in Central and Eastern European Telecommunications." In Bjorn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunica- tions Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Munasinghe, Mohan. The Economics of Power System Reliability and Planning. Bal- timore, Md.: Johns Hopkins University Press, 1979. Munasinghe, Mohan, and Mark Gellerson. "The Determination of Optimal Reli- ability Standards for Electricity Supply." Bell Journal of Economics, vol. 10, no. 1 (Spring 1979), pp. 353-64. Munasinge, Mohan, and Jeremy J. Warford. "Shadow Pricing and Power Tariff Policy." World Bank Staff Working Paper no. 286. Washington, D.C., June 1978. Narula, A. M. "Post Office Savings Bank in India: A Saga of Progress." Union Postale, vol. 4 (1972), p. 49. National Council of Applied Economic Research, India. "Fertilizer Survey, 1975- 76." New Delhi, India, 1976. Processed. . "Survey of Rural Public Call Offices." Study commissioned by India P&T May 1978. Processed. National Research Council. Review of Electronic Mail Service Systems Planning for the U.S. Postal Service. Report to the U.S. Postal Service. Washington, D.C.: National Academy Press, 1981. REFERENCES 443 National Research Council, Board on Science and Technology for International Development. Science and Technology Information Services and Systems in Africa. Report of a workshop held in Nairobi, Kenya, April 1989. Washington, D.C.: National Academy Press, 1990. National Research Council, Board on Telecommunications and Computer Appli- cations, Commission on Engineering and Technical Systems. Office Work- stations in the Home. Washington, D.C.: National Academy Press, 1985. Nettleton, Greta. "Distance Training for Telecommunications Managers: A Case Study." Paper presented at the ICDE conference in Caracas, Venezuela, Novem- ber 1990. Processed. Nettleton, Greta, and Emile McAnany. "Brazil's Satellite System: The Politics of Applications Planning." Telecommunications Policy. vol. 13, no. 2 (June 1989), pp. 159-66. Neumann, Karl-Heinz, and Thomas Schnoring. "Reform and Unification of Telecommunications in Germany." In Bj6rn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunications Sector: Lessons from Recent Ex- perience. Washington, D.C.: World Bank, 1991. Newberg, Paula R., ed. New Directions in Telecommunications Policy, 2 vols. Durham, N.C.: Duke University Press, 1989. Nicol, Lionel Y "Communications, Economic Development, and Spatial Struc- tures: A Review of Research." Working Paper no. 404. Berkeley: Institute of Urban and Regional Development, University of California, 1983a. . "Communications, Economic Development, and Spatial Structures: A Theoretical Framework." Working Paper no. 405. Berkeley: Institute of Urban and Regional Development, University of California, 1983b. Nicolai, C., and Bj6rn Wellenius. "Estudio de inversiones en telefonos rurales, V y VI regi6n." Departamento de Electricidad, Universidad de Chile, Santiago, January 1979. Processed. Nilles, J. M. "Development of Policy on the Telecommunications Tradeoff." Phase 11 Report. University of Southern California, Los Angeles, March 1974. Processed. Nilles, J. M., F. R. Carlson, P. Gray, and G. J. Hanneman. The Telecommu- nications-Transportation Tradeoff: Options for Tomorrow. New York: John Wiley and Sons, 1976. Noll, Roger G. Introduction to Telephones and Telephone Systems, 2d ed. Boston, Mass.: Artech House, 1991. _____. "Telecommunications Regulation in the 1990s." In Paula R. Newberg, ed., New Directions in Telecommunications Policy, vol. 1. Durham, N.C.: Duke University Press, 1989. Nora, S., and A. Minc. The Co?nputerization of Society. Cambridge, Mass.: M.I.T. Press, 1980. Nordlinger, Christopher W. "Users of Public Telecommunications Facilities and Their Benefits in a Developing Country: A Case Study of Senegal." In ITU, In- formation. Telecomniunications, and Development. Geneva, 1986. 444 Nourse, Hugh 0. Regional Economics. New York: McGraw-Hill Book Company, 1968. Nulty, Timothy. "Challenges and Issues in Central and Eastern European Tele- communications." In Bjorn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunications Sector: Lessons from Recent Experience. Wash- ington, D.C.: World Bank, 1991a. . "Emerging Issues in World Telecommunications." In Bjorn Wellenius, Peter Stern, Timothy Nulty, and Richard Stern, eds., Restructuring and Manag- ing the Telecommunications Sector. A World Bank Symposium. Washington, D.C.: World Bank, 1991b. O'Brien, Rita Cruise, E. Copper, B. Perkes, and H. Lucas. "Communications In- dicators and Indicators of Socio-Economic Development, 1960-1970." Insti- tute of Development Studies, University of Sussex, Eng., 1977. Processed. Oeffinger, John C. "Merging Computers and Communications: A Case Study in Latin America." Telematics and Informatics, vol. 4, no. 3 (1987), pp. 195-210. Okazaki, H. "Planning an Urban Area Digital Telephone Junction Network." NEC Research and Development. vol. 73 (April 1984), pp. 43-52. Okundi, Philip 0. "Pan-African Telecommunications Network: A Case for Tele- communications in the Development of Africa." World Telecommunication Forum. Geneva: ITU, October 1975. Okundi, Philip O., and B. G. Evans. "Afrosat: Proposals for an African Domestic Satellite Communication System." Paper presented at the international confer- ence on satellite communication systems technology, London, April 1975. Processed. Okundi, Philip O., Arthur W. Ogwayo, and James P. Kibombo. "Rural Telecom- munications Development in East Africa." Proceedings of the International Tele- communications Exposition (INTELC:OM 77), vol. 1. Dedham, Mass.: Horizon House International, 1977. Organization for African Unity. "Social Aspects of Telecommunications." Paper submitted to the third African telecommunications conference, Monrovia, Li- beria, December 8-19, 1980. Processed. OECD. Information, Technology, and New Growth Opportunities. Information, Com- puter, Communications Policy Series no. 19. Paris, 1989. Telecommunication Network Based Services: Policv Implications. Informa- tion, Computer, Communications Policy Series no. 18. Paris, 1989. . Trends of Change in Telecommunications Policy. Information, Computer, Communications Policy Series no. 13. Paris, 1987. OECD, Working Party on Telecommunications and Information Policies. Perfor- mance Indicators for Public Telecommunications Operators. Paris, 1990. OFTEL. "Responses to OFTEL'S Consultative Document on the Future Regulation of British TELECOM'S Prices." OFTEL Working Paper no. 3. London. Page, Michael. The Flying Doctor Story, 1928-78. Adelaide, Australia: Rigby, 1977. Paine, Nigel. "The Likely Impact of New Forms of Communication and Informa- REFERENCES 445 tion Technology on Distance Education." Document submitted to the Com- monwealth Secretariat, August 1986. Processed. Park, Rolla Edward, and Bridger M. Mitchell. Optimal Peak Load Pricing for Local Telephone Calls. Publication no. R-3404-1-RC. Santa Monica, Calif.: Rand Cor- poration, 1987. Parker, Edwin B. "MicroEarth Station Satellite Networks and Economic Develop- ment." Telematics and Informatics, vol. 4, no. 2 (1987). Parker, Edwin B., Heather E. Hudson, Don Dillman, and Andrew Roscoe. Rural America in the Information Age: Telecommunications Policy for Rural Develop- ment. Lanham, Md.: University Press of America and Aspen Institute, 1989. Perraton, Hilary, ed. Alternative Routes to Formal Education: Distance Teaching for School Equivalency. Baltimore, Md.: Johns Hopkins University Press, 1982. _ . ed. "Distance Education: An Economic and Educational Assessment of Its Potential for Africa." World Bank Discussion Paper, Education and Training Series Report no. EDT 43. World Bank, Washington, D.C., December 1986. Phillips, F., ed. Thinkwork: Working, Learning, and Managing In a Computer- Interactive Society. New York: Praeger, 1991. Pierce, J. "Communication as an Alternative to Travel." Proceedings of the IRE, 50th anniversary issue (May 1962), pp. 225-31. Pierce, William. "A Global-Domestic (GLODOM) Satellite System for Rural Devel- opment." Telecommunications Journal, vol. 46 (December 1979), pp. 745-47. Pierce, William B., and Nicolas Jequier. Telecommunications for Development. Geneva: ITU, 1983. Pierson, J. H. S. "Planning for Telecommunications Expansion: Making the Case for a Telecommunications Budget." Paper presented at INTELCOM '79, Dallas, Tex., March 1979, Pierson Consultants, Inc., New York. Processed. Polishuk, P., R. Guenther, and J. Lawlor. "Cost Comparison of Microwave, Satel- lite, and Fibre Optic Systems." Telecommunications Journal, vol. 54, no. 2 (1987), pp. 114-22. Pool, Ithiel de Sola. "The Communications/Transportation Tradeoff." In Alan Altschuler, ed., Current Issues in Transportation Policy. Lexington, Mass.: D. C. Heath, 1979. . "The Influence of International Communications on Development." Massachusetts Institute of Technology, Cambridge, Mass., February 1976. Processed. _ , ed. The Social Impact of the Telephone. Cambridge, Mass.: M.I.T. Press, 1977. Pool, Ithiel de Sola, and Peter M. Steven. "Appropriate Telecommunications for Rural Development." In Indu B. Singh, ed., Telecommunications in the Year 2000: National and International Perspectives. Norwood, N.J.: Ablex Publishers, 1983. Porat, Marc Uri. The Information Economy: Definition and Measurement. Special Publication no. 77-12(1). Washington, D.C.: U.S. Department of Commerce, Office of Telecommunications, May 1977. 