ENVIRONrM ENTALLY SUSTAINABLE DEVELOPrMVENT 14 ~ STUDIES AND MOIONOGRAPHS SERIES NO. 17 Work in progress for public discussion *;.sIL |qQ X Ex'panCding,) the NlC',sulre ot \\Calth 1111i'alo- ol - - _ §& '\ _ ,l4 ESD Proceedings Series 1 Culture and Development in Africa: Proceedings of an International Conference (also in French) 2 Valuing the Environment: Proceedings of the First Annual International Conference on Environmentally Sustainable Development 3 Overcoming Global Hunger: Proceedings of a Conference on Actions to Reduce Hunger Worldwide 4 Traditional Knowledge and Sustainable Development: Proceedings of a Conference 5 The Human Face of the Urban Environment: A Report to the Development Community 6 The Human Face of the Urban Environment: Proceedings of the Second Annual World Bank Conference on Environmentally Sustainable Development 7 The Business of Sustainable Cities: Public-Private Partnershipsfor Creative Technical and Institutional Solutions 8 Enabling Sustainable Community Development 9 Sustainable Financing Mechanismsfor Coral Reef Conservation: Proceedings of a Workshop 10 Effective Financing of Environmentally Sustainable Development: Proceedings of the Third Annual World Bank Conference on Environmentally Sustainable Development 11 Servicing Innovative Financing of Environmentally Sustainable Development 12 Ethics and Spiritual Values: Promoting Environmentally Sustainable Development 13 The Self and the Other: Sustainability and Self-Empowerment 14 Meeting the Challenges of Population, Environment, and Resources: The Costs of Inaction 15 Rural Well-being: From Vision to Action (forthcoming) ESD Studies and Monographs (formerly Occasional Paper) Series 1 The Contribution of People's Participation: Evidencefrom 121 Rural Water Supply Projects 2 Making Development Sustainable: From Concepts to Action 3 Sociology, Anthropology, and Development: An Annotated Bibliography of World Bank Publications 1975-1993 4 The World Bank's Strategyfor Reducing Poverty and Hunger: A Report to the Development Community 5 Sustainability and the Wealth of Nations: First Steps in an Ongoing Journey 6 Social Organization and Development Anthropology: The 1995 Malinowski Award Lecture 7 Confronting Crisis: A Summary of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities (also in French and Spanish) 8 Confronting Crisis: A Comparative Study of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities 9 Guidelines for Integrated Coastal Zone Management (Continued on the inside back cover) ENVIRONMENTALLY SUSTAINABLE DEVELOPMENT STUDIES AND MONOGRAPHS SERIES NO. 17 Expanding the Measure of Wealth Indicators of Environmentally Sustainable Development The World Bank Washington, D.C. Copyright 0 1997 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing June 1997 This report has been prepared by the staff of the World Bank. The judgments expressed do not necessarily reflect the views of the Board of Executive Directors or of the govermments they represent. Cover art: original painting, artist unknown. Market scene, Haiti. Library of Congress Cataloging-in-Publication Data Expanding the measure of wealth : indicators of environmentally sustainable development. p. cm. - (Environmentally sustainable development studies and monographs series ; no. 17) Includes bibliographical references. ISBN 0-8213-3956-7 1. Sustainable development. 2. Economic indicators. 3. Environmental indicators. L. World Bank. II. Series. HC79.E5E98 1997 338.9-dc2l 97-20084 CEP @3 The text and the cover are printed on recycled paper, with a flood aqueous coating on the cover. Contents Foreword v Acknowledgments vii Chapter 1 Introduction 1 PARr ONE INDICATORS LINKING THE MACROECONOMY AND THE ENVIRONMENT 5 Chapter 2 Are We Saving Enough for the Future? 7 Chapter 3 Measuring the Wealth of Nations 19 Chapter 4 Subsidy Policies and the Environment 40 PART Two EVOLVING INDICATOR THEMES 67 Chapter 5 Land Quality Indicators 69 Chapter 6 Social Capital: The Missing Link? 77 Chapter 7 Poverty and the Environment: Pieces of the Puzzle 94 Chapter 8 Intemational Progress in Indicator Development 99 Boxes 2.1 Pollution and genuine saving in India 10 2.2 Unanswered questions: saving and soil degradation 11 3.1 Interpreting the wealth estimates 20 3.2 Valuing protected areas 22 3.3 The wealth accountant's toolkit 23 3.4 Directions for future work 28 Appendix box 1 Valuing subsoil assets 33 iii iv Expanding the Measure of Wealth 4.1 The simple economics of win-win subsidy reform 42 4.2 Price reform problems in transition economies 47 4.3 Energy in Russia: what lies ahead? 50 4.4 China: subsidy reform in the coal sector 50 4.5 Electricity subsidies 51 4.6 Removing pesticide subsidies in Bangladesh 54 4.7 Removing pesticide subsidies in Indonesia 55 4.8 Reducing fertilizer subsidies in Bangladesh 59 5.1 The land quality indicators program 69 5.2 Examples of soil-related land quality indicators for pressure (P), state (S), and response (R) 70 5.3 Examples of contextual information to assist with the interpretation of land quality indicators 72 6.1 Three views on social capital: common features 79 6.2 Desirable properties of indicators 84 6.3 Indicators of social capital 88 8.1 Trends in passenger road transport in the United Kingdom 102 8.2 Environmental pressure indices 103 8.3 The World Bank's 1997 World Development Indicators (WDI) 105 8.4 Indicators of material flows 106 Figures 2.1 Genuine saving for Tunisia, 1970-94 8 2.2 Genuine savings rates by region 12 2.3 Genuine savings rates by region 13 2.4 Human capital and genuine saving in Chile 14 2.5 Contribution of current education expenditures to genuine saving 15 3.1 Wealth shares across selected regions, 1994 26 3.2 Wealth shares in low-income natural resource exporters, 1994 26 4.1 Estimated changes in energy subsidies in selected countries, 1990-91 to 1995-96 49 4.2 Trends in fertilizer consumption 57 7.1 Effect of poverty on predicted deforestation in Mexico, 1980-90 95 7.2 Emissions intensity of firms in rich and poor areas of Java 97 Tables 2.1 Genuine saving (percentage of GNP) 15 3.1 Countries ranked by wealth per capita, 1994 24 3.2 Increases in country rankings 25 3.3 Wealth per capita by region, 1994 25 3.4 The components of natural capital, 1994 27 Appendix table 1 Country-level natural, capital estimates 34 4.1 Indicative emissions from different fossil fuels 44 4.2 Characteristics of countries in the sample 46 4.3 Subsidy rates and total subsidies for fossil fuels, 1990-91 and 1995-96 48 4.4 Pesticide policies in selected countries, 1993 56 4.5 Explicit subsidies on urea in selected countries, 1971-90 58 4.6 Average annual government expenditures on explicit fertilizer subsidies in selected Asian countries, 1982-94 58 4.7 Irrigation charges and unit costs in selected Asian countries 61 4.8 Estimated subsidies to irrigation in developing countries, 1983-93 62 8.1 Examples of thematic indicator sets 104 Foreword he publication by the World Bank in late progress toward sustainable development. These 1995 of Monitoring Environmental Progress: include new estimates of national wealth and A Report on Work in Progress generated great genuine savings, a detailed analysis of changes interest in the use of indicators to measure the in subsidies with environmental consequences, pace and direction of change in environmentally and progress on the conceptual foundations of sustainable development. In particular, the at- social capital. While many of these new estimates tempts to redefine what it means to be "wealthy" are more refined than those presented in 1995, the or "poor" by recognizing that a country's wealth underlying story remains the same. If anything, is the combination of various forms of capital- the new estimates reinforce the importance of the produced, natural, and human resources-led to natural resource base of all economies as well as new thinking on what constitutes wealth and how the fundamental role of human resources (includ- it could be measured. The emphasis on stocks of ing both human capital and the more difficult to wealth supports a new paradigm for sustainable define but important concept of social capital) in development, as a process of managing a portfo- determining a nation's wealth and, in turn, the lio of assets to preserve and enhance the oppor- opportunities for welfare gains for a nation's tunities people face. population. In response to growing worldwide interest, the The Bank's work on indicators has attracted a World Bank established an Indicators and Envi- high level of interest from many others around ronmental Valuation Unit in the Environment the world-both researchers and policymakers. Department to serve as a focal point for work on We would like to recognize the numerous inputs indicators and related issues. At present work in and comments received and specifically acknowl- this area is proceeding on several fronts: devel- edge the generous financial support from the oping environmental indicators that can be moni- Governments of Norway and Sweden for this ex- tored for Bank investment projects, expanding the panded work on indicators. set of environmental indicators in the newly re- We present our results to provoke discussion vised World Development Indicators produced by and further research. We are convinced that these the International Economics Department of the new approaches to indicators serve the paradigm Bank, and continuing work on resource account- of sustainable development founded on healthy ing and indicators of sustainability. ecosystems, vigorous economies, and equitable This report was produced by the Indicators social systems. and Environmental Valuation Unit. It is appro- priately subtitled Indicators of Environmentally Ismail Serageldin Sustainable Development because it highlights Vice President "portfolio" indicators for tracking a country's Environmentally Sustainable Development v : Acknowledgments his volume is the product of a team effort Dennis Anderson of the Industry and Energy led by the Indicators and Environmental Department; Zmarak Shalizi, Ken Chomitz, Valuation Unit of the Environment Depart- Maureen Cropper, Philip Keefer, Mead Over, ment of the World Bank. The members of the unit Annika Persson, and I. J. Singh of the Policy Re- are John Dixon, chief; Jan Bakkes, Kirk Hamilton, search Department; Bjorn Larsen of the Eastern Arundhati Kunte, Ernst Lutz, Stefano Pagiola, and Europe and Central Asia and Middle East and Jian Xie. Research assistance was provided by North Africa Technical Department; Tjaart W. NMichael Clemens and Andrew Sunil Rajkumar. Schillhorn van Veen and Alexander Borde of the The primary authors are: Agriculture Department; Victor Loksha of the * Chapter 1, "Introduction," Dixon and Asia Technical Department; Jan Bojo of theAfrica Hamilton Technical Department; Dan Biller of the Latin * Chapter 2, "Are We Saving Enough for the America Technical Department; Dely Gapasin and Future?," Hamilton and Clemens Lee Travers of the East Asia Department; Colin * Chapter 3, "Measuring the Wealth of Na- Bruce and Bhaskar Naidu of the South Asia De- tions," Dixon, Hamilton, and Kunte partment; and Florian Grohs of the Europe and * Chapter 4, "Subsidy Policies and the Envi- Central Asia Department. ronment," Pagiola, Rajkumar, and Xie External colleagues include John Kerr and * Chapter 5, "Land Quality Indicators," Sara Mark Rosegrant of International Food Policy Re- Scherr (International Food Policy Research search Institute; Julian Dumanski of Agriculture Institute), Christian Pieri (Agriculture De- Canada;Anne Harrison and ChristianAverous of partment), and Lutz the Organisation for Economic Cooperation and * Chapter 6, "Social Capital: The Missing Development; Allen Hammond, Robert Repetto, Link?" Christiaan Grootaert (Social Policy and Eric Rodenburg of the World Resources In- Division, Environment Department) stitute; Alison Van Rooy of the North-South In- * Chapter 7, "Poverty and the Environment," stitute, Canada; Donald Bleiwas and Lorie Wagner Pagiola and Kunte of the U.S. Geological Survey; John Hartwick of * Chapter 8, "International Progress in Indi- Queen's University at Kingston; Jeff Vincent of the cator Development," Bakkes. Harvard Institute for International Development; Many individuals within and outside the Bank Joy Hecht of the World Conservation Union; and provided information, reviewed material, and Christine Real de Azua of Accounting for the En- served as colleagues in this process. Those inside vironment, in addition to many others who pro- the World Bank include David Cassells, Gloria vided important data and insights into the issues Davis, Gordon Hughes, Kseniya Lvovsky, and discussed in these chapters. We are grateful to all Deepa Narayan of the Environment Department; of these individuals for their support. vii viii Expanding the Measure of Wealth The capable assistance of Isabel Alegre, Sriyani The overall support and guidance of Ismail Cumine, Jim Cantrell, and Gaudencio Dizon with Serageldin, vice president for Environmentally desktop publishing; Jane Whitten, Alicia Hetzner, Sustainable Development, and Andrew Steer, di- and Vrginia fitchcock with editing; and Tme Nielsen rector of the Environment Department, are greatly with coordination is gratefully acknowledged. appreciated. CHAPTER 1 Introduction X atural resources count, but people count include the links between environmental quality even more. This is the main lesson from and economic growth and between the use of re- ,\N the new estimates of the wealth of na- sources and the quality of the resource stock. This tions contained in this report. By expanding the is done because we realize that economic growth measure of wealth to include natural capital and that is paid for by rapid resource depletion or human resources in addition to produced assets degradation and that results in major health and (the traditional measure of wealth) one can ex- productivity impacts on society is probably nei- amine both the total of national wealth using this ther sustainable nor desirable. broader definition, and the differences found be- The seven chapters that follow are presented tween countries. This analysis yields important in two parts, divided by an explicit focus on links insights in the search for development that is eco- between the macroeconomy and the environment nomically and environmentally sustainable. (Part One), and evolving indicator themes (Part It should not be surprising that the role of Two). people and society is the most important in de- termining overall national wealth. Investments Part One: Indicators Linking the in education and health care, the more traditional Macroeconomy and the Environment items associated with the term human capital, are The links between various economic forces and complemented by the more difficult to define con- the environment are the theme of the next three cept of social capital, which relates to how indi- viduals and societies organize and interact, for good or bad. The importance of the human fac- national wealth or rates of subsidies and net sav- tor, manifest as raw labor, as an educated and ings, these chapters highlight new analytical ap- proaches to a better understanding of the healthy population, or as functioning, collective magnitu an dectionoerseaforce an the insttutonshels eplan unxpeteddifereces magnitude and direction of these forces and their institutions helps explain unexpected differences liktoevrnnalqltyndainlwl- between countries. For example, two countries bing. with similar endowments of natural capital (and g sometimes similar levels of investment in pro- Portfolio Indicators duced assets) might have very different growth paths and levels of well-being and might have Chapters 2 and 3 are representative of "portfolio" quite different levels of overall environmental indicators; chapter 2 is concerned with the dynam- quality. ics of creating and maintaining wealth. The indi- This report explores a number of indicators cator chosen to explore this is genuine saving-the of environmentally sustainable development that true rate of saving of a nation after accounting 1 2 Expanding the Measure of Wealth for the depreciation of produced assets, the deple- ciency and reduce deleterious effects on the envi- tion of natural resources, investments in human ronment. In the countries studied, fossil fuel sub- capital, and the value of global damages from sidies have dropped from more than $100 billion carbon emissions. Negative rates of genuine sav- in 1990/91 to nearly $60 billion in 1995/96, with ing must lead, eventually, to declining well-be- the countries of Eastern Europe being the primary ing. The analysis shows that many of the most contributors to the decline. (It should be noted resource-dependent countries do have low or that accurate data from this region are notoriously negative rates of genuine saving and that the difficult to obtain and that these numbers should value of pollution damages may be large in rap- be considered as very preliminary results.) While idly urbanizing and industrializing countries. Be- it is still too early to judge the scale of the positive cause the rate of genuine saving is determined by effects of these reforms on the environment, there a wide range of macroeconomic, resource, and is every reason to be confident that they will be environmental policies, this indicator has the po- substantial. tential to link the concerns of traditional National Environmental Action Plans to the interests of the Part Two: Evolving Indicator Themes key economic ministries. Whereas genuine savings is a flow measure The past few years have witnessed rapidly in- having direct policy links, stock estimates also creasing interest in, and applied work on, the use yield valuable insights. Chapter 3 presents esti- of indicators to monitor change. The chapters in- mates of the wealth of nations in which produced cluded here highlight substantive progress in assets, natural capital, and human resources are thinking on such diverse topics as land quality, all factored into the analysis. The analysis ex- social capital, and poverty and the environment. plores the composition of wealth at a point in time The last chapter presents selected information on (largely for the year 1994). The overall wealth international progress in indicator development, nlumber therefore gives an indication of the which illustrates both what has been accom- amount of these three forms of capital that each plished and the remaining challenges. nation commands. The numbers and groupings Land Quality Indicators of countries are not terribly surprising-the Organisation for Economic Co-operation and De- Land resources form the most basic natural asset velopment (OECD) countries rank high and of most countries. As seen in chapter 3, agricul- many of the poorer countries of Africa and Cen- tural lands and pasture account for up to 80 per- tralAsia (when measured by traditional GNP per cent of many countries' natural capital. Obviously, capita) rank low. What is of note is the composi- it is important to maintain and even enhance the tion of wealth within each country and how it productivity of this resource. Chapter 5 consid- changes as countries develop. These changes are ers the issue of land quality indicators (LQIs)- clearly seen in the analysis of countries by income what they are, how they can be estimated, and groups-both the size of the national wealth "pie" what has been learned from preliminary applica- grows and the distribution among components tions of the LQI approach to selected countries. changes as countries increase in aggregate wealth. The initial country and subregional results pre- sented here indicate that there is considerable potential for use of LQIs to inform decisionmak- Subsidy reform and the potential for this reform ing and that local-level indicators tend to be the most useful for policy formulation. Data con- to benefit the environment is the subject of chap- straints, however, are a very real problem for LQIs. ter 4. Subsidies on fossil fuels and agricultural inputs are a drain on the Treasury, distort invest- Social Indicators ment decisions, and lead to excessive releases of pollutants to the environment. Subsidy reduction We started this introduction by stating that natu- is therefore the classic "win-win" policy reform, ral resources matter, but that people are even more with the potential to both increase economic effi- important. Chapters 6 and 7 consider two distinct Introduiction 3 aspects of the role of people in determining the Other Work on Indicators nature and rate of economic growth. Chapter 6 explores the difficult to quantify concept of social The work at the World Bank is just one example capital, the extra element that defines how indi- of an extensive range of indicator development viduals and societies interact, organize them- efforts around the world. The final chapter in the selves, and share responsibilities and rewards. It report explores a number of the different national is now recognized that social capital is a critical and supranational indicator efforts. The systems explanatory variable in explaining the success of vary depending on the extent of data aggrega- htion involved (the chapter identifies three levels certain countries and the lack of progress of oth- of aggregations: large individual sets, thematic in- ers. Although there are a number of definitions dicators, and systemic indicators) and the in- of social capital available, and a variety of indica- tended end-use of each approach. tors are proposed that can be used to identify and A number of country programs are explored, track changes in social capital over time, we un- and a range of cross-national indicator efforts are fortunately know less about how it is created and also discussed. The portfolio indicators reported how to invest in its development. in chapter 2 (genuine saving) and chapter 3 (na- Chapter 7 explores the links between poverty tional wealth) and the land quality indicators in and the environment. Although it is still not com- chapter 5 are all examples of representative efforts. pletely clear whether poverty causes environ- As will be clear to the reader of this report, mental degradation or whether environmental the indicator arena is full of interesting and chal- degradation causes poverty, there is enough ex- lenging issues. It is also obvious that much of the perieneworldwide to conclude that the two work is in its infancy. Many of the results pre- perience worldwide to conclude that the two sented here are initial efforts at estimating indi- characteristics are commonly associated. As such, cators and are offered in the spirit of transparent the use of indicators of resource quality or deg- exchange of research results and thinking. We radation and indicators of the incidence of pov- hope to provoke dialogue and to advance both erty offer important tools for improved problem the methodologies used and the policy applica- identification and informed decisionmaking. tions of indicators for sustainable development. I PART ONE Indicators Linking the Macroeconomy and the Environment CHAPTER 2 Are We Saving Enough for the Future? D espite the pressing short-term priorities lem squarely in the realm of integrated economic most countries have for controlling infla and environmental accounting-loosely dubbed tion, reducing unemployment and foster- green national accounting (Hamilton and Lutz ing economic growth, the last decade has seen a 1996)-and suggests that expanding our tradi- remarkable upsurge of concern about the sustain- tional national accounting measures of savings ability of economic development over the longer and wealth could be an important step in guid- term. Much of this thinking is rooted in the re- ing policies for sustainable development. Adjust- port of the Brundtland Commission in 1987, ing savings measures to reflect environmental which argued that current development choices, depletion also fits well with many of the tradi- by exploiting and degrading the environment, tional concerns of development economics, which may diminish the well-being of future genera- emphasizes the savings-investment gap and the tions. While it has long been appreciated that critical role that financing investment plays in the economic processes have environmental conse- development process. quences, a crucial message in the Brundtland Developing "greener" national accounts holds Commission report is its emphasis on the extent the additional promise of treating environmen- of overlap between the economic and environ- tal problems within a framework that the key mental spheres, with linkages in both directions. economic ministries in any government will un- The United Nations Conference on Environ- derstand. For too long now ministries of finance ment and Development (the Rio Conference) in and planning have paid scant attention to the ex- 1992 helped to cement this understanding and ploitation of the natural resource base or the dam- prompted most countries to commit to achiev- aging effects of environmental pollution, while ing sustainable development. This led to the not- countries have been developing National Envi- unexpected problems of coming up with an ronmental Action Plans that read as if they were operational definition of sustainable develop- written by the environment ministryfor the envi- ment and designing policies for sustainable de- ronment ministry, with no links to the economics . . ~~~~~~ministries. velopment. Simply defining broad goals is not sufficient, however-we need indicators to mea- Genuine Saving sure progress toward sustainable development. However it is defined in detail, achieving sus- What follows is a new set of estimates of what tainable development is at heart a process of cre- the 1995 report Monitoring Environmental Progress ating and maintaining wealth. Wealth in this termed genuine saving: the true rate of saving in a context is conceived broadly, to include produced nation after due account is taken of the depletion assets, natural resources, healthy ecosystems, and of natural resources and the damages caused by human resources. This places the indicator prob- pollution; this report followed the pioneering 7 8 Expanding the Measure of Wealth work by Pearce and Atkinson (1993). The new Figure 2.1. Genuine saving for Tunisia, 1970-94 estimates of genuine saving feature broader cov- PLrcentage of GNP Goss do res erage of natural resources, improved data and 35 investnent methods of calculation, and significant enhance- 30 ments in the treatment of human resources. 25 The policy implications of measuring genu- 20 / \ ,\ ine saving are quite direct: persistently negative 15 / Net saving rates of genuine saving must lead, eventually, to declining well-being. For policymakers the link- Iv age of sustainable development to genuine rates 5 Gne n of saving means that there are many possible in- 0 terventions to increase sustainability, from the S_ macroeconomic to the purely environmental- 1970 1975 1980 1985 1990 1994 these will be discussed in the concluding section. Note:Gross domestic investment minus net foreign borrowing equals The traditional measure of a nation's rate of gross saving; gross saving minus depreciation equals net saving; net accumulation of wealth, as reported in the World saving minus depletion equals genuine saving. Bank's World Development Indicators for instance, Source: World Bank calculations. is gross saving. This is calculated as a residual: GNP minus public and private consumption. yielding te cuv forntsving.nFially, the bot- Gross saving represents the total amount of pro- tom line (literally and figuratively) is genuine say- duced output that is set aside for the future in ing, which is obtained by subtracting the value the form either of foreign lending or of invest- of resource depletion and pollution damages ments in productive assets. Gross savings rates from net saving. cansaylitle bou th sutaiabiityof eveop- Resource depletion is measured as thre total can say little about the sustainability of develop- rents on resource extraction and harvest. For each ment, however, because productive assets depre- ciate through normal wear and tear: if this resource-bauxite, copper, gold, iron ore, lead, deprciate ronis greater than gross saving, then ag- nickel, silver, tin, coal, crude oil, natural gas, and depreciation is greater than gross saving, then ag- phosphate rock-rents are estimated as the dif- gregate wealth, as measured by produced assets, ference between the value of production at world is in decline. Net saving, total gross saving less prices and total costs of production, including de- the value of depreciation of produced assets, is preciation of fixed assets and return on capital. one step closer to a sustainability indicator, but . . t focuse narrwly o prodced asets.Strictly speaking, this measures economic prof- focuses narrowly on producedlassets. its on extraction rather than scarcity rents, and Measures of genuine saving address a much for technical reasons this gives an upward bias broader conception of sustainability, by valuing to the value of depletion (and a downward bias changes in the natural resource base and envi- to genuine saving) in what follows. No explicit ronmental quality in addition to produced assets. adjustment is made for resource discoveries, since The simplest way to explain the accounting that exploration expenditures are treated as invest- underpins genuine saving is with a picture. Fig- ment in standard national accounting (see ure 2.1 presents the components of genuine sav- Hamilton 1994). ing as shares of GNP for Tunisia. Forest resources enter the depletion calcuia- The starting point in the calculation of genu- tion as the difference between the rental value of ine saving is just standard national accounting. roundwood harvest and the corresponding value The top curve in figure 2.1 is gross domestic in- of natural growth both in forests and plantations. vestment, the total of investments in structures, Only where harvest exceeds growth is a deple- machinery and equipment, and inventory accu- tion charge made for any given country. This mulation. Net foreign borrowing, including net valuation captures the commercial value of for- official transfers, is then subtracted from this top ests, therefore, but ignores the other services curve to give gross saving, the difference between provided by trees, including carbon storage, production and consumption over the years. Next watershed protection, and the supply of the depreciation of produced assets is deducted, nontimber (and nonfuelwood) products. Are We Saving Enoughfor the Fututlre? 9 Pollution damages can enter green national Critical natural resources (for example, accounts in several ways. While damage to pro- rainforests and the biodiversity they harbor) and duced assets (acid rain damaging building mate- critical pollutant levels (destruction of the ozone rials, for example) is in principle included in layer by chloroflorocarbon emissions, for in- depreciation figures, in practice most statistical stance) can be accommodated in the measure of systems are not detailed enough to pick this up. genuine saving. In the case of critical pollutants, The effects of pollution on output (damage to correct measures of marginal damages will pro- crops, lost production owing to morbidity) are vide the necessary deduction from the savings already reflected in the standard national ac- rate. To the extent that rainforests are overex- counts, but not explicitly. The key pollution ad- ploited, the excess clearance should be valued justment is for welfare effects, valuing the based on global willingness to pay for preserva- willingness to pay to avoid excess mortality and tion, a willingness that should rise sharply as criti- the pain and suffering from pollution-linked mor- cal levels of forest are approached. While the bidity. theory of genuine saving can deal quite ad- Underlying the treatment of pollution in green equately with critical stocks of resources or pol- national accounting is an extended Hicksian no- lutants, the practice is more difficult. Values tion of income. First, it is reasonable to assume associated with critical stocks are not measured that people derive welfare from, and have pref- in the empirical analysis below, partly for prag- erences for, both consumption and environmen- matic reasons (many pollutants are local, for in- tal quality. If societies seek to maximize this stance) and partly because there is inadequate extended conception of welfare over long time information to establish good measures of global horizons, then wealth can be conceived as the willingess to pay. present value of this stream of welfare now and There are some obvious assets left out of the in the future. "Green" net national product (NNP) foregoing description of genuine saving. Fish are is the maximum amount of produced output that excluded for a number of practical reasons, in- can be consumed at a point in time while leaving cluding the difficulty of measuring stocks, the this measure of wealth constant, and genuine sav- mobility of these stocks, and the fact that rents ing is the difference between green NNP and con- appear to have been largely dissipated in many sumption. For most significant pollutants fisheries owing to the lack of effective manage- (particulate matter, acid emissions, and lead emit- ment regimes. Soil erosion has figured promi- partculte mtte, acd eissins,and eadnently in many of the empirical green national ted to air; fecal matter, heavy metals, and biologi- cal oxygen demand in water) the adjustment accounting studies carried out for individual required to derive green NNP is the deduction of countries. Attaching a value to soil erosion, how- p nemissions valued at their marginal so- ever, requires detailed local data that are not pialollutio as measured by willingness to pay. widely available. Moreover, as box 2.2 argues, sig- cIan costs,a mealowsurhed wllingnessoto pmays f nificant research is still required to determine the In what follows, the value of damagres from pollution emissions is calculated only for carboni economic costs, as opposed to physical losses, of polltio emssins s clcultedonl fo cabon soil erosion and degradation in developing coun- dioxide, using a figure for marginal global dam- tries. ages of US$20 per metric ton of carbon emitted Ast (Fankhauser 1995). This value serves simply as a As the next chapter on measuring the wealth placeholder for other pollutants, whose effects ablendevelopment prdigm yin expaino woud rquie dtaledcoutryleel nalsisin able development paradigm iS an expansion of would require detailed country-level analysis in economic assets to include both natural and hu- order to be properly reflected in the genuine sav- man resources. While the ramifications for sav- ings calculations. Box 2.1 presents the effects of ings calculations of natural resource depletion valuing air and water pollution in the genuine have just been described, the treatment of human savings calculation for India-the results suggest resources presents some interesting conceptual that pollution damages may be significant in a issues. First, it is clear that educational expendi- wide range of rapidly urbanizing and industrial- tures on capital goods (equipment and buildings izing countries. for schools and universities, for instance) are al- 10 Expanding the Measitre of Wealth F Box 2.1 Pollution and genuine saving in India The inclusion of pollution damages in genuine saving tent with the conclusion in Brandon and Hommann calculations requires detailed country-specific data and that damages for water pollution are considerably careful consideration of which effects of pollution are larger than those for air. Box figure 2.1 shows the de- properly reflected in a savings measure. As a general composition of genuine saving in India in 1991. rule the economic cost of pollution damages falls into In this figure, extended domestic investment is the three categories: effects on economic assets, effects on value of domestic investment in produced assets plus current output, and welfare effects associated with ex- current expenditures on education-the latter were l cess mortality and morbidity. roughly 3.5 percent of GNP in 1991. The gap from ex- The theory of green national accounting (Hamilton tended domestic investment to net saving consists of 1996) suggests that welfare effects should be deducted net foreign borrowing (1.0 percent of GNP) and de- from genuine saving if the appropriate marginal will- preciation of produced assets. l ingness-to-pay measures can be obtained. Interna- Depletion of natural resources (oil, gas, coal, and tional experience in measuring pollution damages minerals) and the value of carbon dioxide emissions indicates that welfare effects are larger in value than amount to roughly 5.6 percent of GNP. With respect to the damage to economic assets (from acid rain, for ex- pollution only further country-level studies, based on ample) and reductions in current output (crops sound epidemiology, will increase our confidence in damaged by acid rain, for instance).Brandon and the transferability of industrial-country willingness- Hommann (1995) value pollution damages for India to-pay measures to the developing country context. in a recent study, which can serve as a rough guide to These crude estimates do suggest, however, that re- the magnitude of damages in a developing country ducing pollution is not simply a luxury for rich coun- with significant pollution emissions. The study first tries. Starting from traditional gross domestic estimates the loss in disability-adjusted life years investment levels of 23 percent of GNP, the analysis of (DALYs) associated with water pollution (largely from genuine saving suggests that India's true rate of sav- sewage) to be on the order 14.3 million DALYs in 1991. ing was on the order of 8 percent in 1991. Using India's annual per capita GNP of US$330 as the lower bound of willingness to pay to avoid the loss of a DALY yields a figure of roughly US$4.7 billion as the Bx Iur 2.1 Genuine ing In Indb, 191 value of the welfare loss. Brandon and Honmmann point Percenge dofGNP out that this value could be nearly twice as large if other 30 possible valuations were applied. The air pollutants with the largest effects in India are particulate matter (especially PM10, particles less 25 _ than 10 microns in size that lodge in the lungs) and Net en borrowing lead from gasoline. By transferring U.S. estimates of 20 and depredation the willingness to pay to reduce the risk of death to the Indian context, Brandon and Hommann suggest _ ---___________ that the welfare loss associated with air pollution could I 5 _Depson d have been as large as US$2.1 billion in 1991.An analy- natural rS sis by Cropper and Simon (1996) of data for New Delhi ------- suggests, however, that owing to different distributions l0 Pfk__ ion dmius of underlying health status between the populations of the U.S. and India, the figure for excess deaths could 5 be one-third of that calculated by Brandon and Hommann. As an illustrative valuation of pollution effects, Extended Net Genuine Genuine therefore, the figure for combined air and water pollu- dames*c saving sav sawvig tion damages of US$6.8 billion is used. This amount investment ecluing $1udbn represents roughly 2.5 percent of GNP and is consis damages damages ready treated as investment in the standard na- Properly valuing human capital is an exercise tional accounts, and so are induded in both gross fraught with difficulty. Afirst approxination presented and genuine savings rates. However current edu- inthegenuinesavingsalculationsbelow,issimplyto cation expenditures, on teachers' salaries and text- treat current education expenditures as investment books for example, are treated as consumption in rather than consumption. This leads to a notion of ex- the national accounts, which is clearly inconsis- tended domestic investment that will serve as the start- tent with the expanded conception of wealth. ing point in the calculation of genuine saving. Are We Saving Enough for the Future? 11 Box 2.2 Unanswered questions: saving and soil degradation Good soil is an important factor of production, and In work on Indonesia Repetto and others (1989) es- maintaining the asset is important for preserving the timated that capitalized losses in future productivity productive potential for the future. The quality or are approximately 40 percent of the annual value of quantity of this asset can be diminished by rain, wind, upland farm production. For Mali, Bishop and Allen cultivation practices, or simply the removal of nutri- (1989), using conservative assumptions of a ten-year ents through harvests. Farmers throughout the world time horizon and a 10 percent discount rate, estimated are counteracting the trend toward degradation gen- that the present value of current and future net farm erally by low-cost means such as mulching, contour income forgone nationwide due to one year of aver- farming, grass strips, and vegetative barriers. More age soil loss amounts to between 4 and 16 percent of costly structures like terracing are seldom adopted. agricultural GDP. The World Resources Institute and Soil degradation for the current period is reflected theTropical Science Center (Sol6rzano and others, 1991) in the national accounts through diminished yields estimated that soil depreciation amounts to almost 10 or through higher costs of production where farmers percent of Costa Rica's annual agricultural production. counteract the degradation, for example with higher These and other efforts to date have been based on fertilizer applications. What is not recorded is the re- weak and incomplete data--ctainly more work on duction in the stock of this natural capital. Inasmuch dat and meodiseqe to-yeld mer e s as such a reduction has a direct impact on future agri- data and dnethods is required to yield lirmer estimates cultural production, it represents a depreciation that In addition to on-site effects of soil erosion and deg- . ., . . ...... ' . .. . . ... radation one must consider off-site effects. Some of should, in principle, be reflected by genuine saving. these show up as costs in the national accounts in the But a smaller stock of soil simply in terms of quantity t periow Example would be countsin chenI has much different implications than a smaller stock current period. Examples would be dredging of chan- t of metals, minerals, forests, or fish stocks. What is cru- nels or decreased fish catch frm water with high tur-(i cial from an economic and accounting perspective is bidity. In other cases costs are more hidden (like l not the stock of remaining soil as such but the longer- term productivity effects resulting from the dimin- irrigation) or not accounted for (like damage of sedi- ished stock. In that regard it is not the amount of soil ments to coral reefs). eroded that is the key but rather the amount (depth) In assessments of the depreciation of a capital stock i and quality of the soil that remains, including the soil -in this case the stock of soil-that are based on dis- ! that is deposited in other fields used for agriculture. counting the future income losses, the discount rate l In addition, one must consider other factors that im- chosen strongly influences the results. For nations as a I pact productivity (like compaction, pH balance, or- whole and for national income and accounting pur- ganic matter content) that generally receive less poses one may assume rates of 5 to 10 percent. For in- I attention. Actual field research that shows the con- dividual farmers these rates tend to be higher, nection between soil loss, degradation and deposition, particularly for smaller farms. and the agricultural productivity in developing coun- Research on soil erosion and degradation is diffi- tries has been rare, and even where such information cult as it is influenced by stochastic variables (such as exists it cannot easily be extrapolated since soil deg- rainfall intensity). Much more work is needed to ex- radation can differ from field to field based on types tend the focus from plot-level experiments to the wa- of soil, slope, vegetation cover, and cultivation prac- tershed and from purely technical aspects to one that tices. Not surprisingly, therefore, estimates of the costs simultaneously considers economic and social aspects. | of soil erosion and degradation that have been made Research is being carried out at present on land qual- | at national levels have varied widely. ity indicators (see chapter 5). j Valuing human capital formation in this way The two broad categories of adjustments to implies that the accounting framework in this standard savings figures will move these mea- chapter differs from that presented in chapter 3, sures in opposite directions, with resource deple- which will emphasise the more inclusive notion tion and environmental degradation decreasing of human resources. Here the emphasis is nar- saving while current educational expenditures in- rowly on human capital, and the measure is a crease it. For expositional purposes, therefore, the gross measure of capital formation. To the extent analysis of regional trends in genuine saving in that education creates disembodied capital in the the next subsection will consider only depletion form of knowledge that is disseminated and and degradation effects, while the discussion of stored, this treatment is appropriate (the altema- human capital in the following subsection makes tive would be to assume that human capital de- the full adjustment for natural resources, carbon preciates when individuals die). dioxide, and investments in human resources. 12 Expanding the Measutre of Wealth Regional Trends in Genuine Saving Figure 2.2 Genuine savings rates by region A calculation of average genuine saving rates- Percentage ofGNP that is, genuine saving as a percentage of GNP- 20 reveals striking differences among various 15 EastAsidFadfic regions of the world. In many developing areas I0 decisive moments in economic performance are L tin Amearibbean reflected in large movements in the genuine sav- . - ings rate, shown in the figures that follow. 0 Sub-Saharan Africa Consumes 'R While East Asia Saves -10 Sub-SaharanAfria -15 The comparison of genuine savings rates reveals 1970 1973 1976 1979 1982 1985 1988 1991 1993 a remarkable trend for the countries of Sub- Sa- Source: World Bank calculations. haran Africa (figure 2.2). In that region average genuine savings rates rarely exceeded 5 percent genuneavinsrtesarey exceededIn stark contrast to the above stands the East of GNP during the 1970s, followed by a sharp Asia and Pacific region, with recent genuine say- negative turn at the end of that decade from which ings rates topping 15 percent of GNP and no signs they have never recovered. Despite slight recov- of reversing a long-term upward trend. High ery in the early 1990s, regional genuine dissaving genuine savings rates since the early 1980s in has recently been near 7 percent. C6te d'Ivoire China, Hong Kong, the Republic of Korea, and Nigeria, which together represent 16 percent Singapore, Thailand, and Taiwan (China) have of regional economic activity, exert great influ- mirrored the recent outstanding economic per- ence on these statistics with their rates of genu- formance of these economies. In the middle range ine dissaving approaching 20-30 percent of GNP of genuine savings rates for the region lie The precipitous drop around 1980 correlates well Indonesia's and Malaysia's-lower than those of with the onset and aftereffects of the second oil the East Asian tigers but still on a par with some price shock, with a negative flip in Nigeria's cur- of the higher rates in Latin America. Regional rent account balance, and with regional GDP dissavers like Lao People's Democratic Republic, growth rates averaging less than half their val- Papua New Guinea, and Vietnam show some ues before the plunge in genuine saving. Equally signs of approaching positive genuine saving important, negative genuine savings rates have rates. been accompanied by persistently low regional indicators of human welfare, including education, Are Fossil Fuel Exporters on a Sustainable Path? nutrition, and medical care (World Bank 1996b). The oil crisis also coincided with a period of Consistently negative genuine saving in the decline in genuine saving throughout Latin Middle East and North Africa region stands in America and the Caribbean, where figures had sharp relief against the policies in other areas of previously remained near 8 to 9 percent of GNP. the world (figure 2.3). Recent genuine savings In 1982, the year of Mexico's debt crisis, regional rates of 25 to 40 percent in Bahrain, Oman, Saudi genuine saving plummeted to 5 percent. As the Arabia, and Yemen have not been offset by the area has emerged from debt, returned to demo- modestly positive savers of the early 1990s such cratic rule, and spurred the vigorous growth of as Algeria, Egypt, Israel, Morocco, and Tunisia. the jaguars, genuine savings rates have shown a What happened to the enormous rents made from consistently positive trend. They remain, how- fuel exports as crude oil prices rose by 150 per- ever, well below 5 percent of GNP, as strong sav- cent between 1978 and 1979? While significant in- ers like Brazil and Chile are offset by the genuine vestments did take place, the net effect on saving dissaving of Venezuela and Ecuador and the was still negative, a problem that was exacerbated near-zero genuine saving of Mexico. by the Iran-Iraq war. Are We Saving Enotugh for the Future? 13 Regional total consumption as a share of GNP the volatility and large rates of genuine dissaving rose from around 50 percent in the 1970s to more seen in other regions. than 70 percent by the end of the 1980s, and im- The above picture of genuine savings rates is ports of food and manufactured goods flowed not complete, however. For just as standard mea- into the region as Saudi Arabia and Iran turned sures of saving ignore consumption bf natural re- their current account surpluses of the 1970s into source assets such as forests, it can be argued that deficits in the 1980s (World Bank 1996a). Despite they also ignore investment in one of a nation's overall regional gains in the early 1980s, recent most valuable assets: its people. rates of genuine dissaving hover between 10 and 15 percent. Investing in Human Capital South Asia and High-Income Countries The process of calculating genuine saving is, in essence, one of broadening the traditional defini- In South Asia near-zero recent genuine saving in tion of what constitutes an asset. Up to this point Bangladesh and Nepal is offset by consistently in the regional examples given above, standard positive rates of genuine savings in India, which measures of net saving (gross saving minus de- float at just under 10 percent and buoy up the en- preciation of produced capital) were adjusted tire region. India's GNP per capita growth rate of downward by an estimate of depreciation in the around 3 percent in recent years is above the av- value of natural capital due to natural resource erage for its income group. depletion and an allowance for pollution damages. Finally, rates of genuine saving in the high- There exist additional stocks, however, whose con- income OECD countries, pushed upward by high tribution to production is essential yet which are investment, lack of dependence on natural re- not considered assets even within the extended source depletion, and strong exports. of high definition of saving presented so far. Perhaps the value-added goods and services, are near 10 per- most important of these is the knowledge, experi- cent for much of the period depicted. Countries ence, and skills embodied in a nation's populace, with greater intensities of natural resource extrac- its human capital tion, such as Australia, Canada, and the United The world's nations augment the stock of hu- States, had the lowest genuine savings rates, re- man capital in large part through their education cently between 1 and 3 percent per year, while systems, into which they collectively pour trillions continental Western Europe and Japan were the of dollars each year. Standard national accounts biggest savers, often exceeding 10 to 15 percent. label as an investment less than 10 percent of this Recent recessions in 1982-83 and 1990 coincided amount, or only that portion which is spent on with downward turns in genuine savings rates, fixed capital such as school buildings. Current (as but the figures consistently exhibit an absence of opposed to capital) expenditures on education include teachers' salaries and the purchase of Figure 2.3 Genuine savings rates by region books and are treated strictly as consumption. Percentage of G NP Within the genuine savings framework, however, I5 this is clearly incorrect. If a country's human capi- I10 Higcome tal is to be regarded as a valuable asset, expendi- 5----__- -^ tures on its formation must be seen as an 0 SoLAhAsia investment (Hamilton 1994). -5 Controversy surrounds the correct method of -10 Middle valuing such an investment in the human capital -15 NorhAi stock, since one dollar's current expenditure on -20 education does not necessarily yield exactly one -25 dollar's worth of human capital (see, for example, -30 Jorgensen and Fraumeni 1992). Traditional sav- -3570 1973 1976 1979 1982 1985 1988 9911993 ings rates should theoretically be adjusted by the change in value of human capital to reflect this investment, but there is as yet no consensus on 14 Expanding the Measutre of Wealth how to carry out this valuation. What can be said Figure 2.4 Human capital and genuine saving with certainty, however, is that current education in Chile expenditures are not consumption, within a Percentage of GNP framework where the notion of wealth is ex- 40 Exended domestc panded to include human resources. As a first ap- investment Gross domestc investment proximation, therefore, rates of genuine saving 30 L\ \ A \ are adjusted upward by the levels of current 20 XiL.,. spending on education. , , 'X ' / /Net saving Chile Invests in Its Young People 0 o v \'-/nun sain The effects of including human capital investment 10 in the genuine savings calculation can be signifi- cant. In Chile, for example, current education ex- -20 penditures represented about 3.1 percent of GNP 1970 1975 1980 1985 1990 1994 in the early 1990s (figure 2.4). Such investment Note: Extended domestic investment is gross domestic investment helped keep genuine savings rates from slipping plus current spending on education; gross domestic investment mi- into the negative realm in the late 1980s, and rep- nus depreciation equals net saving; net saving minus depletion equals into ~~~~~~~~~~~~~~~~genuine saving. resent more than one-third of the high rates seen Source: World Bank calculations. in recent years. In 1993 and 1994 nearly half of the rents from natural resource depletion were, no- genuine savings rates markedly upward (table tionally at least, being reinvested in human capi- 2.1). In Sub-SaharanAfrica accounting for educa- tal. tion investment means that recent genuine sav- ings rates are negative but closer to zero. In the The Rich Get Richer? Middle East and NorthAfrica region genuine sav- ings rates are consistently negative even after the adjustment. Finally, high rates of education invest- ment in genuine saving accentuates the differences ment in high-income OECD countries and the East between countries with strong and those within Asia and Pacific region sharpen the contrast be- weak savings efforts. Without the effects of edu- tween the genuine savings effort'in these areas cation spending, there is only a modest difference between recent average genuine savings rates among low- and middle-income countries, while Conclusions the high-income countries exceed this level by around 5 percent (figure 2.5). Weighing the effects The World Bank study Oil Windfalls: Blessing or of storing up human capital, however, changes the Curse? (Gelb 1988) poses the critical question that savings picture. Large investments in education comes to mind when regional trends in genuine by the most economically successful countries lead saving are examined. It is evident from figures them to exceed the genuine savings rates of their 2.2 and 2.3 that measuring genuine saving tends counterparts in other income groups by around 8 to depress the savings rates of resource-rich coun- percent in the same period. Investments in young tries compared with resource-poor countries. people by middle-income countries were also no- However, it is obviously incorrect to conclude ticeably larger than in low-income nations. This from this that natural resource endowments are mirrors the preponderance of human resources as necessarily detrimental to economic performance a share of wealth in high-income countries shown (for evidence that growth rates have been weaker later in chapter 3. for resource- intensive economies, see Sachs and Warner 1995). Education and Regional Genuine Saving Rates The depressed rates of genuine saving for re- source-rich countries represent an opportunity Adjusting rates of genuine saving to embrace not seized. These countries have the potential to changes in human capital assets shifts regional transform a resource endowment into other eco- Are We Saving Enoughfor the Future? 15 Figure 2.5 Contribution of current education is simply being consumed, to the detriment of fu- expenditures to genuine saving ture well-being. (average 1987-90) A particularly useful way for policymakers to Percentage of GNP think of a genuine savings chart such as figure 20 2.1 is in terms of which policies have an impact on the levels of the individual curves. The level 15 Eof gross saving anchors the curves, in the sense Current spending on education that they will all shift up or down with the fun- 10 Genuine saving without eucab damental saving effort. The determinants of gross saving, basically fiscal and monetary policies but including more micro-level issues such as the vi- 5 ability of the financial sector, therefore play a cen- tral role in determining policy responses to r deficient genuine saving. Average low- Average middle- Average high- The level of gross domestic investment is in- income income income fluenced both by the willingness of the private Source.:World Bank calculations. sector to invest-which is again closely tied to nomic assets that will boost incomes and growth. macroeconomic policies-and by public invest- This potential has not been realized for a variety ment programs. The investment of natural re- of reasons. Closer examination of the data shows source rents, collected through government that it is often the gross savings effort that is in- royalties, is a central issue in this regard. In addi- sufficient in these countries, which points te . .tion to the quantity of public investment, quality sfen squarely at broader macroeconomic policies. is also a critical factor; this is one of the central ger squarely * . mi *oce. lessons of the Oil Windfalls study. Investments in But resource and environmental policies clearly have a role to play as well. human capital represent in many countries some Note that it is perfectly reasonable for coun- of the highest quality outlets for public invest- tries to choose to develop and deplete nonrenew- ments. Extending the measure of domestic invest- able resources as a source of development finance. ment to include current education expenditures, But this is only reasonable if the resource rents as shown in figure 2.4, highlights this issue for are in fact invested rather than being consumed. policymakers. Measuring negative rates of genuine saving is an The ministries of finance, development plan- indication that due prudence is not being fol- ning, and human resources therefore all have key lowed-that some amount of the national wealth roles to play in determining the rate of genuine Table 2.1 Genuine saving (percentage of GNP) Average, Average, Region and income category 1970-79 /980-89 /990 /99/ /992 /993 Region Sub-Saharan Africa 7.3 -3.2 -3.8 -1.2 -0.6 -1.1 Latin America and Caribbean 10.4 1.9 5.5 4.1 4.7 6.1 East Asia and Pacific 15.1 12.6 18.6 18.7 18.7 21.3 Middle East and North Africa -8.9 -7.7 -8.8 -10.8 -6.6 -1.8 South Asia 7.2 6.5 7.6 6.3 7.1 6.4 High-income OECD 15.7 12.4 15.7 14.5 14.0 13.9 Income category Low 9.8 3.3 5.7 7.5 9.0 10.5 Middle 7.2 2.9 10.0 9.7 7.8 8.1 High 15.2 12.3 15.9 14.6 14.1 14.1 Note: Data include an adjustment for current spending on education. Source: World Bank calculations. 16 Expanding the Measure of Wealth saving. The gap between net saving and genuine base. In all cases, the statistics used were in nomi- saving is determined by resource and environ- nal prices. Gross saving was calculated as the mental ministries. It is to the policies of these min- difference between GDI and the amount of bor- istries that we now turn. rowing, that is, the sum of GDI and current ac- For resource ministries the challenge is to en- count balance after net official transfers. Net saving sure efficient resource extraction and harvest. was calculated as the difference between gross Security of tenure for producers is therefore one saving and depreciation of produced capital. De- of the principal concerns, especially for living re- preciation is not widely reported. It was therefore sources such as forests and fish. Rent capture is estimated as follows. Estimates of gross fixed capi- the other main issue. Without sufficient capture tal formation from a World Bank Policy Research of available rents through resource royalties there Department study (Nehru and Dhareshwar 1994) is a strong incentive for producers to overexploit were used together with gross domestic invest- natural resources, depressing genuine saving ment figures to estimate depreciation of produced from its efficient level. capital in each year according to a simple per- For environment ministries the issue is not to petual inventory model reduce pollution damages to zero, but rather to GFC GFC = GDI D reduce them to socially optimal levels. This im- t+l - t -Dt - plies that there will always be some gap between where GFC is gross fixed capital stock in year t, net saving and genuine saving resulting from GDI is gross domestic investment in year t, and D pollution emissions, but there will be an optimal is depreciation of human-made capital in year t, level for this gap. The efficiency of environmen- all in constant prices in local currency. tal policies has clear consequences for the other Genuine saving was calculated as the difference macro variables discussed previously, in particu- between net saving and the sum of rents from all lar the willingness of the private sector to invest. extraction and harvest of natural resources as well The bottom line in the analysis of genuine sav- as damages from carbon dioxide emissions. For ing is that policies leading to persistently nega- each nonrenewable resource, and in all countries tive savings rates must entail, eventually, declines in each of the years 1970-1994, rent was estimated in welfare. The intuition, or hope, that greening as the national accounts could influence policies for R = (PC) Q sustainable development has taken some time to be realized, but the analysis of saving and wealth where R is rent, P is world market price, C is the holds out this possibility. This analysis also em- average unit production cost (cost of raw mate- phasizes that the key economic ministries, the rial extraction and transportation to city or port), human resources ministry, and the resource and and Q is production volume. Strictly speaking, R environmental ministries all have important measures economic profit rather than scarcity rent. policy levers at their command if the goal is to The effect of calculating economic profits rather achieve sustainable development. Whether natu- than scarcity rents (price minus marginal cost) is ral resource endowments and the benefits of a to overstate depletion and understate genuine healthy environment can be translated into sus- savings, which has a clear impact on the figures tained well-being is a question of good or bad for the Middle East and North Africa region in policy choices. this chapter. This method of valuation also assigns a higher value of depletion than that implied by Technical Appendix the valuation of resource stocks described later in chapter 3-this is largely a consequence of not The most basic components of the genuine sav- having time series of resource stock data. The non- ings calculations, namely gross domestic invest- renewable resources thus treated were crude oil, ment (GDI) and net foreign borrowing (current natural gas, hard coal, lignite, bauxite, copper, account balance after net official transfers), are gold, iron ore, lead, nickel, phosphate rock, silver, reported directly to the World Bank and were tin, and zinc. Most figures for production volumes taken from the Bank Economic and Social Data- came from the Bank's internal Economic and So- Are We Saving Enoughfor the Futture? 17 cial Database. Owing to the physical characteris- age tropical hardwood price. The weights in the tics of natural gas and lignite, there is no single, average were the relative contributions of the uniform world price for these commodities; a three types of wood to national roundwood pro- shadow world price was estimated. For natural duction volume. The relative contribution of co- gas this was the average of several free-on-board niferous wood came from the FAO, and the export prices from Algeria, Canada, the Nether- contribution of fuelwood from World Resources lands and Norway. For lignite the shadow price Institute data. Fuelwood price was estimated was estimated from the average international ex- based on internal Bank sector studies. port price for steam coal by observing the rela- Damages from carbon dioxide emissions were tionship between internal prices for lignite and calculated by taking historical carbon emissions steam coal in countries where the market for both estimates from the United States' Oak Ridge Na- is relatively competitive, such as Australia and tional Laboratories Carbon Dioxide Information Canada, and assuming a similar relationship for Analysis Center and multiplying by a conserva- international prices. Production cost estimates tive estimate of $20 marginal global damages per were collected from a wide variety of sources, both ton of carbon emitted (Fankhauser 1995). Note private and government, as well as from sectoral that these are global damages, so that the net dam- experts within the Bank. ages caused worldwide by emissions in a given Because forests are living resources, a differ- country are assigned to the emitting country. ent method for calculating depletion was em- Finally, calculations in the second half of the ployed in the case of roundwood harvest. For chapter include the effects on genuine saving of each country and year a depletion charge was current (nonfixed capital) spending on education. made only when roundwood harvest exceeded For the estimates of genuine saving including natural growth in the country for that year. Many human capital formation, the calculation pro- countries produce roundwood volumes well be- ceeded precisely as above-with the exception low their annual increments, and for them genu- that the starting point was not gross domestic ine saving was unaffected. Estimates of annual investment (GDI) but rather extended domestic in- increment and production costs were obtained vestment, equal to the sum of GDI and current through consultation with a variety of Bank ex- education expenditures. Data on current educa- perts on forestry. Roundwood production vol- ums an woo prce wer obandhrm ton spending came from the United Nations. umes and wood prices were obtained from For each year regional and income-class av- United Nations Food and Agriculture Organiza- Fr each yer regina anincomes av- t tion (FAQ) data. Since roundwood production is quoierage figures for genuine savings rates are the not homogeneous but rather represents a combi- quotient of aggregate genuine savmig and aggre natio of cnifeous wod, felwod, an oth rgate GNP for the set of all those countries in the nation of coniferous wood, fuelwood, and other reinovnoecasfrwihdt nbt types of values, roundwood could not be given a single price. One way to estimate a price would genuine savings and GNP are available for that be to disaggregate roundwood production ac- year. cording to coniferous production volume, fuelwood production volume, and so on. Another Note method would be to create a composite price for 1. Sachs and Warner (1995) show that the resource- roundwood as an average of the prices of the dif- intensiveness of countries is negatively corre- ferent types of wood, weighted by their contri- lated with rates of growth since 1970. They offer butions to total production volume. The two a number of possible explanations, including a methods are mathematically equivalent and for model of "Dutch disease." the sake of convenience the second was chosen. The roundwood price used for each country-year References was a weighted average of coniferous softwood price, an estimated fuelwood price, and a third Bishop, J., and J. Allen. 1989. "The On-Site Costs of price that, depending on the region to which the Soil Erosion in Mali." Environment Department country in question belonged, varied from a WorkingPaperno.21.WorldBank,Environment nonconiferous softwood price to a regional aver- Department, Washington, D. C. 18 Expanding the Measure of Wealth Brandon, C., and K. Hommann. 1995. "The Cost of Nehru, V., and A. Dhareshwar. 1994. "New Esti- Inaction: Valuing the Economy-Wide Cost of En- mates of Total Factor Productivity Growth for vironmental Degradation in India." World Bank, Developing and Industrial Countries." Policy Re- Asia Environment Division. Washington, D. C. search Working Paper 1313. World Bank, Intema- Cropper, M., and N. Simon. 1996. "Valuing the tional Economics Department, Washington,D.C. Health Effects of Air Pollution." DEC Note 7. Pearce, D.W., and G. Atkinson. 1993. "Capital Theory World Bank, Policy Research Department, Wash- and the Measurement of Sustainable Develop- ment: An Indicator of Weak Sustainability." Eco- Fankhauser, S. 1995. Valuing Climate Change: The Eco- logical Economics (8): 103-08. Fankha fhuser,..195 Valonn : ClimateChangerThEca - Repetto, R., W. McGrath, M. Wells, C. Beer, and F. nomics of the Greenhouse. London: Earthscan. Rossini. 1989. Wasting Assets: Natural Resources Gelb, A. H. 1988. Oil Windfalls: Blessing or Curse? in the National Accounts. Washington, D. C.: New York: Oxford University Press. World Resource Institute. Hamilton, K. 1994. "GreenAdjustments to GDP." Re- Sachs,J. D., andA. M. Warner. 1995. "Natural Resource sources Policy (3): 155-68. Abundance and Economic Growth." Development . 1996. "Pollution and Pollution Abatement in Discussion Paper 517a. Harvard Institute for In- the National Accounts." Review of Income and ternational Development, Cambridge, Mass. Wealth 42 (1). Sol6rzano, R., R. de Camino, R. Woodward, J. Tosi, and E. Lutz. 1996. "Green National Accounts: V. Watson, A. Vasquez, C. Villalobos, J. Jimenez, Pocy Use and Empirical Experience.Environ- R. Repetto, and W. Cruz. 1991. Accounts Over- Policy Usesparend Empirical39 Experience," due: Natural Resources Depreciation in Costa Rica. ment Department Paper no. 39. World Bank, En- Wahnt,D.C:ordRsucsItite Washington, D. C.: World Resources Institute. vironment Department, Washington, D. C. World Bank. 1996a. "Bank Economic and Social Da- Jorgensen, D. W., and B. M. Fraumeni. 1992. "The tabase." World Bank, International Economics Output of the Education Sector." In Zvi Griliches, Department. Washington, D. C. ed., Output Measurement in the Service Sectors. . 1996b. World Development Report 1996: From Plan Chicago: University of Chicago Press. to Market. New York: Oxford University Press. CHAPTER 3 Measuring the Wealth of Nations T }t triting some 220 years ago in The Wealth way that individuals and societies are organized of Nations, Adam Smith recognized the are the most important determinant of the wealth TV sfundamental importance of labor and of nations. Building on work presented in the 1995 its skill level in explaining observed differences edition of Monitoring Environmental Progress between the wealth of different countries, even (World Bank 1995) and Sustainability and the those with similar endowments of natural re- Wealth ofNations (Serageldin 1996) new estimates sources. In the opening paragraph to what is fully have been made of the three major capital com- titled An Inquiry into the Nature and Causes of the ponents that determine a nation's wealth: pro- Wealth of Nations, Smith wrote: duced assets, natural capital, and human resources, the last including raw labor, human capital, and The annual labor of every nation is the the elusive, but important, element known as so- fund which originally supplies it with cial capital. This chapter presents the results of all the necessities and conveniences of new estimates of these different components and life which it annually consumes, and discusses the insights gained from this broader which consist always either in the im- analysis of the determinants of national wealth. mediate produce of labor, or in what is Box 3.1 gives an intuitive introduction to these purchased with that produce from expanded wealth measures. other nations. There are close links between wealth mea- sures and sustainable development. While the Smith recognized that a country's endowment Brundtland Commission defined sustainable de- of natural resources was important; he also fo- velopment in terms of meeting the needs of fu- cused on and emphasized the importance of the ture generations, there is growing recognition human factor. In fact, Adam Smith wrote that that these needs may not be comparable across "whatever the soil, climate, or extent of territory countries and across time. An emerging and pow- of any particular nation, the abundance or scanti- erful interpretation of sustainable development ness of its annual supply [output]" fundamentally concentrates on preserving and enhancing the op- depends on its human resources-"the skill, dex- portunities open to people in countries around the terity, and judgment of its labour." world (Serageldin and Steer 1994). From this Adam Smith's insight of 220 years ago con- viewpoint shifting attention from flow measures tinues to hold. Our analysis of the wealth of na- of economic activity, such as GNP, to the stocks of tions indicates that stocks of both produced assets environmental resources, produced assets, and (human-made capital) and natural resources are human resources is crucial. Stocks of wealth un- important components of national wealth. How- derpin the opportunities people face, and the pro- ever, we also find that human resources and the cess of sustainable development is fundamentally 19 20 Expanding the Measure of Wealth Box 3.1 Interpreting the wealth estimates A useful intuition into these wealth estimates comes assets and associated land amounts to nearly $8,000. from interpreting the wealth of the "average" Central This part of your portfolio can be conceived as shares American in 1994. As shown in box table 3.1 below, in a mutual fund, a fund which owns all of the pro- total wealth per capita in Central America was US$52 duced assets in the region. Part of the shares you own thousand. What does this really mean? The first step come from your own savings, while the rest are an in- in interpreting this figure is to abstract completely from heritance from your parents. the existing distribution of wealth-to pretend, in ef- Your natural inheritance as a Central American to- fect, that the average CentralAmerican actually owned tals to nearly $3,000, largely in the form of agricultural an average share of the region's assets. land, forests, and protected areas. Again, this can be As an "average" CentralAmerican, you are 23 years conceived as shares in a mutual fund, one which owns old, and your life expectancy is 68 years. Assuming all of the natural resources in the region, but there is a that you retire at age 65, therefore, a substantial part of critical difference compared with your holdings of pro- your wealth is the present value of the returns to your duced assets. First, where the resources are managed labor in the economy over the 42 years that you expect sustainably (good soil management, sustainable har to work; it is (pessimistically) assumed that you will sust o d inagm e har cotiu to ad thlaevleecya,adte(o vesting of forests), the income from this part of your ciau) dountdrte s asme to e 4earent. fyou portfolio is in effect an unlimited annuity, one which cial) discount rate iS assumed to be 4 percent. If you yields a steady income in perpetuity. However, where were to borrow against your future income this would yieldsla staicome n perpet However, w r be the figure that a prospective lender would need to renewable resources are not managed sustainably, or know. in the case of nonrenewable resources, this portion of You also own, as the representative CentralAmeri- your portfolio is like a sinking fund that is eventually can, all the produced and natural assets of the region. exhausted. Moreover, the amount of the sinking fund As one person in 32 million, your share of produced that is liquidated each year is not income-it must be reinvested in other assets in order to maintain your Box table 3.1 Decomposition of wealth per capita wealth. in Central America, 1994 ~~~The bottom line for Central Americans and virtu- in Central Amernca, 1994 ally all the citizens of the countries analyzed in this Wealth US. dollars Percent chapter is that the majority of your wealth is the re- tum to your effort and the social capital from which Human resources 41,000 79 you benefit. Natural capital 3,000 6 That having been granted, produced assets and ur- Produced assets 8,000 15 ban land are still significant shares of wealth-up to 31 percent for the average Japanese-and natural Total 52,000 100 wealth is important in many countries, up to 42 per- Source:, Authors' estimates. cent for the average citizen of Madagascar, for instance. the process of creating, maintaining, and man- role of natural and human resources in develop- aging wealth. This is one of the prime motiva- ment, much work remains to be done. This is an tions for the World Bank in pushing ahead with area in which analysts working in individual the estimates in this chapter. countries should be able to make significant For thoroughly pragmatic reasons the wealth progress. estimates in this chapter (and the savings esti- In addition, any attempt to estimate wealth mates in the previous one) are based on the in- across the great diversity of countries found in strumental or use values of natural resources, the world today must make many bold and sim- although place-holder values for protected areas lifyi au ti Re izing this, we will and nontimber forest benefits are included as py assptin . Rognizing thi we wil well. There is a considerable willingness world- wide to pay to preserve nature and the critical sumptions and the data used in the calculations. functions that ecosystems provide, but we have The regional results presented in this chapter are no robust means by which to estimate this designed to focus attention on the various com- amount for the whole range of countries pre- ponents of national wealth, and the investments sented here. So while this work takes several im- in and management of these components. This is portant steps in the direction of recognizing the essential if development is to be sustainable. Measuring the Wealth of Nations 21 Components of National Wealth From the viewpoint of measuring the sustain- ability of economic development, the green na- Although the classical economists recognized the tional accounting aggregate with the most policy importance of land, labor, and capital in explain- relevance is "genuine" saving. This represents the ing economic growth and national wealth, in the value of the net change in the whole range of as- post-World War II period national well-being has sets that are important for development: pro- been measured by gross domestic product duced assets, natural resources, environmental (GDP)-the total value of production in an quality, human resources, and foreign assets. For economy)-or gross national product (GNP)- a fuller presentation of the underlying concepts GDP plus net factor income from abroad. Coun- and empirical evidence on genuine saving, see tries were ranked by their level of GNP per capita, chapter 2 in this report. This chapter also makes and few questions were asked about the under- the link from the indicator-genuine saving-to lying resource base for GNP growth and whether the policy issues that are directly relevant in it was sustainable. achieving sustainable development. The World Bank classified countries accord- Both green national accounts and genuine sav- ing to GNP per capita, and used this criterion to ings measures provide important policy guidance divide the countries of the world into four main on the pace and direction of change of a nation's categories: low income (roughly under $750 per economy. They address how a country is man- capita), lower middle income (about $750 to aged and the sources of national growth. Build- $2,900 per capita), upper middle income (about ing on many of these same ideas, the wealth $2,900 to $9,000 per capita), and high income estimates in this chapter take a broader view of a countries (more than $9,000 per capita). In World country and ask the following questions: Development Report 1996 the Bank, for the first * What are the components and contributing fac- time, also ranked countries by Purchasing Power tors to national wealth? Parity (PPP) estimates of GNP in the core basic ' How can these be managed to promote sustain- indicator table. PPP is a rate of exchange for cur- able economic development? rencies that euf Human resources turn out, not unexpectedly, rencies that equates the purchasing pwower Of to be the dominant form of wealth in the major- given units of a particular currency to what one ty of tre We use the tajvi- dollr wold prchae i theUnitd Sttes.The ity of countries. We use the term resources advis- dollar would purchase in the United States. The el,t itnus h siae rsne eo use of PPP estimates had the effect of reducing fo humancait h is gene nsedered the spread etween the oorest and ichest cou- from human capital, which is generally considered the spread between the poorest and richest coun- tobthprdcofeuainHmnrsucs tries, although few countries actually shifted from inues the preturn bo e to euc and touraw on noegou oaohr includes the returns both to education and to raw one income group to another. More recently, several new approaches have labor. This distinction is important in most de- been developed to help address the inherent veloping countries. shortcomings of GDP and GNP measures. These The wealth of nations approach requires the include the development of "green" national ac- explicit inclusion of environmental, social, and counts that take into account the role of the stocks human factors in addition to the more commonly as well as the flows of renewable and nonrenew- measured economic variables. The approach fo- able resources, and the related concept of genu- cuses on the instrumental or use values of natu- ine savings. Green GNP is the informal name given ral resources, partly because the methods of to national income measures that are adjusted for valuation are well established for these uses. For the depletion of natural resources and degrada- the moment at least, this means that many of the tion of the environment. The types of adjustments critically important ecological and life-support that must be made to standard GNP include mea- functions provided by natural systems, as well suring the user costs of exploiting natural re- as existence values and the aesthetic pleasure we sources-the change in value of a copper mine derive from nature, are not measured as part of as a result of extracting ore for a year, for ex- the wealth of nations. Protected areas are, how- ample-and valuing the social costs of pollution ever, included in the wealth estimates, valued at emissions. the opportunity cost of preservation. The value 22 Expanding the Measuire of Wealth of protected areas is closely linked to accessibil- range, the key factor influencing this result is the ity, as is revealed by cases where we have good use of PPP-based GNP to value human resources. data on willingness to pay for these benefits (see The three countries ranking highest in wealth the discussion in box 3.2). In addition, nontimber per capita in table 3.1 are hardly surprising. Swit- forest benefits are also estimated, as described zerland is wealthy in terms of both human re- below. sources and produced assets, while the United This chapter summarizes the results of wealth States and Canada are relatively better endowed estimates for almost 100 countries. Although with natural resources. Norway leads the Scan- there is a natural temptation to compare the dinavians, owing to North Sea oil and gas, while rankings of individual countries-who is first, productive agricultural land raises the rank of which last? or how is a neighboring country New Zealand. The richer middle-income coun- ranked relative to one's own ranking?-The real tries begin to appear in the per capita wealth value of these estimates is in the explicit recogni- tion of the different components of national Box 3.2 Valuing protected areas wealth, and what lessons can be learned from Protected areas bring a wide range of benefits. How- both those countries that have been judged suc- ever, the specific benefits generated by a particular cessful and those that have been less successful area tend to vary substantially, both in nature and in promoting sustainable development. magnitude. Data and resource constraints prevented The next section presents initial findings of this a full examination of the value of protected areas in this version of the wealth of nations estimates. The exercise followed by the broad conclusions and approach adopted, which values protected areas at policy implications of measuring the wealth of their opportunity cost, is an estimate of their mini- nations. In order to interpret these results, how- mum value. ever, it is important that the reader understand In most high-income countries the greatest ben- ever, . . ...efits generated by protected areas are often based the assumptions that underlie the analysis. De- on their existence value and on the recreational op- tails of the methodology and assumptions can be portunities they offer. Surveys and observations of found in the technical appendix at the end of this actual behavior show that many people have a rela- chapter, a summary of which is presented in box 3.3. tively high willingness to pay for these benefits. Al- | though natural tourism is growing in many middle-income countries, both the willingness and Initial Findings and Emerging Pattern the ability to pay for such benefits are often lower in for the Wealth of Nations lower-income countries. By attracting tourists, how- ever, protected areas often generate benefits for The most aggregate results of the new wealth cal- lower-income countries. Indeed, protected areas culations are shown in table 3.1 (note the uneven form the basis of a thriving tourist industry in sev- wealth ranges reported in this table). The only eral countries, such as Kenya and Costa Rica. About used for selecting countries to be included 10,000 tourists visit Costa Rica annually specifically criteria used for selectmg countries to be mcluded to visit the Monteverde Cloud Forest Reserve, for in the analysis are data availability and reliabil- example; each is estimated to generate about $1,000 ity-as a result, countries of the former Soviet in local eamings. Union and Eastern Europe are not included. Protected areas often also provide a range of Among the most notable of the other countries other services to the national economy. In Haiti, for example, a recent economic analysis showed that ex- affected by problems of data quality are Nigeria penditures on the Pic Macaya National Park were and Algeria. One effect of the wealth calculation, amply justified by the reduction in downstream dam- evident in table 3.1, is to compress the ranges be- age from flooding and sedimentation alone. Far from tween countries when compared with standard being a luxury that a poor country could ill afford, exchange-rate-based GNP per capita. Whereas the this protected area played an important role by miti- gating damage to one of the main irrigated areas in ratio of the mean GNP per capita in the upper the country and to its supporting infrastructure. This quintile of countries to the mean in the lowest case also illustrates the difficulty of estimating the quintile is a little more than 100, this shrinks to a value of protected areas except on a case-by-case ba- ratio of roughly 17 when the figures being com- sis; other protected areas in the same region had a pared are wealth per capita. While valuing natu- able infratucture to protesethere was less vulner ral resources plays a role in compressing the Measuring the Wealth of Nations 23 Box 3.3 The wealth accountants toolkit Measuring the total wealth of a country necessarily * Protected areas are valued at their opportunity cost: involves some heroic assumptions. The first choice the per-hectare rate for pasture land. a wealth accountant confronts is the discount rate: a * Produced assets are calculated using a perpetual in- social rate of 4 percent per annum is used through- ventory model, with investment data and an as- out the measurements. Total wealth is the sum of sumed life table for assets being the major inputs. each of the following components. Urban land is valued as a fixed proportion of pro- * Minerals and fossil fuiels are valued by taking the duced assets. present value of a constant stream of resource-spe- * Human resources are measured residually. The cific rents (or, to be more precise, economic prof- wealth value of returns to both labor and capital its, the gross profit on extraction less depreciation is measured as the present value of the following: of produced assets and return on capital) over the noagural GNP plus alural wae mi- lieo. rve eevs nonagricultural GNP, pluts agricultural wages, mi- life of proved reserves. * Timber is valued as the present value of an infinite nus rents on minerals and fossil fuels, and minus stream of constant resource rents where the rate depreciation of produced assets. Agricultural of harvest is less than annual natural growth (the wages include proprietors' income and exclude re- mean annual increment). Where timber harvest is source rent; agriculture includes hunting, fishing, not sustainable, because harvest exceeds growth, and logging. The present value is taken over the a reserve life is calculated and the timber resource mean productive years of the population-the is treated in the same manner as a mineral. lesser of 65 years or life expectancy at age one, * Nontimber benefits of forests are valued by assum- minus the mean age of the population. Subtract- ing that 10 percent of forested area will yield an ing produced assets derived from the perpetual infinite stream of benefits in the form of nontimber inventory model, and urban land from this present products, hunting, recreation and tourism. Per- value yields the value of human resources at cur- hectare values of nontimber benefits vary from rent exchange rates. This is then revalued using $112 to $145 in developing and developed coun- the purchasing power parity rate to obtain the fi- tries. nal value of human resources. * Cropland is valued as an infinite stream of land * Not included: Fish are excluded from the analysis, rents, where land productivity is projected by re- partly for data reasons and partly because poor gion up to the year 2025 and held constant there- partlymfor dat ren and tl because poor after. Individual rental rates for rice, wheat, and of the world's darine fisheries Inland fisheries maize are multiplied by, production values at world prices to arrive at per-hectare unit rents for could be valued but data limitations prevent this cereal lands; other arable land is valued at 80 per- at present. Water is also not included-most wa- cent of this rate. ter use is already included in the value of agricul- * Pasture land is treated similarly to cropland-rental tural and industrial output. Domestic water use rates are derived from the value of beef, mutton, could be valued in the future but was not included milk, and wool production at world prices. here. range of $100,000 to $200,000, again boosted by sources play a considerable role in a number of natural resource endowments: Chile (minerals, cases. One measure of this is provided by the dif- forests), Argentina (agricultural land), Trinidad ferences in country rankings by PPP-based GNP and Tobago (oil), Malaysia (oil, forests), and some per capita and total wealth per capita, as shown of the Middle Eastern oil states. in table 3.2. While rankings are necessarily rela- As a generalization a broad array of South- eatAsian, agendral iou anbroad Ceral America tive, so that a rise by one country must correspond east Asian, and South and Central American to a fall by another, these differences in rankings countries fall in the range of $50,000 to $150,000 to aflb anoter, the inprankings per capita. Of the~ Su-aaa Afia cuti s give some appreciation for the importance of per capita. Of the Sub-Saharan African countries naua reore inavreyofcutis Botswana and Namibia are the highest ranking, natural resources in a variety of countries. falling in the range of $50,000 to $100,000. All The industrial countries most clearly affected other Sub- Saharan African countries fall below are Australia and New Zealand, while develop- US$50,000, as do China and most of South Asia. ing countries where increases in rank are impor- While human resources are the dominant tant include many from Sub-Saharan Africa, source of wealth for most nations, natural re- Central and South America, and the Caribbean. Table 3.1 Countries ranked by wealth per capita, 1994 (thousand US$) 300f400 250-300 200-250 1s0-200 100-So 7.5-100 50-75 25-50 <25 United States Australia Finland Portugal Chile Mauritius Namibia Sri Lanka Mauritania Switzerland France Ireland Saudi Arabia Argentina Panama Dom. Rep. Jamaica Niger Canada Denmark Spain Korea, Rep. of Greece Costa Rica Ecuador Philippines Bangladesh Japan Belgium Malaysia Brazil Jordan El Salvador Centr. Afr. Rep. Norway Austria Trinidad and Tobago Botswana Paraguay P New Guinea Cote d'lvoire Germany Uruguay Colombia Indonesia China India New Zealand Thailand South Africa Peru Bolivia Kenya Netherlands Mexico Tunisia Morocco Honduras Guinea-Bissau United Kingdom Venezuela Turkey Egypt Pakistan Gambia, The Sweden Guatemala Cameroon Vietnam Italy Senegal Togo Congo Nepal Zimbabwe Madagascar Lesotho Uganda Ghana Chad Nicaragua Zambia Benin Burkina Faso Haiti Mali Sierra Leone Burundi Mozambique Tanzania Malawi Rwanda Notes: Countries are ordered in rough wealth per capita intervals based on preliminary results. Note that wealth intervals are uneven. Estimates for Eastern Europe and countries of the former Soviet Union are not included because of uncertainty about data quality. Source: Authors' estimates. Measuring the Wealth of Nations 25 Regional and Income-Level Comparisons Table 3.2 Increases in country rankings (rank by wealth per capita compared wth rank by PPP GNP As seen in the regional breakdown of total wealth per capita) (table 3.3), the overall predominance of human Natural resources in explaining national wealth is strik- Increase in capital ing, followed in importance by produced assets Country rank ($/capita) and then by natural capital. Niger 15 12,000 Madagascar 9 7,000 Regional CcLmparisons Australia 7 35,000 Guinea-Bissau 7 8,000 Human resources account for 60 percent or more Senegal 6 5,000 of total wealth in all regions except the Middle Mali 5 5,000 East and more than 70 percent in 5 of the 12 re- New Zealand 5 51,000 gions presented. Produced assets are generally Central African Rep. 4 6,000 second in importance and vary from a low of 15 Honduras 4 3,000 percent of the total wealth to a high of 30 percent, Canada 3 37,000 with 20 percent or less for the poorest regions. Paraguay 3 7,000 Natural capital is generally third in importance Chad 3 6,000 in explaining national wealth, but again it varies Dominican Rep. 2 8,000 widely, from a low of 2 percent to a high of 39 Saudi Arabia 2 72,000 percent of the total. Natural capital exceeds the El_Salvador_2 ______ value of produced assets in the Middle East and Source: Authors' estimates. West Africa. The country-level natural capital numbers underlying these regional aggregates are productivity tomorrow, for instance, the conse- presented in appendix table 1. - quences for human welfare both locally and glo- Care must be taken in interpreting the low bally would be extremely serious. share of natural capital in the most developed re- Regional variations are illustrated in figure 3.1, gions. This is not the same as saying that natural which gives the shares of total regional wealth. capital is unimportant in these regions, but rather The Sub-Saharan Africa and South America that there is a preponderance of human resources charts present a more balanced breakdown of and produced assets. If the agriculture of North sources of wealth: natural capital accounts for America and Western Europe were halved in about 13 percent in Sub-Saharan Africa and 9 Table 3.3 Wealth per capita by region, 1994 Dollars per capita Percent share of total wealth Total Human Produced Natural Human Produced Natural Region wealth resources assets capital resources assets capital North America 326,000 249,000 62,000 16,000 76 19 5 Pacific OECD 302,000 205,000 90,000 8,000 68 30 2 Western Europe 237,000 177,000 55,000 6,000. 