446 Preece, Robert S. "The Role of Telecommunications in Economic Growth and Income Distribution." In ITU, World Telecommunications Forum 1987, Part IV. Geneva, 1987. Prest, Alan, and R. Turvey. "Cost-Benefit Analysis: A Survey." Economic Journal (December 1965). Pye, Roger. "Communications within the Scottish Office." Communications Studies Group Report no. w/74174/py, University College, London, 1974. Processed. Pye, Roger, and J. B. Goddard. "Telecommunications and Location." Regional Studies, vol. 11, no. 2 (1977). Pye, Roger, and Gillian Lauder. "Regional Aid for Telecommunications in Europe: A Force for Economic Development." Telecommunications Policy (June 1987), pp. 99-113. Pye, Roger, M. Tyler, and B. Cartwright. "Telecommunicate or Travel?" New Sci- entist (September 12, 1974). Pye, Roger, and P. I. Weintraub. "Attitudes toward Business Travel and Teleconferencing." Impact Paper no. 12. In Richard Harkness, Technology As- sessment of Telecommunications/Transportation Interactions, vol. 2. Menlo Park, Calif.: Stanford Research Institute, May 1977. Pye Telecommunications, Ltd. "A Study of the Future Frequency Spectrum Re- quirements for Private Mobile Radio in the U.K." Cambridge, Eng., 1976. Processed. Qvortup, Lars. "Community Teleservice Centres and the Future of Rural Soci- ety." Telematics Project, Odense University, Odense, Denmark, September 1989. Processed. Race, Tim. "Going It Alone, Bit by Bit." New York Times Magazine, December 2, 1990, pp. 19-33. Ramajo, German. "Presentation on Companiia de Telefonos de Chile, S.A." In Bj6rn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommu- nications Sector. Washington, D.C.: World Bank, 1991. Rao, Paladugu V. "Telephone and Instructional Communications." In Ithiel de Sola Pool, ed., The Social Impact of the Telephone. Cambridge, Mass.: M.l.T Press, 1977. Reid, A. A. L. "Comparing Telephone with Face-to-Face Contact." In Ithiel de Sola Pool, ed., The Social Impact of the Telephone. Cambridge, Mass.: M.l.T Press, 1977. Renaud, Bertrand M. National Urbanization Policies in Developing Countries. New York: Oxford University Press, 1981. Research Institute of Telecommunication and Economics. "Analysis of Economic and Social Effects of Telecommunications Centered on Users' Appraisal." To- kyo, 1974. Processed. Rettig, Jack. "Converting to Measured Service the Low-Cost Way." Bell Laborato- ries Record (November 1981), pp. 281-83. Richardson, Harry W. Regional Economics. New York: Praeger Publishers, 1969. REFERENCES 447 Rockoff, Maxine L. "The Social Implications of Health Care Communication Systems." IEEE Transactions on Communications, vol. cou-23, no. 10 (October 1975), pp. 1085-88. Rohlfs, Jeffrey H., and Harry M. Shooshan. 'Will Price Caps Correct Major Eco- nomic Flaws in the Current Regulatory Process?" Paper presented at the twen- tieth annual Williamsburg conference, Williamsburg, Va., 1988. Processed. Roojsma, A. H. "Optimization of Digital Network Structures." In M. Akiyama, ed., Teletraffic Issues in an Advanced Information Society. Amsterdam: Elsevier Press, 1985. Satellite Communication Services. Teleconferencing Newsletter, vol. 1, no. I (June 1981). Saunders, Robert J. "Organizatiornal Structures for Telecommunications." Tele- communications Journal (August 1982), pp. 481-87. . "Rural Telecommunications: Economic and Policy Implications." Seminar on Rural Telecommunications, Neu, Delhi, September 11-22, 1978, vol. 1. Geneva: ITU, 1979. Saunders, Robert J., and C. R. Dickenson. "Telecommunications: Priority Needs for Economic Development." Telecommunications Journal, vol. 46, no. 9 (Sep- tember 1979), pp. 567-68. Saunders, Robert J., and Jeremy J. Warford. "Evaluation of Telephone Projects in Less Developed Countries." Telecommunicatiotis Journal, vol. 46, no. I (Janu- ary 1979), pp. 22-28. ____. "Telecommunications Pricing and Investment in Developing Countries." Proceedings of the International Telecommunications Exposition (INTELCOM 77), vol. 1. Dedham, Mass.: Horizon House International, October 1977. . Village Water Supply: Economics and Policy in the Developing World. Balti- more, Md.: Johns Hopkins University Press, 1976. Saunders, Robert J., Jeremy J. Warford, and Patrick C. Mann. "Alternative Con- cepts of Marginal Cost for Public Utility Pricing: Problems of Application in the Water Supply Sector." World Bank Staff Working Paper no. 259. Washing- ton, D.C., May 1977. Scherer, F. M. Industrial Market Structure and Economic Performance. Chicago, Ill.: Rand McNally College Printers, 1970. Schoppert, D. W., and others. "Some Principles of Freeway Directional Signing Based on Motorists' Experiences." Highway Research Board Bulletin no. 244. National Academy of Sciences, National Research Council, Washington, D.C., 1960. Processed. Schramm, Wilbur. Big Media, Little Media. Beverly Hills, Calif.: Sage Publica- tions, 1976. ____ . Mass Media and National Development. Stanford, Calif.: Stanford Uni- versity Press, 1964. Schultz, Richard J., and Peter Barnes, eds. "Local Telephone Pricing: Is There a Better Way?" Centre for the Study of Regulated Industries, Montreal, Canada, 1984. 448 Secretaria de Comunicaciones y Transportes, Direcci6n General de Telecomuni- caciones, Coordinaci6n General del Plan Nacional de Telefonia Rural, Mexico. "Plan Nacional de Telefonia Rural: Documento basico." Mexico City, June 1979. Processed. Selwyn, Lee L. "A Perspective on Price Caps as a Substitute for Traditional Revenue Requirements Regulation." Paper presented at the twentieth annual Williamsburg conference, Williamsburg, Va., 1988. Processed. Selwyn, Lee L., and Gregory F. Borton. "usp: Will It Really Decrease Local Ser- vice Costs?" Telephony (January 28, 1980), pp. 36-38. Selwyn, Lee L., and Scott C. Lundquist. "Adapting Telecom Regulation to Indus- try Change: Promoting Development without Compromising Ratepayer Protec- tion." IEEE Communications Magazine (January 1989). Shapiro, P. D. "Telecommunications and Industrial Development." IEEE Transac- tions on Communications, vol. coM-24, no. 3 (March 1976). Sherman, Roger. The Regulation of Monopoly. Cambridge, Eng.: Cambridge Uni- versity Press, 1989. Short, J. A. "Long Range Social Forecasts: Congestion and Quality of Service." Long Range Intelligence Bulletin no. 10. British Post Office Telecommunica- tions, London, April 1976. Processed. Short, John, E. Williams, and B. Christie. Social Psychology of Telecommunica- tions. London: John Wiley and Sons, 1976. Simon, Herbert A. "Theories of Decisionmaking in Economics." American Eco- nomic Review, vol. 49 (June 1959). Simonelli, Albert Lozano. "El papel del servicio postal en la vida econ6mica y cultural de los paises en vias de desarrollo." In Uni6n Postal de las Americas y Espafia, Conferencias del I seminario UPLU-AUPAE, Lima. Montevideo: Oficina Internacional de la Uni6n Postal de las Americas y Espania, 1967. Simpson, Alan. "The State of the Optical Art." Telephony (August 27, 1990), pp. 40-44. Singh, Indu B., ed. Telecommunications in the Year 2000: National and Interna- tional Perspectives. Norwood, N.J.: Ablex Publishers, 1983. Skoog, R. A., ed. "The Design and Cost Characteristics of Telecommunications Networks." Bell Telephone Laboratories, 1980. Processed. Smith, Robert. "Telecommunications and Technology: Current Trends." Impacts of Telecommunications on Planning and Transport. Research Report no. 24. Lon- don: Department of the Environment and Transport, 1978. Sorokin, A. L. The Economics of the Postal System. Lexington, Mass.: Lexington Books, 1980. Spence, A. M. "The Economics of Internal Organization: An Introduction." Bell Journal of Economics, vol. 6 (1975), p. 169. Squire, Lyn. "Some Aspects of Optimal Pricing for Telecommunications." The Bell Journal of Economics and Management Science (Autumn 1973), pp. 515-25. REFERENCES 449 Squire, Lyn, and Herman G. van der Tak. Economic Analysis of Projects. Balti- more, Md.: Johns Hopkins University Press, 1975. Srivisal, Chote, and Hiroyoshi Tamura. "Usage Patterns of Residential Tele- phones Based on Customer Interviews." Office of Economic Studies, Tele- phone Organization of Thailand, September 1980. Processed. Stahl, H. J., and A. Burmeister. "Benefits of Telecommunication for Transporta- tion Systems in Developing Countries." In mIl, World Telecommunications Forum 1987, Part IV Geneva, 1987. Steiner, Peter 0. "Peak Loads and Efficient Pricing." Quarterly Journal of Eco- nomics, vol. 71 (November 1957), pp. 585-610. Stephens, B. W., and others. "Third Generation Destination Signing: An Elec- tronic Route Guidance System." Highway Research Record no. 265. National Academy of Sciences, National Academy of Engineers, Washington, D.C., 1968. Processed. Stern, Eliott, and Richard Holti. Distance Working in *Jrban and Rural Settings: A Study for the tEC FAST Program. London: Tavistock Institute of Human Rela- tions, 1986. Stigler, George J. "The Economics of Information." Journal of Political Economy, vol. 69 (1961), pp. 213-25. Stover, William J. Information Technology in the Third World: Can I.T Lead to Hu- mane International Development! Boulder, Colo.: Westview Press, 1984. Sy, Jacques Hababib. "African Nations and Access to Telecommunications Ser- vices: Political Economy and Legal Issues." Paper presented at a seminar on tel- ecommunications restructuring sponsored by the World Bank, the ITU, and ACCT, Tunisia, May 1992. Taylor, Lester D. "Problems and Issues in Modeling Telecommunications De- mand." In L. Courville, A. Fontenay, and R. Dobell, eds., Economic Analysis of Telecommunications: Theory and Applications. New York: North-Holland, 1983. ____ . Telecommunications Demand: A Survey and Critique. Cambridge, Mass.: Ballinger, 1980. Taylor, William E. "Local Exchange Pricing: Is There Any Hope?" In James H. Alleman and Richard D. Emmerson, eds., Perspectives on the Telephone Indus- try: The Challenge for the Future. New York: Harper and Row Publishers, Ballinger Division, 1989. Telecom France, "How France Financed Its Telephone Program." no. 1 (May 1981), pp. 29-32. Telephone Organization of Thailand. "Telephone Usage Survey." Economic Stud- ies Group, Bangkok, October 1980. Processed. Telephony "Using Satellite Teleconference to Spread Marketing Gospel Through- out a Big Company." (January 25, 1982), pp. 34-43. Thery, Gerard. "On the Road to Making France's Telecommunications System the Best." Telephony (November 21, 1977), pp. 60-72. Thomas, Brownlee. "The Role of Communications in Economic Development." 