74 23 2 Middle East 150,000 65,000 27,000 58,000 43 18 39 South America 95,000 70,000 16,000 9,000 74 17 9 North Africa 55,000 38,000 14,000 3,000 69 26 5 Central America 52,000 41,000 8,000 3,000 79 15 6 Caribbean 48,000 33,000 10,000 5,000 69 21 1 i East Asia 47,000 36,000 7,000 4,000 77 15 8 East and Southern Africa 30,000 20,000 7,000 3,000 66 25 10 West Africa 22,000 13,000 4,000 5,000 60 18 21 South Asia 22,000 14,000 4,000 4,000 65 19 16 Note: West Africa does not include Nigeria's data because of data quality issues. Similarly, Algeria's data are not included in the estimates for North Africa. Source: Authors' estimates. 26 Expanding the Measure of Wealth percent in South America, produced assets for Figure 3.2 Wealth shares in low-income natural resource another 23 percent and 17 percent, and returns exporters, 1994 to human resources for the remaining 64 percent _ and 74 percent respectively. In contrast, wealth ;ets shares in the Middle East reflect the primary im- 21% portance of the oil and gas resources found there-fully 39 percent of total wealth comes from natural capital, almost entirely accounted for by oil and gas. Since some low-income countries are highly NHtuuman resource dependent, it is worth examining more capital 59% closely the composition of wealth in these coun- 20 tries, as shown in figure 3.2. Not surprisingly, these economies, which are largely dependent on Source: Authors' estimates. export revenues from primary commodities (other than petroleum), have a large share of their wealth in natural capital (20 percent). aoh r 4 pecn.Tme esucsaeipr tant but not dominant, accounting for 3 to 10 per- Importance of Natural Capital cent of the total value, with non-timber forest benefits and protected areas each adding another Even though natural capital is normally third in few percentage points. A somewhat surprising importance as a source of wealth behind human finding is the importance of metals, minerals, and resources and produced assets, it does form the fossil energy sources. The share of this aggregate ecological basis for life and is a fundamental build- category ranges from a low of 8 percent in the ing block of national wealth. The composition of poorest countries to 48 percent in the upper- natural capital varies from country to country, and middle-income countries that include many of the a disaggregation by income groupings is instruc- energy exporters. tive (table 3.4). Not surprisingly, agricultural lands The lessons for sustainable development (including pasturelands) are the most important based on the analysis of the composition of natu- type of natural capital, often accounting for half ral capital include the importance of maintain- or more of all natural wealth. This is especially ing the potentially renewable resource base in a common in the poorest countries, where agricul- healthy state. This applies to agricultural lands, tural croplands alone account for 80 percent of pastures, and forests and protected areas. Since the total natural capital and pasturelands add the last category in table 3.4, the metals and min- Figure 3.1 Wealth shares across selected regions, 1994 Sub-Saharan Africa South America Middle East Produced Produced Produced assets _assets _ asset __ 17_ __ 23% Natural N6l3%Jr tl=Human an ~~~~~~~~~rsurces Human Human Naturl 43% rsucsresources capirtal 64% 39% Note: Sub-Saharan Africa does not include data on Nigeria. Source: Authors' estimates. Measuring the Wealth of Nations 27 Table 3.4 The components of natural capital, 1994 (percentage shares) Country income Agricultural Pasture Nontimber Protected Metals, minerals, class crop/and land Tlmber forest benefits areas oil, coal, gas High 41 IS 10 4 11 19 Upper middle 28 10 6 5 2 48 Lower middle 56 5 5 3 4 27 Low 80 4 3 2 2 8 Source: Authors' estimates. erals and fossil energy sources, are by definition predominant method of valuing natural re- nonrenewable, the important policy questions for sources employed in these estimates. these resources are those relating to the rate of * Natural capital is important regionally, making depletion and whether the revenues gained from up more than 10 percent of total wealth in the depletion are invested or consumed. Since these Caribbean, East and SouthernAfrica, the Middle nonrenewable resources account for 10 percent East, South Asia, and West Africa. non perenewab ale resources cacot f, 10 pre * Petroleum and mineral resources are important to 50 percent of all natural caital, these are not components of wealth outside of the low-income trivial questions. countries, ranging from 8 percent to nearly 50 percent of natural capital. Conclusions and Policy Implications pecnfnt ralcptl * Human resources, including the retums to raw labor, are the dominant component of wealth, Expandig the measure of wealth sheds new light between 40 to nearly 80 percent of the total in all on the model of development that has guided the regions. lending and advice given by development insti- There is no simple story about natural resource tutions. This fresh analysis of the elements of wealth and growth. While there is some evidence wealth, the determinants of future well-being, that the most resource- intensive economies have suggests a number of conclusions and leads to grown more slowly than their peers since the new policy implications. The conclusions are: 1970s (Sachs and Warner 1995), in the end the • Valuing human resources on the basis of the pur- transformation of resource wealth into income chasing power of currencies (the PPP approach) growth depends on sound policy, in particular rather than nominal exchange rates significantly the effectiveness of public investment of resource decreases the range between low-income and rents. high-income countries, giving a truer picture of It must be emphasized that the analysis of ag- the consumption possibilities that citizens of gregate wealth presented here ignores the distri- these countries enjoy. This has a major impact gete wealth psther ignores th issui- on the estimates of wealth presented here. abuton of wealth within countries. An issue for * This does not mean that poverty has been "as- manygcountrieswi tereor be not ponlytes sumed away" in these estimates: forty two of the management of existing wealth but also policies ninety countries examined have wealth per affecting its distributon. Analyzing aggregate capita figures of less than $50,000 and twenty wealth also masks the contribution of social capi- five of these have per capita wealth of less than tal to economic development (see chapter 6). This $25,000. This compares with an OECD average preliminary exercise suggests new avenues for wealth in excess of $250,000 per capita. broadening and strengthening wealth estimates * Agricultural land is the dominant natural (see box 3.4) as well.- resource across country income classes, making Perhaps the key policy question from the up more than 50 percent of natural capital, with analysis is raised by figure 3.2: how does a low the exception of upper-middle- income countries income resource-exporting country transform it- (where petroleum-exporting countries tend to be self into a high income country? There is no classified). This is especially the case for low-in- simple policy prescription, but elements of the come countries, where agricultural land is more answer must include: by exploiting exhaustible than 80 percent of natural wealth. It must be em- resources and investing the rents effectively; by phasized again, however, that use value is the managing renewable resources (forest, fisheries, 28 Expanding the Measure of Wealth agricultural land) sustainably; by increasing in- for development is created. A key role for vestment in produced assets; and by increasing governments is in establishing the incentive investment in human capital in lockstep with framework within which private saving and in- other investments. vestment are made. But government investment The measurement of the wealth of nations in is critically important as well, whether in infra- total and in its components is largely motivated structure, education, or the sorts of institutional by concerns about sustainable development, in structures within which social capital can grow. particular the notion of sustainable development In terms of development outcomes, the quality as preserving and enhancing opportunity. While of public investment is paramount. There is a this is important in itself, given the commitments wealth of empirical experience to suggest that in- that the World Bank and its client countries have vestments in human capital are a very high qual- made to achieving environmentally sustainable ity outlet for public investments. development, it also suggests a new paradigm The analysis of wealth in this chapter supports for economic development. The traditional ap- this new development paradigm and suggests the proach to development places a heavy emphasis following broad facts: on building infrastructure. The analysis of wealth * Natural resources are an important share of suggests a new model for economic development: wealth in many countries, particularly low-in- development as portfolio management, the process come countries. of transforming an endowment of assets in order o Management of natural resources and the envi- ronment is therefore an important part of the de- to achieve development objectives. velopment process. Nations on the path to development have * Human resources nevertheless form the domi- three broad forms of endowment with which to nant share of wealth, even in low-income coun- work: natural resources, raw labor, and the so- trie& cial capital that is the result of the cultural tradi- * Porifolio balance may be important-that is, tions and historical experience of the nation in while investments in produced assets are a nec- essary part of the development process, con- question. These endowments, plus the historical comitant investments in human capital are accumulation of produced assets and human required, and these represent a highly levered capital, represent the starting point for the de- target for public (and private) investment. velopment process. Partqof this endowment, the Where does this analysis lead? One impact is natural resource base, can be a source of devel- largely didactic: the measurement of wealth pro- opment finance when properly managed; when vides the information needed to change the think- combined with savings by households and insti- ing of policymakers about the nature and tutions, and net foreign saving, a pool of finance composition of wealth. However, reassessing the Box 3.4 Directions for future work In undertaking this exercise, several simplifying but economy has a clear link to its capacity for investing well-founded assumptions have been made. The exer- in the different types of capital for future generations. cise has come up against constraints of data availabil- Second, the estimates of subsoil assets are based ity and quality, some of which can be more easily on proven recoverable reserves. But unproved reserves resolved than others. But there are certain areas in also have a market value and an economic value based which further work could clearly enhance the estimates on option pricing. An estimate of the value of unproved considerably and thus increase their policy relevance. reserves could also be included in total wealth. First, as measured here, total wealth comprises Finally, the value of a nation's water resources en- three types of capital-natural capital, produced as- ters the current calculations implicitly, via the value of sets, and human resources (including human and so- agricultural land. Water is an important input into ag- cial capital). It does not include the value of a nation's ricultural production and its value is included in the stock of financial assets and liabilities. Since the net returns to land. Even though in many countries the value of financial assets and liabilities is equal to net lion's share of water resources go to agricultural uses, foreign indebtedness, there would be considerable water use by industry and households could also be value in including the level of indebtedness in the valued by estimating consumptive uses and estimat- wealth measure. The level of indebtedness of an ing economic values for different uses. Measuring the Wealth of Nations 29 measure of wealth every five years or so would If the focus in measuring wealth had been nar- have practical implications as well, by providing rowly on human capital, working with educa- a set of milestones on portfolio composition to tional attainment data and estimates of the help guide the development planning process. returns to education in different countries would For the year-to-year process of economic and have been the preferred calculation method. This natural resource management, flow measures would have restricted the scope of the wealth es- will likely have the most policy relevance. This is timates considerably, completely ignoring the role the motivation for the previous chapter on the that social capital plays in the development pro- measurement and policy implications of genuine cess. So while the estimation of the returns to human resources as a residual, described below, saving-note however, that saving is the measure i I of changes in wealth, so the conceptual framework iS m one sense second-best in terms of rigor, it is of changer inch weth, soahe. cocpua rmeok the only way to arrive at a more encompassing is very much the same.v measure of the role that human capital, raw la- N lf op bor, and social capital play in the development ment finance. But there is no guarantee that boun- process. However, the residual approach does not tiful natural resources will lead to development permit separate estimates of the individual con- that is sustainable and equitable. Only sound stituents of human resources. policy can transform one into the other. The calculation methods used here are con- Technical Appendix strained both by the goals set out above and by the limitations of the data sets available interna- tionally. Analysts working in individual countries This appendix discusses the methodologies used could improve upon the estimates presented here for estimating the three major components of na- in a number of ways. First, a discount rate that is tional wealth-produced assets, natural capital, more appropriate for the country in question and human resources. The country-level break- could be used. Next, better estimates of total re- down of natural resource wealth appears in ap- source rents would result from site-specific data pendix table 1. on costs of extraction and harvest. Third, more The methods employed to measure the wealth robust values of the stock of produced assets of nations are firmly rooted in economic theory. could be employed. And if land and agricultural Resource rents provide the basis for valuing natu- produce markets are relatively undistorted, more ral resources, and stock values are estimated by direct values for agricultural and urban land taking the present value of flows. There are two could be derived from sales data. Finally, detailed overriding goals in choosing the calculation meth- data on educational attainment and the returns ods used in this chapter: to education could permit a direct estimate of the 1. The methods must permit cross-country compari- value of human capital, leaving raw labor and sons. social capital as the only residual categories in 2. The methods must focus on the economic poten- the total wealth calculation. tial in countries. As a practical matter the second point dictates using world prices to value natural resources and Produced assets, also known as human-made tradable assets, abstracting from whatever dis- capital, have always been an important policy tortions may exist in local markets. It also moti- variable for national economic planning. In the vates the use of purchasing power parity (PPP) wealth calculations the value of produced assets exchange rates for the valuation of human re- is based on the perpetual inventory model (Nehru sources, to better represent the consumption pos- and Dhareshwar 1993). These 1990 estimates have sibilities entailed by the returns to these resources. been extended to 1994 by calculating the net ac- The first point argues for the use of both world cumulation (initial capital stock plus gross do- prices and a common discount rate across coun- mestic investment less depreciation). tries in making the wealth estimations (the choice Note that machinery and equipment are val- of discount rate is described below). ued at the nominal, unadjusted prices reported 30 Expanding the Measure of Wealth in traditional national accounts. For other pro- be complex, depending on the mix of assets that duced assets-buildings and other structures- countries possess, with the effects being most valuation is at PPP exchange rates, since there is extreme for countries that have relatively long- a large non-tradable component to these assets. lived assets (large mineral deposits, sustainably Urban land has been valued as a fixed pro- managed forests, and young populations with portion of the value of buildings and other long life expectancies, for instance). While this structures. Balance sheet accounts for Canada will be the subject of future empirical investiga- (Statistics Canada 1985) reveal that the value of tion, it seems clear that the effect of altering the urban land is roughly 33 percent of the stock choice of discount rate would be to ratchet all of value of structures, and this percentage is used the total wealth values across countries up or as a placeholder in the current estimates. down, with relatively minor changes in country Choice of Discount Rate rankings and asset mixes within countries. Natural Capital Most of the stock estimates derived for wealth measurement are based on taking the present Natural capital includes the entire environmen- value of flows, making the choice of discount rate tal patrimony of a country. The estimates pre- clearly critical. Since the concern here is with sus- sented here are based on a subset of resources tainable development, the appropriate value selected for their general importance and the would be the discount rate that a govermnent availability of data. For any given country cer- would choose in allocating resources across gen- tain important resources may not be included, erations in its quest for sustainable development, pointing out the importance of making more de- This is essentially an argument for using social tailed national wealth estimates to inform policy- rather than private rates of discount, eliminating making at the national level. the effects of taxes and risk that elevate private The elements included in the natural capital discount rates. estimates presented here include agricultural The growth theory literature establishes the land, pasture lands, forests (timber and non- social rate of return on investment (SRRI) to be timber benefits), protected areas, metals and nin- given by erals, and coal, oil, and natural gas. There are two major measurement issues: what is included in SRRI = r + uc each variable, and how are values determined? For all of the elements of natural capital included where r is the pure rate of time preference (the in the wealth estimates, international market rate of impatience), u is the elasticity of the mar- prices were used, adjusted by an appropriate fac- ginal utility of income, and c is the growth rate of tor to represent the rent portion of the traded per capita consumption. Estimates of the SRRI for industrial countries (see, for instance, Pearce and re.eThe eonomicareoany nt resouc Ulph 1995) arrive at values in the range of 2 to 4 (also r re tas betadin rent whenc percent. For the fastest growing developing coun- prisons ar em d e between r cs vary tries it is clear that rates considerably higher than proct is the dif the mar- this would be appropriate, reflecting per capita oket price and the cost of the various inputs needed consumption growth rates that may be as high as to extract, process, and market it. As such, it rep- 7 to 9 percent. For the poorest-performing devel- resents the inherent surplus value in the extrac- oping countries per capita consumption growth tion or harvest of a resource. When there is rates have been near zero for nearly two decades, overexploitation of a resource like a fishery, rents and so a choice of SRRI less than that for devel- may be driven to zero (the entire market value of oped countries would be appropriate. the fish is absorbed by labor and capital costs, Because making cross-country comparisons is for example), thereby dissipating the natural profit one of the goals of this chapter, a single discount inherent in the resource. rate for all countries of 4 percent was chosen. The The approach used here means that all coun- effects of choosing a higher or lower rate would tries are treated equally with respect to the val- Measuring the Wealth of Nations 31 ues assigned to their natural resources. For com- Forests are a potentially renewable resource ponents that are widely traded, such as metals, that can be managed in either a sustainable or an minerals, and hydrocarbons, this is a reasonable unsustainable manner. The value of forestland is assumption. For agricultural products and their based on annual roundwood production valued use in calculations of the values assigned to farm- at timber rents (price minus production costs). land and pasture, this assumption may introduce In countries where timber production is below an upward bias to land values in many of the net annual increments, this annual value is then poorest countries. This is especially true for re- capitalized using a 4 percent discount rate in per- moter areas where transport costs and lack of petuity. However, in countries where roundwood markets mean that there may be only limited op- production is greater than net annual increments, portunities for trade in agricultural products. forests are treated as a mine and the annual value Agricultural cropland is valued based on the av- is capitalized over the time to exhaustion. erage per hectare return (production times world Protected areas are lands that are set aside to commodity price) for three principal grains-rice, protect biodiversity or unique cultural, scenic, or wheat and maize-with the average price historic sites. There are many different types of weighted by the area sown to each crop in any benefits associated with protected areas, some of individual country. A crop-specific adjustment which are easier to value (for example, tourism factor, varying from 30 to 50 percent, is applied or recreation) and others of which are very hard to the gross value to represent the net economic to value (for example, unknown genetic materi- value (the economic rent) of each hectare (World als). Because of the difficulties in placing values Bank 1992, 1993, 1995a). This annual value, ad- on many of the benefits from protected areas we justed for changes in yield and area under culti- have used an opportunity cost approach to rec- vation over time (estimates from the International ognize that there is a positive value associated Food Policy Research Institute, Evenson and with these areas. In this case the values previously Rosegrant 1995, and Rosegrant and others 1995), identified for pastureland are used as a proxy for is then taken into perpetuity at a 4 percent dis- a minimal value for protected areas. This is done count rate to indicate the present and future for two reasons: first, many protected areas are values associated with sustainably managed ag- lands where agriculture, forestry, or pasture are ricultural lands. Other arable land is valued at 80 the most likely alternative uses; and second, the percent of the country-specific average per hect- implicit value of protected areas must be at least are for these crops. This adjustment is made to this large or else the area would not be protected recognize the fact that other crops are likely to and would have been allocated to other, more produce lower returns per hectare. While this is directly productive uses. Of course, in some cases a reasonable assumption for many of the coarse protected areas are areas without any productive grains and tuber crops, this assumption is prob- alternative uses (for example glaciers or deserts) ably not accurate for certain high-valued crops and in these cases the assumption used here may such as coffee, tea, vegetables, or cocoa. There- be an overestimate of their minimum value. fore, for countries that grow a lot of high-value Nontimberforest benefits. In each country one- noncereal crops our assumption would underes- tenth of forest area is multiplied by a per-hectare timate the value of land. estimate of nontimber forest benefits ($145 per Pasturelands were valued in a similar manner, hectare in industrial and.$:112 per hectare in de- whereby the national output of meat, wool, and veloping countries; Lampietti and Dixon 1995). milk is valued at international prices and an ap- This annual value is then capitalized at a 4 per- propriate rental rate (in this case 45 percent) is cent discount rate over an infinite time horizon. applied to estimate returns to pastureland. This This value is used as a "place holder" to recog- value is taken into perpetuity at a 4 percent dis- nize that there are important values associated count rate. In countries where there are signifi- with forests over and above the direct timber val- cant feedlot operations, yielding a very high ues. apparent return to pasture, the per-hectare value Metals and minerals are classic nonrenewable is capped at the level of the return to cropland. resources that can only be managed in a nonre- 32 Expanding the Measure of Wealth newable manner. The estimate of a nation's assets and the depreciation. of produced assets. wealth of metals and minerals is therefore a func- Since the value of resource stocks has been ex- tion of production and reserves, average extrac- plicitly accounted for under natural capital, it is tion rate, and the economic rent from their important to subtract the rents on current pro- extraction. Detailed calculations for the genuine duction of these stocks from current income savings estimates (see chapter 2) are used to esti- (nonagricultural GNP) so as not to double count. mate a nation's mineral wealth. Yearly produc- This amount is then discounted over the average tion (extraction) is valued using estimates of number of productive years of the population, cal- resource rents for the following metals and min- culated by taking the life expectancy of the na- erals: bauxite, copper, iron ore, lead, nickel, phos- tional population at year 1 (or 65, whichever is phate rock, tin, and zinc. The return is smoothed lower) and subtracting the mean (average) age of over the period 1990-94 and then capitalized at a the population. This calculation estimates the re- 4 percent discount rate over the remaining time turns to human resources, produced assets, and to exhaustion. Present value is calculated by as- urban land with the given population or age suming an optimal path for unit scarcity rents and distribution and the current labor force that im- a constant revenue stream (see appendix box 1). plicitly reflects current levels of under- and un- The reserve measure used is for proven reserves, employment. the quantity of resource that can be exploited The lower of age 65 or life expectancy at age 1 profitably at current prices and costs-this is a is used as the cutoff since we assume that indi- conservative measure, therefore, and will tend to viduals are no longer working and producing change over time with new discoveries, new tech- beyond that age. For example, in India life nologies, and changes in world prices. Where data expectancy at age 1 is 61 years, and the average on reserves were not available, time to exhaus- age of the population is 27 years, yielding remain- tion is assumed to be 20 years. ing years of productive activity of 34 years. For Oil, coal, and gas are the main energy resources Sweden, the life expectancy at age 1 is 78 and the included in the wealth analysis. These fuels are mean age of the population is 40. Using an up- also nonrenewable. Yearly production is valued per bo vle pofu65tionreflet woring lf rath using estimates of resource rents. The value of than life expectancy, in Sweden this yields a time returns is smoothed over the period 1990-94 and horizon of 25 years. then capitalized at a 4 percent discount rate over hese annua valsa the remaining time to exhaustion. As with met- using a 4 percent discount rate. From this stock als and minerals, present value is calculated by vue werct the vale Frod assts assuming an optimal path for unit scarcity rents value we subtract the value of produced assets and a constant revenue stream (see appendix box uman resources. 1). Where data on reserves were not available, BeauseGf time to exhausfion is assumed to be 20 years. Because GNP includes net flows of wages and property incomes abroad, the methodology just Human Resources described for arriving at the residual return to human resources implies that these net flows of The most difficult component of national wealth income abroad are all ascribed to human re- to measure is the return to individuals and soci- sources. Some portion at least of these flows is eties from the use of natural capital and produced probably better ascribed to produced assets, capital. In this calculation the value assigned to which will increase or decrease the estimated human resources is a residual value based on the value of human resources according to whether return to a country's population estimated by the net flow into a given country is (respectively) multiplying agricultural GDP (which includes the negative or positive. This forms part of the value added of the forestry and fisheries sectors agenda for refining the wealth estimates in the in the aggregate national accounts data used here) future. by 45 percent to reflect the return to labor com- Unlike the assumptions made for natural capi- ponent, and then adding all non-agricultural GNP tal, for which international prices are used, and and subtracting the economic rents from subsoil for produced assets for which national prices are Measuiring the Wealth of Nations 33 Appendix box 1 Valuing subsoil assets In the absence of competitive markets for stocks of sub- nential rates. The empirical evidence is against the lat- soil assets, estimating their value necessarily entails ter assumption at least-trends in prices for subsoil taking present values of the total rents (defined as the resources have been flat or slightly declining for the economic profits on extraction: net operating surplus past several decades. less a "normal" return on produced assets) generated If resource prices are assumed to continue their flat over the life of the resource deposit. Calculating this trend and unit scarcity rents (price minus marginal ex- present value requires strong assumptions about the traction costs) are assumed to increase as optimization discount rate and the time path of total rents. requires, then at least two other approaches to valua- I Two pragmatic alternative approaches to valuing tion are possible. One is to assume constant produc- I subsoil assets have been suggested in the literature: tion quantities over the life of the deposit; the second net prices and simple present values. The latter is fa- is to assume constant resource revenues, which, when i vored by many national statisticians because it uses combined with rising unit scarcity rents, implies de- | quite neutral assumptions: total rents are assumed to clining quantities extracted. The second of these is constant over a mine life that is determined by the clearly the more attractive: it seems a reasonable propo- be. . sition that in an uncertain world resource owners current reserves to production ratio, and the present woul see stab n r e e value is taken using the social discount rate as the dis- for a path for scarcity rents that is optimal. It is the count factor. The net price approach (used, for instance, "constant revenue" assumption that is used in this chap- in Repetto and others 1989) is even simpler: the re- ter. Empirically, this gives values that are substantially source deposit is valued as the current unit economic greater than the simple present value estimate, but less profit times the quantity of resource in the ground. than the net price approach. The problem with the simple present value ap- Because real mines do not behave like textbook proach to valuation is that it assumes no optimization; mines, the problem of valuing subsoil assets is inher- mine owners would implicitly be holding assets whose ently one in which there are no good solutions, only yield is zero. However, the net price approach faces "less bad" ones. The constant revenue approach used equally serious difficulties: to be consistent with opti- here fits this criterion by being consistent both with mization, it implicitly assumes that unit extraction costs optimization and the empirical fact of a flat long-term are constant and that resource prices rise at near expo- trend in prices of subsoil resources. used, for human resources we have used PPP estimate is $1,280; for Japan the GNP estimate is values rather than unadjusted national prices. $34,630 per capita, while the PPP is $21,140. This adjustment has the effect of flattening the The values for human resources combine distribution of per capita incomes, increasing the many diverse elements, including the impacts of values for the poorest countries, and reducing per investment in education and health, and, most capita incomes for some of the high-income coun- difficult to measure, social capital, the "ether" tries. For example, in World Development Report within which societies use and combine physi- 1996, the GNP per capita figure at nominal cal, natural and human capital. Social capital is exchange rates for India is $320 while the PPP discussed at length in chapter 6. 34 Expanding the Measutre of Wealth Appendix table I Country-level natural capital estimates Nontimber Natural Timber forest Protected Subsoil capital Pastureland Cropland resources resources areas assets Country $per capita (percentage of total) Argentina 9,850 3,270 5,200 280 480 100 520 (33) (53) (3) (5) (1) (5) Australia 35,340 7,270 14,150 1,030 2,150 1,650 9,080 (21) (40) (3) (6) (5) (26) Austria 7,570 1,480 2,410 1,720 150 1,580 230 (20) (32) (23) (2) (21) (3) Bangladesh 3,110 60 3,000 0 0 10 20 (2) (97) (0) (0) (0) (1) Belgium 1,750 470 1,110 100 20 50 10 (27) (63) (6) (1) (3) (1) Benin 1,930 70 1,030 440 250 120 10 (4) (54) (23) (13) (6) (1) Bolivia 6,060 690 2,520 160 1,820 240 640 (I1) (42) (3) (30) (4) (I 1) Botswana 5,620 1,180 260 420 2,700 490 570 (21) (5) (8) (48) (9) (10). Brazil 7,060 1,070 2,740 1,200 960 190 910 (15) (39) (17) (14) (3) (13) Burkina Faso 2,400 210 1,870 100 120 90 (9) (78) (4) (5) (4) Burundi 1,940 90 1,820 10 10 10 0 (5) (94) (0) (1) (0) (0) Cameroon 6,800 270 4,840 650 430 270 340 (4) (71) (10) (6) (4) (5) Canada 36,590 2,310 9,910 6,230 4,560 6,830 6,750 (6) (27) (17) (12) (19) (18) Central Afr. Rep. 6,470 440 2,010 520 2,600 900 (7) (31) (8) (40) (14) Chad 5,550 470 4,110 340 500 120 (9) (74) (6) (9) (2) Chile 14,440 1,100 4,910 1,560 180 1,110 5,580 (8) (34) (I1X) (1) (8) (39) China 2,670 100 2,010 90 30 10 420 (4) (75) (3) (1) (1) (16) (table continues on the followng page) 0 less than $ 10 per capita. (0) less than I percent. .. No data. Note: Numbers in parentheses are percentages of the total. Estimates for Eastern Europe and the countries of the former Soviet Union are not induded because of uncertainty about data quality. Similar problems exist for Algeria and Nigeria. Measuring the Wealth of Nations 35 Appendix table I Country-level natural capital estimates (continued) Nontimber Natural Timber forest Protected Subsoil capital Pastureland Cropland resources resources areas assets Country $per capita (percentage of total) Colombia 6,100 1,160 2,490 390 410 270 1,380 (19) (41) (6) (7) (4) (23) Congo 4,420 20 200 1,040 2,200 0 960 (i) (4) (24) (50) (0) (22) Costa Rica 7,860 1,480 5,690 180 100 410 (19) (72) (2) (1) (5) C6te d'lvoire 3,790 80 2,870 570 210 10 30 (2) (76) (IS) (6) (0) (1) Denmark 11,070 270 7,210 380 30 1,930 1,260 (2) (65) (3) (0) (17) (11) Dominican Rep. 8,380 560 7,310 90 30 280 100 (7) (87) (1) (0) (3) (1) Ecuador 11,330 1,160 4,880 440 270 2,610 1,970 (10) (43) (4) (2) (23) (17) Egypt 2,360 420 1,540 0 0 70 330 (18) (65) (0) (0) (3) (14) El Salvador 1,150 250 890 10 10 0 (22) (77) (1) (0) (0) Finland 15,930 90 4,670 6,970 1,660 2,420 110 (1) (29) (44) (10) (15) (1) France 8,120 1,350 5,210 700 90 700 60 (17) (64) (9) (1) (9) (1) Gambia, The 2,120 190 1,850 10 20 50 (9) (87) (1) (1) (2) Germany 4,150 430 2,100 490 30 750 350 (10) (SI) (12) (1) (18) (8) Ghana 1,920 60 1,510 190 150 10 10 (3) (78) (10) (8) (1) (1) Greece 5,210 1,490 3,080 170 90 60 320 (29) (59) (3) (2) (1) (6) Guatemala 1,720 300 930 170 110 150 60 (18) (54) (10) (6) (9) (4) Guinea-Bissau 7,970 200 7,440 330 .. (2) (93) (4) .. Haiti 840 110 720 0 0 0 0 (13) (86) (0) (0) (0) (0) Honduras 3,380 410 1,610 820 210 230 100 (12) (47) (24) (6) (7) (3) (table continues on the following page) 36 Expanding the Measuire of Wealth Appendix table I Country-level natural capital estimates (continued) Nontimber Natural Timber forest Protected Subsoil capital Pastureland Cropland resources resources areas assets Country $per capita (percentage of total) India 3,910 90 3,440 50 20 110 210 (2) (88) (1) (0) (3) (5) Indonesia 7,480 60 5,780 720 150 100 670 (1) (77) (10) (2) (1) (9) Ireland 17,780 11,770 4,810 510 40 120 530 (66) (27) (3) (0) (1) (3) Italy 3,400 430 2,430 110 40 230 160 (13) (71) (3) (1) (7) (5) Jamaica 3,080 110 280 50 10 0 2,630 (4) (9) (2) (0) (0) (85) Japan 2,300 120 1,360 220 70 490 40 (5) (59) (10) (3) (21) (2) Jordan 1,020 260 360 0 0 100 300 (26) (35) (0) (0) (9) (29) Kenya 1,730 740 840 10 10 120 0 (43) (49) (1) (1) (7) (0) Korea, Rep. of 2,940 50 2,290 120 40 390 50 (2) (78) (4) (1) (13) (2) Lesotho 940 340 600 0 0 0 (36) (64) (0) (0) (0) Madagascar 6,510 500 5,350 310 330 20 (8) (82) (5) (5) (0) Malawi 880 60 600 90 80 40 (7) (68) (I1) (10) (4) Malaysia 11,820 20 6,190 1,310 230 840 3,230 (0) (52) (I1) (2) (7) (27) Mali 4,840 530 3,620 270 340 70 (I11) (75) (6) (7) (1) * Mauritania 5,100 1,060 2,270 0 70 50 1,640 (21) (45) (0) (1) (1) (32) Mauritius 1,240 20 1,180 10 10 10 (2) (95) (1) (1) (1) * Mexico 6,630 810 1,520 200 140 110 3,860 (12) (23) (3) (2) (2) (58) Morocco 2,210 480 1,480 60 100 10 80 (22) (67) (3) (5) (0) (4) Mozambique 1,130 90 360 400 280 0 0 (8) (32) (35) (25) (0) (0) (table continues on the folowing page) Measuring the Wealth of Nations 37 Appendix table I Country-level natural capital estimates (continued) Nontimber Natural Timber forest Protected Subsoil capital Pastureland Cropland resources resources areas assets Country $per capita (percentage of total) Namibia 7,180 1,400 1,230 2,310 380 1,860 (20) (17) (32) (5) (26) Nepal 2,900 380 2,150 90 60 210 10 (13) (74) (3) (2) (7) (0) Netherlands, The 4,140 560 1,020 80 10 230 2,250 (14) (25) (2) (0) (6) (54) New Zealand 51,090 22,130 12,600 4,340 770 9,950 1,300 (43) (25) (9) (1) (19) (3) Nicaragua 3,690 540 2,110 580 360 90 0 (15) (57) (16) (10) (2) (0) Niger 12,340 310 11,600 50 80 300 0 (3) (94) (0) (1) (2) (0) Norway 30,220 110 1,680 2,520 700 5,110 20,090 (0) (6) (8) (2) (17) (66) Pakistan 1,880 140 1,480 0 0 100 150 (7) (79) (0) (0) (6) (8) Panama 6,300 930 3,960 270 310 830 (I5) (63) (4) (5) (13) Papua New Guinea 7,490 10 560 1,550 2,370 20 2,980 (0) (7) (21) (32) (0) (40) Paraguay 6,990 1,490 3,590 1,150 650 100 (21) (51) (16) (9) (1) . Peru 4,630 350 2,770 220 800 50 430 (8) (60) (5) (17) (1) (9) Philippines 2,730 50 2,400 140 30 30 80 (2) (88) (5) (1) (1) (3) Portugal 4,040 280 2,140 1,140 110 190 190 (7) (53) (28) (3) (5) (5) Rwanda 1,110 100 930 0 10 70 0 (9) (84) (0) (1) (6) (0) Saudi Arabia 71,880 330 3,600 20 20 67,910 (0) (5) (0) (0) (94) Senegal 5,300 290 4,180 310 250 210 60 (6) (79) (6) (5) (4) (1) Sierra Leone 3,040 60 2,570 180 110 0 120 (2) (84) (6) (4) (0) (4) South Africa 4,200 880 1,790 90 30 80 1,340 (21) (43) (2) (1) (2) (32) (table continues on the following page) 38 Expanding the Measure of Wealth Appendix table I Country-level natural capital estimates (continued) Nontimber Natural Timber forest Protected Subsoil capital Pastureland Crop/and resources resources areas assets Country $per capita (percentage of total) Spain 5,740 940 3,690 430 140 390 140 (16) (64) (8) (3) (7) (3) Sri Lanka 3,480 140 2,970 90 30 250 0 (4) (85) (3) (1) (7) (0) Sweden 14,590 440 4,390 5,890 1,160 2,300 410 (3) (30) (40) (8) (16) (3) Switzerland 3,050 950 820 600 50 620 0 (31) (27) (20) (2) (20) (0) Tanzania 2,200 310 920 530 310 120 0 (14) (42) (24) (14) (6) (0) Thailand 7,600 110 6,270 110 50 980 80 (1) (83) (1) (1) (13) (1) Togo 2,670 50 2,250 0 90 170 120 (2) (84) (0) (3) (6) (4) Trinidad and Tobago 12,110 70 2,540 40 40 100 9,310 (1) (21) (0) (0) (1) (77) Tunisia 6,370 550 5,070 10 20 10 710 (9) (80) (0) (0) (0) (11) Turkey 3,940 490 2,950 170 90 40 200 (12) (75) (4) (2) (1) (5) Uganda 2,230 120 1,680 210 90 130 0 (5) (75) (9) (4) (6) (0) United Kingdom 4,940 1,540 1,820 110 20 710 730 (31) (37) (2) (0) (14) (15) United States 16,500 2,570 7,210 1,730 410 1,400 3,180 (16) (44) (I0) (2) (8) (19) Uruguay 14,810 6,040 8,530 160 60 10 (41) (58) (1) (0) (0) Venezuela 20,820 860 3,130 40 570 1,270 14,960 (4) (15) (0) (3) (6) (72) Vietnam 3,990 70 3,490 70 30 260 70 (2) (87) (2) (1) (7) (2) Zambia 5,490 160 3,330 660 940 30 360 (3) (61) (12) (17) (1) (7) Zimbabwe 2,520 450 990 400 220 280 170 (18) (39) (16) (9) (I1) (7) Measuiring the Wealth of Nations 39 References Evenson, R., and M. W. Rosegrant, 1995. "Produc- Serageldin, I. 1996. Sustainability and the Wealth of Na- tivity Projections for Commodity Market Mod- tions: First Steps in an Ongoing Journey. Environ- eling." Paper presented at the final workshop of mentally Sustainable Development Studies and the International Cooperative Research Project Monographs no.5. Washington D.C.: World Bank. on Projections and Policy Implication of Me- Serageldin, I., and, A. Steer. 1994. "Epilogue: Ex- dium- and Long-term Rice Supply and Demand, panding the Capital Stock." In Serageldin and April 23-26, Beijing. Steer, eds.,Making Development Sustainable: From Lampietti, J. A., and J. A. Dixon. 1995. "To See the Concepts to Action. Environmentally Sustainable Forest for the Trees: A Guide to Non-Timber For- Development Occasional Paper no. 2. World est Benefits." Environment Department Paper 13. Bank: Washington D. C. World Bank, Environmnent Department, Wash- Smith, Adam. 1776. An Inquiry into the Nature and ington, D. C. Causes of the Wealth of Nations. Chicago: Univer- Nehru, V., and A. Dhareshwar. 1993. "A New Data- sity of Chicago Press, 1977. base on Physical Capital Stock: Sources, Meth- Statistics Canada. 1985. National Balance Sheet Ac- odology, and Results." Revista de Andlisis counts 1961-84. Financial Flows Section. Ottawa. Econ6mico 8 (1): 37-59. World Bank.1992. "Arab Republic of Egypt:An Ag- Pearce, D. W., and D. lJlph. 1995. "A Social Discount ricultural Strategy for the 1990s." Sector Report Rate for the United Kingdom." Centre for Social and SR-11083. Middle East and North Africa Region, Economic Research on the Global Environment. Agriculture Operations Division. Country De- Working Paper 95-01. University College London. partment II, Washington, D. C. Repetto, R., W. McGrath, M. Wells, C. Beer, and F. _ . 1993. "Republic of Yemen: Agricultural Sec- Rossini, 1989. Wasting Assets: Natural Resources tor Study: Strategy for Sustainable Agricultural in the National Accounts. Washington, D. C.: Production." Sector Report SR-11126. Middle World Resources Institute. East and North Africa Region, Agriculture Op- Rosegrant, M. W., M. Agcaoili-Sombilla, and N. D. erations Division, Country Department II, Wash- Perez. 1995. "Global Food Projections to 2020: Im- ington, D.C. plications for Investment." Food, Agriculture, _ . 1995a. "Mongolia: Prospects for Wheat Pro- and the Environment Discussion paper 5. Inter- duction. " Sector Report SR-13882. EastAsia and national Food Policy Research Institute, Wash- Pacific Regional Office, Agriculture Operations ington D.C. Division, China and Mongolia Department, Sachs, J. D., and A. M. Warner. 1996. "Natural Re- Washington, D.C. source Abundance and Economic Growth." _ . 1995b. Monitoring Environmental Progress: A Development Discussion Paper 517a. Harvard Report on Work in Progress. Washington, D.C. Institute for International Development, Cam- _ . 1996. World Development Report 1996: From Plan bridge, Mass. to Market. New York: Oxford University Press. CHAPTER 4 Subsidy Policies and the Environment t has long been recognized that government public goods such as infrastructure, which tend interventions often harm efficiency. In recent to be under-provided by private sector activity, years it has also been increasingly recognized or attempt to remedy perceived market failures. that many government interventions can result Governments also intervene to pursue noneffi- in substantial environmental damage. This situa- ciency objectives such as income distribution or tion creates the potential for win-win policy re- regional development objectives.And some inter- forms that both increase efficiency and reduce ventions are based on considerations of political environmental problems. This potential is particu- patronage. Whatever their intent, policies often larly great when policies subsidize environmen- persist long after the original reasons have been tally damaging activities. fulfilled (or not). They also tend to create vested This chapter examines recent trends in sub- interests. sidy policies in the energy and agricultural input Government interventions can take a variety sectors. These sectors are thought to have con- of forms: excise taxes and subsidies affect the siderable potential for win-win policy reforms be- prices of goods directly; tariffs increase the price cause of the high levels of inefficiency caused by of imported goods directly while import quotas subsidy policies and the extent to which they have increase them indirectly; exchange rate policy af- exacerbated environmental problems. fects the value of all tradable commodities; and The chapter begins with a discussion of the rules and regulations limit the choices economic different forms of government interventions and agents can make. In many countries governments the problem of measuring their effect on eco- have also been directly involved in the produc- nomic incentives. Because interventions can en- tion and distribution of important commodities. courage or discourage activities in many ways, a It is not uncommon for policies to work in oppo- broad definition of subsidies is needed. The dis- site directions. While some policies might encour- cussion then turns to the effect of subsidy poli- age an activity, others might discourage it. cies on the environment. The nature and extent of subsidy policies in the energy and agricultural Efficiency Effects input sectors are then discussed in detail. In par- ticular the chapter examines the extent to which Except where policies correct market failures, they subsidy policies have changed over time. always result in a loss of efficiency. This is a trade- off that may well be worth making; societies have Government Policies and the Environment many worthy objectives that are not related to economic efficiency. Often, however, policies are Governments intervene in their economies for a enacted without sufficient consideration-or with variety of reasons. Some interventions provide blithe disregard-of their efficiency consequences. 