450 M.A. thesis, Department of Political Science, McGill University, Montreal, Canada, 1984. Thorngren, Bertil. "How Do Contact Systems Affect Regional Development?" Environment and Planning, vol. 2 (1970). ____. "Silent Actors: Communication Networks for Development." In Ithiel de Sola Pool, ed., The Social Impact of the Telephone. Cambridge, Mass.: M.I.T Press, 1977. Tiene, Drew, and Shingenari Futagami. "Educational Media in Retrospect." World Bank Discussion Paper, Education and Training Series Report no. EDT 58. Washington, D.C., February 1987. Tietjen, Karen. AMD Rural Satellite Program: An Overview. Washington, D.C.: Academy for Educational Development, 1987. Tomey, J. F. "The Field Trial of Audio Conferencing with the Union Trust Com- pany." Report of contract no. H2104R. New Rural Society Project, Fairfield University, Fairfield, Conn., 1974. Processed. Tomlinson, M., and Bj6rn Wellenius. "Towards Economic Pricing of Telecommu- nications in the Developing World." In ITU, World Telecommunications Forum 1987, Part [V. Geneva, 1987. Tornato, P. "Substitution Transport Telecommunication." Ministere de I'Equipe- ment, Paris, 1974. Processed. Townsend, David, and Patricia D. Kravtin. "Industry Structure and Competition in Telecommunicacions Markets: An Empirical Analysis." Paper presented at the seventh international conference of the International Telecommunications Society, Boston, Mass., July 1988. Processed. Trebbing, Harry M., ed. Performance under Regulation. East Lansing, Mich.: Insti- tute of Public Utilities, 1968. Trevains, S. J. "Audio Conferencing in the Civil Service: An Evaluation of the Use of the Remote Meeting Table." Report no. csD/78100/TR. Communica- tions Studies and Planning, Ltd., London, 1978. Processed. Turvey, Ralph, and Dennis Anderson. Electricitv Economics: Essays and Case Stud- ies. Baltimore, Md.: Johns Hopkins University Press, 1977. Tyler, Michael. "Implications for Transport." In R. C. Smith, ed., Impacts of Tele- communications on Planning and Transport. Research Report no. 24. London: Department of Transport and Department of the Environment, July 1978. . "A Note on the Interpretation of the NASA Teleconference Usage and Im- pact Data." Communications Studies Group, University College, London, 1976. Processed. Tyler, Michael, B. Cartwright, and D. Bookless. "The Economic Consequences of Energy Scarcity." Intelligence Bulletin, vol. 1. Long Range Studies Division, Post Office Telecommunications, Cambridge, Eng. (1974). Tyler, Michael, B. Cartwright, and G. Bush. "Interaction between Telecommuni- cations and Face-to-Face Communication, The Energy Factor." Intelligence Bul- letin, vol. 3, Long Range Studies Division, Post Office Telecommunications, Cambridge, Eng. (1974). REFERENCES 451 Tyler, Michael, B. Cartwright, and H. A. Collings. "Prospects for Teleconference Services." Intelligence Bulletin, vol. 9, Long Range Studies Division, Post Office Telecommunications, Cambridge, Eng. (May 1977). Tyler, Michael, Martin Elton, and Angela Cook. The Contribution of Telecommu- nications to the Conservation of Energy Resources. Special Publication no. 77- 17. Washington, D.C.: U.S. Department of Commerce, Office of Telecommu- nications, July 1977. Ungerer, Herbert. "The European Situation: An Overview." In Bj6rn Wellenius and Peter Stern, eds., Implementing Reforms in the Telecommunications Sector: Lessons from Recent Experience. Washington, D.C.: World Bank, 1991. Union Postale. "Swedish Rural Postmen Also Provide Social Service." Vol. 10 (1974), pp. 148-52. United Nations. International Comparisons of Inter-Industry Data. Industrial Plan- ning and Programming Series no. 2. New York, 1969. . "Radio-guidage automobile presque pas a pas." Science et Vie, Paris (Feb- ruary 1977). Uno, Kimio. "The Communication Sector in Japan and Its Role in Economic Development." In Meheroo Jussawalla and D. M. Lamberton, eds., Communi- cation Economics and Development. Elmsford, N.Y: Pergamon Press, 1982. U.S. Congress. Development of the RFD System. Congressional report. Washington, D.C.: GPO, 1971. _______ RFD Roads. Hearings before the Joint Committee on Federal Aid in the Construction of Postal Roads, December 20, 1913. Washington, D.C.: GPO, 1913. U.S. Federal Communications Commission. "Further Notice of Proposed Rulemaking." cc Docket no. 87-313. Washington, D.C. U.S. Government. Report of the Postmaster-General. Washington, D.C.: GPO, an- nual, 1900, 1901, 1902. U.S. House of Representatives. Rural Delivery Systems, no. 54. Hearings before the Committee on Expenditures in the Post Office Department, February 14- 17, 21-23, 1912 (Washington, D.C.: GPO, 1912). Universal Postal Union. "Memorandum on the Role of the Post as a Factor in Economic, Social, and Cultural Development." Berne, Switzerland, 1974. Processed. . "Study on Postal Development: Report by the Executive Council of the 18th uPu Congress, Rio de Janeiro, 1979." Document no. 17(e). Berne, Swit- zerland, 1980. Processed. Vignon, H. "La restructuration des telecommunications au Benin." Paper pre- sented at a seminar on telecommunications restructuring sponsored by the World Bank, the ITU, and ACCT, Tunisia, May 1992. Vogelsang, Ingo. "Price Cap Regulation of Telecommunications Services: A Long-Run Approach." Paper no. N-2704-MF. Rand Corporation, Santa Monica, Calif., 1988. Wall, S. D. "Four Sector Time Series of the U.K. Labor Force, 1841-1971." Long 452 Range Studies Division, Post Office Telecommunications, London, 1977. Processed. Wall Street Journal, European ed., vol. 1, no. 2 (February 1, 1983), p. 1. Walsham, Geoffrey. "Models for Telecommunications Strategy in the LDCs." Tele- communications Policy June 1979), pp. 109-10. Walters, A. A. An Introduction to Econometrics. London: Macmillan, 1968. Watson, F. V. V. "Restructuring of the Telecommunications Sector in Sri Lanka: Review of Progress, 1988-9 1." In Bj6rn Wellenius and Peter Stern, eds., Imple- menting Reforms in the Telecommunications Sector: Lessons from Recent Experi- ence. Washington, D.C.: World Bank, 1991. Waverman, Leonard. "The Regulation of Intercity Telecommunications." In Almarin Phillips, ed., Promoting Competition in Regulated Markets. Washington, D.C.: Brookings Institution, 1975. Webber, Melvin M. "A Telecommunications Strategy for New Cities of the 21st Century." Working Paper no. 330. Berkeley: Institute of Urban and Regional Development, University of California, October 1980. Wedemeyer, Dan J., and M. R. Ogden, eds. Telecommunication and Pacific Devel- opment: Alternatives for the Next Decade. New York: Elsevier Science Publishers North-Holland and Prc, 1988. Wein, Harold H. "Fair Rate of Return and Incentives ... Some General Conclu- sions." In Harry M. Trebbing, ed., Performance under Regulation. East Lansing, Mich.: Institute of Public Utilities, 1968. Weisbrod, Burton A. "Concepts of Costs and Benefits." In Samuel B. Chase, Jr., ed., Problems in Public Expenditure Analysis. Washington, D.C.: Brookings In- stitution, 1968. Wellenius, Bj6rn. "Apuntes: Curso de planificaci6n de sistemas de telecomuni- caciones EL720." Departamento de Electricidad, Universidad de Chile, Santiago, 1971. Processed. . "Beginnings of Sector Reform in the Developing World." In Bj6rn Wellenius, Peter Stern, Timothy Nulty, and Richard Stern, eds., Restructuring and Managing the Telecommunications Sector. A World Bank Symposium. Wash- ington, D.C.: World Bank, 1989. . "The Changing World of Telecommunications: Policy Options in Deve- loping Countries." In ITU, World Telecommunications Forum 1987, Part IV Geneva, 1987. . "Communication, Systems, and Society: An Interdisciplinary Founda- tion for Telecommunication Studies." Ph.D. diss., Department of Electrical En- gineering Science, School of Physical Science, University of Essex, Colchester, Eng., 1978. . "Concepts and Issues on Information Sector Measurement." In Meheroo Jussawalla, Donald Lamberton, and Neil D. Karunaratne, eds., The Cost of Thinking: Information Economies of Ten Pacific Countries. Norwood, N.J.: Ablex Publishing, 1988. . "Financing Telecommunications in the Developing World-Issues and Opportunities." Proceedings of ITU TELECOM 86. Nairobi, Kenya, 1986. REFERENCES 453 ______. "Hidden Residential Connections Demand in the Presence of Severe Supply Shortage." IEEE Transactions on Communications Technology (June) 1969a. ______. "Income and Social Class in Residential Telephone Demand." Telecom- munications Journal. vol. 36 (May) 1969b. ______. "On the Role of Telecommunications in Development." Telecommunica- tions Policy (March 1984), pp. 59-66. ______. "The Role of Multilateral Development Financing Agencies." Paper pre- sented at the eighth international conference of the International Telecommu- nications Society, Venice, Italy, March 1990a. Processed. . "Telecommunications and Third World Development." In Maurice F. Estabrooks and Rodolphe H. Lamarche, eds., Telecommunications: A Strategic Perspective on Regional. Economic, and