40 Subsidy Policies and the Environment 41 These consequences can prove quite dispropor- longer flow to the use in which they have the tionate to the non- efficiency objectives achieved- highest return. if, indeed, the non-efficiency objectives are achieved at all. It is not uncommon for policies to Subsidies and the Environment fail to achieve their objectives or even for them to have perverse effects. In recent years it has become increasingly clear that misguided government policies and distorted Subsidy Policies markets that set inappropriate prices for natural resources can result in substantial environmental The focus of this chapter is on policies that subsi- damage (World Bank 1992; Panayotou 1993; dize environmentally damaging activities. In this Munasinghe and Cruz 1995). The specific impact context subsidy policies are broadly defined as of particular policies on the environment is not any policies whose effect is to reduce the costs of always clear, however. Some policies have unam- an activity relative to what they would have been biguously harmful effects on the environment, but in the absence of the policies. For example, gov- others have uncertain environmental effects. Even emnment agencies might sell pesticides to farm- when the direction of environmental impact is ers for less than their cost of production or provide cear, the magnitude of the effect is often difficult credit for buying pesticides at below-market in- . terest rates. In both these cases government funds to establish withoutelaborte empiicl anaies - are ein usd t redce he ostof te pstiide an effort often inhibited by the deficiencies of are being used to reduce the cost of the pesticide available data. Likewise, analysis of the likely ef- to farmers. We shall call such policies explicit sub- to fames We sh' alsc oiexlctsb fects of policy reform is made difficult by the com- sidies. The cost of pesticides to farmers can also fety of policicl by the be reduced by other, more indirect means, such plexity of linkages between policies and the as policies that allow pesticide imports at prefer- environment and the scarcity of data on their mag- ential exchange rates. Although such policies do nitude. In a second-best world in which multiple not rely on explicit expenditures of government aistortions and market failures exist, it cannot be funds, they artificially reduce the cost of pesti- automatically assumed that all policy reform is cide use to farmers and thus have exactly the same beneficial. effect on behavior as an equivalent explicit sub- There is nothing ambiguous about the harm- sidy. ful impacts of subsidizing environmentally dam- Under this approach the extent of subsidies aging inputs, however. Since any increase in use for input use is given by the difference between of such inputs increases environmental damage, what users pay for that input and what they and since subsidies result in increased use, the would have paid in the absence of all policies, direction of impact is clear (box 4.1). The magni- multiplied by the quantity used. The subsidy rate tude of the impact is often uncertain, however. is the difference in price expressed as a propor- The extent of damage depends on two factors: tion of the unsubsidized price. Because subsidy the responsiveness of input use to changes in sub- rates do not depend on quantities used, they are sidies, and the amount of damage caused by each often better guides to policy changes than subsi- unit of input use. dies. The amount of explicit subsidies is given by government expenditures on subsidies. This is Win-win policy reform. Although design of only a subset of all subsidies, and thus is a poor policy reform that fully incorporates environmen- indicator of the extent to which incentives are af- tal considerations is a complex task, considerable fected. However, the magnitude of explicit sub- improvements can be achieved with measures sidies is often of interest to decisionmakers who that fall well short of this ideal. A full analysis of have to allocate scarce budgetary resources. the effects of policy reform on the environment Whether subsidies are direct or indirect, they is outside the scope of this chapter. In cases where reduce the cost of inputs. Use of subsidized in- environmentally damaging inputs are subsidized, puts is encouraged and will be higher than it however, the linkages between policies and en- would have been in the absence of the subsidies. vironment are clear; the required reforms, there- This creates an inefficiency because goods no fore, are equally clear. Because such subsidy 42 Expanding the Measuire of Wealth Box 4.1 The simple economics of win-win subsidy reform The double benefits of subsidy reform in the case of Efficiency effects of subsidies. If the government sub- environmentally-damaging inputs are easily demon- sidizes use of the input, farmers will perceive a lower strated. Consider an input such as a fertilizer or a pes- price for the input, P.. Accordingly, they will increase ticide. In deciding how much of that input to use, their use of the input to quantity Q5. This results in a farmers consider both the benefit they expect to receive social inefficiency. Because of the subsidy the price of and the cost of the input. The marginal return to addi- the input no longer reflects its social opportunity cost, tional units of the input generally diminishes as use of and more of it is used than would be socially optimal. the input increases. Farmers use the input until the mar- Removing the subsidy would increase efficiency ginal return obtained from using an additional unit of REvirngthe ould inrea. sewefficiency. the input equals the cost of the input. Box figure 4.1 ve nvinmetalso subsidieo sentat S s s * * r s . 1 ~~~use of this input also results in environmental dam- shows this by the intersection of the marginal revenue (MR) line and the price (F) line. On this basis farmers age, as shown in the lower panel of the figure. In the would use quantity Q of the input, case of fertilizers or pesticides this damage could take the form of polluted runoff, for example. The greater the use of the input, the greater the environmental dam- nBox figure 4.1 Links between prces, input use, age. Since subsidizing an input induces farmers to use and environmental damage more of it, the resulting environmental damage will increase, from D to D1. Price of input Win-win reforms. Reducing or eliminating subsidies on environmentally damaging inputs would therefore, be a win-win reform since there would be both effi- ciency and environmental benefits. Efficiency would | PG \ \ A improvebecause overuse of the input would decrease. "Green tax" The environment would benefit because the reduced P \: \ T use of the input resulting from removing the subsidy would reduce environmental damage, from D, back to D. Subsidy Limitations of subsidy reforms. This is not to say, how- } PS * \ \ T ever, that simply removing subsidies would result in a socially optimal outcome. The pollution costs of input use are an extemality that farmers have no incentive M[R to take into account, so they would still be overusing the input even in the absence of subsidies. If the value . \ MRE of the environmental damages could be calculated, then *______________________ the socially optimal level of input use could be calcu- QE: Q* Q: . Quantity lated. This is shown by curve MRE, which shows the Q, Q: Qs of input marginal revenue resulting from use of the input tak- Envirownmental * ing environmental costs into account. Using this curve, damage the socially optimal level of input use is QE. At this level the marginal benefits derived from use of the input exactly offset its cost and that of the resulting environmental damage, DE. To ensure that farmers | Ds . * . w *; * * . . . A choose this input level, a green tax would have to be applied, increasing the price they perceive to PG. Al- D.. */ . vtematively, measures could be taken to encourage a D *........,/ *- ' ' aag u shift to an alternative, cleaner technology-that is, one Damage du in which less or no environmental damage is caused /~ ........ per unit of output. Removing subsidies, therefore, is often only a first step toward ensuring that environ- Quantity mentally damaging inputs are used in a socially opti- _______ ______ ______of input mal way. policies create both economic inefficiency and in- Limitations. While removing subsidies will creased environmental damage, removing them both improve efficiency and reduce environmen- is a win-win reform. tal damage, it will not necessarily result in so- Stubsidy Policies and the Environment 43 cially optimal outcomes. Indeed, it is often safe (Burniaux and others 1992). (All values are given to assume that it will not-removing subsidies in 1995 U.S. dollars, unless otherwise indicated.) will reduce incentives to overuse environmentally In 1990-92 total fossil fuel subsidies were esti- damaging inputs, but it will not create incentives -mated to be $235-245 billion (Larsen and Shah to take environmental externalities into account 1992; Larsen 1994) not including OECD countries, (box 4.1). Additional measures may also be nec- in which subsidies were estimated at $12 billion essary to complement the subsidy reforms and in 1990-91 (IEA 1995). ensure that they have the desired effect. This section focuses on fossil fuel subsidies. Care is also necessary when, as is often the Fossil fuels are the main forms of energy used in case, there are multiple distortions. Removing one the world today, accounting for about 90 percent distortion may not result in improved efficiency of total world primary energy supply (including if other distortions remain. The same is true of fossil fuels used to generate electricity). The cur- environmental problems. For example, removing rent state of fossil fuel subsidies is examined and subsidies on relatively clean fuels such as natu- compared with the situation five years ago, based ral gas may worsen problems if subsidies persist on a recent draft study (Rajkumar 1996). for other, dirtier fuels, since the change in rela- tive prices will make the latter more attractive. Energy Use and the Environment Measuring Subsidies Energy use is, of course, necessary for economic activity. But energy use can also have a variety of Measuring subsidies is complicated by the great harmful effects on the environment. variety of policy instruments used by govern- * Air pollution. Fossil fuel combustion releases pol- ments and by the poor quality of available data. lutants such as sulphur dioxide (SO2), nitrogen Given the broad definition of subsidies adopted oxides (NOX), and particulates (TSP) into the at- here, the approach taken to subsidy measurement mosphere. Particulates and sulphur dioxide can consists of comparing the prices of the goods un- cause acute and chronic respiratory conditions der consideration to reference prices that repre- such as pharyngitis, bronchitis, and asthma, as ser the prices these goods would have had in well as lung cancer and abnormal physiological senthe price thes. This woach cadtin development. Nitrous oxides can cause respira- the absence of policies. This approach captures tory disease in children. Acid rain (from sulphur the net effect of all the different policy instruments dioxide), soot, and ash from incomplete combus- that affect a good's price. tion can also cause damage to forests and biodi- For tradable goods (goods that are or could versity. be traded internationally) the reference price used * Water pollution. Water quality can be adversely is the world (or border) price, corrected for trans- affected by mine drainage, oil pollution, and acid port costs into or out of the country (depending deposition from excessive fuel use. on whether the country is an importer or ex- * Global effects. Fossil fuel combustion and trans- porter). port result in emission of greenhouse gases such For nontradable goods the reference price as carbon dioxide and methane. used is the long-run marginal cost of production. Safety concerns. Some forms of energy produc- The technical notes to this chapter provide de- tion and distribution pose large safety risks. tails on the actual procedures and data sources Different forms of energy harm the environ- tails on the actucale. procedures anddatasources ment in different ways, and to different degrees. Table 4.1 provides indicative levels of emissions Energy Subsidies associated with different types of fuels. Coal burning in thermal power stations and in domes- Subsidies for energy production or use have been tic stoves is the chief source of particulates and a common form of intervention in both industrial sulphur dioxide. Diesel trucks and buses are also and developing countries. In 1985, total fossil fuel responsible for high emissions, with biomass subsidies (for both developing and industrial combustion playing a role in some rural areas. countries) were estimated to be US$330 billion, Carbon monoxide, lead, and nitrous oxides are $130 billion of which came from the Soviet Union associated primarily with gasoline and diesel con- 44 Expanding the Measure of Wealth sumption. In general, natural gas is the cleanest than they would have been in a noninter- fossil fuel, while coal results in the highest levels ventionary situation. It is possible and common of emissions and pollution. Hydroelectric power for both types of subsidies to coexist, creating a does not produce emissions, but it can profoundly wedge between producer prices and consumer modify ecosystems affected by dam and reser- prices. In general, energy subsidies in develop- voir construction as well as aquatic life and riv- ing countries tend to favor consumers, while sub- erine ecosystems downstream; it may also force sidies in industrial countries tend to favor resettlement of people. The impact of energy use producers. on the environment is also affected by the nature Whatever form energy subsidies take, the re- and location of energy use as well as by the type sult is prices that fail to reflect the true economic of fuel. costs of supply. Low consumer prices result in Because the harmful effects of energy use are overuse, inefficient use, and wastage of energy. externalities from the perspective of individual High producer prices encourage excessive users, they are typically not incorporated into production and the operation of high-cost, consumption decisions, thus resulting in exces- uneconomic units that would otherwise be sive levels of damage. Given the existence of ex- tenliis it wol.sal edsrbet a uncompetitive. Production patterns become ternalities, it would usually be desirable to tax more energy- and capital-intensive and less energy use. Conversely, subsidies on energy use labor-intensive. Subsidies also tend to drain gov- exacerbate environmental problems. Removal of emient budgets, often resulting in higher taxes, subsidies, therefore, is only a first-albeit ., , . . ~~displacement of private investment by govern- crucial-step toward socially efficient energy use. ment borrowing, and higher external debt lev- Patterns of Energy Subsidies els, all of which can have negative effects on economic output and growth. Energy is subsidized in many ways, both direct The pattern of subsidization suggests that, in and indirect. Among the more explicit are direct most cases, energy use would fall substantially if grants and tax breaks to producers and distribu- subsidies were reduced. This is not always the tors. State-owned or state-managed companies case, however. are often heavily involved in the energy sector. * Removal of distortions may, in the long run, re- Price controls are common. Countries with exten- sult in incrieased production efficiency, higher sive energy resources often impose export restric- economic growth, and increased energy con- tions, thus keeping domestic prices artificially low. sumption. There is substantial potential for A distinction can be made between end-user cost-cutting in many subsidized coal sectors in or consumer subsidies and producer subsidies. developing countries, for example. Consumers are subsidized if consumer prices are * Because energy plays such an important part in kept lower than free-market levels, while produc- economic activity, subsidy removal is likely to ers are subsidized if producer prices are higher have substantial general equilibrium effects, making prediction of the impact of reforms dif- Table 4.1 Indicative emissions from different fossil fuels ficult. (tons of emissions per petwoule of energy content) Where inter-fuel substitution is possible, price Fuelype SO, TSP NOx CO, reductions may affect the composition rather Lignite 21-1,900 11-2,500 43-440 93-99,000 than the quantity of fuel use. Since the environ- Hard coal 14-1,300 11-2,500 43-440 85-91,000 mental effects of fuels differ, such changes in the Natural gas 1 2-3 45-290 38-40,000 composition of fuel use can affect the level of Liquid fuels 10-1,400 0.5-93 100-400 62-66,000 environmenl:al damage. Others 2,000 3,000 1,600 460,000 * Subsidy removal affects use and production pat- Note: These values are only indicative and vary according to the par- terns only to the extent that individuals and firms ticular characteristics of the fossil fuel, the use to which it is put, and respond to prices. In many transition economies the efficiency of energy conversion. this is still far from being the case, especially for I petajoule = I x IOE 15 joules. Source:IEA 1989; Homer 1993. state enterprises. Sutbsidy Policies and the Environment 45 Estimating Fossil Fuel Subsidies can be prohibitive in large countries, thus effec- in Developing Countries tively isolating much of their domestic markets from the world market. The last five years have seen significant changes * Much of the coal produced and used in Eastern in the magnitude of energy subsidies throughout Europe is brown coal, which has relatively low the world. To evaluate these changes a compari- heat content. Because of its high transport costs son of fossil fuel subsidies in 1991-92 and 1995- per unit of useful heat value, there is virtually 96 in a sample of nineteen developing countries no trade of this coal on world markets. and Russia was carried out (Rajkumar, 1996). The In a domestic market that is isolated from countries in the sample account for 77 percent of trade, prices in the absence of intervention would the world's total energy consumption outside the converge to the long-run marginal cost of pro- industrial world, and for 38 percent of global car- duction. At prices below marginal cost, higher- bon dioxide emissions (table 4.2). The sample also cost mines would shut down, thus reducing includes the countries known to be large energy production and forcing consumer prices to rise subsidizers. To ease comparability and take ad- until demand was met by the new level of out- vantage of recent improvements in available data, put. Thus, a gap between consumer prices and mar- 1990-91 subsidies were re-estimated using the ginal costs indicates a subsidy to both producers same methodology, assumptions, and data and consumers, and the size of the gap indicates sources. the size of the (combined) subsidy per unit. Estimates of subsidies are based on two ap- Limitations. Given the relatively poor quality proaches. Technical details are provided in the of available data, the estimates should be inter- appendix to this chapter. A full description of the preted as order of magnitude. There is also some methodology and the assumptions made can be pretedias ordervof m nitu e isalso some found in Rajkumar (1996), which also provides a subjectivity involved in the estimation of distri- foun in Rajkumar 1996), which also provides a bution costs (for natural gas in particular) and complete breakdown of subsidies by country, fuel long-run marginal production costs (for coal). type, and sector. ln-u aga rdco ot frca) However, since the same methodology and data Petroleum products and natural gas. Petroleum collection methods were used to derive estimates products and natural gas are fairly transportable for 1990-91 and for 1995-96, any bias is likely to and are widely traded. In the absence of govern- be similar for both sets of numbers. The estimated ment intervention domestic prices would con- change in subsidies and subsidy rates is likely to verge to world (or border) prices, corrected for be more reliable, therefore, than the actual levels transport and distribution costs. If domestic or rates themselves. prices are kept below border prices, consumers Transition economies. Additional caveats apply are being subsidized. Thegap between domestic and for the transition economies. The interpretation border prices indicates the size of the per-unit sub- of subsidies as encouraging increased use of an sidy, including the effect of both implicit and ex- input is based on the premise that prices act as plicit subsidies. Because subsidies on petroleum signals. This has not been the case in centrally products and natural gas in developing countries planned economies, and is still not always the benefit mostly consumers, measuring consumer case in the transition economies. This is particu- subsidies is equivalent to measuring total subsi- larly true in parts of the former Soviet Union, es- dies in most cases. pecially for natural gas and coal (box 4.2). The Coal. The same approach can, in principle, be estimates for these sectors should be considered used to measure coal subsidies (and is in fact used as indicative subsidies rather than actual subsi- in the case of China, India, and the Republic of dies as understood in the developed-country Korea). The characteristics of coal and its mar- sense, therefore. Continued movement toward kets often limit its applicability in some cases, market reform and enforcement of pricing mecha- however. nisms is required for reductions in subsidies to * The relative bulkiness of coal limits trade except be as meaningful in these economies as in mar- for certain grades. The cost of transporting coal ket economies. 46 Expanding the Measure of Wealth Table 4.2 Characteristics of countries in the sample Carbon dioxide emissions (million metric tones) /992 GNP per capita Population Solid Liquid (/995 dollars) (mi/lions) Energy intensity Country orgroup fuels fuels Gas 1994 /994 /993 Russia 607 683 758 2,650 148 2.06 Eastern and Central Europe Bulgaria 29 15 8 1,250 8 1.20 Czech Rep. 100 19 12 3,200 10 1.40 Hungary 20 23 16 3,840 10 0.89 Poland 283 37 16 2,410 39 1.11 Romania 39 35 43 1,270 23 1.43 Asia China 2,088 398 30 530 1,191 1.37 India 552 161 22 320 914 0.65 Korea, Rep. of 93 167 9 8,260 45 0.37 Thailand 16 73 15 2,410 58 0.36 Oil producers Egypt, Arab Rep. of 3 57 16 720 57 0.87 Indonesia 15 97 40 880 190 0.51 Iran, Islamic Rep. of 6 152 48 2,120 63 0.58 Mexico 13 251 53 4,180 89 0.39 Nigeria 0 37 9 280 108 0.43 Saudi Arabia 0 124 66 7,050 l8 0.70 Venezuela 1 57 47 2,760 21 0.80 Others Argentina 3 59 48 8,110 34 0.29 Brazil 40 154 8 2,970 159 0.21 South Africa 239 48 0 3,040 41 0.88 World 8,588 9,050 3,828 Note: Energy intensity is defined as total energy use divided by GDPF measured in tons of oil equivalent per thousand US$. Source: World Bank data; World Resources Institute 1996; IEA 1995. Finally, three more points should be noted. ucts therefore account for part of the measured * Between 1990 and 1995 total energy use declined change in sulbsidies. (When this effect is corrected by more than 25 percent in the former Soviet for, the qualitative results do not change substan- Union and by about 20 percent in Central and tially, as discussed in the technical notes.) Eastern Europe as a result of economic condi- * The estimates are for aggregate subsidies for each tions. Total subsidies would have declined, there- fuel type, for any particular country.They do not fore, even if the subsidy rates remained the same. cpueteefc fdsotosdet ifrn Changes in subsidy rates provide a better indi- tial pricing among sub- categories of fuels or Chatornofgrefr in s athese cnrvides tan btota i - among different types of users. Developing cator of reform in these countries than total sub- countries often price fuels consumed widely by sidies. households at lower levels relative to supply * Changes in measured subsidies depend partly costs than fuels consumed by industries, partly on movements in border prices; if domestic for political reasons and partly out of concern prices remain the same and border prices fall, for distributional consequences. Subsidies on the subsidy rate would appear to fall even in the kerosene, a fuel widely consumed by Asian absence of policy reforms. Recent declines in households, tend to be higher than subsidies for world prices for natural gas and petroleum prod- other petroleum products, for example. Subsidy Policies and the Environment 47 Box 4.2 Price reform problems As part of their transition from planned to in transition economies I market economies these countries have made . . ~~~~~~~~substantial moves toward subsidy reduction. Comparisons of subsidies in centrally planned and substantial o ves towardsbidy redcto transition economies with those in market economies While levels of itervention remai high, prces must be made with care, since prices can have very have generally been allowed to rise in both nomi- different meanings. nal and real terms. Subsidy reduction has been In centrally planned economies, particularly in especially dramatic in Russia, where subsidies on the Soviet Union, production and consumption were i determined by central planners. While prices existed and payments were made, they had no effect on de- declined by almost $20 billion over the period cisions; they were purely accounting prices. More- 1990-94 (Gurvich and others 1996). While some over, since any deficits were covered by the state, of the reduction in the transition economies can costs were also of little concemn. Similarly household I be attributed to the fall in real GDP they have energy use was often not metered-a problem that still persists-so consumers were equally insensitive experienced, subsidy rates have also fallen i to prices. practically all countries sampled and for all fuel As the transition to market economies proceeds, types. Despite these improvements much remains prices are becoming more meaningful. Currently, to be done. however, many elements of the old system still re- Many transition economies have large coal I main, and are hampering the effectiveness of sub- sectors. Since coal use is especially harmful to the Isidy reform. Bill collections for utilities are low, particularly environment (see table 4.1), reform in this sector in the former Soviet Union. In the Russian natural is particularly important. Employment consider- I gas sector, for example, it is estimated that 50 per- ations, however, have impeded prog-ress. The cent of bills were not paid in 1994, with state enter- Czech Republic, Hungary, and Poland have al- prises the main culprits. Moreover, many state ready taken significant steps toward reforming enterprises still operate on the basis of soft budget their coal sectors, while Bulgaria and Romania constraints, with deficits being covered by the state. In Ukraine official coal prices have been raised are further behind. In Russia and Ukraine, how- to the point that they are only about 10 percent be- ever, restructuring has hardly begun. According low official costs (official costs, however, comprise to recent World Bank studies the future viable mainly operating costs, grossly underestimating core of the industry in Russia will be only about capital costs). The effects of this have been limited, two-thirds its current size. In Ukraine at least a however, by nonpayment problems. Debt is grow- third of all coal mines need to be closed, and non- ing, with accumulating arrears in payments both by consumers to mines and by mines to employees. payment of bills, wages, and input prices is a par- ticularly severe problem (box 4.2). Although Central and Eastern Europe and Trends in Fossil Fuel Subsidies Russia account for the bulk of the estimated fos- sil fuel subsidy reduction between 1990-91 and The results of the study are summarized in table 1995, the results must be interpreted with care. 4.3 and figure 4.1. The results show a substantial Because of soft budget constraints in transition reduction in total subsidies since 1990-91. Total economies, the effect of subsidy reduction on in- fossil fuel subsidies fell from about $119 billion in centives for use remains unclear. If reforms con- 1990-91 to about $58 billion in 1995-96. Part of tinue and payment of bills is legally enforced, the this reduction, however, has resulted from declin- reduction in subsidies will result in increased ing energy use. Unit subsidy rates, shown in table energy efficiency and substantial reductions in 4.3, are thus more indicative of the extent of sub- environmental pollution (box 4.3). sidy reform. Average subsidy rates fell in all re- gions and for all fuel types. Total subsidies fell by China and India. Both these Asian giants have about $60 billion. significantly reduced subsidies since 1991. In China subsidy rates fell from 42 percent to 20 Central and Eastern Europe and Russia. Available percent and total subsidies fell from $24 billion evidence indicates that energy subsidies in Cen- to $10 billion, while in India subsidy rates fell tral and Eastern Europe and in the Soviet Union from 25 percent to 19 percent and total subsidies were high throughout the 1970s and 1980s. fell from $4.2 billion to $2.6 billion. 48 Expanding the Measure of Wealth Table 4.3 Subsidy rates and total subsidies for fossil fuels, 1990-91 and 1995i-96 Subsid rates {96 Totalsubsidies Petroleum products Naturalgas Coal Totals (1995 US$millions) (96 GDP) Country orgroup 1990-91/1995-96 1990-911995-96 1990-91 1995-96 1990-91 1995-96 1990-91 1995-96 /995-96 Russia, 38 16 67 42 48 25 45 31 28,797 9,427 1.50 Eastern Europe Bulgaria 43 24 27 23 65 33 54 29 2,003 733 7.05 Czech Rep. 0 39 29 29 28 24 22 1,173 978 2.96 Hungary 2 2 28 34 na na 13 16 548 560 1.47 Poland 28 1 40 6 63 26 50 18 4,653 1,692 1.97 Romania 27 2 64 54 80 33 54 37 4,743 1,876 7.24 Total 22 4 48 37 53 26 42 23 13,120 5,838 3.19 Asia China 55 2 na na 37 29 42 20 24,545 10,297 2.42 India 21 15 na na 32 27 25' 19 4,250 2,663 1.06 Korea, Rep. of 0 na na na na na 0 0 42 12 0.00 Thailand 10 9 na na na na 10 9 524 459 0.37 Total 30 5 na na 35 29 33 16 29,362 13,430 1.19 Oil producers Egypt, Arab Rep. of 50 33 70 56 na na 55 40 2,299 1,336 3.39 Indonesia 28 21 43 17 na na 29 21 2,071 1,333 0.92 Iran, Islamic Rep. of 88 77 82 75 na na 86 77 13,076 9,622 8.68 Mexico 32 12 31 39 na na 32 16 5,403 2,271 0.66 Nigeria 60 38 na na na na 60 38 928 592 1.87 Saudi Arabia 63 28 76 61 na na 66 34 3,837 1,720 1.42 Venezuela 70 55 86 85 na na 76 66 3,455 2,397 . 4.00 Total 54 38 64 59 na na 56 42 31,067 19,272 2.26 Others Argentina 4 3 24 3 na na 12 3 659 150 0.06 Brazil 26 0 na na na na 26 0 2,193 11 0.00 South Africa 8 6 na na 20 na 12 4 981 367 0.31 Total 16 3 24 3 20 na 17 2 3,833 528 0.06 OECD na na na na na na na na 12,453 9,890 0.05 Total 36 16 58 44 57 44 45 28 118,632 58,385 0.27 Note: A zero indicates less than I percent. a. Estimates for Russia are for 1990 and 1994 in PPP dollars and are taken from Gurvich and otihers 1996. See technical appendix for details of derivation. Source: Rajkumar 1996 for developing countries; IEA 1995 for OECD; Gurvich and others 1996i. Changes in real energy prices in China have prices have covered operating costs, with some occurred as part of the ongoing economic reforms allowance for capital depreciation. Furthermore, (box 4.4). Reforms began in the 1980s and have prices for higher grades of coal are no longer fixed continued into the 1990s. While most energy by the central government. prices are still set by the government, efforts are Because the pricing system does not suffi- made to base them on economic costs. Energy ciently differentiate between mines, however, sig- intensity in China has fallen by about 30 percent nificant cross-subsidization of higher-cost mines since 1984, which can be at least partly attributed by lower-cost ores remains, thus sustaining many to these reforms. India has also continued to reduce energy sub- uneconomic mines. Similarly, the degree of dif- sidies, albeit more slowly than China. All petro- ferentiation among grades of coal is low, result- leum products are at or above world prices, with ing in cross-subsidization among lower grades. the notable exception of kerosene, a fuel widely Removal of subsidies and market reforms may consumed by households. Coal pithead prices also promote coal-washing, which would reduce have also been raised, and since 1991 average emissions per unit of coal used. Suibsidy Policies and the Environment 49 Figure 4.1 Estimated changes in energy subsidies in selected countries, 1990-91 to 1995-96 Billions of 1995 US dollars 30 Coal 25 Natural gas 20 Petroleum I10 5 0 1990 1994 1990-91 1995-96 1990-91 1995-96 1990-91 1995-96 1990-91 1995-96 Russia Eastern Europe Asia Oil producers Others Note: See table 4.3 for countries in each group. Source: World Bank estimates. Other Asian countries. Asian countries gener- ers in the sample increased real prices for petro- ally experienced a reduction in subsidies in the leum products substantially. Prices more than early 1990s, although quantities consumed in- doubled in Iran and almost tripled in Saudi creased significantly as a result of high economic Arabia over the past two years, for example. Ven- growth. ezuela raised gasoline prices by more than 500 Oilproemain oil-producing coun- percent in April 1996, partly due to IMF pressure. Oil producers. The had oilyrhig en- Average subsidy rates for this group have now tries have historically had relatively high energy fallen from 56 percent to 42 percent. A greater subsidies. Prices were set adminstratively and concern for efficiency has also developed. Before kept below border prices andrsoetime even 1995, for example, Saudi Arabia maintained the below cost. These countries argued that energy .' should be priced at low levels for the benefit of price of diesel, a refined petroleum product, be- should~ ~ ~ ~ ~ ~ ~~~ ~o tha ofe fue oil aeel resdua rodct Asneatre- all, but particularly the poor (Kosmo 1987). Data low that of fuel oil, a residual product. As a re- sult diesel oil was often used instead of fuel oil to from a sample of oil exporters showed average energy subsidies in 1985 as being 5.6 percent of generate power, a gross waste of resources. This GDP (Kosmo 1987). These subsidies persisted imbalance has now been corrected, with diesel through the 1980s despite falling oil prices and oil prices rising from 3.2 cents to 11.2 cents per growing budgetary difficulties, with some excep- liter. (For comparison, the price of diesel to end tions, such as Indonesia, which eliminated gaso- users in the United States, before tax, was 18 cents line and fuel oil subsidies by 1985. In Mexico, in 1995). Venezuela, and to a lesser extent Egypt, nomi- Despite recent reductions subsidy rates among nal price increases were eroded by inflation. Real oil producers remain high compared with those domestic prices in the Islamic Republic of Iran in other countries. This group still has some of remained among the lowest in the Middle East, the largest subsidizers in the world. The rate of even after it became a net importer of petroleum reform has also been uneven. Natural gas subsi- in the 1980s during the war with Iraq; in 1991 its dies are still high, with little upward movement total subsidies amounted to 9 percent of GDP. in real prices (except in Indonesia). Most of these Domestic consumption responded strongly to countries plan to reduce subsidies further, how- the highly subsidized prices, growing almost ever. Iran, for example, has stated that it intends twice as fast as real GDP on average. to eliminate petroleum product subsidies within A combination of budget problems, high five years. debt-service ratios, and the desire to remain net exporters caused a change in attitudes toward Latin America. In LatinAmerican countries as energy subsidies in the 1990s. All the oil produc- a whole real domestic prices for fossil fuels fell 50 Expanding the Measutre of Wealth Box 4.3 Energy in Russia: what lies ahead? Russia has dramatically reduced its energy subsidies, Box figure 4.3 Projected change in energy and there is potential for further reduction. Most ef- consumption and emissions in Russia, 1997-2010 fects of these reductions are still to be felt, however. The World Bank, in collaboration with Russian ex- Change relative to 1990 perts, has developed a model to forecast energy con- Percent sumption and emissions over the next fifteen years. Pc These forecasts incorporate a number of assumptions, including annual GDP growth of 5 percent after 2000, o a long-term recovery of the Russian coal sector, and 10 reduced reliance on nuclear power. According to these forecasts, complete subsidy -20 elimination would reduce energy consumption by _ _ more than 10 percent relative to its 1990 level, whereas maintaining prices at their 1990 levels would result in _40 l High subsidies (1990 prices) slightly increased consumption. With complete sub- U - Eliminating subsidies sidy removal, in 2010 emission of particulates would 50 be 40 percent lower and emission of carbon dioxide 14 -60 percent lower than they would be in the absence of price reform after 1990. Complete subsidy removal 70 Energy Total Sulfur Nitrogen Carbon would reduce global carbon dioxide emissions by use suspended dioxide oxide dioxide about 290 million tonnes, or about 1.3 percent of cur- partculates rent world emissions. Source: World Bank projections. slightly in the 1980s as a result of high inflation In most of Sub-Saharan Africa energy subsidies rates in many of the region's countries, with es- have not been important-indeed, gasoline, die- pecially large falls in ultra-inflationary Brazil. sel, and, to a. lesser extent, kerosene have often Argentina was a notable exception; despite gal- been explicit]y taxed for revenue purposes. South loping inflation, nominal prices outpaced infla- Africa began a process of deregulation in the late tion, so that by 1990 only kerosene and light fuel 1980s, privatizing portions of the coal sector and oil remained subsidized. In the 1990s a stabiliz- reducing subsidies to the synthetic petroleum ing of inflation rates was accompanied by a rise product sector. The industry is now completely in real fuel prices. Brazil experienced a strong rise deregulated and privatized, and foreign owner- in real petroleum product prices. ship is permitted. Subsidies persist for synthetic petroleum product manufacturing, in which coal Other developing countries. In the rest of the is an input, but they are due to be phased out by developing world price movements were mixed. the end of 1998. Box 4.4 China: subsidy reform in the coal sector China has made remarkable progress in reforming en- 1996). Price controls on coal from state mines have also ergy price over the last five years, with fossil fuel sub- been partially lifted, so that less than 20 percent of all sidies falling from $24 billion a year $10 billion. This coal is now sol.d at controlled prices. Moreover, the reform continues a process begun in the 1980s. problems of nonpayment of bills, which have plagued The steady reduction in subsidies to the huge coal the coal sectors in the former Soviet Union, have been sector, which accounts for 73 percent of China's com- overcome in China. A strict contract system has been mercial energy needs, has been particularly impres- in sive. Coal subsidy rates fell from 61 percent in 1984 oeration since 1994 whereb state mines only sup- (Kosmo 1987) to 29 percent in 1995. These reforms have ply coal that has been paid for in advance. been achieved by allowing the gradual entry into the Available data indicate that China's reforms have sector of collective and individual mines (operating had very significant effects on energy efficiency. En- alongside state rnines). Currently about 50 percent of ergy intensity in China-once among the highest in production is carried out by these private mines (Wang the world-has fallen by about 30 percent since 1985. Source: Kosmo 1987; Wang 1986. Subsidy Policies and the Environment 51 Box 4.5 Electricity subsidies About 34 percent of the world's commercial energy pro- subsidies further. These high power subsidies encour- duction is used to generate electricity and heat; about age overuse of power and consequently of the inputs two-thirds of this comes from the combustion of fossil to power, resulting in high emissions levels. To make fuels. The fossil fuel and power sectors are thus matters worse, coal is commonly used to generate interlinked, and policies affecting one will affect the other. power in the former Soviet Union, Central and East- TIhis has special relevance in the context of subsi- em Europe, China, and India. Moreover, this coal is dies to the power sector. Heat and electricity can be often of low quality and has a high sulphur content. subsidized in a number of ways. Capital inputs to the Over the past few years, however, electricity prices sector can be provided below cost; electricity prices can have risen significantly in real terms in most countries. be kept low, with the state bearing the resulting losses; Reductions in fossil fuel subsidies have played an im- and fossil fuels can be provided to the power sector at portant role in this increase. Other factors include re- subsidized prices. duced subsidies to other inputs and a reduced Power subsidies are often high in developing coun- willingness to use state funds to offset power compa- tries. Total electricity subsidies in the early 1990s have nies' losses. This is encouraging from an environmen- been estimated at about $54 billion, or about 3.6 per- tal perspective, although the effects will be felt only in cent of GNP on average for developing countries; add- the long run and are dependent on a continued com- ing heat subsidies would increase estimated power mitment to market reform. OECD countries. Among OECD countries, en- indicate that the effect of prices on energy effi- ergy subsidies are primarily targeted to produc- ciency is likely to be substantial, although it may ers rather than consumers. Despite some notbeimmediate.Energyintensityfellby38per- downsizing of coal mines, total subsidies re- cent in OECD countries between 1971 and 1988. mained high in Germany, Japan, and Spain. The For developing countries various researchers United Kingdom, however, adopted a vigorous have found that, in the long run, energy demand coal restructuring program in the late 1980s, is likely to fall by half a percentage point for each which reversed the steep rise in subsidies it ex- percentage point price rise (Anderson 1995), al- perienced in the early 1980s. According to though individual estimates vary from country estimates for 1993, total energy subsidies to pro- to country (see, for example, Hope and Singh ducers amounted to $9.5 billion, with Germany 1995; Julius 1986; Tybout and Moss 1992; accounting for 70 percent of this total (IEA 1995). Eskeland and others 1994). As a whole, therefore, industrial countries have Because coal is the dirtiest of the fossil fuels, been less successful at reducing energy subsidies the greatest environmental benefits would result than developing countries in recent years. from coal subsidy reforms. Eliminating coal sub- sidies in OECD countries alone would result in a Effects of Subsidy Reform on Energy Use reduction of global carbon dioxide emissions of and the Environment about 1.5 percent (DOE 1989; Okugu and Birol There is a clear negative relationship between en- 1992), while eliminating coal subsidies in Russia hergy prics an cleanergyaintivensitiohichpis b n s e would reduce carbon dioxide emissions by 0.4 ergy prices and energy intensity which is a simple percent. indicator of energy efficiency. Energy intensity is In some countries the possibility that remov- defined as the ratio of energy use to GDP evalu- ing subsidies on fossil fuel prices could cause con- ated at purchasing power parity GDP. Prices sumers to use fuelwood instead provides a caveat explain roughly half of the differences in to the expected environmental benefits. Increases cross-country energy intensities, the other half in fuelwood consumption would accelerate de- being attributable to differences in climate, income forestation, exacerbating problems of land deg- level, and the composition of final output (Kosmo radation and habitat loss. However, fuelwood is 1987). a possible substitute for other fuels in a narrow Because the process of subsidy reform is a re- range of uses-primarily household consumption cent one it is too early to detect any resulting re- for heating and cooking. Furthermore, the scope duction in environmental damages. Studies of for substantial increases in fuelwood consump- energy use following the oil shock of the i970s tion by urban households would be limited by 52 Expanding the Measure of Wealth space and transport cost considerations. A study discussed first, followed by subsidies on fertiliz- of Indonesia found that in most areas of the coun- ers and subsidies on irrigation water. try consumers' substitutability between kerosene and fuelwood was almost zero (Pitt 1985). Pesticide Subsidies Energy subsidies are often claimed to have Until recently pesticides were considered an im- positive redistributional effects. If this is true, any I environmental benefits derived from subsidy re- portant component of improved farming prac- duction might be offset by increases in poverty, tices. In an effort to encourage farmers to adopt Studies have shown, however, that the benefits them, subsidies for pesticide use were widely of energy subsidies tend to accrue disproportion- adopted among developing countries. Subsidies oftenery soricherubsders. In Poland, forexample,the were both direct and indirect. Partly as a result of ately to richer users. In Poland, for example, the teesbiispsiieuehsicesdsgii wealthy spend a larger fraction of their income these subsdes pescide use has increased signifi- on energy consumption than the poor, indicat- cantly in the past decades. In Indonesia, for ex- ing that subsidies would be regressive (Freund ample, pesticide use increased by more than one and Wallich 1995). Targeting cash relief to the and a half times between 1975 and 1985. poor through social assistance programs is pref- Pesticide use and the environment. There is sub- erable to energy subsidies. Another possible so- stantial evidence that pesticide use can pose im- lution would be to provide vouchers, transfers, portant threats to the environment and public or lifeline pricing for a small block of a commod- health. Common problems include groundwater ity such as natural gas. contamination, public health problems, losses of Conclusions nontarget crops and other species, and pesticide resistance. An estimated 10,000 people die and There have been substantial reductions in energy 400,000 suffer from acute pesticide poisoning an- subsidies in recent years. Although it is still too nually in developing countries (Repetto 1985). In soon for most of the efficiency and environmen- the Philippines prolonged exposure to pesticides tal benefits of these reforms to be experienced, it has been found to cause chronic health problems is likely that both will be significant. However, a for farmers, including eye, dermal, pulmonary, significant number of problems need to be ad- neurologic, and kidney problems (Antle and dressed in transition economies to ensure that Pingali 1995). Pesticide residues on food can also these reforms achieve these goals. Needs for sub- have adverse effects on the health of consumers. sidy reform also remain in the OECD countries. Pesticide policies. Many developing countries Agricultural Input Subsidies heavily subsidized pesticide use in the 1970s and 1980s.According to a nine-country study (Repetto Subsidies for agro-chemicals have been a frequent 1985), subsidy rates in the early 1980s ranged from form of intervention in developing countries. The 20 to 80 percent. Several countries, including use of agro-chemicals, including fertilizers and Egypt, Indonesia, and Senegal had subsidy rates pesticides, has grown rapidly in recent decades. of more than 80 percent. Despite the high level of National efforts to promote them, especially subsidy rates, however, expenditures on explicit through subsidies, have played an important role subsidies tended to be relatively small compared in that increase. Because subsidies have made to expenditures on fertilizer subsidies. In Indo- them appear cheaper to producers, agricultural nesia annual expenditures on explicit pesticide inputs have been used more heavily than they subsidies during the 1980s were about $120 mil- would have been in the absence of subsidies. This lion, compared with about $500 million on fertil- has resulted in inefficiency and has increased the izer subsidies. [n India annual pesticide subsidies extent of environmental damage caused by use were only $35 million, compared with almost of these inputs. This section examines subsidies $1,500 million for fertilizer subsidies. on agricultural inputs, their changes over time, Since the mid-1980s the extent of pesticide sub- and their implications for agricultural production sidies has been substantially reduced in many and the environment. Subsidies on pesticides are countries. Bangladesh was one of the first to elimi- Subsidy Policies and the Environment 53 nate pesticide subsidies; fiscal problems forced it In many cases pesticide use has been replaced to do so in the mid-1970s (box 4.6). Indonesia by integrated pest management (IPM) tech- eliminated most pesticide subsidies between 1986 niques. IPM techniques attempt to manipulate the and 1988 (box 4.7). The Philippines followed suit crop environment by taking advantage of its in 1988 and India in 1992. biodiversity to reduce the chances of damage. The Many African governments also moved to re- actual techniques vary depending on the crop and duce pesticide subsidies in the late 1980s, often the ecosystem, but they typically include use of in the context of structural adjustment programs. insect-resistant varieties, changes to planting Pesticide subsidies had often been provided to times and other operations to exploit insect life producers of export crops to stimulate produc- cycles, encouragement (or introduction) of natu- tion. Falling world prices for these crops was an- ral enemies (biocontrol agents), and mixed crop- other factor that often led to subsidy reductions ping. Pesticide use, if any, is limited to reacting (Farah 1994). Togo, for example, phased out its to particularly high levels of infestation, rather pesticide subsidy rates of 100 percent for cotton than prophylactic applications. A number of producers beginning in 1988 (Kiss and Meerman countries have made ,PM a national policy, in- 1991). Similarly, Cameroon abruptly reduced pes- cluding Bhutan, China, Indonesia, Madascar, Lao ticide subsidy rates of more than 100 percent for Peoples Democratic Republic, and Vietnam, cocoa producers following the collapse of the co- (Fleischer and Waibel 1993). The World Bank has cooa eproducers marke ing 1990 collapse 1994). the been a strong supporter of IPM since the early coa export market in 1990 (Farah 1994). 19S The trend toward removal of pesticide subsi- dies has also been observed in transition econo- Remaining problems. Substantial strides have mies of Eastern and Central Europe and the been made toward reducing subsidies on pesti- former Soviet Union. For example, Poland com- cides. Many countries have eliminated explicit pletely abolished pesticide subsidies (which rep- pesticide subsidies, including Bangladesh, Egypt, resented 1.1 percent of total agro-food-related Ghana, India, Indonesia, Pakistan, Vietnam, and expenditures) in 1991 (OECD 1995). In Latin most LatinAmerican countries. Despite these im- America explicit subsidies have also largely been provements, many countries stllf implement poli- phased out. Indeed, in several countries domes- cies that directly or indirectly favor pesticide use tic pesticide prices are now higher than border and impede the adoption of ecologically sound prices (Valdes and Schaeffer 1995a, 1995b). pest control techniques (table 4.4). Nevertheless, a variety of direct and indirect * Through pest surveillance systems, outbreak subsidies to pesticide use remain in many coun- budgets, and calendar spraying, governments tries. Table 4.4 summarizes pesticide-related poli- still subsidize pesticide use indirectly and exert cies in a number of countries in 1993. strong influences on farners' decisions related Although pesticides were once seen as a criti- to pesticide applications. cal component of intensive agricultural produc- * Agricultural credit requirements often conflict tion, a growing consensus is emerging that they with efforts to reduce pesticide use as credit is tion,oa growing consensusta isiemergrogett tey aoften conditioned on use of specified amounts may not only play just a minor role but may ac- of pesticides (Thrupp 1990). In Costa Rica, for tually be detrimental. Several studies have shown example, farmers had to prove they used pesti- that reducing pesticide use does not necessarily cides to qualify for agricultural loans. In addi- result in decreasing agricultural production. In tion, some countries provide subsidized credit Indonesia substantial reductions in pesticide use for pesticide purchases. In C6te d'Ivoire, for ex- beginning in the late 1980s have been accompa- ample, credit for pesticide purchases is available nied by an increase in rice production (see box at rates of 4 to 5 percent below market rates 4.7). In cases where productivity does decline be- (Adesina 1994). *To encourage pesticide use some importing cause of reduced pest control, the costs of lower courage lertriffse so p oetic - countries charge lower tariffs on pesticide im- yields are often largely offset by the gain result- ports. Import tariffs for pesticides in Thailand, ing from improved farmer health (Antle and for example, are only 5 percent, compared with Pingali 1995). about 20 percent for machinery (Waibel 1990). 54 Expanding the Measure of Wealth Box 4.6 Removing pesticide subsidies in Bangladesh As in other green revolution countries pesticide use use remains low by regional standards and only 10 to was an integral part of the high-yield variety technol- 20 percent of the area planted to high-yield rice is ogy package adopted by Bangladesh, which called for treated. The perceived high cost of pesticides has made three scheduled pesticide applications. Use was fur- farmers use them in a purely reactive way: pesticides ther encouraged by subsidies. As a result pesticide use are applied only when pest infestations are detected grew rapidly in the early 1970s. (Pagiola 1995). Fiscal problems led to the subsidies being halved Groundwater tests during 1994-95 in areas with a in 1973-74, however, and then removed entirely in high potential for pesticide contamination found traces 1978. Pesticide use fell dramatically in the wake of the in only 13 percent of samples, and most were longer- reduction in subsidies. The treated area fell from 3 to lived organo-cilorines used in the past rather than the 5 million hectares in 1971-74 to less than 1 million hect- moderately persistent and less toxic organo-phos- ares two years later, and then further to around 0.5 phates in current use (Pagiola 1995). The pollution million hectares after the complete elimintation of sub- record of other countries in the region that continued sidies (Duloy and Nicholas 1991). to subsidize pesticide use heavily well into the 1980s Although total pesticide use has since recovered and makes it clear that, for Bangladesh, removing pesti- now exceeds the levels of the early 1970s, per hectare cide subsidies was clearly a win-win reform. Source: Pagiola 1995. Countries with pesticide production often have million metric tonnes in 1960 to 146 million met- low tariff rates on imported active ingredients ric tonnes in :[988 (figure 4.2). Growth has been and high rates on imported products. These dis- especially rapid in the developing world, par- criminatory trade policies result in inefficiency ticularly in Asia. Fertilizers have been one of in pesticide production and use. the driving forces in the growth of agricultural Many donors and international agencies have output. Only in recent years has the growth in become very cautious in providing pesticide as- world fertilizer consumption abated, largely as sistance, and some have strongly supported IPM a result of reduced consumption in Eastern Eu- approaches. Nevertheless, problems remain with rope and the former Soviet Union. donations of pesticides from a few countries such as Japan according to Matteson and Meltzer Fertilizer use and the environment. The envi- (1995). Donated pesticides are often sold to farm- ronmental consequences of fertilizer use are ers at heavily discounted prices. much more ambiguous than those of pesticides. Eliminating subsidies is only a first step in en- While excessive or inappropriate fertilizer use suring that pesticide use does not exceed socially can lead to environmental damage, fertilizers optimal levels. Given the important negative ex- properly usedL can have beneficial environmen- temalities often associated with pesticide use, the tal effects. Indeed, failure to use fertilizers can principleoftaxingpesticidesisbecomingincreas- often have adverse environmental conse- ingly accepted. Several OECD countries have lev- quences. ied taxes on pesticide use. Denmark raised pesticide taxes to as much as 37 percent of the Positive environmental effects. Without fertil- retail price in 1996. Sweden has coupled high izers sustained high-yield agriculture would taxes on pesticides with subsidies to help farm- not be possibl]e. Low use of fertilizer can have ers switch to pesticide-free farming techniques, two kinds of adverse environmental effects. No developing country government has yet un- First, it can result in degradation of cultivated dertaken the next logical step and taxed pesti- fields, and perhaps ultimately to their abandon- cide use. Indeed, a number of countries, such as ment. Second, a much larger area would be re- the Philippines, waive sales tax on pesticides. Ippmes, waiv ~~quired to pr-oduce any given amount of Fertilizer Subsidies agricultural products. This would accelerate the already rapid conversion of natural habi- Total worldwide use of fertilizers measured in tat to agricultural use, which is thought to be terms of the three main nutrients-nitrogen (N), the driving force behind much current habitat phosphorus (P), and potassium (K)-rose from 27 and species loss (Pagiola and others 1997). Subsidy Policies and the Environment 55 Box 4.7 Removing pesticide subsidies in Indonesia Agriculture employs 50 percent of the labor force in Box figure 4.7 Effect of removing pesticide Indonesia and accounts for about 20 percent of GDP. subsidies in Indonesia Rice is the most important agricultural product. Rice self-sufficiency has been the focus of agricultural policy Milled rice production since the 1960s. To increase rice output the government Million metric tons has aggressively promoted the use of high-yielding 40 varieties (HYVs) and the associated use of fertilizers and pesticides. As a result the area planted to HYVs 20 increased to more than 90 percent in 1990 and rice pro- duction doubled between 1970 and 1990. After having been the world's largest rice importer for years the 0 1 country attained self- sufficiency in rice in 1985. 1981 1984 1987 1990 1994 Prior to 1986 pesticides were an integral part of gov- Pesticide production emment efforts to boost agricultural production. Pes- Million mpetric tons ticide use was promoted in many ways, including 60 direct subsidies on pesticide sales, government spray- so ing, and favorable credit packages. The pesticide sub- sidy rate was about 85 percent in the early 1980s (Pincus 40 1994). Expenditures on pesticide subsidies reached 151 30 _ _ billion rupiah ($126 million) in the 1986-87 fiscal year. 20 As a result of these policies domestic pesticide pro- 1981 1984 1987 1990 1994 duction soared from 6,000 metric tons in 1972 to 53,100 metric tons in 1985 (Pincus 1994) and substantial addi- Pesticide subsidy rate tional quantities were imported. Percent Massive use of pesticides caused considerable 100 harm. Pesticide pollution was a major cause for con- 80 cern in Indonesia's densely populated village commu- 60 nities, particularly where water for drinking and 40 I bathing was in limited supply. Problems were further 20 compounded by the disposal of hazardous wastes pro- 0 duced by the pesticide industry. Ironically, heavy pes- 1981 1984 1987 1990 1994 ticide use also came to undermine the very purpose Source: World Bank data; Pincus 1994; Pearson 1991. for which it had been promoted: stabilizing agricul- tural production. Targeted pests soon developed re- sistance to pesticides, leading to widespread pest metric tons in 1990, while imports fell to a third of the resurgence problems. In addition pests such as the mid-1980s levels. This reduction in use has been ac- brown planthopper, which had not previously been complished without adverse effects on rice production. considered a problem, became serious threats after Total milled rice production rose from 27 million met- overuse of pesticides killed their natural enemies. ric tons in 1986 to 30 million metric tons in 1990. Al- However, as the applications of insecticides went up, though no data exist to quantify this effect, the so did the brown planthopper infestation. A brown reduction of pesticide use is also thought to have alle- planthopper outbreak in 1976 resulted in the loss of a viated damage to the environment-particularly to million tons of rice, an amount sufficient to feed more public health and to biological diversity. than 2.5 million people (Kenmore 1991). Despite the improvements problems have not been These problems led Indonesia to drastically modify entirely eliminated. Several indirect subsidies remain- its pesticide policies in 1986 (see box figure 4.7). Many for example, through the government's emergency out- pesticides were banned, and integrated pest manage- break budget, which provides govemment sprays or ment (IPM) was officially adopted as a national policy. free pesticides in response to reported pest outbreaks. All direct subsidies on pesticides were eliminated by Many credit packages still require pesticide use. Vari- 1989, and the extent of indirect subsidies was substan- ous measures to protect the domestic pesticide indus- tially reduced. As a result of these changes pesticide try also remain in place, and enforcement of the use fell dramatically. Production dropped to 22,100 pesticide ban is weak. Adverse environmental effects. Fertilizers can plications are taken up by crops, meaning that have adverse environmental effects if they leave substantial quantities of nutrients escape into the the fields on which they are applied. By some environment. This can at times result in health measures only about 50 percent of fertilizer ap- problems. Nitrates leached from fertilizer can be 56 Expanding the Measure of Wealth Table 4.4 Pesticide policies in selected countries, 1993 Favorable Sales below Subsidized exchange Preferential Outbreak Aid-in-kind Adoption Country market price credit rate tariff rate budget donations of lPM Asia Bhutan Yes Yes No Yes Yes Yes Yes China No n/a No No Yes Yes Yes Indonesia No No No No Yes No Yes Laos Yes Yes No Yes No Yes Yes Malaysia No No No No Yes No Yes Nepal No No n/a Yes Yes Yes Yes Sri Lanka No Yes Yes No No No No Vietnam No No n/a Yes Yes No Yes Africa Cameroon Yes Yes No Yes No Yes No Egypt No Yes Yes Yes Yes No No Ghana No Yes No Yes Yes Yes Yes Kenya No Yes No No Yes No No Madagascar No Yes n/a Yes Yes Yes Yes Malawi No Yes n/a n/a Yes Yes No Nigeria No Yes No Yes Yes No No Sudan Yes Yes No No Yes Yes No Tanzania No No No n/a No Yes No Latin America Brazil No No n/a No No No No Colombia No No No No Yes No No Costa Rica No Yes n/a Yes Yes n/a n/a Cuba No Yes Yes n/a Yes No Yes Peru No No No Yes No No No Source: Fleischer and Waibel, 1993. carcinogenic, for example, and can cause blue application. In many rice-producing countries baby syndrome. Unlike pesticides, however, the shortages of rnicronutrients such as sulphur and primary concern is not over possible health ef- zinc have become important constraints to yield fects but over the ecological consequences of in- growth (Pagiola 1995). Use of some fertilizers can troducing large quantities of nutrients into increase the level of cadmium and other harmful ecosystems. This can result in degradation of wa- heavy metals in agricultural soils, polluting agri- ter quality and eutrophication. Phosphorus-laden cultural products and posing a danger to public runoff from agricultural areas in central Florida, health. for example, are resulting in extensive damage to the ecosystem of the Everglades. Because the Fertilizer policies. The importance of fertilizers Everglades are naturally very low in phospho- to agricultural production has made promotion rus, most of its biota are very sensitive to phos- of fertilizer use an important aspect of national phorus additions. Fertilizer use also contributes policy in many countries. Almost all developing to climate change problems, since gaseous nitro- countries have, at various times and to different gen losses add to total emissions of nitrous ox- degrees, subsidized fertilizers. Subsidies have ides. been widely used to stimulate increased fertilizer use and thereby bring about increased crop yields. Adverse productivity effects. Although fertilizer Fertilizer subsidies were considered particularly use increases yields, misuse can also have adverse important in inducing farmers to adopt improved impacts, in the long term, on soil productivity by practices, which often depended heavily on fer- increasing the soil acidity and reducing the use tilizers. Subsidies appear to have been success- of organic manures and crop residues. Unbal- ful in this regard: between 1971 and 1980 fertilizer anced fertilizer application can lead to excessive use increased more rapidly in countries with sub- depletion of certain nutrients; micronutrients, for sidies than in countries without subsidies example, have often been neglected in fertilizer (Couston 1985). Suibsidy Policies and the Environment 57 Trends in fertilizer policies. Although the lack rapid growth of fertilizer use resulted in very high of data for many countries prevents calculation budgetary outlays (table 4.6). For example, In- of total subsidies at the global level, the trend in donesia spent $1 billion on explicit fertilizer sub- total subsidies can be discerned. Fertilizer subsi- sidies in 1984, Nigeria spent $270 million (32 dies increased rapidly during the 1960s and 1970s, percent of its agricultural budget) in 1985, and peaked in the late 1970s and early 1980s, and then India spent more than $2.8 billion in 1988. started to diminish (table 4.5). Fertilizer subsidies in many countries began In the early 1970s explicit subsidy rates for fer- to experience a reduction in the 1980s (Maene tilizers usually ranged from 25 to 50 percent of 1994; Isherwood 1996).Average unit subsidies on unsubsidized retail costs (Repetto 1988). For ex- urea decreased in many countries (see table 4.5). ample, subsidy rates for urea in 1968-72 were 56 Budgetary expenditures also tended to decline percent in Bangladesh (box 4.8), 50 percent in (see table 4.6). In Indonesia total subsidies fell Chile, 36 percent in Indonesia, 50 percent in Ni- substantially. Several countries, including geria, 47 percent in Senegal, and 81 percent in Bangladesh, the Philippines, Sri Lanka, and most Uruguay. Many countries also provided addi- Latin American countries, have eliminated ex- tional indirect subsidies through favorable ex- plicit fertilizer subsidies. This declining trend con- change rates, tariffs, and foreign exchange tinues today, although it has levelled off in many allocations, preferential domestic taxation, and countries. Iran is one of the few exceptions to this subsidized credit. For example, credit for fertil- trend, having experienced an increase in total fer- izer purchases was interest-free in Brazil in the tilizer subsidies. late 1970s; in Nicaragua fertilizer imports were In the centrally planned economies of the East- subject to extremely favorable exchange rates. ern Bloc, fertilizers were allocated rather than Even though the goal of encouraging adop- sold, and were heavily subsidized. With the tran- tion of fertilizer use had long since been achieved sition to a market economy fertilizer subsidies in in most countries, subsidies remained high these countries have been reduced or, in many through the early 1980s. Subsidy rates of 50 to 70 cases, eliminated (Maene 1994). In OECD coun- percent for fertilizers were common in the early tries explicit subsidies on fertilizers have been 1980s (World Bank 1986). In 1983-84, for example, rare; producer subsidies have generally taken the explicit subsidies for nitrogenous fertilizers were form of output rather than input subsidies. In fact, 36 percent in Burkina Faso, 61 percent in Iran, several countries are taxing fertilizer use (Maene 33 percent in Pakistan, and 60 percent in Togo 1994). (Repetto 1988). High subsidies coupled with A variety of factors have prompted countries to reduce or curtail fertilizer subsidies: Figure 4.2 Trends in fertilizer consumption * Their financial burden was a major factor in Million metric tons many African and Latin American countries. 160 n ' The need for improved efficiency has often been Latin America Eastern Europe, \ ^ a motivation. Because of subsidies, fertilizers 140 former Soviet Union tended to be overused and misused. Moreover, 120 Afiathe mechanisms required to i-mplement subsi- 100 Cther dized input policies-often based on govern- \00 \ - ' -ment or parastatal distribution systems-tended 80 \ to become increasingly bloated and inefficient Organization l.m t: over time. 60, Cooperation aMn . o 60 ^ Development International organizations, including the 40 - P f World Bank and the IMF, and international do- Asia nors have long advocated the removal of subsi- 20 KA_ia dies to correct price distortions and improve 0 1 8 production efficiency. Subsidy reform has been a 1961 1971 1981 1993 major component of structural adjustment Source: FAO data. programs. 58 Expanding the Measure of Wealth The conditions created by the transition to Table 4.6 Average annual government expenditures on market economies and the lack of financial re- explicit fertilizer subsidies in selected Asian countries, sources were important factors in the centrally 1982-94 planned economies of the former Soviet Union (mon 95 US$) and Eastern Europe. Country 1982-84 1985-97 1988-90 1991-93 1994 Recent trends in fertilizer subsidies have not been uniformly downward, however. Fluctua- Bangladesh 56 21 68 19 0 tions and even increases in subsidies have been India 1,194 2,006 2,833 2,010 1,685 observed. Many countries still provide direct or Indonesia 732 530 515 333 96 indirect subsidies. Governments have resisted Iran 478 432 505 658 n/a removing subsidies for various reasons, includ- South Korea 106 387 15 n/a n/a ing economic incentives, protectionism, social Nepal 9 6 13 17 n/a equity, or pressure by special-interest groups. In Pakistan 178 156 102 33 2 Nigeria fertilizer subsidies remain the single most Philippines 48 46 20 0 n/a expensive agricultural program, accounting for SriLanka 64 44 12 0 n/a 31 percent of the government's budget allocation Thailand 5 3 3 n/a n/a to agriculture. Fertilizer subsidies are reportedly Source: FADINAF' database, except Indonesia 1991-94 from Indone- being reintroduced in several countries of Cen- sia Center for Policy and Implementation Studies. tral Europe and the former Soviet Union. Environmental effects of subsidy reform. Fertil- Table 4.5 Explicit subsidies on urea in selected Eruerovdes cla oductivy benefita countries, 1971-90 izer use provides clear productvity benefits and, (/995 US$ per metric ton of urea) in many developing countries, often causes only Country 1971-75 /976-80 /98/.85 1986.90 relatively low levels of environmental damage. Asia Because of this, the potential for win-win reforms Bangladesh 90 99 17 19 is less marked than that for pesticides. The ex- tent to which current subsidy reforms will result Indonesia n/a n/a 51 66 in environmental improvements is, therefore, dif- Iran 411 294 239 101 ficult to predict. The environmental problems as- Pakistan 108 69 77 n/a sociated with overuse and misuse of fertilizers Oman 335 n/a n/a 52 have seldom been systematically studied. Even Turkey 164 n/a n/a 55 where data exist, they are generally insufficient Africa to determine the long-term impacts of changes Burkina Faso 253 177 136 12 in fertilizer use. Nigeria n/a 60 59 n/a From the efficiency perspective the need to South Africa 38 20 n/a n/a overcome farmers' risk aversion has often been Togo 367 299 200 99 cited as justifying subsidies. Infant-industry ar- Tunisia(*) 38 132 181 100 guments have also been made for protection of domestic fertilizer industries. Whatever merit South America these arguments may once have had, they are Venezuela n/a 152 n/a 188 clearly no longer applicable. After decades of ex- Eastern Europe perience, fertilizers have been widely accepted Hungary 57 57 112 67 by farmers andi the fertilizer industry is a mature OECD sector in many countries. Poverty considerations Australia 90 42 12 n/a are also seldom a factor, since fertilizer subsidies New Zealand 84 73 16 n/a are an extremely inefficient way of aiding poor Portugal n/a 197 308 n/a farmers, who account for only a fraction of total Spain 106 95 98 52 fertilizer use. Notes. Because of missing observations in the FAO database, some Irrigation Water Subsidies values shown are for slightly different periods. The FAO ceased col- lecting these data in 1990. (*) Data for subsidies on ammonium nitrate shown. The area irrigat:ed worldwide grew by about 2 per- Source: FAO database. cent a year during the 1960s and 1970s, and by Subsidy Policies and the Environment 59 Box 4.8 Reducing fertilizer subsidies in Bangladesh Bangladesh is one of the poorest and most populous types of subsidies existed in this period: a direct price countries in the world. Agriculture is a domninant sec- subsidy when border prices exceeded government tor, accounting-for 38 percent of GDP and employing sales prices, an indirect production subsidy when ex- 70 percent of the workforce. In the past decade agri- factory prices exceeded border prices, an indirect dis- culture has grown by about 2 percent annually thanks tribution subsidy when distribution costs exceeded to widespread adoption of improved, intensified crop- private sector costs, and a production subsidy through ping practices. Increased fertilizer use has been a ma- low-price sales of natural gas to urea producers. jor component of this growth. Between 1971 and 1993 Although subsidies contributed to the promotion total fertilizer use increased from 0.3 to 2.3 million met- of fertilizer use, they imposed heavy financial burdens ric tonnes. Urea, which is produced locally from natu- on the government. By the late 1970s fertilizer subsi- ral gas, is the main fertilizer used, with a 67 percent dies accounted for 4 percent of the national budget market share. (Mokarrum 1994). The government's control of the fer- Until 1978 the government had a monopoly on fer- tilizer market also resulted in misallocation of re- tilizer production, procurement, and distribution sources and inefficient production and distribution. through two parastatals: the Bangladesh Agricultural The BADC's marketing and distribution costs rose by Development Corporation (BADC) and the Bangladesh 240 percent in 1973-78 even though sales increased Chemical Industries Corporation (BCIC). Four basic by only 90 percent (Mokarrum 1994). In addition con- Box figure 4.8 Effect of removing urea subsidies cern had arisen over environmental pollution result- in Bangladesh ing from fertilizer production and use, including the Urea use possible contamination of surface and drainage water Klogram per hectare and lake eutrophication (Karim n.d.; Huq and Wheeler 500 1993; Pagiola 1995). In 1978 Bangladesh began experimenting with par- tial deregulation of urea sales. Retail prices were de- 250 regulated completely in 1983. The number of fertilizer retailers and wholesalers increased rapidly, and sub- stantial competition soon resulted. In 1988 the gov- 0 emment attempted to encourage further competition 970 1975 1980 1985 1990 1995 in the fertilizer wholesale market through the system Real price of urea of transportation discount points. Direct factory sales Thousands of 1995 taka of urea to private wholesalers at the same price as I_ 0charged to government agents were allowed in 1989. Since 1991-92 almost all urea has been marketed by 5 /\_/ the private sector. In parallel with price deregulation, subsidies for urea were gradually phased out by 1990. The savings resulting from the removal of fertil- 0 izer subsidies were estimated at $294 million in 1990- 1970 1975 1980 1985 1990 1995 93:$253 million from saving in fertilizer transportation Subsidy rate and movenent, $33 million from reductions in direct Percent subsidies, and $7 million from procurement through 75 private sector imports. ,5 Despite the abolition of price control and the re- l 50 V\rs duction in subsidies real prices for urea have declined 25 over time thanks to improved efficiency in distribu- 0 tion, increased domestic production, and decreases in world urea prices in the mid-1980s. Together with a -251 substantial increase in the area planted to improved 1970 1975 1980 198S 1990 1995 varieties, this has resulted in a sustained increase in Source: Renfro, 1992; Bangladesh Bureau of Statistics. various years. fertilizer use of 10 percent annually from 1970 to 1990. about 1 percent a year during the 1980s and early have substantial irrigated areas. In recent years, 1990s, reaching about 250 million hectares. Three however, growth in irrigated areas has occurred countries-China, India, and Pakistan-account primarily in other countries. for nearly half of the world's irrigated area. The Irrigation and the environment. Although irri- United States and the former Soviet Union also gation can result in substantial increases in ag- 60 Expanding the Measure of Wealth ricultural production, it can also have substan- during 1983--93; 0.24 million hectares in Africa, tial adverse environmental impacts, both on-site 2.28 million hlectares in Asia, and 0.28 million in and off-site. Latin America (FAO 1994). The average unit cost * On-site impacts. Irrigation can result in substan- of supplying irrigation was about $5,300 per hect- tially higher yields by guaranteeing the timing are inAsia, $23,400 inAfrica, and $11,600 in Latin and quantity of water available to crops. It can America (Jones 1995). Average annual invest- also make crop production possible in dry seasons. ments in irrigation, therefore, are about $21 bil- Misuse or overuse of irrigation can, however, also lion. Operations and maintenance costs can be lead to long-term damage toproductivitybecause roughly estimated as 10 percent of the annual of salinization or waterlogging (Umali 1993). capital investment (Repetto 1986), bringing total * Off-site impacts. Agriculture is the single biggest annual costs to about $23 billion. Assuming that user of freshwater worldwide. Water use for ir- cost recovery is only 65 percent of operations and rigation often has important adverse effects on maintenance costs, total annual irrigation subsi- both the quality and quantity of water available dies in the developing countries are about $22 bil- to other uses. Return flow from irrigated fields lion (table 4.8). Note that this estimate does not often carries pesticides, fertilizers, and other include the apportunity cost of water in altema- pollutants. Management of water for agricultural Pi t purposes can substantially affect the timing, .. .s. volume, and velocity of water flow and ground- The massive underpricing of irrigation water water recharge, thereby altering natural lake, rv- has resulted in substantial overuse. The efficiency erine, estuarine, and marine habitats. Water use of water use has been extremely low-as little as for agriculture can also affect upstream habitats 30 percent of water made available for irrigation because aquatic systems are often altered far into in developing countries is actually consumed (Xie their watersheds to regulate downstream water and others 1993). A study of eleven major irriga- delivery. tion systems in China showed that farmers often use more subsidized irrigation water than neces- irrigution subsidies. Irrigation subsidies have sary for crop growth and substitute water for been pervasive in developing and industrial na- other inputs (Repetto 1986). Overuse of irrigation tions alike. Many irrigation systems have been water has bee!n a major factor behind the water- built and operated by government agencies. Typi- logging and salinization problems being experi- cally, water users are charged only a fraction of enced in many countries. the cost of supplying water to them (table 4.7). Limited cost recovery has meant that exten- In many cases charges even fail to cover op- sive contribulions from governments' public in- erations and maintenance (O&M) costs, and they vestment and current expenditure budgets have almost never cover any of the substantial capital been necessairy to maintain irrigation systems. costs incurred in developing water collection and The failure of charges to cover operations and distribution systems (Repetto 1986; Tsur and maintenance expenses has often resulted in a Dinar 1995). Charges are often based on the area long-term deterioration of irrigation systems. In irrigated rather than on the amount used, further Mexico the irrigated area declined during the late reducing disincentives to overuse. For privately 1980s because of the dependence on subsidies (89 financed irrigation, energy for water pumping is percent subsidy rate) and government spending often subsidized (for example, in India and the cuts. Lack of maintenance funding in China has United States). kept more than 930,000 hectares of irrigated farm- Extent of subsidies to irrigation. A rough esti- land out of production since 1980. Worldwide mate of the extent of subsidies to irrigation wa- an estimated 150 million hectares, more than 60 ter can be made from available information. Since percent of the world's total irrigated area, need data on annual depreciation of the capital stock some form of upgrading to remain in good work- are not available, average annual capital invest- ing order (Gleick 1993). ment is used to approximate annual depreciation. Although subsidies to irrigation have often About 2.8 million hectares of new irrigation land been justified as being needed to assist poor farm- was created in the developing world annually ers, in many instances it is medium-size and rich Subsidy Policies and the Environment 61 farmers who have reaped most of the benefits Table 4.7 rrigation charges and unit costs (Briscoe and Garn 1994). in selected Asian countries rotal capkat and Summary O&M recurrent Water Cost costs costs charges recovey There has been substantial progress in reducing Country ($/ha) ($/ha) ($/ha) ($/ha) the heavy subsidies that once characterized pes- Indonesia 33 387 26 78 Korea, Rep. of 210 1,523 192 9 1 ticide and fertilizer markets in developing coun- Nepal 16 207 9 57 tries, and which have led to excessive and Philippines 14 166 17 121 Thailand 30 272 8 28 inefficient use of agrochemicals and consequent Bangladesh 21 375 4 18 environmental damage. These problems, along Source: Repetto 1986. with the subsidies' budgetary costs, have raised serious concern over the subsidy policy and led Conclusions to the reduction of subsidies in many countries. Compared with the situation in the early stage In recent years substantial progress has been of their introduction, the use of fertilizers and pes- achieved in reducing the extent of policies on en- ticides has been widely accepted by farmers and vironmentally damaging inputs. Trends toward the agrochemical sector today has become ma- subsidy reduction are particularly marked in the ture. There seems to be little ground for justify- energy sector, where total subsidies have been ing the subsidies. Removing explicit pesticide more than halved in the past 5 years alone. subsidies should not be the last but rather the first Substantial scope remains for additional re- step in correcting the policy and market failure. ductions in subsidies on environmentally dam- It is time for governments to move further toward aging inputs in many countries. In others, complete elimination of subsidy policies. however, most of the obvious and "easy" reforms The long-term effect of reductions in subsidies have been accomplished, and additional reform to agricultural inputs will also be affected by efforts might require trade-offs between differ- changes in other policies that affect agriculture. ent objectives. Despite the often high subsidies provided for ag- Removal of subsidy policies is not always suf- ricultural input use, the vast majority of devel- ficient to ensure that both efficiency and the en- oping countries have until recently had policies vironment improve. Additional reforms are that, on balance, discriminated heavily against omeimes needed to ensure that subsidy reform aitragain- has the intended effect. This is seen most clearly agriculture. Resources were extracted from agre- in the case of energy subsidies in the transition culture in a variety of ways: overvalued exchange economies of the former Soviet Union. rates, protection of competing sectors, price con- Even where progress is only partial, the re- trols, and high direct taxation. A sample of eigh- forms detailed in this chapter are likely to be good teen developing countries found that transfers news for the environment. Although data to con- out of agriculture averaged 46 percent of agri- firm and measure these effects are scarce, there cultural GDP during 1960-84 (Schiff and Valdes are good reasons to believe that environmental 1992). These policies have substantially slowed damages are being reduced. This should not be agricultural growth, and many countries are cause for complacency, however. While subsidy gradually moving away from them. As agricul- reform will reduce environmental damage, it will tural activities become more attractive as a result, remain excessive unless further, proactive steps increased demand for agricultural inputs is likely are taken to discourage the use of these inputs. to be induced. Total use of these inputs, there- Removing subsidies is only the first step in the . . . ~journey toward envirornmentally sustainable in- fore, may well increase even as subsidies for their use decline. This has occurred with fertilizers in put use. Bangladesh, for example (see box 4.8). Neverthe- Technical Notes less, input use will be lower than it would have been if subsidies had been kept in place while The estimates of fossil fuel subsidies are based on overall policies were reformed. a recent draft study carried out by the Pollution 62 Expanding the Measutre of Wealtht Table 4.8 Estimated subsidies to irrigation in developing countries, 1983-93 Annual Capital cost increase in per unit of Annual captal Annual C)&M Total annual Aninualirrgation imngatedland imgated land inestrment cost; costs subsidies Region (million ha) (US$per ha) (million US$) (million US$) (million US!) (eiillion US$) Africa 0.24 23,400 5,710 571 6,281 5,909 Latin America 0.28 11,600 3,271 327' 3,598 3,386 Asia 2.28 5,300 12,058 1,206 13,263 12,480 Total 2.80 21,038 2 104 23,142 21,775 Note: Assumptions: operations and maintenance is 10 percent of annual capital investnnents; cost recovery is 65 percent of operations and maintenance. Source: Irrigated land from FAO 1994; unit costs from Jones 1995 (adjusted to 1995 prices'. and Environmental Economics Division of the ists within the World Bank and with other avail- World Bank Environment Department (Rajkumar able data sources. Nevertheless, there is a high 1996). The study estimated subsidies for natural degree of uncertainty associated with many data gas, coal, and a range of petroleum products for on developing economies. each of the nineteen countries in the sample (ex- cluding Russia). Calculations were done sepa- Definitions. For each product in each sector the rately for the industrial, power, and residential subsidy estinnates are based on the difference be- sectors. tween the reference price and the actual end user price, multiplied by the quantity consumed. If the Data sources. Price and cost data were obtained reference price is less than the actual price paid, primarily from sources within the World Bank the subsidy is assumed to be zero. and the International Monetary Fund, and from The total subsidy estimates in table 4.3 and International Energy Agency reports, supple- figure 4.1 were derived by summing subsidies mented by data from various government and across sectors and fuel types. The subsidy rates international organization sources. presented were derived by summing all subsi- Quantity data were obtained from the Inter- dies and then dividing by total dollar consump- national Energy Agency. For petroleum products tion at referertce prices. data on traded prices in world markets were used as border prices: Rotterdam spot prices for Eu- Tradable products. For readily tradable goods rope, Singapore markets for Asia, New York the reference price used was the border price, prices for Latin America, Gulf prices for Africa adjusted for (a) transport and distribution costs, and the Middle East. These prices were obtained estimated from data from countries with mini- from the International EnergyAgency and the Oil mal market intervention (such as the United and Gas Journal. For natural gas and coal more States) and adjusting for country-specific condi- country-specific border prices were needed, given tions with assistance from Bank energy special- the variability across countries. These were Ob- tained farombilintacrnal WountrldiBeso s and w ists; and (b) excise, sales, value-added and other tained from internal World Bank sources and Itn- taeondnsiclyodgos,wch er ternational Energy Agency reports. A complete taxes on domroestically sold goods, which were list of sources is available in Rajkumar (1996). subtracted from domestic prices if they were im- Price data for the following were available posed at the end user level and only on domestic only up to 1994: coal (Czech Republic, Russia, In- sales (taxes tlhat also apply to exports, such as dia), natural gas (Iran, Czech Republic), and pe- excise taxes on natural gas in Russia, were not troleum products (China). A special effort was taken out). made to obtain data for 1996 in cases where sig- This approach is not well suited to measuring nificant changes occurred this year, for example, short-run natural gas subsidies because the in- in Venezuela, where gasoline prices rose by more frastructure necessary for exporting more (or at than 500 percent in April 1996. all) is lacking in many countries (for example, in To ensure consistency and relative accuracy, Saudi Arabia). The border price is still a suitable the data were cross-checked with energy special- reference price for measuring long-run implicit Subsidy Policies and the Environment 63 subsidies, however. Estimates of natural gas sub- References sidies in countries such as SaudiArabia take capi- tal investment in infrastructure into account, Adesina, A. 1994. "Pesticide Use and Policy in West including it (as capital costs) in the transport/ Africa: An Overview." In Agne (1994). distribution margin, which is subtracted from the Agne, S., G. Fleischer, and H. Waibel, eds. 1994. Pro- border price. ceedings of the Gottingen Workshop on Pesticide Poli- cies. Gottingen: University of Gottingen, Effects offalling world prices. Changes in mea- Institute of Agricultural Economics. sured subsidy rates are partly the result of fall- Anderson, K. 1995. The Political Economy of Coal Sub- ing world prices, which reduce the gap between sidies in Europe. Nota Di Lavoro 50.95. Milan: domestic and border prices even if the former re- Fondazione ENI Enrico Mattei. main constant. To allow for this effect, 1995-96 Antle, J., and P. Pingali. 1995. "Pesticides, Produc- subsidy rates can be re-calculated holding bor- tivity, and Farmer Health: A Philippine Case der prices at their 1991 levels (in real terms). The Study." In P. Pingali and P. Roger, eds., Impact of estimated t 11esin real trm.e Pesticides on Farmer Health and the Rice Environ- ment. Los Banios: International Rice Research In- products and natural gas would then be 19 per- stitute. cent and 48 percent in Central and Eastern Eu- Briscoe, J., and M. Gamn. 1994. "Financing Agenda rope; 51 percent and 72 percent for oil producers; 21: Freshwater." World Bank, Transport, Water, 12 percent and 14 percent for other countries; and and Urban Development Department, Washing- 18 percent for petroleum products in Asia. The ton, D.C. main qualitative difference between these num- Burniaux, J. M., J. P. Martin, and J. Oliveira-Martins. bers and those in table 4.3 is that natural gas sub- 1992. "The Effect of Existing Distortions in En- sidy rates for the oil producers and other ergy Markets on the Costs of Policies to Reduce countries rise slightly here rather than falling be- CO. Emissions: Evidence from GREEN." OECD tween 1990-91 and 1995-96, while natural gas Economic Studies (19, Winter). subsidy rates for Eastern Europe remain almost Couston, J. 1985. "Implications of Crop and Fertil- unchanged. For natural gas in the other regions izer Relationships in Developing Countries." and for the other two fossil fuels, however, sub- Paper presented at the IFA Annual Conference, sidy rates still fall on average. Thus, the qualita- Munich, Germany. tive results remain essentially unchanged, with DOE (U.S. Department of Energy). 1989. "The Im- some qualifications to be made for natural gas. pact of Eliminating Coal Subsidies in Western Europe." Office of Coal, Nuclear, Electric andAl- Nontradable products. For goods with limited ternative Fuels, Washington, D. C. trade possibilities, the reference price used was Duloy, J., and P. Nicholas. 1991. "Controlling Pesti- the long-run marginal cost of production. How- cide Pollution: The Choice between Regulation ever, reliable estimates of this are difficult to ob- and Taxation." World Bank, South Asia Country tain. Thus, long-run average cost data (which Department II, Washington, D. C. Eskeland, G., E. Jimenez, and L. Liu. 1994. "Energy probably understate marginal costs) were used Pricing and Air Pollution: Evidence from Manu- in most cases. Available average cost data tend facturing in Chile and Indonesia." Policy Re- to grossly underestimate capital costs, however, search Working Paper 1323. World Bank, Policy especially in Eastern Europe. Capital costs thus Research Department, Washington, D.C. had to be estimated based on the available break- FAO (Food and Agriculture Organization). 1994. down of reported costs (with the help of Bank FAO Production Yearbook. Rome. energy specialists), using the replacement cost Farah, J. 1994. Pesticide Policies in Developing Coun- method. tries. World Bank Discussion Paper 238. Wash- ington, D. C. Exchange rate. Where available, market rates Fleischer, G., and H. Waibel. 1993. Survey on National as reported in the IMF International Financial Sta- Pesticide and Pest Management Policies. Gottingen: tistics were used; otherwise, official rates were University of Gbttingen, Institute ofAgricultural used. Economics. 64 Expanding the Measure of Wealth Freund, C., and C. Wallich. 1995. "Raising House- . 1989. "Commercial Energy Subsidies in Devel- hold Energy Prices in Poland. Who Gains? Who oping Countries: Opportunity for Reform." En- Loses?" Policy Research Working Paper 1495. ergy Policy (June). World Bank, Europe and Central Asia Country Larsen, B. 1994. "World Fossil Fuel Subsidies and Department II, Washington, D. C. Global Carbon Dioxide Emissions in a Model Gleick, P., ed. 1993. Water in Crisis: A Guide to the with Interfuel Substitution." Policy Research World's Fresh Water Resources. Oxford: Oxford Working Paper 1256. World Bank, Policy Re- University Press. search Department, Washington, D. C. Grubb, M. 1994. "Global Policies for Global Prob- Larsen, B., and A. Shah. 1992. "World Energy Sub- lems: The Case of Climate Change." In T. Sterner, sidies and. Global Carbon Dioxide Emissions." ed., Economic Policiesfor Sustainable Development. World Development Report 1992. Background Pa- Amsterdam: Kluwer Academic Publishers. per 25. World Bank, Washington, D. C. Gurvich, E., and G. Hughes. 1996. "The Environ- Maene, L. 1994. "Recent Experience with the mental Impact of Energy Subsidies: A Case Phasing-out of Fertilizer Subsidies with Particu- Study of Russia." World Bank, Environment lar Reference to Developing Countries." Agro- Department, Pollution Mangement and Environ- chemicals News in Brief 17 (1). mental Economics Division, Washington, D. C. Matteson, P., and M. Meltzer. 1995. Environmental Homer, J. 1993. Natural Gas in Developing Countries. and Economic Implications ofAgricultural Trade and World Bank Discussion Paper 190. Washington, Promotion Policies in Kenya: Pest and Pesticide Man- D. C. agement. Washington, D. C.: U.S. Agency for In- Hope, E., and B. Singh. 1995. "Energy Price Increases ternational Development. in Developing Countries: Case Studies of Colom- Mokarrum, S. 1994. "Privatization of Bangladesh bia, Ghana, Indonesia, Malaysia, Turkey, and Fertilizer Marketing and Distribution: A Case Zimbabwe." Policy Research Working Paper Study." Agro-chemicals News in Brief 17 (1). 1442. World Bank, Policy Research Department, Munasinghe, M., and W. Cruz. 1995. Economywide Washington, D. C. Policies and the Environment: Lessonsfrom Experi- Huq, M., and D. Wheeler. 1993. "Pollution Reduc- ence. World Bank Environment Paper 10. Wash- tion without Formal Regulation: Evidence from ington, D. C. Bangladesh." Environment Department Divi- OECD (Organisation for Economic Co-operation sional Working Paper 1993-39. World Bank, En- and Development). 1995. Review of Agricultural vironment Department, Washington, D. C. Policies: Poland. Paris: OECD. IEA (International EnergyAgency). 1989. Energy and Okugu, B. E., and F. Birol. 1992. "Curbing Carbon the Environment: Policy Overview. Vienna: OECD. Dioxide Ermissions byAxing EC Coal Subsidies." 1995. Energy Policies of IEA Countries, 1994 Re- OPEC Bulletin (November/December). view. Vienna: OECD. Latin American Energy Organization. 1992. Study Isherwood, K. 1996. "Fertilizer Subsidy Policies in of Pricing Policies of Oil and Products for Latin Regions Other Than Asia and the Pacific." Pa- America and the Caribbean. Quito. per presented at the Regional Workshop on Fer- Pagiola, S. 1995. "Environmental and Natural Re- tilizer Policies and Subsidies, Bali, Indonesia. source Degradation in Intensive Agriculture in Jones, W. 1995. The World Bank and Irrigation. Wash- Bangladesh." Environment Department Work- ington, D. C.: World Bank. ing Paper 15. World Bank, Environment Depart- Julius, D. 1986. "Domestic Pricing of Petroleum ment, Washington, D. C. Products: Efficiency and Equity Impacts in De- Pagiola, S., J. Kellenberg, L. Vidaeus, and J. Srivastava. veloping Countries." OPEC Review (Spring). 1997.MainstreamingBiodiversity in Agricultural De- Karim, Z. N.d. Fertilizer in the National Economy and velopment: 'Toward Good Practice. World Bank; En- Sustainable Environmental Development. Dhaka: vironment Paper 15. Washington, D. C. Bangladesh Agricultural Research Council. Panayotou, T. 1993. Green Markets: The Economics of Kenmore, P. 1991. Indonesia's Integrated Pest Manage- Sustainable Development. ICEG Sector Studies Se- ment: A Modelfor Asia. Jakarta: FAO. ries no. 7. San Francisco: ICS Press. Kiss, A., and F. Meerman. 