Business Development. Moncton, Can- ada: The Canadian Institute for Research on Regional Development, 1986. _____. "Telecommunications in the Developing World: Current Trends and New Issues." Paper presented at a symposium to honor Professor Kenneth Cattermole, University of Essex, Colchester, Eng., September 1990b. Wellenius, Bj6rn, R. Budinich, and P. Moral. "Estimaci6n prelinminar de la demanda de servicio telex en Chile." Fundaci6n Chile, January 1979. Processed. Wellenius, Bj6rn, and Peter Stern, eds. lniplementing Reforms in the Telecommuni- cations Sector: Lessons Jrom Recent Experience. Washington, D.C.: World Bank, 1991. Wellenius, Bj6rn, Peter Stern, Timothy Nulty, and Richard Stern, eds. Re- structuring and Managing the Telecommunications Sector. A World Bank Sympo- sium. Washington, D.C.: World Bank, 1989. Wenders, John T. The Economics of Telecommunication: Theory and Policv. Cambridge, Mass.: Ballinger Publishing Company, 1987. Whitlock, Erik, and Emilia Nyevrikel. "The Evolution of Hungarian Telecom- munications Policy." Telecommunications Policy, vol. 16, no. 3 (April 1992), pp. 249-58. Wilkinson, G. F. "The Estimation of Usage Repression under Local Measured Service: Empirical Evidence from the GTE Experiment." In Leon Courville, Alain De Fontenay, and Rodney Dobell, eds., The Economic Analysis of Tele- communications: Theory and Applications. New York: Elsevier Science Publish- ers, 1983. Williams, E., and 1. Young. "The Choice of Teleconference or Travel amongst Loudspeaking Telephone Users." Report no. E/77077/wL. Communications Studies Group, University College, London, 1977. Processed. Williams, F. C. W. "The Rural Mail Delivery System in New Zealand." Union Postale, vol. 4 (1972), pp. 51-54. Williams, Frederick. The New Telecotnmunications: Infrastructure for the Informa- tion Age. Belmont, Calif.: Wadsworth Publishers, 1991. Williams, Howard, and Andrew Gillespie. "Telecommunications and the Recon- struction of Regional Comparative Advantage." In ITU, W/orld Telecommunica- tions Forum 1987, Part W. Geneva, 1987. 454 Wilson, Carol, and Czatdana Inan. "New Markets Developing." Telephony Oanu- ary 22, 1989), pp. 32-48. World Bank. Annual Report. Various years. Washington, D.C. . World Bank Atlas 1988. Washington, D.C., 1989a. . World Development Report 1980. New York: Oxford University Press, 1980. . World Development Report 1990. New York: Oxford University Press, 1990. . World Development Report 1994-Infrastructure for Development. New York: Oxford University Press, 1994. _. "Telecommunications Development, Investment Financing, and the Role of the World Bank." Paper presented at the !TLJ plenipotentiary conference, Nice, France, 1989b. Processed. . "Telecommunications: World Bank Experience and Strategy." Washing- ton, D.C., February 1992. Processed. Wurtzel, Alan H., and Colin Turner. "Latent Functions of the Telephone: What Missing the Extension Means." In Ithiel de Sola Pool, ed., The Social Impact of the Telephone. Cambridge, Mass.: M.I.T. Press, 1977. Yap, Lorene Y L. "Internal Migration in Less Developed Countries: A Survey of the Literature." World Bank Staff Working Paper no. 215. Washington, D.C., September 1975. Yatrakis, P. G. "Determinants of the Demand for International Telecommunica- tions." Telecommunications Journal, vol. 39 (December 1972). Yoshita, M., and H. Okazaki. "Digital Telephone Junction Network Planning Sys- tem Considering Logical and Physical Network Aspects." In M. Akiyama, ed., Teletraffic Issues in an Advanced Information Society. Amsterdam: Elsevier Press, 1985. Index Access charges, 259, 271, 275, 290, 293 emergency use of telecommunications in, Accounting systems of operating entities, 26, 27, 141-42, 167, 178, 220-21, 229; 71, 72 Pcos in, 24, 169, 236, 238-40, 243, 244- Activist view of telecommunications devel- 49, 361; radio telephone systems in, 168- opment, 21-22 69, 319; residential telephone lines in, Administration. See Operating entities 214, 215, 227-29; rural use of telecom- Advanced services, 9, 32, 45, 74, 316, 329- munications in, 23, 24, 207, 220-21; tar- 30. See also specific services iffs and tariff setting in, 56-58; telecom- Advisory board, proposed, for regulatorV re- munications reform in, 319; telephone form, 328 density in, 5, 227, 233 n12, 251 n7, 319. Advisory Group on Telecommunications See also specific countries Policy, 21, 33 nl Asian Development Bank, 76 Africa. See Sub-Saharan Africa ATS.T (American Telephone and Telegraph African Development Bank, 76 Company), 19, 61 n22, 298 Agglomeration economies, 126, 129 Audio conferencing. See Teleconferencing Agriculture, use of telecommunications in: Auditing of operating entity, 72, 73 in Asia, 23, 102-06; benefits of, 17; com- Australia, 39, 112, 265, 279, 342, 354 pared with other sectors, 99, 223-24; in Automatic toll ticketing, 276 Latin America, 23-24, 102-06; postal Autonomy of operating entities, 44, 65-68, systems and, 367-68; in Sub-Saharan Af- 70 rica,. 118 n16, 119 n25, 185-86, 232 Average incremental cost method of tariff nll; in United States, 102-06 setting, 298 n7 AID (Agency for International Develop- ment), 349-51 Backward linkages in input-output analysis, Aid, foreign, 34 n18, 51, 76 107, 113 Alaska, 27, 28, 60 nl9, 343, 346, 348-50, Bangladesh, 168-69, 203, 209, 343 358 Banking by post, 371-73 American Telephone and Telegraph Com- Banking, electronic, 22, 23, 364 pany (AT&T), 19, 61 n22, 268 n22, 80 n8, BDT (Bureau for Telecommunications Devel- 268 n24, 298 opment), 34 n21 Argentina, 5, 81 n15, 279 Bebee, E. L., 90, 91 Arrow, Kenneth, 119 n24, 119 n28 Bell Canada, 61 n23, 154 n46 Asia: agricultural use of telecommunica- Benefits of telecommunications use. See tions in, 23, 102-06; business use of tele- Cost-benefit analysis communications in, 23, 25, 186-88, 204, Benin, 321-22 207, 215, 219-23, 226; information sec- Best-alternative method of consumer sur- tor in, 113, 333 n2; educational use of plus estimation, 164-69 telecommunications in, 214, 215, 349; Billing and collection, 44, 66, 72 455 456 INDEX Black markets in celephones. 133 n31, 173. Canadian Radio-Television and Telecom- 283 nlS munications Commission, 299 n12 Border pricing, 413 Capital indivisibilities, 46, 288, 298 n7 Borrowing of telephones, 227, 228 Capital intensity of telecommunications, Bower, L. B., 115 67, 74, 80 n8 Brazil, 25, 99, 34 nl2, 182, 267 nZQ, 279, Capital structure and requirements of oper- 309 ating entities, 73-74. See also Investment British Telecom. 67, 281, 298, 307-08 in telecommunications Broadcasting, 32, 36 n47, 340. See also Capital subscription charges. See Subscriber Radio telephone systems financing Bryan, Elizabeth B., 252 n22 Caribbean countries. See specific countries. Build-operate-transfer arrangements, 312 Carsberg, Sir Bryan, 281, 283 n14 Bulgaria, 318 n4 (:C[R (International Consultative Commit- Bureau for Telecommunications Develop- tee for Radio Communication), 34 n21 ment (BDr), 34 n21 cciTT (International Consultative Commit- Burkina Faso, 80 n7, 203, 208 tee on Telephone and Telegraph), 34 Burma. See Myanmar n21, 60 n18, 86-90, 93, 96 n3, 100, 262 Businesses, locational decisions of, 126, ciD-Ro, 338 130-31 Cellular telephones, 39, 45, 60 n19, 316, Businesses, telecommunications use by: in 319, 327, 337, 343 Asia, 23, 25, 186-88, 204, 207, 215, 219- Central America. See specific countries 23, 226; consumer surplus associated with, Central and Eastern Europe. See Eastern 179-82; for coordination of activities, 18: Europe. See also specific countries. in Greece, 23; in Egypt, 208, 219; in Latin Chile: business telephone use in, 206-07, America, 23, 206-07, 219-20, 238; on res- 220; cellular telephones in, 327; competi- idential lines, 118 n6, 200, 217 n4, 227, tion in, 327; consumer surplus estimation 250; rcos for, 236, 238, 401; purposes of, in, 166, 174: cost-benefit analysis of tele- 2 19-24; in rural and urban areas, 203-11, communications in, 166, 174-75; excess 220-21; in Sub-Saharan Africa, 23, 183- demand in, 33 n8; local calling in, 133 85, 200, 222-24; telephone density and, n26, 212; private networks in, 327: 200-03; and use of postal systems, 367. See privatization in, 79, 326; purposes of tele- also specific industries communications use in, 212-13, 220, 224, 227; reform and regulation of tele- Cable television, 36 n48, 50 communications in, 326-27; spatial order Call attempts, 34 nIO, 173, 177, 275. 279 in, 122-23; telecommunications use in Call duration. 271, 383 urban areas in, 212-13; terminal equip- Call metering, 44, 249, 265, 267 n12, 269 ment market in, 327; use of rCos in, 174, n36, 275-76, 283 n18 239, 244-46 Call-Telated externalities, 283 nil China, 319, 349, 352 Cameroon, 24 Christaller, Walter, 122 Canada: audio conferencing in, 353; econo- Classification of telephone lines, 201-03 mies of scale in, 46; educational use of Clawson, Marion, 194 ni telecommunications in, 353, 355-57; Collection and billing, autonomy in, 44, 66, eniergency use of telecommunications in, 72 28; information sector in, 112; radio and Colombia, 28, 102-06, 214, 284 n22 satellite systems in, 353. 355-57; substi- Combined telephone-telegraph systems, 363 rution of telecommunications for travel Commercialization of operating entities, in, 149-50; tariff analysis in, 293; tele- 70-72, 311, 317, 319-20. See also conimunications regulation in, 293, 299 Privatization; Telecommunications reform n12; telephone density in, 5; transport Communications input coefficients, 101, and energy consumption in, 136 102-06 INDEX 457 Communications output distribution coeffi- Cost-benefit analysis, 30-31, 157-58, 164- cients, 102 66, 187-88 Community Telephone Centers, in Peru, 24 Cost reduction effect, 180 Community Teleservice Centres (CTCCS), Costs: of access, 282 n4, 424; accounting 355 for, 71; of cellular systenms, 39, 60 nl9; Compania de Telefonos de Chile, 326 depreciation and, 282 n6; and economies Competition: in advanced services, 330; ef- of scale, 46-53; as factor in tariff setting, fects of mobile radio service on, 312, 260, 265, 270-71, 286-90, 298 n7; fall- 329; in industrial countries, 19, 35 n28, ing, 17, 37, 54-55, 56, 136; information 307, 308; in Latin America, 315-16, 327; and transaction, 126-, of inputs, 287-88; in local and long-distance services, 264, of introducing time-of-day metering, 266, 312, 316; regulation of, 259, 296, 311- 269 n36; of local and long-distance serv- 13, 327; between telecommunications ices, 61 n25, 263-64; location decisions and postal systems, 68, 373, 375; in tele- and, 122; of microwave and satellite sys- communications equipment, 267 n12, tems, 38. 40, 60 nI 1-12, 60 n19, 359; of 312, 316, 318 nl; X-efficiency and, 281 optical fiber systems, 41, 59 n9; organiza- Complementarity of telecommunications tional efficiency and, 52; of rco use, 248; and transport, 136 of software, 41; technological innovation Computer conferencing, 21, 337, 339 and, 45; total long-run incremental, 289- Computer software, 41, 43 90; traffic volume and, 264; of transport Congestion of telecommunications net- and personal travel relative to telecommu- works, 15, 34 niO, 177, 275 nications, 17, 133 n20, 136, 192 Connection charges, 271, 290 C5te d'lvoire, 324, 374 Consumer surplus, measurement of: best- Courier systenms, 369 alternative method of, 164-69; for busi- Cross-sectional analyses, 86-88, 94 nesses, 179-82; definition of, 160; in Cross-subsidies, 266 n2, 293-94 Egypt, 176, 177; expenditure method of, Currency overvaluation, marginal cost esti- 173-95; in Latin America, 161-64, 166, mation and, 287 174-75; in Kenya, 179-82; rcos and, Customer premises equipment. See Termi- 160, 166, 169, 174-75, 188-89; price- nal equipment change method of, 161-64; in rural areas, 175; and shape of demand curve, 171 n9; Data bases, on-line, 268 n20, 338, 340 travel costs and, 173 Data communications, 9, 22, 36 n48, 139, Continuing education, 348, 354 318 nl, 327, 339-40 Corruption, 310 Demand: for advanced services, 9; backward Cost allocation categories, 422-25 linkages and, 107; derivation of function Costa Rica: consumer surplus in, 161-63, of, 164-66, 167, 171 n9, 174-78, 191; 175, 188; education and telephone use diurnal variation in, 271; price elasticity in, 215; growth rate of telecommunica- of, 161, 260, 267 n15, 268 n26, 273-74, tions services in, 309; Pco use in, 128, 381-83; in rural areas, 125; by stage of 132 n8, 161-63, 175, 236-38, 244, 247. development, 118 n4. See also Consumer 361, 379-412; purposes of telecommuni- surplus, measurement of; Excess demand cations use in, 228; residential use in, for telecommunications 215, 228; rural use of telecommunica- Density. See Telephone density tions in, 132 n8, 215, 236-38; spatial Depreciation costs, 282 n6 order in, 122; substitutability between Deregulation. See Regulation; Teleconmmu- posts and telecommunications in, 361; nications reform telephone density in, 251 n7; transport Desktop publishing, 339, 340 cost savings in, 175 Development support index, 90-91 Cost-based pricing, 260, 265, 270-71, 286- Digital remote stand-alone line units, 62 90, 298 n7 n38 458 INDEX Digital switching systems, 41, 50 telecommunications use in, 219; repeat Discounting, 159, 271, 283 n17 call attempts in, 177; rural use of tele- Diseconomies of scale, 52-53, 62 n36 phones in, 208, 232 nl 1; use of alterna- Distance of call, 188, 263-64, 268 n25. See tives to telephones in, 231. 232, 361 also Local calling; Long-distance services Elasticities. See Price elasticities Distance traveled to use Pcos, 238, 398 Electricity consumption, telephone use and, Distributional issues, 20, 126, 158, 199- 90 217, 217 nl, 250, 276, 284 n22, 284 n26 Electronic analog exchanges, 40 Diversity of supply, 311-13 Electronic banking, 22, 23, 364 Duopoly system in United Kingdom, 307, Electronic blackboards, 154 n43, 352 308 Electronic Computer Oriented Mail (E- Duration charges, 271 coM), 364 Electronic message transfer systems (elec- Earth stations. See Satellite systems tronic mail), 22, 68, 337, 339, 347, 363- East Asia, information sector in, 113, 333 65, 376 n 10 n2 Electronic route guidance, 139, 153 n13 Eastern Europe, 94, 309, 317-18. See also El Salvador, 132 n8, 163-64, 219 specific countries Emergency use of telecommunications: and East Germany (former), 318 n2 in insolation, 226-28; in Asia, 26, 27, Ecological issues, 27, 136 141-42, 167, 178, 220-21, 229; cellular, Econometric method of marginal cost esti- radio, and satellite systems and, 342-44; mation, 289 in Egypt, 233 n20; in Fiji, 27; by health Economic efficiency of tariffs, 270-82, care services, 26-27, 342-44; in indus- 291-93 trial countries, 28, 228, 229; in Latin Economic rate of return, 16, 165-67 America, 27, 28; Pcos and, 238, 243, Economic reform in developing countries, 247, 250, 411; in Sub-Saharan Africa, 81 nlS, 310. See also Telecommunica- 27, 222, 243, 323 tions reform Empresa Nacional de Telecomunicaciones Economies of scale, 46, 47-50, 53, 55, 59 (Chile), 326 nl, 61 nn2l-26, 62 n36, 62 n39, 263 Energy use: savings from telecommunica- Economies of scope, 62 n35 tions, 17, 136-37, 149-51, 154 n45, Educational use of telecommunications, 24, 183-85; by telecommunications systems, 214, 215, 267 n20, 336, 348, 349, 353- 42-43, 151 57, 368-69 Engineering cost approach to economies of Education of Pca users, 207-08, 214-15, scale, 46 236, 238, 240, 248-49, 361, 388, 410 Engineering-economics model of marginal Efficiency: of billing and collection, 72; of en- cost analysis, 290 ergy use, 154 n45; organizational, 52; tariff Engineering planning model of marginal setting and, 270, 291-93; transport, 24- cost analysis, 289 26, 137-41, 175-75, 185-86; X-efficiency, ENTEL (Peru), 27 281-82, 285 n29, 285 n31, 298 Environmental issues, 27, 136 Efficiency prices. See Shadow prices Equipment, telecommunications: competi- Egypt: agricultural use of telephones in, 232 tion among suppliers of, 267 n12, 312, nil; borrowing of telephones in, 228; 316, 318 nl; costs of, 38-42, 422-25; business use of telephones in, 208, 219; domestic manufacture of, 79 nl; in East- combined telephone-telegraph systems in, ern Europe, 317; inventories of, 72; mar- 363; congestion of telecommunications kets in used, 54; scarcity of appropriate, systems in, 177; consumer surplus in, 63 n41; switching, 40-41, 43, 44, 46, 50; 176, 177; education of PcO users in, 208; terminal, 41, 312, 316, 318 nl, 327; emergency use of telephones, 233 n20; transmission, 38-40, 43, 46, 47-50. See excess demand for telephones in, 176; also specific types postal savings banks in, 372; purposes of Ethiopia, 25-26, 80 n7, 201, 209, 222, 234 INDEX 459 European Community, 318 nl operating entities by, 65, 78, 295, 305; Evans, D. A., 252 n22 revenues from telecommunications serv- Excess demand for telecommunications: in ices, 16, 34 nI 2, 56; role in telecommuni- developing countries, 9, 15-16; in Egypt, cations sector of, 66-67; use of tele- 176; in Latin America, 33 n8, 163; in phones by, 26, 204, 252 n25, 403. See Myanmar, 178; price rationing and, 274; also Operating entities; Regulation; Taxa- residential pricing under, 276; subsidies tion; Telecommunications reform and, 256; tariff policy and, 65, 73; travel Great Britain. See United Kingdom substitution and, 148. See also Waiting Greece, 23, 101, 125 lists for telephones Green Paper on Telecommunications, 317, Expenditure method of estimating con- 318 nl sumer surplus, 173-95 Growth rates of telephone systems, 9, 56 Externalities, 170 n7, 271-72, 283 nll, Guatemala, 213-14, 219-20 330 Guinea, 322, 324 Guyana, 27, 344, 346, 348 Facilities sharing, 52 Facsimile machines, 9, 24, 33 n3, 33 n7, Halina, J. W., 112 45, 144, 338, 339, 350, 376 n7 Hlarkness, R. C., 131 Factor analysis of Pco use, 389-94, 409 Health care delivery, use of telecommunica- Fairness issues in telecommunications pric- tions for, 26-27, 340-47, 366 ing, 274 Healthnet, 347 Falling costs of telecommunications, 17, 37, Heymann, H. J., 98 nl9 56, 136 Historical accounting costs, 270, 293 Families, benefits of telecommunications Hudson, Heather E., 22 to, 18 Human resources. See Staffing issues Federal Communications Commission Hungary, 318 n3 (United States), 61 n22, 299 n20 Fiber optic systems, 39-40. 