1991. Integrated Pest Man- Pearson, S. R. 1991. Rice Policy in Indonesia. Ithaca: agement and African Agriculture. World Bank Tech- Cornell University Press. nical Paper 142. Washington, D. C. Pincus, J. 1994. "Discussion Points on Pesticide Kosmo, M. 1987. Money to Burn? The High Costs of Policy and Rice Production in Indonesia: Com- Energy Subsidies. Washington, D. C.: World Re- bining Reform from Above and Action from Be- sources Institute. low." In Agne (1994). Sutbsidy Policies and the Environment 65 Pitt, M. M. 1985. "Equity, Externalities and Energy Valdes, A., and B. Schaeffer. 1995a. Surveillance of Subsidies: The Case of Kerosene in Indonesia." Agricultural Price and Trade Policies: A Handbook Journal of Development Economics (The Nether- for Chile. World Bank Technical Paper 291. Wash- lands) 17. ington, D. C. Rajkumar, A. S. 1996. "Energy Subsidies." Environ- Vald6s, A., and B. Schaeffer. 1995. Surveillance ofAg- ment Department Working Paper. World Bank, ricultural Price and Trade Policies: A Handbookfor Environment Department, Washington, D. C. Colombia. World Bank Technical Paper 268. Wash- Renfro, R. 1992. "Fertilizer Price and Subsidy Poli- ington, D.C. cies in Bangladesh." World Development 20 (3): Waibel, H. 1990. "Pesticide Subsidies and the Diffu- 437-55. sion of IPM in Rice in Southeast Asia: The Case Repetto, R. 1985. Paying the Price: Pesticide Subsidies of Thailand." FAO Plant Protection Bulletin, 38 (2): in Developing Countries. Research Report 2. Wash- 105-11. ington, D. C.: World Resources Institute. Wang, X. 1996. "China's Coal Sector: Moving to a 1986. Skimming the Water: Rent-Seeking and the Market Economy." World Bank, China and Performance of Public Irrigation Systems. Washing- Mogola Department, Infrastructure Division, ton, D. C.: World Resources Institute. Mongon, D. C. .1988. Economic Policy Reform for Natural Re- Washington, D. C. sour88Ec onservaton.cy Ren rirnment DeaturtmRen World Bank. 1986. World Development Report 1986: Working Paper 4. World Bank, Environment Trade and Pricing Policies in World Agriculture. Department, Washington, D.C. New York: Oxford University Press. Rola,A., and P. Pingali. 1993. Pesticides, Rice Produc- * 1992. World Development Report 1992: Develop- tivity, and Farmers' Health. Los Bafios: Intema- ment and the Environment. New York: Oxford tional Rice Research Institute and World University Press. Resources Institute. World Resources Institute. 1996. World Resources Schiff, M., and A. Valdes. 1992. The Political Economy 1996-97. Washington, D. C. of Agricultural Pricing Policy. Baltimore, Md.: Xie, M., U. Kuffner, and G. Le Moigne. 1993. Using Johns Hopkins University Press. Water Efficiently: Technological Options. World Thrupp, L. 1990. "Inappropriate Incentives for Pes- Bank Technical Paper 205. Washington, D. C. ticide Use: Agricultural Credit Requirements in Developing Countries." Agriculture and Human Selected additional references Values (Summer-Fall): 62-69. Tsur, Y., and A. Dinar. 1995. "Efficiency and Equity Srivastava, J.P., N.J.H. Smith, and D.A. Forno. 1996. Considerations in Pricing and Allocating Irriga- Biodiversity and Agricultural Intensification: Part- tion Water." Policy Research Working Paper 1460. ners for Development and Conservation. Environ- World Bank, Agriculture and Natural Resources mentally Sustainable Development Studies and Department, Washington, D.C. Monographs no. 11. Washington, D.C.: World Tybout, J. R., and D. Moss. 1992. "The Scope for Bank. Interfuel Substitution in Manufacturing Indus- Schillhorn van Veen, T.W., D.A. Forno, S. Joffe, D.L. tries: A Case Study of Chile and Colombia." The Umali-Deininger, and S. Cooke. 1997. Integrated World Bank Economic Review 8 (1). Pest Management: Strategies and Policiesfor Effec- Umali, D. L. 1993. Irrigation-Induced Salinity: A Grow- tive Implementation. Environmentally Sustainable ing Problem for Development and the Environment. Development Studies and Monographs no. 13. World Bank Technical Paper 215. Washington, D.C. Washington, D.C.: World Bank. PART Two Evolving Indicator Themes I I I CHAPTER 5 Land Quality Indicators hapter 2 highlights the role of good qual- (UNDP), the United Nations Environment ity agricultural land in terms of its asset Programme (UNEP), and the Consultative Group value and income-generating capacity. for International Agricultural Research (CGIAR). This suggests the importance of developing ap- In many developing countries fears that land propriate methods to measure land quality. How- degradation threatens food production capacity ever, preliminary investigation reveals the have stimulated projects and policies to encour- complexity of developing meaningful indicators age more sustainable land-use practices. Land of land quality, particularly at a national level degradation problems include soil erosion, salin- where averages generally have little meaning. It ization, nutrient mining, compaction, organic appears that work at the level, of (relatively ho- matter loss, deforestation and forest degradation, mogeneous) agro-ecological zones or the local and pasture and rangeland degradation (Scherr level is more promising. and Yadav 1996). However, taking action to im- The purpose of this chapter is to take stock of prove natural resource management is hampered preliminaryworkthathasbeenundertakenonland by a shortage of reliable, easily accessible infor- quality indicators (LQIs) and to describe planned mation on the nature and extent of land degra- work to be undertaken by the World Bank together dation problems and their productivity impacts. with the Food andAgriculture Organization (FAO), This has been a primary motivation for initiating the United Nations Development Programme the land quality indicators program (see box 5.1). Box 5.1 The land quality indicators program To better understand land degradation problems, a coa- The LQI program is conceptually similar to earlier lition of international agencies initiated the land qual- national and international efforts that led to the devel- ity indicators (LQI) program in 1994. This initiative opment of standardized indicators of economic and i seeks to develop a set of natural resource indicators: environmental performance. Gross national product, statistics or measures that help characterize the condi- for example, is measured the same way throughout tion of natural resources related to land. The organiz- the world, facilitating comparisons across countries. ing agencies for the LQI program include the World tn ou,aciamgopnssarssonre. ing agencies for the LQI program include the World Air and water quality are measured by standard indi- I ?Bank, the Food and Agriculture Organization (FAO), I Bank, the Food and Agricult ure Organization (FAO), cators that provide an objective assessment that is ap- jthe United Nation as Development Programme (UNDP), the United Nations Environment Programme plicable n widely differig countries and contexts. The (UNEP), and the Consultative Group on International LQI program seeks to develop a set of standardized Agriculture Research. Several other organizations, in- indicators (mainly focused on the local and district lev- cluding the World Resources Institute (WRI), the In- els) to provide concise, reliable information about the ternational Food Policy Research Institute (IFPRI), and condition of land, including the combined resources t other centers of the Consultative Group on Agricul- of soil, water, vegetation, and terrain that provide the l ture Research (CGIAR) are also participating. basis for land use (Pieri and others 1995). 69 70 Expanding the Measure of Wealth Selecting Land Quality Indicators levels that are considered high for a desert eco- system may be very low for a humid area. Like- The key natural resources that are influenced by- wise, all ecosystems are not prone to the same and in turn influence-agricultural activity, are threats. For example, soil salinization is a prime soils, surface and groundwater, pasture and concern in many irrigated areas, but is irrelevant range, forests and trees, and agro-biodiversity. For in most tropical hillside environments. Accord- each resource policymakers and land managers ingly, different LQIs may be needed for distinct are often interested in monitoring both quality and agro-ecological conditions. The specific indica- quantity and their changes over time. tors that are ultimately selected by different us- The complexity of land resources poses chal- ers are likely to vary considerably, depending on lenges that make the development of standard- the scale, ecozone, agricultural system, available ized indicators more difficult than for air and resources, available data sources, and other water quality. Weather, terrain, and other bio- factors. logical and physical factors combine to determine Box 5.2 presents a set of pressure, state, and production potential of land, which consequently response indicators for selected land quality is- varies by ecosystem. For example, production sues (see Adriaanse 1993 for a more detailed dis- Box 5.2 Examples of soil-related land quality indicators for pressure (P), state (S), and response (R) Soil erosion Soil quality P: * Cultivated area in zones of high erosion P: * Ratio of cultivated to cultivable land potential or impact * Net nutrient extraction in cropland, pastures, * Pasture areas grazed in zones of high erosion and forests potential * Area under nonconservation tillage practices * Forest understory exploited in zones of high * Area under poor crop residue management erosion potential * Density of roads in zones of high erosion S: * Level of o:rganic matter, by soil type or area potential * Level of nutrients, by soil type, area, or use * Area deforested in zones of high erosion (direct measures or models) potential * Area occupied by indicator plants * Area without perennial soil cover during * Water-holding capacity critical erosive periods * Land productivity adjusted for input use * Lack of vegetative protection of waterways : Annual soil loss (measured, or predicted by model) R: * Fertilizer and manure applications, by soil type * Incidence of gullying or area 0 Incidencetof glelying * Area covered by sustainable nutrient *Sedimentation levels in rivers or behind dams applications *Deposit of sediment in coastal areas apiaiii , Deposit of sediment in coastal areas * Farmers, groups, or area participating in soil * Amount of silt deposited on agricultural lands improvement programs * Area with soil depth reduced to less than 10 , Farmland abandoned centimeters * Change to crops with lower nutrient R: * Area protected by conservation practices or requirements landscape features * Adoption of conservation and zero tillage * Revegetation and structural protection of practices waterways * Farmers served by soil conservation programs Similarly detailed preliminary indicators of pres- * Farmers using selected soil conservation sure-state-response have been identified for other practices characteristics (soil salinity and alkalinity, ground * Change in land use to more or less erosive and surface water supply, water quality related to practices land management, pasture quality, forest area, for- * Total value of soil conservation programs est quality, and agrobiodiversity) but are not shown * Operational area under soil management plans here due to space constraints. Land Quality Indicators 71 cussion of the Pressure, State, and Response may grow only on degraded soils. Thus, the in- Framework). These are examples only, and nor- dicator for soil degradation might be the extent mally not all of these indicators will be relevant of area growing that indicator plant, which is in any particular situation. fairly easy to monitor, rather than soil nutrient Theories of causality about land degradation status, which is more expensive. Local people are related mainly to human management. Some may have common indicators of natural resource of these theories are not verified empirically; conditions. Use of participatory methods to de- hence the selection of pressure indicators is an fine and collect indicator information could serve area for further research, and the current think- not only to decrease the costs of monitoring, but ing presented here is preliminary. National popu- also to engage local people more deeply in re- lation density, for example, is not used as a source assessment. pressure indicator because increases in density The indicators presented in box 5.2 are in ge- do not directly place pressure on resources. neric terms. A particular indicator, for example, Rather, it is land-use practices actually used that "sedimentation levels in rivers," is likely to be result in the pressure. People could be moving to important in many land quality monitoring cities or to nonagricultural occupations. Or the schemes. However, the particular indicator cho- presence of more people may lead to more re- sen may vary, depending on the relative impor- source-conserving investment. tance of that variable, and the degree of precision The state indicators were selected as represent- that monitors can afford or need to have. That ing the condition, at a point in time, of the re- is, the indicator may range from a qualitative as- sources that are of greatest economic, ecological, sessment by river basin experts (high, moderate, or social importance. The response indicators low), to a quantitative measure of sediment lev- were designed to include both formal program els behind important dams, to a measure of sedi- responses and individual or household manage- ment levels in different parts of a river basin based ment responses, as well as other types of adjust- on sophisticated sampling and instrumentation ments that result from the change in resource and transformed into an index of river-basin-wide state (for example, changes in diet, changes in fuel status. use, emigration, or land abandonment). Interpreting Land Quality Indicators Selecting Indicators The policy implications of changes in LQIs can- Policymakers (at local, project, national, or inter- not be judged on the basis of individual indica- national levels) are the main audience for LQIs. tors. Rather, the monitor must assess trends in sets Indicators should be easily understood and trans- of LQIs in order to determine whether or not the parent. Together the set should tell a "story" about indicator is signaling a significant change. For this the pressure and response sequence and let poli- reason box 5.2 suggests a number of different in- cymakers know whether policies are working or dicators for pressure, state, and response for each not. Indicators should give reliable signals of im- land quality issue; the appropriate choice depends portant trends in the actual situation. They need on the local situation, available data, and sensi- to be comparable across time. It would be useful to hve n inicaionnot nlyof he "verge" tivity to possible confounding factors that would to have an indication not only of the "average" afetitrpeain state, but also of variability around that average. affect interpretationt In some cases it may be useful to use thresholds, LQIs should be interpreted in the context of that is, it may only be important to know whether broader trends in the region. A major increase in the condition of a resource falls below a particu- fertilizer use might be seen as a positive indica- lar level. Nutrient balance is an example of such tor of land improvement under conditions of an LQI (see box 8.5). stable agricultural land area. But if land expan- An objective is to seek the lowest-cost indica- sion is taking place at a much faster rate than fer- tors that serve the purpose. In some cases these tilizer use, then net nutrient depletion may still may be surrogates or proxies of actual land qual- be occurring. Also, one would expect to see quite ity variables. For example, certain plant species different problems and potentials in areas of the 72 Expanding the Measure of Wealth same agro-ecozone that had different levels of primarily on FAO statistics and statistics collected population density and market access. by WRI for its annual report on the state of world For these reasons contextual information must resources. also be collected to interpret LQIs (box 5.3). The This search found only a limited number of focus is on key population, economic, and land indicators tha't could be used. The only pressure base characteristics. Wherever possible this infor- indicators available were related to changes in mation-as with the LQIs-should distinguish land use and levels of forest extraction and live- between different agro-ecological zones so as to stock population. With the exception of forest tell a more accurate and policy-relevant story of area and crop yield there were no data available land quality pressures, state, and response, from international databases to be used as state indicators. Testing the Land Quality Indicators Crop yield is a somewhat problematic indica- Approach tor for land quality since yield has so many de- terminants. Given that the key land-quality policy Case studies were undertaken to see to what ex- question is to assess the extent to which yield tent LQIs can be developed on the basis of exist- changes are alfected by changes in underlying ing data and to identify key issues for the land quality, relative to other factors, it is gener- development of LQIs and the direction in which ally not appropriate. With the exception of fer- we need to move in the future to improve them. tilizer use for soil quality and the establishment of protected areas for forestry, no internationally National-level Case Studies available response indicators were identified. The preliminary conclusion from the case study work The World Resources Institute (WRI) explored the at the national level is that not many useful in- use of LQIs at the national level for three coun- sights can be gained from existing information tries-Honduras, Niger, and Vietnam-using in- in international databases. However, there is ternationally available data sets. These depended much more national-level data available that could be adapted for international comparisons. Box 5.3 Examples of contextual information District-level Case Studies to assist with the interpretation of land quality indicators The International Food Policy Research Institute Population (IFPRI) explored LQIs at the district level for the * Rural population density for total or arable land Central Hillside!s Region of Honduras and Jhabua * Rural poverty index District of India. Both regions are homogeneous Markets agro-ecological zones in which IFPRI researchers * Road density (all-weather, feeder) had field experience. District data were drawn * Degree of commercialization of agriculture * Density of agricultural input outlets from national, subnational, and local sources. * Real agricultural, livestock, forest input to LQIs were developed for land quality problems output price ratio (for key products) identified to be most important in each region. * Value of total production (agriculture, livestock, Information from Honduras showed that for- forestry) * Farm size distribution (for example, Gini est area declined by more than a third in the stud- coefficient) ied region between 1965 and 1982, although the Land proportion of forested area on farms remained * Area of cropland, pasture, forest quite stable. Protected areas, which cover over 8 * Intensity of agricultural land use (for example, percent of the land area, are mainly forested. For- 'R ratio)pecnoftelnaraarmalyfrse.F- * Agricultural land under irrigation est quality is a serious problem. Commercial log- * Livestock population density ging concessions were stopped after 1992, and * Livestock feed resources timber production is fairly low. Data from the * Average yields for key products H * Land tenure categories Honduran Forestry Development Bureau, how- ever, suggest a sharp increase in timber produc- Land Quality Indicators 73 tion in the Department of Francisco Morazan af- surveys, however, there was little accessible in- ter the 1993 Forest Law was enacted. Forest den- formation on farm-level response in terms of ei- sity is low over a large area (often reflecting ther resource management or adjustments in transition to silvopastoral systems, usually with consumption or livelihood due to changes in the little management). Resin production is declin- resource base. ing. A low but increasing proportion of resin col- lectors are using less damaging methods. Data Issues In the Jhabua district in India soil erosion ap- be wiespred, asmightbe exected It is premature to consider using existing national- pears to be widespread, as might be expected level LQIs for international comparisons. It would from the pressure indicators. However, no data be most useful to begin developing LQIs at the are available concering the extent of continuing subnational level, and relate them to particular soil erosion and little is known about the extent agro-ecological zones. Meanwhile, work can be- of soil conservation practiced in this area. Soil loss gin on a flexible framework for national report- on eroded land is estimated to have reduced pro- ing of such LQI sets. ductivity by up to a third (Sehgal and Abrol 1994). The subnational case studies illustrated that Some evidence suggests that most soil conserva- there are data that could be mobilized for the de- tion programs have had limited effectiveness, in velopment of LQIs. Many national agencies, de- which case they provide a poor indicator of farm- velopment programs, projects, donors, and level response (Pretty and Shah 1993). Decline scientists regularly or irregularly collect data on in soil nutrients is a common problem when tropi- different pressure, state, and response variables. cal forests with naturally infertile soils are cleared Access to these data sets, however, is often diffi- for agriculture. Originally, soil nutrients were re- cult. Data collected by consultants and donors stored through extended fallow periods. Over may be considered confidential or proprietary, the last several decades, however, even short- and for some public agencies information is term fallowing has been nearly abandoned due power, not to be widely circulated. Public agen- to the high demand for land. cies have weak archiving capacity, and important Fertilizer use is growing rapidly in Jhabua, but data sets are often found only with the individual remains very low. Data on organic matter appli- who was in charge of collecting them. Data sets cations are not available. The overall trend in start and stop with access to funding. India is of declining manure applications due to Finding out the proper definition (and hence, the high value of dung for use as fuel. Limited interpretation) of particular indicators can also data on the state of soil nutrients are available be difficult for subnational case studies, even for from measurements conducted by the National basic data sets such as the Agricultural Census. Centre for Human Settlements and Environment This is particularly important in the development (1993) study, although the sample is small (only of historical data sets, as definitions and data col- 22 observations for the whole district). Organic lection methods often change over time, affect- matter and nitrogen content are uniformly low, ing interpretation of trends. while phosphorus content is highly variable. An important contribution of the LQI program During these preliminary case studies it was would be to identify a defined set of indicators, not possible to estimate many of the indicators formalize responsibility to particular groups for presented in box 5.2. Data on the state of land the collection and reporting of information on an resources were very limited and usually taken ongoing basis, and compilation at regular inter- from small samples. Vegetative cover data from vals for public access. Availability of such LQI remote sensing were found for both regions but series could substitute for much of the effort that remnotbesensing lywerefound fora useabothlegion currently goes into baseline and monitoring data had not been analyzed in a useable way. Re- colcto fo prjcs sponse variables were widely found for responses related to government programs, NGO projects, Existing Data Sources and other formal interventions. With the excep- tion of Agricultural Census data on selected prac- Agricultural censuses and surveys provide a key tices and special one-time community and field source of data related to land quality, particularly 74 Expanding the Measutre of Wealth pressure and response variables, although they group interviews, or systematic observation often need additional analysis. They also offer a points. If these are not done, then key LQIs can- potential for eventually integrating additional not be constructed. LQIs as part of ongoing national data collection programs. Further Development of Land Quality Another source of data for LQIs is remote sens- Indicators ing. Satellite imagery can be interpreted to as- If more robust and policy-relevant LQIs are to be sess vegetative cover, moisture levels, forest constructed, then further efforts will be required. density, agroforestry practices in farmland, and O s othe vaiabes. eril p6toraphseres ver Of particular importance are development of spa- otmer varianbleus.dto asseri photographes ovser tially explicit clata sets, better integration of LQIs time cand bnveustmedntoass changesdn land uge,t at different sc,ales, further research on cause-ef- some and cnvtntspand landemag e fect relationships, and institutional innovations. practices, and changing spatial patterns. Missing Data Spatial and Geographic Issues Land quality issues, by their nature, are geo- A comparison of the LQIs available for the case studies and box 5.2 with the preferred LQI vari- graphically specific. "Average" fertility levels pro- vide a poor basis for policy if in some areas soils ables highlights several critical gaps. For certain are very feri anr eshere they area deple types of pressure variables, few direct, indirect, or Identification of the a hot spots" for land degrada- proxy indicators were found. Most important was tion is especially important. The ecological func- the lack of data to link pressures (land-use, man- tion and impact of land management practices agement) to particular types of land geographi- will differ depending on where in the watershed cally. In Honduras and Jhabua it is critical for the or forest these are used. New techniques for assessment of erosion hazard to determine how georeferencing data collected in the field, for ana- much cropping takes place on the steeper slopes. lyzing remotely sensed images, and for analyz- But without georeferenced land-use information, ing 'layers" of spatiallyexplicit data through even existing geo-referenced data sets on topog- geographic information systems, offer the poten- raphy and land use potential cannot be used. A tial for lower-cost and more powerful LQIs. As a second category of missing information relates to matter of fact, TQI data are already being used to biomass removals from pasture and forests and prepare land systems maps for agro-ecological use and extraction of water. zoning purposes. For example, in Indonesia both Direct measures of the resource state variables the World Bank-supported National Master Plan are particularly weak. These are variables such for Forest P`i;ntations and some Overseas Devel- as soil loss, sedimentation, organic matter, and opment Administration supported projects have groundwater levels that require technical moni- resulted in the detailed mapping of soil suitabil- toring. This does not mean that detailed, quanti- ity and erodability for most of Sumatra and tative technical surveys are necessary; in some Kalimantan. These data can be used for planrning cases one can use low-cost proxy indicators (for purposes. example, number of communities reporting wa- ter quality problems, rather than direct sampling Integrating LQIs at Different Scales and laboratory analysis of water quality). Finally, a critical state variable is vegetative cover for as- A critical challenge in the development of LQIs is sessing soil, pasture, and forest condition. their definition and integration at different scales. The main missing LQIs for the response vari- In this chapter we have principally addressed the ables are related to actual resource management LQIs needed for agro-ecological zones or district practices by farmers or other users. While there scales. Indicator data collected at local levels were are many proxies or qualitative indicators that aggregated for analysis at the higher scale. Yet might be applied to some of these variables, they one of the most important potential uses of LQIs still require field data collection through surveys, is at the local or community level. Potentially, Land Quality Indicators 75 data collected at this level could be relatively low Conclusions cost and higher quality than data collected by dis- interested regional agents. "Grassroots" indicators This chapter represents an initial effort to explore could prove to be inexpensive proxies for key the feasibility of land quality indicators and to LQIs. assess their potential usefulness based on a num- ber of case studies. The work showed that cur- Research Needs rently available data are seriously deficient. In particular, existing information on pressure indi- Obtaining usable and reliable LQIs requires de- cators is problematic because so little of it is geo- velopment and testing. In some cases this will graphically focused and georeferenced. In the mean testing different indicators against one an- case of most state indicators data simply do not other, in terms of effectiveness cost, and accuracy. exist, or exist only for very small and nonrepre- In other cases studies will need to compare alter- sentative areas. native measurement methods or assess the reli- Any new data collection effort must be care- ability of proxies. Conceptual and experimental fully targeted to produce policy-useful indicators. work will probably be needed to sort out the sam- Specifically, will the proposed indicator flag the pling strategies for LQIs at different scales. Coor- need for a possible change in the type of inter- dinated pilot efforts to develop LQIs for a range vention, the location of intervention, the target of agro-ecological zones, at various scales, should population for intervention, the target beneficia- permit fairly rapid development of basic indica- ries of intervention (for example, if externalities tor sets. In the long run more basic research is are the main policy issue) or the intensity of in- needed to evaluate the relationships between par- tervention? Ideally, LQIs would flag "hot spots" ticular pressure or response factors and the state where land quality problems are having, or are of the resource and to identify biophysical and likely to have, significant productivity, social economic thresholds beyond which key land welfare, or ecological effects. qualities can be expected to change qualitatively. Expectations about LQIs should be realistic: The LQI process itself may help point to priority LQIs provide the raw data for analysis, but even issues for these types of research. To maintain when sophisticated indices are used, they do not both the technical integrity of the indicators and provide analysis and interpretation. Internation- their relevance for policymakers, the main re- ally comparable national data do not, at present, search thrust for LQIs should remain interdisci- provide useful LQIs. It is therefore sensible to seinarcthsfry.sudea start at the subnational level and identify the most plinary. relevant, reliable, and cost-effective approaches Institutionalizing Land Quality Indicators to indicator development. Aggregation to the na- tional level would be the next step. Meanwhile, a The development, collection, analysis, and regu- meaningful framework for international compari- lar distribution of LQIs will require some reor- son, on an agro-ecological zones basis, can be de- dering of institutional arrangements. Data veloped as lessons are learned from pilot efforts collection for LQIs could be integrated in or at subnational LQI development. "piggy-backed" onto already existing institutions The process of developing LQIs also highlights such as census bureaus; household, farm, or com- shortcomings in available data and knowledge munity surveys; or watershed management moni- of production systems. As such the LQI approach toring. The analyses and reporting of the LQIS potentially provides a useful framework for iden- could then be achieved inexpensively through ex- tifying new research priorities, and the potential isting institutions. However, this would require context for applying research results. training and national capacity building on sam- References pling and data collection, as well as on data stan- dards, data compatibility, and development and Adriaanse,A. 1993. Environmental Policy Performance maintenance of computerized information Indicators. The Hague: Ministry of Housing, systems. Physical Planning, and the Environment. 76 Expanding the Measure of Wealth NCHSE (National Centre for Human Settlements Scherr, S., and S. Yadav. 1996. "Land Degradation and Environment). 1993. Sustainable Utilization in the Developing World: Implications for Food, of Natural Resources in Jhabua District. Bhopal. Agriculture and the Environment to 2020." A Pieri, C., J. Dumanski, A. Hamblin, and A. Young. 2020 Vision for Food, Agriculture, and the Envi- 1995. Land Quality Indicators. World Bank Dis- 2020 Discussior 14d IntemationalF cussion Paper no. 315. Washington, D. C. ronment Discussion Paper 14. International Food Pretty, J. N., and P. Shah. 1993. Soil and Water Con- Policy Research Institute, Washington, D. C. servation in the 20th Century: A History of Coer- Sehgal, J., and I. P. Abrol. 1994. Soil Degradation in cion and Control. Rural History Center Series. India: Status and Impact. New Delhi: Oxford and Reading, U.K.: University of Reading. IBH. CHAPTER 6 Social Capital: The Missing Link? S ustainable development has been defined as different levels of economic performance. So have a process whereby future generations re regions or cities within countries, and even com- ceive as much capital per capita as-or munities within regions or cities. more than-the current generation has available * Case 1: The high growth rates of the East Asian (Serageldin 1996a, 1996b). Traditionally, this has "miracle" economies, relative to other parts of included natural capital, physical or produced the world, can only in part be explained by con- capital, and human capital. Together they consti- ventional factors such as investments in human tute the wealth of nations and form the basis of and physical capital and technology. Govern- economic development and growth. In this pro- ment policies provided an enabling environ- cess the composition of capital changes. Some ment, characterized by institutional arrange- natural capital will be depleted and transformed ments and organizational designs that enhanced into physical capital. The latter will depreciate, efficiency, exchange of information, and coop- and we expect technology to yield a more effi- eration between government and industry cient replacement. This century has seen a mas- (World Bank 1993; Stiglitz 1996). sive accumulation of human capital. * Case 2: In a study of Italy Putnam argues that It has now become recognized that these three the higher density of voluntary associations types of capital determine only partially the pro- among people in northern Italy explains the cess of economic growth because they overlook region's economic success relative to southern the way in which the economic actors interact and Italy, where such associations are less frequent organize themselves to generate growth and de- (Putnam and others 1993). velopment. The missing link is social capital. At * Case 3: After the 1991 fall of Somalia's govern- this broad level of conceptualization there is little ment civil disorder prevailed and incomes de- disagreement about the relevance of social capi- clined throughout most of the country. An tal. There is, however, no consensus about which exception was the port city of Boosaaso, where aspects of interaction and organization merit the a local warlord organized a security force and a label of social capital, nor in fact about the valid- council of clan elders with support from local ity of the term capital to describe this. Least people. Trade flourished and incomes improved progress has been made in measuring social capi- (Buckley 1996). tal and in determining empirically its contribu- * Case 4: In Gujarat, India, violent confrontations tion to economic growth and development. This between local people and government officials chapter will address each of these issues in turn.1 over the way forests were managed led to eco- nomic stagnation. After communities were mo- The Concept of Social Capital bilized and joint forest management was instituted, conflicts declined and land produc- Countries with similar endowments of natural, tivity and village incomes rose (Pathan and physical, and human capital have achieved very others 1993). 77 78 Expanding the Measure of Wealth Each one of these cases displays an aspect of ment. In principle it is fairly straightforward to social capital and its contribution to economic count civic associations, their membership, and growth. the number of times they meet. However, if such The term social capital has found its way into associations a.re only considered as social capital economic analysis only recently, although vari- if they achieve desirable outcomes, this implies ous elements of the concept have been present that somehow agreement must be obtained on under different names for a long time. The eco- what constitutes desirable outcomes. Further, it nomic literature on the role of institutions, which needs to be determined whether each association goes back at least to the 1920s, is especially rel- strives in effect for such an outcome. In practice evant. The focus on institutions has been revived there are likely to be major judgment and con- recently in the "new institutional economics" lit- sensus-building problems associated with con- erature. In the political science, sociological, and structing a list of desirable outcomes and the anthropological literature social capital generally effort may well not be feasible in most settings. refers to the set of norms, networks, and organi- A second and broader concept of social capi- zations through which people gain access to tal was put forth by Coleman (1988),3 who de- power and resources, and through which deci- fines social capital as "a variety of different sionmaking and policy formulation occur.' entities, with two elements in common: they all Economists have added the focus on the contri- consist of some aspect of social structure, and they bution of social capital to economic growth. At facilitate certain actions of actors-whether per- the microeconomic level this is seen primarily sonal or corporate actors-within the structure" through the ways social capital improves the (p. 598). This broadens the concept to include functioning of markets. At the macroeconomic vertical as well as horizontal associations, and level institutions, legal frameworks, and the also the behavior among other entities such as government's role in the organization of produc- v~ ~ ~~~mareooi pefr firms.4 Vertical associations are characterized by tion are seen as affecting macroeconomic perfor- hierarchical relationships and an unequal power mance. distribuhon among members. Clearly, this wider The most narrow concept of social capital is ribution among memers a wider associated with Putnam (Putnam 1993; Putnam range of associations covers a wider range of ob- and others 1993). He views it as a set of "hori- jectives-positive as well as negative. Coleman zonta assoc s b e p : sis explicit about this: "A given form of social capi- tal consists of social networks ("networks of civic tal that is valuable in facilitatin certain actions engagement") and associated norms that have an may be useless or even harmful for others" (p. effect on the productivity of the community. Two 598). In fact, this view of social capital captures empirical presumptions underlie this concept: social structure at large, as well as the ensemble networks and norms are empirically associated, of norms governing interpersonal behavior. and these have important economic conse- A third and most encompassing view of so- quences. While originally this concept of social cial capital includes the social and political envi- capital was limited to associations having posi- ronment that enables norms to develop and tive effects on development, recently it has been shapes social structure. In addition to the largely relaxed to include groups that may have unde- informal, and often local, horizontal and hierar- sirable outcomes as well, such as associations chical relationships of the first two concepts, this with rent-seeking behavior (for example, the Ma- view also includes the more formalized institu- fia in southern Italy) and even militia. The key tional relationships and structures, such as gov- feature of social capital in this definition is that it ernment, the political regime, the rule of law, the facilitates coordination and cooperation for the court system, and civil and political liberties. This mutual benefit of the members of the association focus on institutions draws on North (1990) and (Putnam 1993). Olson (1982), who have argued that such institu- This relaxation of the concept brings it a big tions have an important effect on the rate and step closer to operationalization and measure- pattern of economic development. Social Capital: The Missing Link? 79 How Does Social Capital Affect future state of the world and the response of other Economic Outcomes? agents to that state. These situations represent cases of market failure.5 Social capital, under any There is growing evidence that social capital, un- of its definitions, can contribute to alleviating der any of the three definitions, can have an im- such failures. Social capital does not remove the pact on development outcomes-growth, equity, uncertainty, but it may create mutual knowledge and poverty alleviation. Associations and insti- about how agents will respond to different states. tutions provide an informal framework to It may also serve as an enforcement mechanism organize information sharing, coordination of ac- to ensure that these expectations about mutual tivities, and collective decisionmaking. Bardhan behavior are in fact realized. This reduces con- (1995) has argued that what makes this work is tracting costs. peer monitoring, a common set of norms and lo- Group-based lending schemes-from tontine cal-level sanctions. The well-known case of the in West Africa to the Grameen Bank-work be- Grameen Bank in Bangladesh illustrates how cause members have better information on each these factors can be used to overcome the poor's other than banks do. In general, information lack of access to credit markets. However, problems can be particularly severe in capital Bardhan also warns against overstating the role markets. One response was the so-called Delib- of local institutions: local overlords may find it eration Councils in Japan and Korea, which man- easy to capture local institutions, externalities of aged competition among firms for credit and development may surpass the working area of foreign exchange in a transparent process that en- locar te couraged cooperative behavior and information localre nstituaion, andty may hav it no sharing among firms by taking away incentives revenue-raisvig capacity. Hence wnat iS neeaed for rent-seeking behavior (World Bank 1993; is a balanced view of the role of the central, state, Campos and Root 1996). The rule of law and a and local-level institutions. This suggests that well-functioning court system (elements of social three definitions of social capital are not really 7capital in its broadest definition) also contribute alternative views, but rather complementary di- to reducing uncertainty by enforcing contracts mensions of the same process (box 6.1). and thus providing a priori information to con- Ifmbeconomic tracting parties about the penalties for noncom- Information sharing. Decisions by economic npliance. In the absence of effective courts many agents are often inefficient because they lack ad- informal associations internalize this policing role equate or accurate information. In some circum- for their members. A striking example is that of stances one agent derives a benefit from relaying diamond merchants, who often trade millions of incorrect information to the other. Credit or em- dollars worth of diamonds with a handshake. ployment applications are examples of such a However, failure to deliver will irrevocably lead situation. In other circumstances "optimal" deci- to expulsion from the group-and all members sions are difficult because of uncertainty over a realize this. Unfortunately, this mechanism also Box 6.1 Three views on social capital: common features The three views on social capital broaden the concept * All imply that "desirable" social relationships and from mostly informal and local horizontal associations institutions have positive externalities. Because these to include hierarchical associations and formalized cannot be appropriated by any one individual, each national structures, such as government and the rule agent has a tendency to underinvest in social capi- of law. Yet they share several common features: tal; hence, there is a role for public support of social S All link the economnic, social, and political spheres. capitail building.- They Allhl the beliefomic, social, relandpoticalnshipser- All recognize the potential created by social relation- They share the belief that social relationships affect ships for improving development outcomes but also economic outcomes and are affected by them. recognize the danger for negative effects. Which * All focus on relationships among economic agents outcome prevails depends on the nature of the rela- and how the formal or informal organization of those tionship (horizontal versus hierarchical) and the can improve the efficiency of economic activities. wider legal and political context. 80 Expanding the Measure of Wealth works for groups pursuing "undesirable" out- The example cited at the beginning of this chap- comes-which makes criminal mafias more effi- ter relating to forest management-whereby vil- cient as well. lages and government jointly manage a common The information-sharing role of social capital resource also falls in this category. 