41, 44, 47, 59 iRM (International Business Machines Corp.), n7, 59 n9 144 Fiji, 27, 319, 349 nCE (Instituto Costarricense de Electrici- Finance. See Investment dad), 379 Financial institutions, telephone use by, 210 II:RP (International Frequency Registration Financing sources for developing countries, Board), 34 n21 74-79 Improved vehicle use. See Transportation Foreign aid, 34 nI8, 51, 76 (use of telecommunications to improve) Foreign exchange, 75-76, 80 n12, 179-80, Income distribution. See Distributional 287 issues Foreign investment, 81 nlS, 310, 333 n9 Income, household: backward linkages to, Foreign personnel in operating entities, 69, 107; business and residential telephone 71 use and, 207, 212-14; Pcos and, 236, Forest-resources management, 27 238, 243, 387, 397, 410; tariff setting France, 5, 19, 22, 35 n28, 99-100, 151 and, 260; telecommunications density Frequency of use of PCos, 398-99 and, 86-88, 90-91, 93 Incremental capital-output ratio, 74 Gellerman, Robert F., 90 Independent Commission for World Wide German Democratic Republic, 318 n2 Telecommunications Development, 33 Germany, 99, 112, 122, 318 n2 India: banking in, 23, 372; business tele- Ghana, 27 phone use in, 23, 25, 207, 219-21, 236; Gilling, E. T W., 90, 91 diseconomies of scale in, 62 n36; emer- Goddard, J., 131 gency calling in, 26, 27, 167, 178, 220- Gordon, D. A., 139 21; estimation of consumer surplus in, Government: ownership and financing of 167, 169, 178-79; government revenues 460 INDEX from telephone service in, 34 n12; in- Interest rates, marginal cost estimation and, comes of telecommunications users in, 287 207; input-output analysis in, 100; long- Internal rate of return, 15, 159, 163-66, distance calling in, 220; off-peak dialing 170 nn2-3, 177, 182 in, 221; rao use in, 169, 236, 239, 240, International Business Machines Corp. 243, 244; purposes of telecommunica- (IBM), 144 tions use in, 26, 27, 219-21, 240, 243, International Extension College (United 351; rural use of telephones in, 207, Kingdom), 369 220-21; subscriber trunk dialing in, 178- International Frequency Registration Board 79; tariffs and tariff setting in, 57, 284 (IFBR), 34 n21 n19; telecommunications reform in, 319; International Telecommunication Union transport and telecommunications use in, (IL:), 20, 33 nl, 34 n21, 80 n6, 96 n3, 140; two-part call charge in, 284 n19 141, 320 Indonesia, 319, 333 n2, 350, 369 Inventory management, 72, 193 Industrial countries: competition in, 19, 35 Investment in telecommunications: auton- n28, 307, 308; educational uses of tele- omy in planning for, 66; foreign, 81 nIS, communications in, 353-57; emergency 310, 333 n9; indivisibilities in, 46, 288, use of telecommunications in, 28, 228, 298 n7; in industrial countries, 5, 19; 229; growth rates of telephone systems in, input-output analysis of, 115-16; internal 9; improved use of vehicles in, 138-39; generation of funds for, 74; marginal cost information sector in, 112-13; level of analysis and, 288; in newly industrializing telecommunications investment in, 5, 19; countries, 19; rates of return on, 15-16, peak pricing in, 265; postal services in, 158-60; restrictive and activist views on, 366, 370, 371; privatization in, 19, 281, 19-22; shortage of capital for, 309; 307-08; radio telephone systems in, 138, sources for developing countries of, 74- 355-57; regulation and reform of tele- 79; strategy to complement efficiency communications in, 19, 293, 299 n12, pricing for, 279-81; tariff setting and, 307; residential telephone lines in, 200, 279-81 217 n3, 228, 229; teleconferencing in, sDNt (Integrated Services Digital Network), 144-45, 147; telephone density in, 5, 99, 45, 60 n18, 328, 330 376 n4; use of telecommunications for Isolation, geographic, 26-28 health care delivery in, 342; utilization Ivory Coast. See Cote d'lvoire factors in, 90. See also specific countries Industrial use of telecommunications serv- Jamaica, 29 ices, 101-06 Japan, 19, 102-06, 112, 229, 279, 284 n24, Inflation, 59 n3, 163-64 371 Informal sector, 132 n, 134 n42 Jipp, A., 86 Information gap, 114 Joint ventures, 312 Information sector, 107-16, 205-07, 249, Jonscher, Charles, 112, 119 n29 304, 332 n2 Journalists, subsidies to, 261 Information technologies, telecommunica- Jussawalla, Meheroo, 113 tions and, 329-31 Input-output analysis, 100, 101-07, 113, Kahn, Alfred E., 282 n2 114-17 Kaul, S. N., 140 Inputs, costs of, 287-88 Kenya: agricultural use of telephones in, Integrated services digital networks (IsDmNs), 118 n16; alternatives to telephones in, 44 231, 361; business telephone use in, 23, Intelpost, 364 222-24; consumer surplus in, 179-82; ef- INTELSAT, 40, 352-53 ficiency gains of businesses in, 180-82; Interactive data communications, 339 rco use in, 240; purposes of telecommu- Inter-American Development Bank, 34 nl8, nications use in, 222, 240, 346; services 76, 234 sector use of telephones in, 210; spatial INDEX 461 order in, 123-25; telex in, 179-80; travel of scale in, 53, 62 n39; elasticity of de- substitution in, 146 mand for, 268 n26; as indicator of spatial Kilgour, Mary Cameron, 122 order, 122-25; productivity gains from, Klein, A., 133 nlO 145; rates of return on, 277-78; subsidi- Knowledge Network (Canada), 353 zation of local calls by, 277; tariffs for, Kochen, Manfred, 21 263-64, 277-78; underconsumption of. Korea, Republic of, 24, 101, 118 n4, 247 278; vehicle use and, 139; waiting times for, 252 n20 Lathey, C. E., 138, 151 Latin America: agricultural use of telecom- Madagascar, 322 munications in, 23-24, 102-06; business Mail. See Postal services use of telecommunications in, 23, 206- Mail, electronic. See Electronic message 07, 219-20, 238; competition in, 315-16, transfer systems 327; consumer surplus estimation in, Maitland Commission, 3, 33 nl 161-64, 166, 174-75; educational use of Malawi, 344 telecommunications in, 267 n20, 348; Malaysia, 201, 319, 333 n2 emergency use of telecommunications in, Mali, 203, 208 ,27, 28; excess demand for telecommuni- Management. See Operating entities cations in, 33 n8, 163; Pcos in, 128, 132 Management checklist for operating enti- n8, 161-63, 174-75, 236-39, 244-46, ties, 70-73 248, 361, 379; privatization in, 79, 314, Marginal cost analysis, 271-73, 282 n2, 282 315-16, 326; radio telephone systems in, nn6-7, 286-90, 293, 415 25, 27, 344; residential telephone lines Marginalization of nonbusiness and rural in, 200, 212-14, 215, 224, 227, 228; users, 124, 262 rural use of telecommunications in, 23, Market distortions in developing countries, 24, 166, 215, 236-38; telecommunica- 287 tions regulation and reform in, 314, 315- Market information, use of telecommunica- 16, 326-27; telephone density in, 5, 99, tions to improve, 16-18, 23-24, 179-80 251 n7. See also specific countries Market-response view of telecommunica- Learn Alaska, 349, 358 tions development, 21 Learning Channel (United States), 354 Mauritius College of the Air, 369 Leibenstein, Harvey, 281 Mayer, Martin, 217 n3 Lesotho, 346 Medical use of telecommunications. See Lexis Financial Information Service, 268 Health care delivery, use of telecommuni- n20 cations for Lightning (emergency) calls, 167. See also Mercury Communications, Ltd., 307 Emergency use of telecommunications Metering of calls. See Call metering Ling, S., 90 Mexico, 34 n12, 81 nIS, 238, 279, 315-16 Literacy, 368-69 Meyer, John R., 61 n 22, 266 nl, 268 n22 Local calling: competition in, 312; costs of, Microwave systems, 38, 47 61 n25; determinants of traffic of, 93; Migration, rural-urban, 124, 128-29 metering for, 267 n12; service area for, Minitels, 22, 35 n28 127, 262-63; subsidization by long- Mitchell, Bridger M., 267 n12, 268 n24 distance services of, 277; tariffs for, 262- Mobile communication terminals, 152 n6 63, 275, 290 Modems, 33 n3, 144 Locational decisions of businesses, 121-34 Monopoly structure of operating entities, Locationalist view of regional imbalances, 259, 270, 279-82, 296, 305-16, 325-31, 126 373-76. See also Competition; Operating Long-distance services: breakdowns in, 60 entities n17; business use of, 268; competition in, Moscow Electrotechnical Institute of Com- 264, 312, 316; costs of, 61 n25, 263-64; munication, 145 determinants of traffic in, 93; economies Multiaccess radio systems, 39, 186 462 INDEX Multilateral development banks, 34 n18, tion of, 70-72, 311, 317, 319-20; de- 76-78, 234. See also World Bank mand responsiveness of, 80 n5; financial Multiple tariff systems, 276 viability of, 15, 256; foreign exchange re- Myanmar, 25, 26, 177-78, 363 quirements of, 75-76; government fi- nancing and ownership of, 65, 78, 295, 305, 310-11, 313; management checklist Ntata Te 35 o for, 70-73; organizational efficiency of, Stateso, 354 52, 64-65, 280; overexpansion of, 53; Natonal Telephone Program (Philippines), planning and quality control by, 70-71; Natural monopoly, teleconimunications as, separation of policy and regulatory func- 296 tions of, 313, 314; staffing issues of, 54, Nepal: business telephone use in, 221-22; 62 n33, 66, 67, 69, 70, 72, 80 nn6-7; classification of lines in, 201; educational subscriber financing of, 278; value added ctasifiatio ofline in 201 edcatinai services provision by, 330 use of telecommunications in, 214; emer- Operations research, 193 gency use in, 229; purposes of telecom- Opertions capitrch 4J4 muniatios us in 22122, 27, 29; Opportunity cost of capital, 414 residentials tselepone usein, 214, 227, Optical fiber cable systems, 39-40, 41, 44, residential telephone use In, 214, 227 4 5n7599 229; services sector use in, 210, 217 n16 47, 59 n7, 59 n9 telephone density in, 233 n12; telex in, Optimization model of marginal cost analy- 217 n16; use of telephones in tourism in, sis, 289 *28; waiting lists for telephones in, 210, Organizational efficiency and costs, 52 221 Organization of African Unity, 29 Newly industrializing countries, 19. See also Outsourcing, 313 specific countries New Zealand, 373 Packet switching, 23, 35 n28, 365 NICNET (India), 339 Pakistan, 154 n37, 164-65, 319 NICNicol, (Indio , 339 Pan-African Telecommunications Network, Nicol, Lionel Y., 126 21 Niger, 322-23 Panama, 90 Nigeria, 369 Papua New Guinea: business use of tele- Nonhomogeneity J. telephone calls, J90 communications in, 223; emergency use Nonhomogeneity of telephone calls, 190, of telephones in, 229; Pco use in, 236, 34 .238, 240, 243, 244-49, 361; purposes of North American Free Trade Agreement, telecommunications use in, 223, 224, 316 tlcmulain s n 2,24 227, 229, 240; residential telephone use in, 224, 227, 229; substitutability of pos- O'Brien, Rita Cruise, 98 n19 tal services and telecommunications in, Office of Telecommunications (OFTEL), 307. 361-62; telephone density in, 227, 233 308 n12, 251 n7; telex in, 28 Off-peak dialing, 221 Paraguay, 23-24 Off-peak pricing, 271, 283 nI7, 293 Park, Rolla Edward, 267 n12 Offshore data processing, 144 Parker, Edwin, 21 Oil-surplus developing countries, telecom- Pcos. See Public call office (PCO) tele- munications investment programs in, 18 phones Okundi, Philip, 21 Peak pricing, 73, 265, 275, 269 n31, 275, On-line data bases, 268 n20, 338, 340 282 n5, 283 n17 OPEc countries, telephone density in, 94 Peak use periods, 268 n29 Open Network Provision, 318 nl People's Democratic Republic of Yemen, Operating entities: accounting systems for, 141-42, 149 71-72; auditing of, 72, 73; autonomy of, People's Republic of China. See China 44, 65-68, 70, 72; capital structure and Personnel. See Staffing issues requirements of, 73-74; commercializa- Peru: audio conferencing in, 27, 351; busi- INDEX 463 ness use of telephones in, 238; Commu- Africa, 324; tariff setting and, 295; nity Telephone Centers in, 24; educa- X-efficiency and, 281 tional use of telecommunications in, 348; Procurement, 50, 51, 79 nl emergency use of telecommunications in, Productivity, 16, 17, 56, 61 n23, 145, 296, 27; health care use of telecommunica- 297 tions in, 27; PCO use in, 238, 245, 248; Project Access, 352 radio and satellite systems in, 27, 350- Project SHARE, 350, 352 51; rural use of telecommunications in, Public call office (Pco) telephones: in Asia, 132 n8, 238; teleconferencing in, 27, 24, 169, 236, 238-40, 243, 244-49, 361; 351; travel reduced by telecommunica- business use of, 236, 238, 401; call dura- tions in, 24 tion at, 383; call metering and, 249; con- Philippines, 79, 102, 186-88, 363 straints on use of, 246-49; consumer Planning issues, 70-71, 73, 129, 257. See surplus from, 160, 166, 169, 174-75, also Investment in telecommunications; 188-89; costs of, 248; determinants of Operating entities use of, 334-94; distance traveled to use, Political use of telecommunications, 29 238, 398; education and use of, 208, 215, Pool, Ithiel de Sola, 21, 130, 142, 228 236, 238, 240, 248, 388, 410; emergency Porat, Marc Uri, 112, 119 n29 use of, 238, 243, 247, 250, 411; employ- Postal savings banks. 371-72 ment of users of, 129, 395-97; factor Postal services: banking and business serv- analysis of use of, 389-94, 409-10; fre- ices provided by, 367, 371-73; competi- quency of use of, 388; government use of, tion between telecommunications and, 401; importance to users of, 244-46; in- 68, 373, 375; in developing countries, come and use of, 236, 238, 240, 243, 364-65, 369, 373; electronic message 397, 410; in Latin America, 128, 132 n8, transfer and, 363-64, 376 nlO; employ- 161-63, 174-75, 236-39, 244-46, 248, ment by, 374; independence of telecom- 361, 379; location in postal service facili- munications services from, 67-68, 375; ties, 68; migration, rural-to-urban, and, in industrial countries, 366, 370, 371; lo- 128; price elasticity of demand for, 381- cation of PCOs in, 68; promotion of 83; pricing strategy for, 280; purposes of health, education, and literacy by, 366, use of, 240-43, 383-410; quality of serv- 368-69; services to agriculture by, 365- ice of, 234; radio as alternative to, 245; 66; subsidies to and by, 68, 377 n30; sub- rates of return on, 166; rural and urban stitution of telecommunications for, 167, use of, 128, 132, 166, 175, 188-89, 236- 361-62; transport and, 369-71, 377 n30 40, 379; social use of, 246, 248, 401, Power sources for telecommunications sys- 410; in Sub-Saharan Africa, 234, 239, rems, 39, 42, 59 n6 240, 246, 248; use by women of, 238, Price-cap regulation, 296-98, 308, 316 240, 248; use in reducing transport costs Price-change method of consumer surplus of, 141, 166, 169; use in tourism of, 382; estimation, 161-64 in Vanuatu, 188-89 Price elasticities, 165, 171 nlO, 189-93, Public goods, externalities and, 170 n7 267 nIS, 268 n26, 273, 381-83 Public radiocall service, 186 Price rationing, 274-75 Pulse code modulation systems, 41, 60 n 15 Pricing. See Tariffs and tariff setting Purposes of telecommunications use, 219- Principal component analysis, 98 n19 24, 227-29, 240-43, 249, 383-410; in Private automatic branch exchanges, 33 n9 Chile, 212-13; in India, 26, 27, 219-21, Private networks, 65, 327 240, 243, 349; in Kenya, 222, 240, 344 Privatization: conditions for success of, Pye, R, 131 324-25; in industrial countries, 19, 281, 307-08; in Latin America, 79, 314, 315- Quality control by operating entities, 70 16, 326; and other forms of private sector Quality of service, 56, 195 n22, 234, 298, participation, 313; regulation under, 295, 316 314, 327; steps in, 324; in Sub-Saharan Queueing, 274 l 464 INDEX Radio telephone systems: as alternative to Rural areas: in Asia, 23, 24, 207, 220-21; rCos, 245; antimonopoly effects of, 312, business use of telephones in, 207-11, 329; in Asia, 168-69, 319; for control of 220-21; consumer surplus from tele- vehicles, 138; educarional use of, 348; for phone use in, 175; danger of marginaliza- emergency service, 342-43; in industrial tion of, 124; demand for telecommunica- countries, 138, 355-57; in Latin Amer- tions in, 125; in Egypt, 208, 232 nil; in ica, 25, 27, 344; rnultiaccess, 186; social Ethiopia, 209; in Europe, 125; govern- services delivery via, 335, 337; in Sub- ment use of telephones in, 252 n25; in Saharan Africa, Z5-27; in USSR, 138. See Latin America, 23, 24, 166, 215, 236-38; also Cellular telephones local service area size in, 263; migration Ramsey pricing, 273-74, 283 nl4 to cities from, 128-29; rcos in, 128, 132, Rate averaging, 259 166, 175, 188-89, 236-39, 379; residen- Rate-of-return regulation, 259, 296 tial use of telephones in, 201, 215; satel- Rates. See Tariffs and tariff setting lite systems in, 35 n26; subsidies to, 127, Rates of return. See Economic rate of re- 252 n26, 285 n28; tariff policy for, 127; turn; Internal rate of return telephone density in, 5; value-of-service Rationing, price, 274-75 pricing in, 267 n8 gCA, 145 Rural Communications Services Project Reform of telecommunications sector. See (Peru), 350 Telecommunications reform Rural free delivery, 366, 367. 370, 373, 377 Regulation: of access charges, 259; of ad- n33 vanced services, 329-30; advisory board Rural Satellite Program, 349-51 proposed for, 328, of competition and Rural Telecommunications Development monopoly, 259, 296; to extend access to Project (Philippines), 186 telecommunications, 259; in industrial countries, 19, 293, 299 nI2, 307; in Latin America, 315-16, 326-27; price- Salaries, autonomy In setting 66 cap, 296-98, 308, 316; privatization and, SatelLi'e s47 295, 314, 327; productivity and, 296, Satelite systems: advantages over terrestrial 297; quality of service and, 298; rare-of- systems of, 336; applications of, 45; costs return, 2, 2' s o oof, 40, 60 nl 1-12, 60 n19, 263, 357; de- return, 259, 296; separation of operations livery of social services by, 335-37, 348- trom, 313-14; tariff setting and, 259, 50, 353-54; emergency use of, 342-43; in technological innovation and, 329 Indonesia, 350; in industrial countries, Reliability of telecommunications systems, 60 nI9, 343, 353-54; in Peru, 27, 350; 43-45 pilot projects in, 349-53; for rural and Rental of telephone lines and equipment, isolated areas, 35 n26, 336, 349-51; in 63 n42, 176, 267, 284 n21 Sub-Saharan Africa, 21, 35 n23; subsi- Repeat call attempts, 34 nlO, 173, 177, 275 dies for, 358; teleconferencing via, 337; in Resale of telephones and connections, 173, West Indies, 351 176, 177-78, 194 nIl , 275 Scale and scope economies. See Economies Residential telephone lines: in Asia, 214, of scale; Economies of scope 215, 227-29; business calls on, 118 n6, Scandinavia, 5. See also Sweden 200, 217 n4, 227-28, 250; in industrial Scherer, F. M., 285 n31 countries, 200, 217 n3, 228, 229; in School of the Air, 354 Latin America, 200, 212-14, 215, 224. Schoppert, D. W., 139 227, 228; social use of, 224-27; subsidies Secretaria de Comunicaciones y Trans- to, 260; tariffs for, 260-62, 276-77; tele- portes (Mexico), 315 phone density and, 200-01; in urban and Selling price effect, 180 rural areas, 201, 213-15 Senegal, 239, 240, 243, 246, 248 Restrictive view of telecommunications de- Sensitivity analysis of internal rates of re- velopment, 20-21 turn, 170 n2 INDEX 465 Separation of telecommunications regula- Subscriber apparatus. See Terminal equip- tion from operations, 318 nl, 325 ment Services sector, use of telecommunications Subscriber financing, 259, 278, 284 n24, by, 99-105, 203-11, 217 nI6, 249 309, 313 Shadow prices, 159, 287-88, 413-17 Subscriber trunk dialing, 97 nl1, 178-79 Shortages of telephone lines. See Excess de- Subsidization: arguments against, 256-57; mand for telecommunications; Waiting cross, 266 n2, 293-94; under declining lists for telephones average costs, 273; and excess demand, Singapore, 113, 333 n2, 376 n14 256; externalities and, 170 n7, 272; and Slowscan video, 154 n43, 347, 351 investment planning, 257; of journalists, Social constraints on PCo use, 248-49 261; of local by long-distance service, Social services, use of telecommunications 277; of postal services by telecommunica- in delivery of, 335-37, 348-49, 353-54 tions, 68; of residential users, 260; of Social use of telecommunications: in Chile, rural areas, 127, 252 n26, 285 n28; of 213, 220; importance of, 18; PCOS and, satellite systems, 358; targeting of, 261; 246, 248, 401, 410; by residential users, of transport by postal services, 377 n30 224-27; in United Kingdom, 228 Substitution: between postal services and Software, computer, 41, 43 telecommunications, 167, 361-62, 410; South America, residential lines in, 200 between telecommunications and per- Southeast Asia, telephone use in, 215. See sonal travel, 149-50; between telecom- also specific country munications and transport, 17, 76, 136, Space requirements of telecommunications 168-69 systems, 42-43 Sweden, 88-90, 99, 127, 138, 200, 370 Spatial analysis and spatial order, 22, 121- Switching equipment, 40-41, 43, 44, 46, 50 25, 130, 188-89 Switzerland, 5, 200 Special Telecommunications Action for Re- Syria, 211. 