6 is of key importance for poverty alleviation. The Associations reduce opportunistic behavior by case of mutual credit groups is one important creating repeated interaction among individuals, example. These groups permit the poor to over- which enhances trust (Dasgupta 1988). This come one of their main constraints, namely ac- "backward-looking" motivation for trust has been cess to credit. The poor also generally have discussed in the social psychology literature. limited ability to cope with risk and hence are Trust can also be "forward looking" and based on more vulnerable to income fluctuations. Access the perception. of retaliation in case of untrust- to insurance against future calamities is an im- worthy behavior. This basis for trust has been portant aspect of their survival strategy. As with the subject of game theory. It is not always nec- credit, the poor can usually not access the formal essary (though often preferred) that interaction insurance market, and local associations can pro- among individuals occurs in the same activity vide an informal substitute. Often this occurs where a risk for opportunistic behavior exists. For within a kinship context and consists of a series example, repeated credit transactions among the of mutual obligations and rights-the rural-to- members of a tontine will strengthen trust, re- urban migrant sends money back to his village duce uncertainty about repayment behavior, and of origin, but can count on food shipments from in effect reduce transaction costs overall. How- the village if he encounters hard times. A farm- ever, members of a soccer club who regularly play ers' cooperative agreeing to share resources in together may still be more inclined to lend money case of a harvest failure is another example. Do- to each other than to strangers, even though here nor agencies concerned with poverty alleviation the potential penalty is less: exclusion from the can enhance the effectiveness of their strategies soccer club may have less serious consequences by stimulating the formation of social capital as than exclusion from the tontine. A cohesive as- part of these strategies. Nevertheless, it needs to sociation creates trust7 and changes, in a way, the be pointed out that mutual credit and insurance identity and thfe objective function of the eco- groups have limitations. In particular, they are nomic agent: utility of the group is partly or vulnerable to groupwide shocks (for example, wholly substituted for personal utility, and it is weather problems), which more formal credit and the former that: is maximized. Although this can insurance mechanisms would be able to diver- occur in horizontal and vertical associations, the sify against. creation of trust and reciprocity is more likely in horizontal groups, especially those based on kin- Coordination of activities. Uncoordinated or ship or other dense networks (for example, based opportunistic behavior by economic agents can on gender, ethnicity, or caste). For that reason, also lead to market failure. This can occur as a for example, the Grameen Bank as well as many result of imperfect information (see above) but informal rotational savings and credit associa- also simply because the benefits of not comply- tions rely primarily on groups of women. ing with an agreement or an expected line of be- One implication of a trust-based functioning havior (a "norm") are greater than the expected of networks and associations is that it is helped penalty. This has been the main reason why a by stability of mnembership and damaged by mo- number of irrigation projects have failed-some bility. A development path characterized by mas- farmers could successfully deviate water to the sive rural-to-urban migration thus runs the risk detriment of others or did not contribute suffi- of eroding social capital. Rural associations may ciently to maintenance costs because formal or be thinned out by departing members and lose social means of imposing equal sharing were ab- critical density. Urban areas may not readily pro- sent. Effective social capital in the form of water vide a suitable environment for recreating these user groups can overcome such problems associations. HIowever, in some cases migrants (Meinzen-Dick and others 1995; Ostrom 1995). have formed urban groups (often along ethnic Social Capital: The Missing Link? 81 lines or by common region of origin) to share in- A caveat. A word of caution is necessary fol- formation about available jobs and to channel this lowing this discussion. Social capital is no pana- information back to their place of origin to help cea for all market failures or impediments to prospective migrants. Like migration, involun- development due to information, coordination, tary resettlement due to dams or other develop- or collective decisionmaking problems. There are ment projects can damage social capital. many examples of local associations that have Ironically, the efficiency of markets itself may made a positive difference. However, by them- selves these associations may not always matter. alsound-ermine the existencemof netorkes in te If a village lacks economic opportunities, credit long-term. Large anonymous marketsc be associations may not be able to raise incomes. more efficient than networks because the best This is simply to say that social capital-like natu- buyer or seller may not be part of the network. If ral, physical, and human capital-has limited the development path is supported by a solid value if not combined with other forms of capi- court system and contract enforcement, anony- tal. One important attribute of social capital is mous markets will replace the "named" transac- that it can make the other types of capital and tions within networks over time, with gains for their productive combination more efficient. We all participating economic agents. If one adheres return to this point in the next section. to a narrow definition of social capital, this will be registered as a decline of social capital. But in Macro-level social capital. Civic associations the broader concept the same phenomenon will promote efficient market outcomes by sharing information, aligning individual incentives with be seen as substitution of one form of social capi- group objectives, and improving collective deci- tal (the rule of law) for another (horizontal asso- sionmaking. This microfocus on markets is, how- ciations). This provides one reason why a broader ever, only part of the story. Even if social capital definition of social capital seems preferable-to is defined only at the microlevel (the narrow defi- better understand the social dynamics that ac- nition, of horizontal associations, or the interme- company economic development. diate definition, which includes hierarchical associations), it must be recognized that market Collective decisionmaking. This is a necessary outcomes are influenced by the macroeconomic condition for the provision of public goods and and political environment as well. The latter can the management of externalities to market pro- be an enabling environment, enhancing the ef- cesses. It is a basic raison d'etre for government. fect of civil associations (as was arguably the case However, just as not all government decisions are in the East Asian success stories) but the macro in the best interest of their constituencies, local environment can also damage or undo the effect and voluntary associations do not always effec- of local-level social capital. Just as it makes little tively maximize their joint utility. The extent to sense to assess a given investment project with- which they do this depends upon how well they out looking at its sector as a whole and the rel- address the information and incentive problems evant macroeconomic policy environment, it discussed earlier. One important aspect of this is makes little sense to consider local associations equity. in isolation. There is some evidence that local institutions The ways in which social capital affects mac- roeconomic outcomes have been investigated in are more effective at enforcing common agree- theconew .insiutional .eoois fraework as- ments and cooperative action when the local dis- sociated with North, and with the "aggregated tributon of assets is more equal and the benefits social capital" argument associated with Olson. are shared more equally. This then provides a The basic argument is that differences in per "local" case of how efficiency and equity go to- capita incomes across countries cannot be ex- gether: better sharing provides an incentive for plained by the distribution of productive re- better coordination in managing local public sources per capita: land and natural resources, goods, which increases productivity for everyone. human capital, and produced capital (including 82 Expanding the Measure of Wealth technology). However, countries also differ in Is It Capital? institutions and other forms of social capital and in public policies. These determine the returns The examples given so far make it clear that so- that a country can extract from its other forms of cial capital is an input into the development pro- capital. Olson argues that low-income countries cess together with the other forms of capital. cannot obtain large gains from investment, spe- However, it is also an output of this process-a cialization, and trade (even if they have a large feature it shares with human capital. Education resource base) because they lack the institutions is worth pursuing for its own sake, and a well- that enforce contracts impartially and secure educated population is an important outcome of property rights over the long run, and because successful development. Likewise, many people they have misguided economic policies. would agree that a rich network of civic associa- The importance of macro-level social capital tions and a well-functioning set of government is illustrated dramatically in some of the transi- institutions are worth having, independent of tion economies of Eastern Europe and the former their effect on future economic growth. Human Soviet Union. The sudden disappearance of gov- and social capital thus share the attribute that they erinent from many social and economic func- are simultaneously a consumption good and an tions has led to a collapse of trust and forced investment. The critical difference is that educa- people to rely on local networks and informal tion can be embodied in one individual and can associations. Richard Rose's "new democracies be acquired by one individual regardless of what barometer" has attempted to measure this phe- other people do. By definition social capital can nomenon (Rose 1995b). One effect is the with- only be acquired by a group of people and re- drawal of people from the "official" economy and quires a form of cooperation among them. the reliance upon multiple informal economies This gives social capital an inevitable public to satisfy most needs. Informal activities include good character and has implications for its pro- growing food, repairing houses, and exchanging duction (Colerrtan 1988, 1990). In particular, like help with friends. While in a well-functioning all public goods, it will tend to be underproduced market economy these activities may be a hobby relative to the social optimum unless the group or a reflection of friendship, in transition this "so- responsible for its production can internalize the cial economy" exists out of necessity. In Ukraine, externality involved. This is why horizontal as- for example, three-fourths of the households are sociations, characterized by equitable power shar- involved in such activities (Rose, 1995a). In Rus- ing among mernbers, tend to be more successful sia the transition has led to what Rose has called at generating social capital. Members are more an "hourglass society" (Rose 1995c). At the base likely to contribute because they have a better there is a rich social life, consisting of strong in- chance of obtaining their fair share of the benefit. formal networks relying on trust between friends Clearly then, in most cases it is not costless to and on face-to-face interaction. The Russian prov- produce social capital. It requires resources- erb, "A hundred friends are worth more than a especially time. The amount of social capital that hundred rubles," epitomizes the importance of will be produced is therefore in part a function of these networks. At the top there is also a rich po- the opportunity cost of time and the expected re- litical and social life among the elite, who com- turn from the social capital (that is, the extent to pete among themselves for power and wealth. which an economic agent will enjoy the public However, the links between top and base are very good that is created). Which group is best suited limited and are characterized by civic distrust by to producing social capital depends largely on the the base. Fewer than one in three Russians ex- scope of the created externality and thus the size pect fair treatment by the police or their munici- of the group needed to internalize it effectively pal office (the post office is the most trusted and avoid free r iders: in the case of tontine it is institution). It is difficult to conceive of sustain- local; in the case of the rule of law it is national able economic growth in Russia without a change and the central government needs to play the es- for the better in the linkages between micro- and sential role. However, as an externality, social macro-level institutions. capital can come to be a by-product of other pro- Social Capital: The Missing Link? 83 duction processes. For example, if the workers historically when human capital accumulation in a factory develop associations for leisure ac- started: initially, when education was an elite tivities, which in turn reduce crime in the neigh- privilege, it led to increased inequality of eco- borhood, the "cost" of bringing people together nomic outcomes. The more the acquisition of and developing trust is largely absorbed in the education became universal, the more the distri- factory's production process. Certainly, the cost bution of economic benefits became equalized. is less than if those associations were to be cre- Although this has not yet been demonstrated ated by people not knowing each other through empirically, the same process is likely to happen a common workplace. with social capital. The more widespread it is, While there is a growing body of empirical the more likely it will contribute to achieving eq- evidence on the benefits of social capital (see next uity. section), there are very few data on the cost side. Although, as we have argued, social capital In part this has to do with the difficulties of mea- can promote and sustain economic development suring social capital. It is easier to determine the in many ways, it should be emphasized that the cost of a machine or building, because it is well historically and cross-sectionally strong correla- defined, than the cost of creating an association tion between human capital acquisition and or a certain level of trust. However, investment levels of development has not yet been demon- decisions for social capital, like any form of capi- strated empirically for social capital. No coun- tal, require a comparison of costs and benefits. try has achieved sustained economic growth This comparison is done implicitly by the indi- without high levels of education, but some highly viduals who set up or join an associatiort, but developed economies have low and arguably de- governments, donors, or other institutions who clining levels of social capital, as measured, for wish to invest in social capital may well want to example, through rising crime rates, declining make more explicit calculations. As a first step family and kinship cohesion, and falling trust in this will require improved operational definitions government, and participation in the political of social capital. process. Analysis by the Inter-American Devel- The process of producing economic growth opment Bank suggests that in some Latin Ameri- requires the combination of different types of can countries the paths of economic development capital. Social capital is one of them, but it has a have so weakened prior social and institutional unique feature in that it also enhances the effi- relationships that the resulting potential for so- ciency of the combination process itself. In cial conflict may undermine sustainable economic Putnam's (1993) words: "Social capital enhances growth. Such situations underline the need for the benefits of investment in physical and human interventions aimed at offsetting eroding social capital" (p. 36). In other words it is not just an capital and creating new forms of social capital input into the production function, but it is also conducive to sustainable development. a shift factor (or exponent) of the entire produc- tion function. As such it is more akin to technology. Measuring Social Capital and Its Impact The application of social capital in develop- ment is not a distribution-neutral process. The Measurement presupposes that one can define better organized segments of society may well fairly well what needs to be measured. As the succeed in affecting economic policy to their own previous section indicated, definitions of social advantage and to the detriment of other groups capital vary greatly. This makes it inherently dif- or even to society at large. There can be signifi- ficult to propose a list of indicators for social capi- cant concentrations of social capital in some com- tal. They will have to evolve as the conceptual munities with few ties to other communities. definition and, more important, the operational Social capital accumulation can be segmented definition of social capital are developed. Mean- along spatial or ethnic lines (Fox 1995). In prin- ingful use of indicators requires a conceptual ciple there is thus no guarantee that enhancing framework within which they can serve to assess social capital will lead to a more equitable soci- a current state, to measure linkages between ety. This is similar, of course, to what has happened policy and outcome variables, and to assess policy 84 Expanding the Measure of Wealth Box 6.2 Desirable properties of indicators Work is currently under way at the World Bank to collect data for a profile of local institu- The following is a nonexhaustive list of proper- tions in twenty countries. The context is how de- ties that indicators should possess. Indicators centralization can help in implementing effective must: rural development strategies and whether it re- * Be developed within an agreed on conceptual suits in better targeting to the poor. The study and operational framework s * Be clearly defined and easy to understand aims to collect data on the institutions that func- . Be subject to aggregation (from household to tion at the local level in the provision of various community, from community to nation) services (health, education, agricultural exten- * Be objective (be independent of the data collec- sion, water supply, forestry). These institutions tor) include nongovernental organizations (NGOs) * Have reasonable data requirements-either available data or data that can be collected at as well as the local political and administrative limited cost and within the capacity of the organization. In three countries (Bolivia, Burkina country's statistical apparatus Faso, Indonesia) this profile will be further de- * Have "ownership" by users tailed and supplemented with data at the house- * Be limnited in number hold level. Specifically, the extent of household * Reflect input, process, or outcome (or, as used . . in the environment literature, pressure, state, participation i different local organzations, and response). the way in which the organizations contribute to different dimensions of household well-being will be explored through a household survey. In or- options (box 6.2). Such a framework is further der to find out which associations have the most advanced for the narrower definitions of social impact on development outcomes and how they capital (micro-institutional in focus) than for the do so, detailed information on the associations broader definitions, which attempt to link insti- will be collected: type of organization (formal or tutions at the macrolevel with economic out- informal), degree of internal homogeneity (by comes. One can thus expect to define a more gender, ethnicity, occupation), membership re- suitable set of indicators for the notions of social quirements, and the type of services provided. capital that focus on horizontal and hierarchical The availability of measures of economic and local associations. Of course this does not imply social outcomes, at both the village and the house- that actual data exist more abundantly at that hold level, will make it possible to test empiri- level. Practically, the selection and development cally hypotheses regarding the impact of social of indicators for social capital can proceed along capital on poverty, effectiveness of public pro- two lines: (1) according to the breadth of relation- grams and projects, and access to health, educa- ships and institutions involved; and (2) accord- tion, and credit. ing to the types of impact social capital has on For the intermediate conceptualization of so- the development process, in which the key dimen- cial capital, which includes hierarchical associa- sions are growth, equity, and poverty alleviation. tions, the same indicators of associations can be For the narrowest concept of social capital, one used, except that they now cover a wider range can inventory civic associations and their at- of social relationships. In the macro concep- tributes (number of members; frequency of meet- tualization, legal and judicial systems and aspects ings; dimensions of membership along ethnic, of government functioning (such as the ability to kinship, or other lines; type of decisionmaking). enforce contracts) all become part of social capi- In practice such inventories are rarely available, tal. This approach solves, in part, the measure- and one typically gains insights into them only ment problem because the social norms and through anthropological or sociological case stud- networks are anchored in an institution or other ies, which tend to have a limited geographical "visible" structure that can be identified. focus. The value of such studies lies primarily in Clearly, to capture the full scope of social capi- analyzing the dynamics of creating associations tal, measurement has to occur and indicators (how, why, and by whom are they created?) and need to be developed at the micro and their effectiveness.As such they are a complemen- macrolevels. At the microlevel the impact of so- tary effort to that of inventorying associations. cial capital is to be assessed primarily by the ex- Social Capital: The Missing Link? 85 tent to which the association or institution con- There is much evidence that local associations tributes to making more efficient market out- play a key role in environmental management, comes possible (by reducing information or especially where common property resources incentive problems) or to providing the "opti- (water, forests) are concerned. This has been es- mum" amount of public goods (by making pecially well documented in the case of irriga- collective decisionmaking more efficient). Obvi- tion and water supply projects. In Cote d'Ivoire, ously, it is very difficult to measure this empiri- for example, rural water supply improved sig- cally since so many different factors affect market nificantly after responsibility for maintenance outcomes simultaneously, but this problem is re- was shifted from the national water distribution ally external to the definition and measurement company to community water groups. Break- of social capital per se. An analogy: it is possible down rates were reduced from 50 percent to 11 to measure land even if it is difficult to measure percent at a third of the cost. However, these re- how much the land contributes to value added sults were sustained only in villages with a high in agriculture. demand for water and where well-functioning Significant and growing evidence exists that community organizations already existed (Hino local associations and networks have a positive 1993). Experiences with water user associations impact on local development and the well-being in countries as diverse as Pakistan and the United of households. Work in India has shown that such States, have indicated that sustainability depends social capital enhanced the ability of the poor to on empowering of the participating farmers to allocate resources efficiently and increased their negotiate with the relevant water authority and resilience to hazards (Townsend 1994). A study the installation of a framework that clearly sets of 750 households from 45 villages in Tanzania out rights and benefits as well as duties and re- suggests that social capital makes a significant sponsibilities (Meinzen-Dick and others, 1995; contribution to household welfare. Social capi- Narayan 1995; Ostrom 1995). Because the tal was measured by membership in groups and formation and maintenance of a water user group networks. Multivariate regression analysis estab- (or any group) demands effort and resources from lished that village-level social capital was a key the participants, the sustain- ability of a group is contributor to household welfare even after tak- further enhanced when the stakes are high, as in ing into account the size of household, male the case of C6te d'Ivoire, where the demand for schooling, female schooling, household assets, and value of water are high. market access, and agro-ecological zone. In some Regarding forests, we cited earlier the role cases, the effect of village-level social capital out- played by community groups in Gujarat to end weighed that of market access or female school- violence over forest management issues and to ing. On the other hand household-level social develop a joint approach between government capital appeared to be less significant than vil- and local people (Pathan and others 1993). In a lage-level social capital (Narayan and Pritchett remote area of Zimbabwe a small community 1996). took over management of wildlife resources. Social capital can also improve the quality of They negotiated the revenue-sharing process and education. A study of U.S. schools showed lower division of responsibilities with government. The dropout rates in religious schools in tight com- results have been better wildlife protection and munities than in other public and private schools, increased revenue from safaris and tourism for even after controlling religion and household fi- both the local community and the government nancial position (Coleman 1988). In a similar vein (Scoones and Matose 1993). As in the case of wa- the breakdown of networks can lead to crime and ter associations, the keys to success include clear violence. A study of urban communities in Ecua- rules of membership, accountability, and sanc- dor, Hungary, the Philippines, and Zambia tioning developed jointly between the local com- showed that depletion of economic assets led to munity and the central authority (Narayan 1995). lower involvement in community organizations, The links between civil society and govern- weaker informal ties among residents, and in- ment also affect the outcomes of government pro- creased crime and violence (Moser 1996). grams. A study of municipal government in 86 Expanding the Measure of Wealth northeast Brazil showed how the creation of re- zations in Bolivia succeeded in pushing up pro- lationships between civil servants and local ducer prices in local markets so that members as associations enhanced the effectiveness of munici- well as nonmembers benefited (Tendler and oth- pal programs. When the staff of a Cerea health ers 1982). This impact depends, however, on the program focused on building trust with clients nature of the organization. In Bolivia and Ecua- in the communities in which they worked, the dor rural organizations enhanced the poor's ac- quality and impact of the program increased- cess to markets and government institutions, but more families were served, and infant mortality the membership principles and barriers to entry declined (Freedheim 1988). of each organization influenced who did and did The garment industry in Brazil and Chile is not benefit from the links with government agen- an example of the role of professional associations cies (Bebbington 1996). More generally, Olson has in solving information and incentive problems shown how strong lobbying organizations can (Stone and others 1992). Brazil has a complex benefit their own members, but can have adverse regulatory system, laws are sometimes inconsis- impacts on economic development through spe- tent with one another, and courts are very expen- cial interest g;roup influence on policymaking. sive. For day-to-day operations business has to Evidence on the positive impact of hierarchi- rely on informal alternatives to govern transac- cal associations is more limited and ambiguous. tions with customers or suppliers, especially In Nigeria such groups reduced food insecurity when credit is involved. Brazilian garment en- by giving poor people moral entitlements on trepreneurs worked out an effective informal which to draw during famine years (Watts 1983). credit information system, which places a pre- In the case. of Boosaaso, Somalia, cited at the be- mium on an untarnished reputation. Neverthe- ginning of this chapter, hierarchical clan-based less, contracts remain insecure and are frequently and intergenerational relationships became the renegotiated, even up to the very moment of de- means by which local stability and security were livery. Brazilian entrepreneurs must therefore ensured, which resulted in a big upturn of eco- adopt risk-reducing strategies, such as produc- nomic activity. ing only noncustomized items and reducing the At the macrolevel social capital becomes the size of orders, which ultimately hurts the expan- fourth category of capital in the production func- sion of business. In contrast, in Chile legal sim- tion (with physical, natural, and human capital). plicity and consistent enforcement of contracts Its contribution to economic growth, investment, have led to a more secure contracting process and or equity can then be assessed in two ways. A very few renegotiations. This has reduced the first approximation is obtained from accounting- default rate on debt. This case study clearly sug- type production function models, which explain gests that there is a limit to the extent to which GDP growth as a function of growth of labor, capi- informal associations can replace the rule of law tal, and technology. After accounting for physi- and a formal court system. This underlines both cal and natural capital, a "residual" is obtained the role of social capital in making business pos- that lumps together social and human capital. sible and the role of government in providing an Separating social from human capital requires a enabling environment. Simplicity, transparency, direct estimation of human capital. This has not and consistency need to be the key features of yet been done successfully. The growth account- this environment. A study of Peru further illus- ing models routinely rely on enrollment figures, trated how the sheer complexity of laws and regu- but this has been criticized. In principle the ad- lations can undermine their effectiveness and vantage of the residual approach is that it identi- provide strong disincentives to economic agents fies the contribution of social capital in its entirety to adhere to formality. In Peru it led to the shift- (at least if one accepts the assumption that growth ing of economic transactions to an informal sec- is a sole function of the four identified factors of tor not protected by formal law, but functioning production). An example of this approach is the thanks to informal substitutes (de Soto 1989). earlier cited jEast Asian Miracle study, which Research also shows that the impact of social found that growth accounting models could only capital can be indirect. Rural economic organi- explain 17 to 36 percent of the difference in Social Capital: The Missing Link? 87 growth performance between EastAsia and other financed projects showed that in countries with parts of the world (World Bank 1993).S the best civil liberties the economic rate of return An alternative method is the direct estimation of these projects was significantly higher than in of the impact on growth, investment, or equity the countries with the worst civil liberties, after of specific components of social capital. There is controlling for a variety of other determinants of a rapidly growing body of literature that has at- project performance. However, political regimes tempted to do this, often focusing on the politi- (democracy versus non- democracy) and politi- cal or democratic aspects of society. Indicators cal liberties did not play a significant role in include measures of political instability (govern- project performance (Isham and others 1995). ment changes, coups); Gastil's measures of civil Economists often argue that government's and political liberty; measures of expropriation main economic role is the enforcement of prop- risk, corruption, enforcement of contracts and erty rights and the management of externalities property rights; and measures of political and to economic processes (including public goods). economic discrimination and social disintegration It stands to reason, therefore, that economic (crime, suicide, riots, illegitimacy, divorce, and so growth is likely to be hampered when these func- on). Box 6.3 contains a partial list of such vari- tions are not properly undertaken. Some of the ables, which have been used in cross-country transition economies are a case in point, because analyses and are therefore available for a num- they have not yet had sufficient time to establish ber of countries. a well-functioning and well-integrated legal One study of twenty nine countries included framework and judicial system to enforce con- direct measures of trust and civic cooperation tracts and property rights. Likewise, the manage- (taken from the World Values Survey) in a cross- ment of externalities is more of a challenge in a country growth equation, and found that each market economy than in a controlled economy variable had a significant positive effect, after con- (where, by definition, they are internalized since trolling for other determinants of growth. The the government controls the economic processes importance of trust was found to be greatest in that generate them). Some empirical studies have low-income countries, where it is assumed to found that variables measuring contract enforce- operate as a substitute for formal institutions that ment, expropriation risk, corruption, and qual- enforce property rights and contracts (Knack and ity of government bureaucracy are powerful Keefer 1996). explanatory factors of growth rates, in some cases Analyses of cross-country data have also been with effects as strong as those of education (Knack undertaken in which the regression model in- and Keefer 1995). Furthermore, countries with cludes the political regime as an explanatory vari- formal institutions that effectively protect prop- able. Civil and political liberties were found to erty and contract rights provide a more condu- have a positive correlation with growth, while cive environment for trust and civil cooperation repressive regimes yield lower growth rates to develop (Knack and Keefer 1996). (Scully 1988; Grier and Tullock 1989; Barro 1989). Other studies have gone beyond the political Political instability has also been associated em- and government factors and looked at the effects pirically with lower growth. However, these cor- of social integration and disintegration on eco- relations have failed to support conclusively a nomic performance. Again, the transition econo- causal model and competing theories abound. mies of Eastern Europe and the former Soviet For example, it has been argued that freedom en- Union provide some powerful evidence of the hances market efficiency and economic perfor- catastrophic interaction that can occur between mance, while others claim that rapid growth economic and social decline. The trouble is that requires controls and reduced freedom. Still oth- here too, to cite Klitgaard and Fedderke (1995), ers see freedom affecting growth mairly through "there is no agreed upon theory to apply." Hence, investment behavior (de Haan and Siermann investigations are largely inductive, seeking for 1996). Political factors can influence not only na- meaningful correlation. The evidence so far seems tional economic performance, but also that of to indicate that good economic performers score development projects. A study of World Bank- higher on some dimensions of social integration 88 Expanding the Measutre of Wealth Box 6.3 Indicators of social capital The following indicators have all been used in empirical studies. Indicators of horizontal associations take a microperspective and typically have been collected for analysis within a country. The other sets of indicators have been calculated at the national level and have been used in cross-country research Horizontal associations Number and type of associations or local Extent of trust in trade unions institutions Perception of extent of community i Extent of membership organization * Extent of participatory decisionmaking Reliance on networks of support * Extent of kin homogeneity within the Percentage of household income association from remittances * Extent of income and occupation homogeneity Percentage of household expenditure within the association for gifts and transfers * Extent of trust in village members and Old-age dependency ratio in households government Civil and political society * Index of civil liberties (Gastil, Freedom House) Index of democracy * Percentage of population facing political Index of corruption discrimination Index of government inefficiency * Index of intensity of political discrimination Strength of democratic institutions -Percentage of population facing economic -Measure of "human liberty" discrimination Measure of political stability Index of intensity of economic discrimination Degree of decentralization of govemment Percentage of population involved Voter turnout in separatist movements Political assassinations * Gastil's index of political rights Constitutional government changes * Freedom House index of political freedoms Coulps Social integration * Indicator of social mobility Other crime rates * Measure of strength of "social tensions" Prisoners per 100,000 people * Ethnolinguistic fragmentation Illegitimacy rates * Riots and protest demonstrations Percentage of single-parent homes * Strikes Divorce rate * Homicide rates * Youth unemployment rate * Suicide rates Legal and governance aspects * Quality of bureaucracy Repudiation of contracts by governnent Independence of court system Contract enforceability * Expropriation and nationalization risk Contract-intensive money (currency / M2) and stability of social institutions but not on oth- capital, and dissolves social capital. More than ers. Those aspects of social integration linked to half the world's low-income countries have ex- the political process seem to matter most, which perienced conflict during the past ten years. Of is in line with the previously discussed studies those countries, thirty have had more than 10 focusing explicitly on the political regime percent of their population dislocated, and in ten (Klitgaard and Fedderke 1995). countries 40 percent of the population has been One of the most disruptive forms of social dis- dislocated. Ethnicity-based conflict is emerging integration is ethnic conflict. It destroys physical as a key factor contributing to Africa's economic capital, disrupts the economy, deteriorates human decline. Studies have suggested that ethnic di- Social Capital: The Missing Link? 89 versity may lead to increased civil strife and po- the standard GDP aggregate will suggest. Useful litical instability. Ethnically fragmented societ- lessons can be learned from the environment ac- ies are prone to competitive rent-seeking behavior counts for the valuation of social capital. It is by the different ethnic groups and have difficulty likely that the construction of a satellite account- agreeing on public goods like education, infra- ing system for social capital would present the structure, and good policies. The role of ethnic most feasible way of linking social capital to the diversity in affecting growth performance has SNA. The key account is the capital account, been quantified in a cross-country study of Sub- which must meet the following identity: Saharan Africa (Easterly and Levine 1995). The study controlled for a wide array of growth de- Closing stock opening stock terminants, such as initial income level, school- + production of capital ing; political stability, and monetary, fiscal, trade, - consumption of capital exchange rate, and financial sector policies. It also ± revaluation of capital. included measures of infrastructure develop- ment, cultural diversity, and economic spillovers The latter account adjusts the value of assets fromneighbors' growth. Yet the studyfound that due to price changes. Progress on the environ- ethnic diversity independently accounted for ap- mental accounts was only made possible after proximately 35 percent of Africa's growth differ- inventories of national capital stocks were mea- ential with the rest of the world. When the sured and progress was made on valuing differ- indirect effects on policies were also considered, ing types of natural capital. The same lines of this figure rose to 45 percent. ' inquiry will be needed for social capital before thife iniaors to social percaita used in these advances will be possible toward an integrated the indicators of social capital used in these syte of soiladeooi°conig studies all represent quantitative or qualitative system of shold beoinomi aountinug measures of social capital, but without any at Finally, it should be pointed out that a num- measua ,- ber of authors purposely blur the distinction be- tempt at direct valuation. Only the "residual tween social and human capital-taking both to approach derived from growth accounting mod- be embodied in people and hence using indica- els is an indirect valuation method. While we tors about people as opposed to measures about have argued that social capital affects market out- institutions or associations. The context is the comes and macroeconomic outcomes, the absence concern with social welfare or human develop- of a market for social capital, due to its public ment. In this context many conventional human good character, makes valuation inherently dif- capital indicators are given social connotations. ficult.' However, valuation is inevitable if social For example the U.S. President's Council on Sus- capital is to be seen as a final or intermediate eco- tainable Development (1996) lists access to edu- nomic output and to be included in the national cation and health care as indicators of social accounts. One of the innovations of the 1993 equity, and the percentage of population attend- United Nations System of National Accounts ing college as a measure of participation in deci- (SNA) was the inclusion of satellite accounts that sionmaking. Likewise the United Nations incorporate the depletion of natural resources and Development Programme's Human Develop- selected environmental costs. This makes pos- ment Index has been used as a measure of social sible the calculation of "environmentally adjusted development. In effect, social indicators are used net national income" and so-called "genuine sav- (almost) synonymously with social capital indi- ing" (the residual of production minus consump- cators. This approach, of course, enhances greatly tion, depreciation of produced assets and the number of available measures of social capi- drawing down of natural resources). The key tal, but this direction of operationalizing social purpose is to recognize that GDP will overstate capital is not consistent with the definitions of economic gain if output is achieved by depleting social capital discussed earlier in this chapter. natural capital (Serageldin 1996a, 1996b; Role of Donors Hamilton and Lutz 1996). Similarly, if economic growth severs social relations and depletes so- It has been argued in this chapter that certain cial capital, "genuine" growth will be lower than forms of social capital can have strong positive 90 Expanding the Measture of Wealth effects on economic growth and can contribute to Create Enabling Environments creating sustainable development. At the same time warning was given that an inappropriate The scope for effective use and strengthening of path of development can destroy social capital, social capital depends critically on the nature of setting off a vicious circle of social and economic the wider political and policy environment. The decline. There is thus clearly a role for govern- latter can encourage or discourage local organi- ment in promoting "desirable" forms of social capi- zation and provide incentives or disincentives for tal. The public good nature of social capital further people to participate. An enabling environment underlines this role, as does the fact that the func- is characterized by general good governance, en- tioning of government itself is a part of social capi- forcement of property rights, an independent ju- tal in its broadest sense. dicial system, a competent and transparent It is logical therefore to ask what the role is of bureaucracy, and mechanisms to promote dia- donor agencies that are concerned with develop- logue and resolve conflict among economic ment-national as well as international, public agents. and private-in creating or stimulating the cre- ation of social capital. The World Bank has Invest in Social Capital recently examined its own potential role, conclud- ing that it needs to increase the extent to which it Direct investment in social capital means direct takes social relationships and local and national support to exicting and emerging organizations. networks and institutions into account in its In practice, nongovernmental organizatons and project design and policy advice. Five areas were local government may often be in the best posi- identified for action. tion to do so, given that most civic associations are small and local. National or international do- Do Your Homework, Do No Harm nors can support international NGOs and confed- erations of local associations. Participatory The Bank needs to better understand existing so- processes in project design often contribute to cial capital in a country prior to developing poli- social capital building by inducing the formation cies and as part of the process of project design. of local user groups. Assessments of social capital could be combined with poverty and social assessments, and would Promote Research and Learning identify existing institutions, social relationships, and networks that contribute to growth and pov- As this chapter illustrated earlier, the measure- erty alleviation, and also those that impede it. ment of social capital and the empirical assess- Such assessments would prevent projects from ment of its contribution to achieving growth and weakening existing positive social capital and sug- equity objectives is only just beginning. Such gest ways to strengthen it. work, including new data collection, needs to be promoted further. Research is also needed on the Use Local-level Social Capital to Deliver Projects most appropriate strategies for working with lo- cal organizations. Finally, the micro-macro link- Existing associations and organizations can be age between social capital and macro- economic called upon to take part in the delivery of a project. performance needs to be explored further concep- This has the potential to improve beneficiary tar- tually as well as empirically. geting, reduce project cost, and enhance the sus- tainability of projects by increasing "ownership." Notes Furthermore, the participating institutions can become strengthened due to their involvement, This chapter draws heavily from the work and thus enhancing social capital. This will require report of the Satellite Group on Social Capital of care in the selection of organizations so that they the Social Development Task Force, which con- are truly inclusive of the intended beneficiaries vened regularly at the World Bank in the spring (especially the poor) and have objectives in line of 1996. It also incorporates some of the discus- with the project. sions of the first Social Capital Workshop, orga- Social Capital: The Missing Linzk? 91 nized by the Environmentally Sustainable Devel- Others, relying on endogenous growth models, opment vice presidency of the World Bank on argue that the low-income inequality that char- April 17-18, 1996. Intellectual debt to the partici- acterized the East Asian economies was in itself pants of both working groups is herewith ac- an important stimulus for growth (Birdsall and knowledged. Special thanks are due to Gloria others 1995). For a recent review and interpreta- Davis, John Dixon, Philip Keefer, Deepa Narayan, tion of the evidence see Stiglitz (1996). Deavis Over, Ismail Serageldin, Zmarak Shalizi 9. One aspect of social capital is routinely valued Mead Over, Ismail Serageldin, Zmarak Shalizi, and traded in the market, namely, the "goodwill" and Alison Van Rooy for very helpful comments of a corporation. Essentially, goodwill is the net- and suggestions on an earlier draft. work of clients and suppliers of the corporation. 1. It is not our objective, however, to provide a full literature review. References 2. Some social scientists claim that the term social capital has been coined only to make the under- Bardhan, P. 1995. "Research on Poverty and Devel- lying concepts acceptable to economics. Econo- opment Twenty Years after Redistribution with mists reply that institutions and other aspects of Growth." In M. Bruno and B. Pleskovic, eds.,An- social capital have always been present in eco- nual World Bank Conference on Development Eco- nomic analysis. nomics 1995. Washington, D.C.: World Bank. 3. Coleman has been credited with introducing the Barro, R. 1989. "A Cross-Country Study of Growth, term social capital into the sociological literature Saving and Government." National Bureau of in his 1988 article, "Social Capital in the Creation Economic Research Working Paper 2855. Cam- of Human Capital," American Journal of Sociol- bridge, Mass. ogy. However, Loury introduced the concept of Bebbington, A. Forthcoming. "Organizations and social capital into economics in 1977 in an analy- Intensifications: Small Farner Federations, Ru- sis of racial inequality, to describe the social re- ral Livelihoods, and Agricultural Development sources of ethnic communities. in the Andes and Amazonia." World Develop- 4. This concept of social capital leans closely to the ment. treatment of firms and other hierarchical orga- Birdsall, N., D. Ross, and R. Sabot. 1995. "Inequality nizations in institutional economics, where the and Growth Reconsidered: Lessons from East purpose of the organization is seen as to mini- Asia." The World Bank Economic Review 9 (3): mize transaction costs (Williamson 1985, 1993). 