215, 228 gional Development (STAR), 112 Sri Lanka, 23-25, 214, 215, 227, 319-20 Taiwan, 56, 90, 279 Sri lanka Telecom, 320 Tanzania, 26, 183-85, 344, 373 Staffing issues, 54, 62 n33, 66, 67, 69-70, Targeted subsidies, 261 8u n7 ' Tariffs and tariff setting: in Asia, 56-58, 80ate en7erprises, operating entitiesas,65 284 n19; autonomy of operating entity State enterpris, in, 66; for business and residential lines, 305, 3et1-11, 313 260-62, 276-77; costs as basis of, 260, Steven, Peter M., 228 265, 270-71, 286-90, 298 n7; distance Stored program control, 44, 45 called and, 263-64; efficiency and, 270, Structuralist view of regional imbalances, 279-81; under excess demand, 73; fair- 126 ness issues associated with, 274; house- Subcontracting, 333 n7 hold income and, 260; in industrial Sub-Saharan Africa: advanced services in, countries, 35 n28, 293; international dif- 9; agricultural use of telecommunications ferences in, 57-59; investment strategies in, 118 n16, 119 n25, 185-86. 232 nil ; to complement, 279-81; for local calling, business use of telecommunications in, 262-63, 275; local service area size and, 23, 183-85, 200, 222-24; emergency use 262-63; for long-distance services, 263- of telecommunications in, 27, 222, 243, 64, 277-78; multiple systems for, 276; 323; PCos in, 234, 239, 240, 246, 248; peak and off-peak, 73, 265-66, 269 n31, political use of telecommunications in, 269 n36, 271, 275, 282 n5, 283 n17, 29; postal systems in, 373; privatization 293; practical aspects of, 286-99; princi- in, 324; satellite systems in, 21, 35 n23; ples and objectives of, 255-60; privatiza- telecommunications reform in, 320-24; tion and, 295; Ramsey pricing approach telephone density in, 5. See also specific to, 273-74, 283 n14; for rural areas, 127; countries in Sub-Saharan Africa, 57, 58; techno- 466 INDEX logical innovation and, 44; for telex, 268 Thailand: advanced services in, 9; alterna- n27; through two-part call charge, 284 tives to telephones in, 229; business use n19; value-of-service principle and, 260, in, 204, 219; classification of telephone 262-63, 265, 267 n8 lines in, 201; emergency use of telecom- Taxation, 34, 56, 67, 257, 280 munications in, 229; government use of Technical assistance from World Bank, 78 telecommunications in, 204; residential Technological innovation, 17, 37-45, 54- telephone use in, 227; subscriber financ- 55, 262, 329 ing in, 279; substitution between posts Telecommunications operating companies. and telecommunications in, 361; tele- See Operating entities communications reform in, 319; tele- Telecommunications reform: in Asia, 319; phone density in, 233 n12; travel substi- competition and, 327; in Eastern Europe, tution in, 146 317-18; in Fiji, 319; future of, 331-32; in Thomas, Brownlee, 98 n19 industrial countries, 19; in Latin Amer- 3M Company, 145 ica, 314, 315-16, 326-27; need for prag- Time-of-day pricing, 265-66 matism in, 332; political economy of, 334 Time required to reach a telephone, 174-75 n15; in Sub-Saharan Africa, 320-24; Time savings from telecommunications use, World Bank lending for, 78 185-86 Telecommuting, 143-44, 153 n24 Time-series analyses, 88-90, 94 Teleconferencing: in Cameroon, 24; by Togo, 57 computer, 21, 337, 339; educational use Toll ticketing, 73, 276 of, 349; energy savings from, 150; in in- Touch-tone telephones, 42, 284 n21 dustrial countries, 144-45, 147; in Peru, Tourism, telecommunications use and, 28, 27, 351; via satellite, 337; for social serv- 29, 93, 210, 382 ices delivery, 337-39; as substitute for Trade negotiations over telecommunica- travel, 36 n48, 144-45, 147; use of video tions, 310 with, 145, 154 n43, 352; in West Indies, Trading in "telephone rights," 275 351 Training of operating entity employees, 69- Telecottages, 354-55 70, 72, 80 n6 Telefonos de Mexico, 79, 315-16 Transaction costs, 126 Telemedicine. See Health care delivery, use Transmission equipment, 38-40, 43, 46, of telecommunications for 47-50 Telephone density: in Asia, 5, 227, 233 Transportation: complementarity of tele- n12, 251 n7, 319; business use and, 200- communications with, 136; congestion of 01; correlation with national wealth and - systems of, 130, 133 n20; costs of, 17, income of, 86-91; in Eastern Europe, 94, 136, 165; energy use of, 136; postal serv- 317, 318 nn3-4; as indicator of supply, ices and, 369-71, 377 n30; relecommuni- 97 n12; in industrial countries, 5 99 cations as substitute for, 17, 148-49, 376 n4; in Latin America, 5, 99, 251 n7; 168-69, 210; use of mobile radio for con- in OPEC countries, 94; residential use and. trol of, 138; use of Pcos and, 141; use of 200-01; in rural areas, 5; in Sub-Saharan telecommunications to improve, 24-26, Africa, 5 137-41, 174-75, 185-86 Telephone equipment. See Equipment, tele- Transport economics, 193, 194 n2 communications Travel: consumer surplus and costs of, 173; energy use for, 149-51; estimating costs Telephone lines, definition and classifica of, 192; and excess telecommunications tion of, 33 n2, 200-03 demand, 148; generated by telecommuni- Television University, 352, 358 cations, 147-48; telecommunications as Telex, 9, 28, 93, 179-80, 217 n16, 268 n27 substitute for, 24, 142-51, 154 n37, 164- Terminal equipment, 41, 312, 316, 318 nl, 66, 168-69, 174-75, 185-86, 188 319, 327 Trunk calls. See Long-distance services Tertiary sector. See Services sector, use of Two-part call charging, 284 n19 telecommunications by Tyler, Michael, 151 INDEX 467 Uganda, 26, 34 n12, 58, 185-86 persal of activity in, 129-30; Pcos in, 239, Unauthorized transfers. See Resale of tele- 240 phones and connections Urban planning, 129 UNESCO, 261 Urgency of use. See Emergency use of United Kingdom: competition in, 19, 307- telecommunications 08; duopoly system in, 307-08; emer- Uruguay, 5, 23, 201, 214, 219 gency use of telecommunications in, 228; U.S. Agency for International Development energy savings from travel substitution in, (AIL)), 349-51 149-51; growth in use of telephones in, USSR, 138, 145-46 361; improved vehicle use in, 138; infor- Usage-sensitive pricing. See Call metering mation sector in, 112; privatization in, Utilization factor analysis, 90 19, 281, 307-08; purposes of use in, 228; U'WIDITE, 349 Ramsey pricing in, 283 n14; residential telephone use in, 228; social use in, 228; Value added services, 312, 318 nl, 330. See substitution of telecommunications for also Advanced services travel in, 143; telecommuting in, 143; Value-of-service principle of tariff setting, teleconferencing in, 143, 145; telephone 260, 262-63, 265, 267 n8, 270 density in, 376 n4; travel generated by Vanuatu, 188-89 telecommunications in, 147 Vehicle use. See Transportation United Nations Development Programme, Venezuela, 130 69, 320 Video conferencing, 145, 352 United States: agricultural use of telecom- Videograph, 154 n43 munications in, 102-06; breakdowns in Video, slowscan, 154 n43, 329, 347, 351 service in, 60 n17; call completion rate Video telephones, 36 n48 in, 279; deregulation and divestiture in, Videotex, 22, 35 n28 307; economies of scale in, 46; educa- Vogelsang, Ingo, 268 n24 tional use of telecommunications in, 354; VSAT (very small aperture terminals), 40, 60 electronic message transfer in, 363; gene- nI2, 338, 339 ration of travel by telecommunications in, 147; improved vehicle use in, 138, 139; Waiting lists for telephones, 9, 33 n8, 163, indirect energy use by telecommunica- 178, 210-11, 221, 256, 274, 283 nlS tions in, 151; information sector in, 112; Warren, Colin, 118 n12 input-output analysis for, 115; mobile Webber, Melvin M., 21, 22, 130 communication terminals in, 152 n6; Wildlife protection, 27 postal services in, 366; residential use in, Willingness to pay, as measure of consumer 217 n3; satellite systems in, 354; size of surplus, 158-60, 170 n6 local service area in, 268 n22; spatial Women, Pco use by, 238, 240, 248 order in, 130; telecommuting in, 144; Wood, H. C., 139 teleconferencing in, 144, 147; telephone World Bank, 34 nll, 34 n18, 75, 76-78, 97 density in, 5, 201; telephone rental in, nll, 234, 419-23 284 n21; use of alternative media in, 251 World Health Organization, 343 nlS University of the South Pacific, 349, 352 X-efficiency, 281-82, 285 n29, 285 n31, University of the West Indies, 349, 351 298 Uno, Kimio, 119 n34 Upper Volta. See Burkina Faso Yatrakis, P. G., 93 Urban areas: business and residential use of Yemen, People's Democratic Republic of, telecommunications in, 201, 203-07, 141-42 212-14; diseconomies of scale in, 53; dis- I _~~~~~I - I *_ :11 41..1 11 114 ==4 * 1 . - u *i ~~1 4*_ I*'. *- ~ t- i1' d ' S * * ' 9. .a *'e *W - e =85~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~5 - - -- . - 4" -I'-41 1e4 WIs|i74-i E' 4 4.4.4-a.4. Bl4lulle ,.4 * * 4-'S^erlle* - i-I * -. l .40 - I *41*: 4-.44I - . 44@wl' S * §S' 4 6' - 44s ' ] |lllu -l|tl 4' -_ *44*em|xf,-eis4s"1<^-n"l I 4 4 * 0 3 - * . - , * 4 . - 4 * a- . * --' -- ,.44 -.4 - * 4 S 4 l 4 - 9 4 4 -' -. I -i |' 4 2 4 4 - 4 *~~~~ - - S. IS~ - .[ > ;,,,,,,, , ,_ g * * - * g 44. '9' - 4 ''484~~~~~~~~~* ].404' -B III- , w- N . *i i 3 S * B l 465 DEVJ 350 * . 4 ' . - a - . 4 * *- 4 1 . I aTELEC'CMMUNICATICN & ECCN 11~~ ~~~~~ - 44 * l 4 4 - ' ' I ll 4QQQ0Q0107O0 I * *~~~~~~~~~~~ $299 * - i, , * -;