477-508. 5. Applications of game theory to economics, as Bradley, P., and K. McNamara, eds. 1993. Living with well as the literature on the economics of infor- Trees: Policies for Forestry Management in Zimba- mation, have attempted to address such situa- bwe. World Bank Technical Paper 210. Washing- tions. Bkeilp20 an 6. It needs to be pointed out that the existence per ton, D. C. se of a group (such as a water user group) does lem.1" Washington Post. not imply social capital. It is the mechanism un- lem.1 Wasinto Post. derlying the group's ability to enforce group Coleman, J. 1988. "Social Capital in the Creation of normns that constitutes the social capital in that Human Capital."American Journal of Sociology. 94 setting. (Supplement) S95-S120. 7. According to Putnam, it is preferable to speak .1990. Foundations of Social Theory. Cambridge, about trustworthiness rather than trust. Mass: Harvard University Press. Dasgupta considers that trust is merely confi- Campos, J., and H. Root. 1996. The Key to the Asian dence in an expected outcome (based either on Miracle: Making Shared Growth Credible. Wash- past experience or on anticipated contract en- ington, D. C.: Brookings Institution. forcement). For further discussion of trust see Dasgupta, P. 1988. "Trust as a Commodity." In Gambetta (1988) and Fukuyama (1995). Gambetta (1988). 8. There are dissenting views on the role of social de Haan, J., and S. Siermann. 1996. "Political Insta- capital in the explanation of East Asian growth bility, Freedom, and Economic Growth: Some rates. Some authors argue that most or even all Further Evidence." Economic Development and of these growth rates can be explained by in- Cultural Change 44: 339-50. creased mobilization of resources (increases in de Soto, H. 1989. The Other Path: The Invisible Revo- labor force participation rates, education, and lution in the Third World. New York: Harper and investment in physical capital) (Krugman 1994). Row. 92 Expanding the Measure of Wealth Easterly, W., and R. Levine. 1995. "Africa's Growth Moser, C. 1996. Confronting Crisis: A Summary of Tragedy: A Retrospective, 1960-89." Policy Re- Household Responses to Poverty and Vulnerability search Working Paper 1503. World Bank, Policy in Four Poor Urban Communities. Environmentally Research Department, Washington, D. C. Sustainable Development Studies and Mono- Fox, J. 1995. "How Does Civil Society Thicken? The graphs no. 7. Washington, D.C.: World Bank. Political Construction of Social Capital in Rural Narayan, D. 1995. "Designing Community-Based Mexico." Paper presented to the Government Development." Environment Department Paper Action, Social Capital Formation, and Third 7. World Bank, Environment Department, Wash- World Development Conference, May 5-6. Cam- ington, D. C. ord ge, Mass. Narayan, D., and L. Pritchett. 1996. "Cents and So- Freedheim, S. 1988. "Why Fewer Bells Toll in Ceara: ciabiity: Household Income and Social Capital in Rural Tanzania." World Bank, Environment Success of a Community Health Worker Program in Ceara, Brazil." Master's thesis, Massachusetts Department, Washington, D. C. Institute of Technology, Cambridge, Mass. North, D. 1990. Institutions, Institutional Change, and Insttute of Technology, Camriluge, M dtass. Economic Performance. New York: Cambridge Fukuyama, F. 1995. Trust: The Social Values and the University Press. Creation of Prosperity. New York: Free Press. Olson, M. 198'2. The Rise and Decline of Nations: Eco- Gambetta, D., ed. 1988. Trust: Making and Breaking nomic Grozvth, Stagflation, and Social Rigidities. Cooperative Relations. Oxford: Blackwell. New Haven: Yale University Press. Grier, K., and G. Tullock. 1989. "An EmpiricalAnaly- Ostrom, E.1995. Incentives, Rules of the Game, and sis of Cross-National Economic Growth, 1951- Development." In M. Bruno and B. Pleskovic, 80." Journal of Monetary Economics 24: 259-76. eds.,Annual Bank Conference on Development Eco- Hamilton, K., and E. Lutz. 1996. "Green National nomics 1995. Washington, D. C.: World Bank. Accounts: Policy Uses and Empirical Experi- Pathan, R., N. Arul, and M. Poffenberger. 1993. "For- ence." Environment Department Paper 39. est Proteci:ion Committees in Gujarat-Joint World Bank, Environment Department, Wash- Management Initiatives." Reference Paper 8. Pre- ington, D.C. pared for Sustainable Forest Management Con- Hino, T. 1993. "Community Participation in ference, sponsored by the Ford Foundation, 'Programme de Restructuration de Delhi. l'Hydraulique Villageoise' in C6te d'Ivoire." The President's Council on Sustainable Develop- World Bank, Africa Country Department 11, ment. 1996. Sustainable America: A New Consen- Washington, D. C. sus for Prosperity, Opportunity, and a Healthy Isham, J., D. Kauffman and L. Pritchett. 1995. "Gov- Environmentfor the Future. Washington, D.C. ernance and the Returns to Investment: An Em- Putnam, R. 1993. "The Prosperous Community- pirical Investigation." Policy Research Working Social Capital and Public Life." American Pros- Paper 1550. World Bank, Policy Research Depart- pect (13): 35-42. ment, Washington, D. C. Putnam, R., with R. Leonardi, and R. Nanetti. 1993. Klitgaard, R., and J. Fedderke. 1995. "Social Inte- Making Dernocracy Work: Civic Traditions in Mod- and Disintegration: An Exploratory ern Italy. Princeton: Princeton University Press. gration and Crss try Da W orld Rose, R. 1995a. "Adaptation, Resilience, and Desti- Analysis of Cross-Country Data." World Devel- tution: Alternative Responses to Transition in opment 23 (3): 357-69. Ukraine." Problems of Post-Communism (Novem- Knack, S., and P. Keefer.1995. "Institutions and Eco- ber/Decem.ber): 52-61. nomic Performance: Cross-Country Tests Using _- 1995b. "New Russia Barometer IV-Survey Re- Alternative Institutional Measures." Economics sults." Studies in Public Policy 250. University and Politics 7 (3): 202-27. of Strathclyde, Glasgow. .1996. "Does Social Capital Have an Economic - 1995c. "Russia as an Hour Glass Society:ACon- Payoff? A Cross-Country Investigation." World stitution without Citizens." East European Con- Bank, Policy Research Department, Washington, stitutional Review 4 (3): 34-42. D. C. Scoones, I., and F. Matose. 1993. "Local Woodland Krugman, P. 1994. "The Myth of Asia's Miracle." For- Management: Constraints and Difficulties for eign Affairs 73 (6): 62-78. Sustainable Resource Use." In Bradley and Meinzen-Dick, R., R. Reidinger, and A. Manzardo. McNamara (1993). 1995. "Participation in Irrigation." Enviromnent Scully, G. 1988. "The Institutional Framework and Department Paper 3. World Bank, Washington, Economic Development." Journal of Political D. C. Economy 96: 652-62. Social Capital: The Missing Link? 93 Serageldin, I. 1996a. Sustainability and the Wealth of Na- Tendler, J., and others. 1983. What to Think about Co- tions: First Steps in an Ongoing Journey. Environ- operatives: A Guide from Bolivia. Rosslyn, Va.: mentally Sustainable Development Studies and Inter-American Foundation. Monographs no.5. Washington, D.C.: World Bank Townsend, R. 1994. "Risk and Insurance in Village . 1996b. "Sustainability as Opportunity and the India." Econometrica 62: 539-91. Problem of Social Capital. " Brown Journal of World Watts, M. 1983. Silent Violence: Food and Famine in North- Affairs 3 (2): 187-203. ern Nigeria. Berkeley: University of California Press. Stiglitz, J. 1996. "Some Lessons from the East Asian Williamson, 0. 1985. The Economic Institutions of Miracle. " The World Bank Research Observer 11 (2): Capitalism: Finns, Markets, Relational Contracting. 151-77. New York: Free Press. Stone, A., B. Levy, and R. Paredes. 1992. "Public In- . 1993. "The Economic Analysis of Institutions stitutions and Private Transactions: The Legal and Organizations-In General and with Re- and Regulatory Environment for Business Trans- spect to Country Studies." Economics Depart- actions in Brazil and Chile." Policy Research ment Working Paper 133. Paris: OECD. Working Paper 891. World Bank, Policy Research World Bank. 1993. The East Asian Miracle. New York: Department, Washington, D. C. Oxford University Press. CHAPTER 7 Poverty and the Environment: Pieces of the Puzzle 'T A.. that is the relationship between pov- Environmental damage tends to affect the \/ \' erty and the environment? Does pov- poor particularly severely for several reasons. erty lead to environmental damage? First, they tend to rely heavily on fragile natural If so, in what ways? Do environmental problems resources for their livelihood. Poor people, almost exacerbate poverty? If so, to what degree? How by definition, have few assets. Much of their do poor people respond to environmental prob- wealth, therefore, derives from their own labor lems? What policy instruments are available to and any natural capital that they have access to. address poverty and environmental problems to- Since they often have little education, their hu- gether? To what degree are there synergies be- man capital tends to have a low return, making tween addressing poverty and environmental any natural capital all the more important. How- problems, and to what degree are there trade-offs? ever, the poor often only have access to poor-qual- Although much has been written about the rela- itv and fragile natural resources, and their claim tiyand the envirounment, to these resources is often tenuous. Poor people tionship btwee poerty also tend to have fewer ways of responding to ansowfaremostdof.these questionsremainlarg - environmental problems. Polluted areas where asinswerest in the linkages between poverty land is cheap or free may be the only areas they As interest in the linkages between poverty can afford to live in. They are also less likely to and environment and awareness of their impor- have the resources to buy bottled water when tance has grown, partial and preliminary answers sewage and industrial waste are polluting the to many of these questions have begun to emerge. drinking water. This chapter reports on the results of several re- The poor are also thought to be important cent World Bank studies that shed some light on agents of environmental damage. With little other these important issues. While these studies pro- land available to them, poor farmers may resort vide certain valuable pieces of the puzzle, how- to cultivating steeply sloped erosion-prone hill- ever, many others remain missing. Much remains sides or to clearing tropical forests. Inability to to be done before the overall picture becomes afford other fuel sources may make them cut clear. down trees at umsustainable rates. High discount rates may make them value short-term benefits Poverty and the Environment far ahead of long-term losses. Whether, in what ways, and why poor people might act in ways Much has been written about the relationship that are damaging to the environment remain im- between poverty and the environment. The poor portant, but generally unanswered, questions. are thought to be both victims and agents of en- The relationship between poverty and the en- vironmental damage. vironment is likely to vary considerably from case 94 Poverty and the Environment: Pieces of the Puzzle 95 to case, especially since a host of other important Figure 7.1 Effect of poverty on predicted deforestation factors, including government policies, institu- in Mexico, 1980-90 tional structures, and the specific characteristics Predicted deforestation of the environmental and natural resources in- Percentage of 1980 forest cover volved, will also vary substantially from case to 50 case. No easy generalizations on the relationship between poverty and the environment are likely, 40 Many depots therefore. Rather, detailed case studies are nec- Few depots _ _ - essary. The difficulty of understanding the rela- 30 tionship between poverty and the environment is compounded by the scarcity and poor quality 20 of available data. This chapter presents the results of ongoing I0 research undertaken by the World Bank on a number of aspects of the relationship between 00 20 40 60 80 00 poverty and the environmnent:° 20 460 810 povAreyand thepoor morenikomelnto deforePoverty (percent of population below minimum wage) * Are the poor more likely to deforest? * Do envirorunental problems result in higher fer- Source. Deininger and Minten 1996. tility, inducing a vicious cycle of rising popula- tion pressure, increasing poverty, and growing tors, such as crop price supports and the avail- envirornental problems? ability of extension services, also play important * Are the poor able to respond to pollution? roles, however. Figure 7.1 shows that in areas with Poverty and Deforestation: a higher density of government maize-purchase Lessons from Mexico and Indonesia depots (which indicates higher maize prices) the level of deforestation is higher at all poverty lev- In many parts of the world shifting cultivation els. Data for subgroups of municipalities provide by poor farmers is thought to be the most impor- additional insights. Poverty is the major pressure tant cause of deforestation (Reardon and Vosti on deforestation in municipalities where pastures 1995). Although logging may play a catalytic role, comprise more than half of agricultural area. Con- most tropical deforestation is thought to result versely, in areas with high initial levels of forest from conversion to agriculture. Actors at the for- cover (more than 70 percent in 1980), poverty is est frontier include very poor, subsistence-ori- not a significant factor influencing changing for- ented shifting cultivators; smallholders growing est cover, and it is in fact commercial motives that export crops; large, commercial farmers and pose the primary threat. ranchers; and government-sponsored tree crop or Indonesia timber estates. Mexico Estimates of deforestation in Indonesia range from 300,000 to 1 million hectares a year. Even at the During the 1980s about 19 million hectares were lower figure deforestation is cause for concern deforested in Mexico, an annual deforestation rate because Indonesian forests are disproportionately of about 2.9 percent. A recent study uses socio- rich in biodiversity. Much deforestation is associ- economic and physiogeographic data for 2,400 ated with the establishment of large palm oil or municipalities to examine the factors behind this pulpwood plantations or with shifting cultivation. high rate of change in forest cover (Deininger and Although there is considerable logging, it is often Minten 1996). The results suggest that poverty is thought that the resulting threat to deforestation associated with higher levels of deforestation. As arises less from clear-cutting than through in- can be seen in figure 7.1, municipalities with creased access to these areas for shifting cultiva- higher levels of poverty lost a greater proportion tors. Shifting cultivators are usually assumed to of their forest cover during the 1980s. Other fac- be subsistence-oriented-poor cultivators seeking 96 Expanding the Measure of Wealth to meet their basic food requirements by exten- population growth (Dasgupta 1995; Cleaver and sive cultivation into the forest fringes. To deter- Schreiber 1994,). mine the extent to which poor subsistence farmers In many areas children contribute to the affect deforestation, a recent study examines crop household through the collection of firewood. As choice at the forest frontier using village census population pressure results in scarcer fuelwood, data (Chomitz and Griffiths 1996). Crop choice do families respond by having more children, or has implications for the nature of deforestation, fewer? If the benefits of additional children out- its relation with poverty, and appropriate policy weigh the costs of childrearing to the household, responses. deforestation might result in greater population Evidence from this study suggests that export- growth and a vicious cycle is likely A recent study oriented tree crops play an important role in de- of common property forests and fertility in Paki- forestation relative to subsistence-oriented stan attempts to address these questions using shifting cultivation. The share of households en- data from the 1991 Pakistan Integrated House- gaged in shifting cultivation at the forest frontier hold Survey (Filmer and Pritchett 1996). is surprisingly small. Even in high deforestation The study finds that the likelihood of a family areas of Kalimantan, only 36 percent of house- having had a child in the last five years increased holds are shifting cultivators-and some of these as the households' perception of problems with households may be using shifting cultivation to firewood supply increased. Similarly, households produce export-oriented crops. Conversely, the living further away from wood sources also had proportion of households at the forest frontier cul- a higher likelihood of having had a child in the tivating tree crops ranges from near half to more last five years. These results seem to support the than three-quarters on all major outer islands hypothesis that environmental degradation may except Irian Jaya. Rubber is the dominant small- lead to higher fertility. However, firewood prices holder crop, with coffee, cocoa, and coconuts be- have a negative impact on fertility, which pro- ing important in some parts of the country. Some vides counter-evidence to this hypothesis. The coconut cultivation, however, may be for subsis- same patterns hold when the data are disaggre- tence consumption rather than for sale in the gated by region, except for the Sindh region. market. There, the data show a strong positive correla- The importance of tree crop cultivation in de- tion between fertility levels and firewood scar- forestation trends indicates that a more nuanced city, supporting the "vicious circle" hypothesis. approach to deforestation might be required, as These results are consistent with the findings some tree crops-including jungle rubber, one of of Cleaver and Schreiber (1994) that deforesta- the most important smallholder tree crops-tend tion rates are positively related to fertility in to be less damaging than other agricultural land thirty-eight African countries, although in this uses. Conversion of forest to jungle rubber is less study as well this effect was small and not always damaging to biodiversity than cultivation of ei- significant. ther field crops or monoculture tree crops, since jungle rubber tends to preserve a significant Poverty and Informal Regulation amount of biodiversity of the primary forest and of Industrial Pollution: Lessons from Asia can relatively easily convert back to natural for- ests. Rapid industrialization and the weakness of for- mal regulatory systems have resulted in substan- Poverty, Environment, and Population: tial levels of pollution in many developing Is There a Vicious Cycle? countries. Surprisingly, however, instances can be found of clean producers. In some cases it appears The problems of poverty, population, and the en- that community pressure has played an impor- vironment are often thought to be linked. In par- tant role in firms' pollution abatement decisions- ticular, it has often been suggested that they are a process that has been referred to as informal linked together in a vicious cycle in which greater regulation. Several recent studies carried out in population results in a worsening environment South and Southeast Asia provide insights into and a worsening environment leads to increased the role of such community pressure in pollution Poverty and the Environment: Pieces of the Ptuzzle 97 abatement and mitigation (Hettige and others Figure 7.2 Emissions intensity of firms forthcoming). in rich and poor areas of lava Emissions per kilogram of biological Bangladesh oxygen demand/million rupiah 1.5 A study of manufacturing plants in the fertilizer and wood pulp sectors in Bangladesh suggests that community pressure has apparently been _ quite effective in many instances (Huq and l.0 Wheeler 1993). The technical level of most plants was largely exogenous, because many were fi- nanced by foreign aid and, as such, reflected the 0.5 environmental standards of the donors. Commu- Poor nity pressure often resulted in responses such as areas ich monetary compensation for damage to fisheries 0 oeg _ubli and paddy fields and installation of end-of-pipe Foreign Public Private firms firms firms treatment equipment. However, in areas where other sources of employment were scarce, plants Source Pargal and Wheeler 1995. faced little community pressure even if they were example, pollution loads and the size of the ex- clearly identifiable as important polluters, posed population) and development indicators Indonesia such as income and education. The correlation with income and education reflects both higher An analysis of the level of water pollution emis- willingness and ability to pay for environmental sions in Indonesia based on a large sample of data quality and greater enforcement capacity in more from Indonesian factories (Pargal and Wheeler highly developed regions. 1995) indicates that pollution intensity was sub- The evidence from these studies suggests that stantially higher in poorer, less-educated commu- the level of income and education of communi- nities (figure 7.2). This differential appears to be ties surrounding industrial plants are powerful too large to attribute to preferences alone and in- predictors of the effectiveness of informal regu- dicates that the ability to pressure polluting firms lation in controlling emissions. Several reasons may be important in explaining the observed dif- have been advanced to explain this (Hettige and ference in abatement performance. others 1995). First, poor communities may be less aware of environmental risks because of their lack China of access to various channels of information and their lower levels of education. Second, and more Analysis of variations in the enforcement of the controversial, it has been suggested that poor Chinese pollution levy system leads to similar communities may be more willingt trade off results (Wang and Wheeler 1996). Although a environmentalquality forincreased employment. countrywide uniform levy is set by the Chinese Thvirdore c ualityieseted temlesaento Natioal Eviromentl Prtecton Aency Third, poor communites tend to be less able to National Environmental Protection Agency bring effective pressure to bear on polluters be- (NEPA), enforcement varies widely across prov- cause of factors such as illiteracy, lack of re- inces. Moreover, provinces can legally augment sources, or lack of influence over government (but not undercut) the national standard with officials. NEPA's permission. Analysis of panel data for twenty nine provinces between 1987 and 1993 in- Lessons and Conclusions dicates that China's pollution levy has been ef- fective in promoting reduction of water pollution. The studies reported here are helping to build up The study also finds that collection rates for pro- a more detailed and sophisticated understanding vincial levies are significantly affected by the se- of the relationship between poverty and the en- verity of local environmental problems (for vironment. They are only small pieces of the over- 98 Expanding the Measuire of Wealth all puzzle, however, and much more work re- Cleaver, K. M., and G. A. Schreiber. 1994. Reversing mains to be done before we can be confident that the Spiral: The Population, Agriculture, and Envi- we understand the relationship between poverty ronment Nexus in Sub-Saharan Africa. Washing- and the environment. ton, D. C.: World Bank The studies reported on in this chapter pro- Dasgupta, P. 1995. An Inquiry into Well-Being and vide mixed support for the hypothesis that Destitution. Oxford: Clarendon Press. poverty leads to environmental degradation. Evi- Deininger, K., and B. Minten. 1996. "Poverty, Poli- dence from Mexico and Indonesia suggests that cies, and Deforestation: The Case of Mexico.' deforestation is related to poverty, but many other Research Project on Social and Environmental factors are also involved. In Indonesia tree crops Consequences of Growth-Oriented Policies. seem to be a more important factor behind de- Policy Research Working Paper 5. World Bank, forestation than subsistence cultivation. Some of Policy Research Department, Washington, D.C. the households planting tree crops are undoubt- Filmer, D., and L. Pritchett. 1996. "Environmental edly poor. But is the deforestation they cause a Degradation and the Demand for Children: reflection of their poverty or of the economic in- Searching for the Vicious Cycle." Research Project centives driving the production of those crops, on Social and Environmental Consequences of which are common to all households? Growth-Oriented Policies. Policy Research There is also mixed support for the notion that Working Paper 2. World Bank, Policy Research environmental degradation results in higher fer- Department, Washington, D. C. tility. This appears to be the case for deforesta- Hettige, H., M. Huq, S. Pargal, and D. Wheeler. tion in Pakistan's Sindh region. In Pakistan as a Forthcoming. "Determinants of Pollution Abate- whole, however, the available evidence is incon- ment in Developing Countries: Evidence from clusive. If nothing else, this example illustrates South and South-EastAsia." World Development. the notion that the relationship between poverty Huq, M., and D. Wheeler. 1993. "Pollution Reduc- and the environment is case-specific and that gen- tion without Formal Regulation: Evidence from eralization is dangerous. Bangladesh." Environment Department Working The studies of urban pollution in Asia suggest Paper. World Bank, Environment Department, that although the phenomenon of informal regu- Washington, D. C. lation is a real and important one, poorer com- Pagiola, S. 1995. "The Effect of Subsistence Require- munities are less able to bring the necessary ments on Sustainable Land Use Practices." Pa- pressure to bear on polluters and, therefore, live per presented at the Annual Meetings of the in more polluted environments. American Agricultural Economics Association, While these studies yield important insights August 6-9, Indianapolis, Ind. and can, to an extent, inform appropriate policy Pargal, S., H. Hettige, M. Singh, and D. Wheeler. actions, they certainly do not provide universally 1995. "Formal and Informal Regulation of applicable explanations of the relationship be- Industrial Pollution: Comparative Evidence tween poverty and the environment. They do re- from Indonesia and the U. S." World Bank, Policy inforce the importance of analyzing the precise Research Department, Washington, D.C. nature of the relationship, and highlight the im- Pargal, S., and D. Wheeler. 1995. "Informal Regula- portant role for indicators of environmental qual- tion of Industrial Pollution in Developing Coun- ity in tracking changes over time. tries: Evidence from Indonesia." Policy Research Working Paper 1416. World Bank, Policy References Research Department, Washington, D.C. Reardon, T., and S. Vosti. 1995. "Links between Ru- Chomitz, K. M., and C. Griffiths. 1996. "Deforesta- ral Poverty and the Environment in Developing tion, Shifting Cultivation, and Tree Crops in In- Countries:Asset Categories and Investment Pov- donesia: Nationwide Pattems of Smallholder erty." World Development (9). Agriculture at the Forest Frontier." Research Wang, H., and D. Wheeler. 1996. "Pricing Industrial Project on Social and Environmental Conse- Pollution in China: An EconomicAnalysis of the quences of Growth-Oriented Policies. Policy Levy System." Policy Research Working Paper Research Working Paper 4. World Bank, Policy 1644. World Bank, Policy Research Department, Research Department, Washington, D. C. Washington, D. C. CHAPTER 8 International Progress in Indicator Development he development of useful environmental gional conventions, requiring formally defined indicators requires not only an under- quantitative outcomes, has bred useful indicators standing of concepts and definitions, but such as a measure for determining when critical also a good knowledge of policy needs. In fact, loads have been exceeded (Hettelingh and oth- the key determinant of a good indicator is the link ers 1995). Integrated assessments such as the up- from measurement of some environmental con- coming Global Environmental Outlook (see also ditions to practical policy options. Environmen- Swart and Bakkes 1995; UNEP forthcoming), or tal indicators can be used at both the international the evaluation of the Fifth Environmental Action and the national level as a tool for, state-of-the- Plan of the European Union (European Environ- environment reporting, measuring environmen- ment Agency 1995) typically feature sets of indi- tal performance, and reporting on progress cators that are derived from current best practice. toward sustainable development. At the national The indicators presented in this chapter are level they can also be used for- clarifying objec- grouped according to the degree to which they tives and setting priorities. condense information. The three major catego- The World Bank is mainly a user of indicators, ries of indicator efforts are atomistic, individual not a developer. It builds on various worldwide indicator sets; thematic indicators; and systemic indicator efforts and data sets, which are the sub- indicators. The presentation is selective (not com- ject of this chapter. Pioneering work by the prehensive) and the examples are chosen to il- Organisation for Economic Co-operation and De- lustrate the various approaches being tried. velopment (OECD) on a conceptual framework, namely the pressure-state-response model, has in- Individual Indicator Sets fluenced the indicator activities of a number of countries (such as Canada, Denmark, Finland, Individual indicator sets represent the least France, Hungary, Japan, the Netherlands, the amount of data aggregation and feature large lists United Kingdom, and the United States; and vari- or menus of indicators, the number of which may ous international organizations (for references see easily exceed one hundred. section on selected indicator publications at the Of the three programs discussed here the first end of this chapter). two (the OECD and United Nations Department Whereas some of the measures mentioned in for Policy Coordination and Sustainable Devel- this chapter have been developed recently as "in- opment (CSD) programs) offer a logical frame- dicators," others, like nutrient balances, have been work and a multitude of potential indicators that used in the field for a long time but are only now users can select from according to their needs. In being used in a broader sense. As another ex- contrast the third program (in the United King- ample model-based support to global and re- dom) is an example of a large set that is intended 99 100 Expanding the Measure of Wealth to be used in its entirety to give a rounded picture The approach adopted by the OECD and its of sustainable development within a country. member countries relies on the assumption that Programs with large lists of indicators typi- there is no unique set of indicators, and that the cally do not define benchmarks against which the appropriate set depends on its particular use and indicator values should be compared. This is not the different needs of the users. This has led to to say that their promoters do not acknowledge the development of a core set of environmental the essential role of benchmarks in the construc- indicators that are linked to the monitoring of en- tion of an indicator. Indeed, some programs with vironmental progress and the measurement of large lists do make an effort to compare indica- environmental performance. These are supple- tor values to pre-existing objectives such as air mented by various sets of sectoral indicators to quality guidelines or commitments to interna- help improve the integration of environmental tional conventions. The U.K. program is a case in concerns into sectoral policies. The OECD core set point. However, for a majority of the potential is of limited size (around forty core indicators) and indicators no ready standards exist. covers a broad range of environmental issues. The joint program of the United Nations En- vironment Programme, the United Nations De- Commissionfor Sustainable Development velopment Programme, the Food andAgriculture Indicators Program Organization and the World Bank on land qual- ity indicators (LQIs) is another example of the Another exarnple of the "menu approach" is the "large list" approach to indicators (see chapter 5). United Nations-sponsored program of the Com- One of the reasons the LQI program suggests so mission for Sustainable Development (CSD) on many candidate indicators is that land manage- indicators (DPCSD 1996). The CSD program has ment has to be tailored to local conditions rang- developed methodology data sheets drafted by ing from pasture quality to waterlogging (Pieri twenty one international organizations. Each and others 1995). Box 5.2 shows the menu from methodology data sheet provides guidance on the which the user can select LQIs. indicator's significance, how it is calculated, its scientific background, and potential data sources. Organisahionfor Economic Co-operation These serve as guides for countries interested in building their own capacity in the measurement The OECD initiated a specific program on envi- and reporting of indicators ronmental indicators in 1990 following a request The menu is very large and at present con- at the 1989 G-7 summit. This program produced tains 142 indiicators. The program is in the early the following results: stage of development and it hopes to arrive even- * Agreement on a common terminology and con- tually at a much smaller set of indicators for each ceptual framework for all OECD countries (for subject area. example, the pressure-state- response model and The structure of the CSD indicator menu may the core themes) well be as important as the specific indicators * Identification and definition of a core set of in- proposed. It uses the conventional pressure-state- dicators on the basis of three major criteria: response framework and expands it to encom- policy relevance, analytical soundness, and mea- pass not only environmental but also social, surability economic, and institutional issues, all of which * Measurement of these indicators for a number are important for sustainable development. of countries Above all the ongoing CSD program has suc- * Regular use of these indicators in the OECD's ceeded in bringing together many international analytical work and environmental performance organizations-ranging from various UN bodies reviews. to nongovernmental organizations-to lend their The results of this work, in particular the con- expertise to the program. Sixteen countries' have ceptual framework using the pressure-state- volunteered to test the indicators developed by response model, have influenced similar activi- this program., and their experience will provide ties launched by a number of countries and in- essential guidance for future development of the ternational organizations. program. International Progress in Indicator Development 101 United Kingdom the Netherlands, and the Nordic countries. (The Nordic Environmental Indicators are compiled The preliminary set of indicators developed for for Denmark, Finland, Iceland, Norway, and Swe- the United Kingdom (UK Department of the En- den.) All three approaches are based on the pres- vironment 1996) also features a large number of sure-state-response scheme, in combination with indicators-a total of 118. The compilers have sys- a list of established policy themes. Moreover, all tematically related indicators to key issues and three approaches come from countries with ex- objectives of sustainable development. In all, 21 tensive experience with the use of indicators in "families" of issues have been distinguished. This policy formulation. Given this similarity in back- has been combined within a pressure-state-re- ground and analytical framework, it is interest- sponse framework, much like in the OECD and ing to note the differences in the indicator CSD programs. Box 8.1 discusses the use of indi- collections. cators in grappling with the issue of road trans- First, as shown in table 8.1, there is variation port and the enviromnent in the United Kingdom. in the overall structure. The Canadian indicator Future work will focus on reducing the num- collection is grouped using a systems approach ber of indicators. However, priority is given to at a highly aggregate level (for example, life sup- disaggregating information on pressures so as to port systems). The Dutch and the Nordic sets are show people where they can make a difference. structured by policy themes. Within this primary For example, the indicator for energy intensity structure all three apply a breakdown along the would be broken down to show where gains have causality chain and distinguish among indicators led to an improvement in the overall trend (for for pressure, state, and response. In this chain the example, changes in industrial energy consump- Canadian collection has additional indicators for tion) and where problems remain (for example, impacts of environmental changes. It is easy to road transport and domestic energy consump- speculate that this difference reflects, at least in tion). An even more rigorous effort in this direc- part, the degree to which the indicators play a fion is Eurostat's program that aims at pressure formal role in reporting to political bodies. Al- indices (see box 8.2). In addition the U.K. pro- though the Canadian indicators program is linked gram tries to satisfy the needs of environmental to the national Green Plan, the Dutch and Nor- managers who rely heavily on information about dic sets discussed here have been designed for the state of the environment. official reporting.2 The Dutch structure primarily reflects important processes (such as flows of Thematic Indicators pollution). The Nordic indicator collection is Many countries tend to a "mid-size" approach in partly structured by processes and partly by re- terms of information reduction in the construc- sources, such as fish and forests. These sorts of tion of environmnental indicators. This thematic variations reflect differences in national endow- approach involves developing a small set of idi- ment, problems, and tradition. Nevertheless, cators for each of the major issues in environmen- however complex the background, the resulting tal policy. In fact, the examples that were lists and structures apparently provide an discussed for the menu approach also structure adequate fit for national or regional needs and the indicator array by recognized policy issues. possibilities. The CSD program does so in terms of Agenda 21. Second, in the three indicator sets different However, many practitioners have gone a step indicators have been chosen to monitor progress further and are now supplying policymakers with for similar environmental issues. For example, smaller sets of indicators. The U.S. Agency for In- urban environmental quality in the Nordic set is ternational Development (USAID) has compiled exclusively represented by indicators for pres- such a set for each of its program objectives to sure by traffic. In the Canadian set a state vari- evaluate the environmental performance of its de- able, the measure for determining when ambient velopment program (USAID 1995 and 1996). air quality objectives have been exceeded, forms Examples of the thematic approach are pre- the indicator under the corresponding theme. The sented in table 8.1: indicator sets from Canada Canadian set does provide indicators for trans- 102 Expanding the Measture of Wealth Box 8.1 Trends in passenger road transport in the United Kingdom A. Passenger miles per head by vehicle B. Real price of different modes of transport Miles per person Index(1974= 100) 7,000 180 6,000 160 Rail far-es Bus fares -Cars, vans, taxis /-\/. - ' 5,000 14\ . w 4,000 120 I i~ / Disposable income 3,000 i;- All motoring 100' 2,000 Other ~Rail8 1,000 th _ Petrol/oil 0 60 1970 1978 1986 1994 1970 1978 1986 1994 Since 1970 the amount of car travel has almost doubled Since 1974, the real price of rail and bus fares has in- to an average of 6,500 miles per person per year, while creased faster than the 50 percent growth in real dis- travel on other road passenger vehicles, mainly buses posable incomes. The real cost of motoring has fallen, and coaches, has reduced to less than 600 miles per making car travel much more affordable. Changes in person per year. price and mode of transport illustrate the difficulty in distinguishing pressures from responses. C. Passenger transport energy consumption and D. NOX emissions from road passenger transport miles traveled Million tons of oil equivalents Billion miles Thousand tons Billion miles 30 475 950 650 Ntrogen oxide _ Energy consumption emissions 25 750 / / / ~~~375 //475 / / . /~~~~37 Passenger miles 47 20 /< Passenger miles 550 , P 275 . 300 1 5 4~350 ~ I10 175 150 125 1970 1978 1986 1994 1970 1978 1986 1994 Over the last twenty years, there has been no improve- NO. emissions from road passenger transport are now ment in fuel efficiency in passenger transport, despite more than two and a half times higher and have out- increases in efficiency of individual vehicles. Fuel use stripped growth in actual traffic. Recently, emissions for road transport (passenger and freight) has risen by have declined d.ue to increased use of diesel cars and nearly 90 percent since 1970, and now accounts for a an increase in road fuel duty. The downward trend quarter of energy consumption. will continue because of the introduction of catalytic converters for n,ew cars. Source: U.K. Department of the Environment 1996. port, but because it clusters stressors and stressed Forests are another area in which one sees systems at a much more general level, it does so more similarity between the Canadian and the under a separate category not narrowed down Nordic sets. This close correspondence is as ex- to urban traffic. The Dutch set notably does not pected, given the similarity in forest cover be- include urban air quality. This is because the tween Canada and the Nordic region. Both sets policy themes and indicators focus on pressures.3 include pressure (harvesting) indicators. On the International Progress in Indicator Development 103 Box 8. Environmental pressureindicethe case even between countries that could be {Box 8.2 Environmental pressure indices classified as similar in many respects. The Statistical Office of the European Union The differences between the three comparable (Eurostat) is conducting a large-scale project to sys- indicator sets of table 8.1 suggest at least two ob- tematically relate trends in environmental pressures *ective reasons wh users and suplers of envi- to activity levels in the various sectors of the economy. Panels of national and sectoral experts will ronmental information reach different solutions formulate indicators. Other projects are scheduled in different countries. First, the environment is to devise methods to supply the required statistical different between countries, and indicators will series and coefficients, and the panels will assign vary to reflect these differences. Second, because weights to aggregate the expected 50 to 100 indica- tors into a pressure index for each of the-ten themes p , d of the Fifth Environmental Action Plan of the Union tors are required to respond to these new needs. (European Commission 1996). Another example of the thematic approach is In total between 100 and 200 experts will be in- the just-published World Development Indicators volved, spread over fourteen countries. The project (WDI) report of the World Bank (1997). Grouped is now about half completed. The most striking re- sult thus far is the overwhelming number of differ- ito a number of discrete sectons, the WDI con- ent indicator proposals put forward by sectoral tains indicators on economic, social, environmen- experts. tal, and institutional variables. The environment Although it is unlikely that such a massive ef- section, for example, contains nine tables with fort and its updating can be afforded by other coun- approximately forty indicators (box 8.3). A ma- tries, its formal structure does highlight several apoate foty i s (box 83 Arma- important steps: first, the different roles for experts jor advantage of the WDI is the ability it provides to identify indicators in each policy field and the in- to link economic, social, and environmental indi- formation needed to compile them; and second, the cators in a common database (covering about 150 weighting that is required to aggregate different countries). measures into an aggregate index. Systemic Indicators response side, reforestation is used as an indica- Systemic indicators have been designed to derive tor in the Canadian set but not in the Nordic set one number to indicate whether a complex sys- because reforestation with a narrow selection of tem is in difficulty. Such indicators are the most commercial species is not a good indicator of ad- ambitious in terms of information reduction. equate responses. Other names for systemic indicators indude "port- In recent additions to the Dutch set, forest area folio indicators" and "synoptic indicators" (Rump is included as an indicator of forest state. How- 1996). Hammond and Adriaanse use "aggregate ever, in this case key pressures on forests are not indicators" (Hammond and others 1995), although harvesting but acid deposition, human-induced aggregation is often not the precise method ap- drought, and excess nutrients. Indicators for all plied. An analogy from medicine is to human three are included in the Dutch set, but not spe- body temperature: if it deviates from the normal cifically for forests. value, something is probably wrong. Once the I Third, the Dutch set is smaller than the other indicator has identified a potential problem, fur- two. In order to provide parliament with a one- ther analysis is needed to find out the precise na- dimensional performance rating on progress to- ture of the problem. ward the objectives of environmental policy, the The wealth and genuine savings indicators, set has been narrowed to one indicator per theme presented in some detail in chapters 2 and 3 of (selected from pressure or state variables), allow- this report, belong to this category of indicators. ing explicit comparison with predetermined They are good examples of systemic indicators goals. that have useful policy implications: persistently What do these differences tell us? Obviously, negative savings or declining total wealth per the same basic approach-the "common gene"- capita clearly indicate that the development of the can lead to the selection of different indicators country as a whole is not sustainable. However, depending on the policy link. This appears to be the opposite, positive saving and constant or 104 Expanding the Measure of Wealth Table 8.1 Examples of thematic indicator sets Nordic countries Netheniands- Canada Climate change Climate change Life support systems Emissions of carbon dioxide Weighted emissions of Climate change Global mean temperature greenhouse gases Stratospheric ozone Annual mean temperatures in the Nordic Capitals Ozone layer depletion Toxics in the environment Changes in the use of fossil fuels Weighted emissions of Biodiversity Ozone layer depletion CFC's and halons Human health and well-being Consumption of ozone-depleting substances Eutrophication Urban air quality Total ozone Weighted gross loads of Urban water Targets for reduction of ozone-depleting substances nitrogen and phosphorus Natural resource Eutrophication Acidification sustainability Net supply of commercial fertilizer and manure Weighted emissions of Marine resources Algal chlorophyll ammonia, sulfur dioxide Forests Light penetration and nitrogen oxides Agricultural soils 'Wnter green' crop land Acidifying deposition Influencing factors Connection to chemical waste water treatment plants Dispersion of toxic Energy consumption Acidification substances Transportation Deposition of acidifying substances Weighted releases of Waste generation/ Area where the critical load for sulfur has been exceeded priority substances, production Quantity of lime used radioactive substances and Population growth and Toxic contamination pesticides lifestyle patterns Cadmium in moss Disposal of solid waste Heavy metals Mass dumped Organic hazardous substances Disturbance of local Collection of Hg, Cd, and PCB environments Urban environmental quality Fraction of the population Number of private cars and light commercial vehicles in the reporting nuisance from capitals odor or noise Number of persons in the capitals who are exposed to noise Sectoral indicators from road traffic Contribution to above Number of vehide-kilometers of public transport in the indices from the following capitals sectors: Biodiversity Agriculture Total kilometers of road per unit of land Manufacturing Endangered and vulnerable species Refineries Protected areas Electricity production Cultural and natural landscapes Transport Forest ditches and drained land Consumers Total wetland area Restoration of wetlands Waste Quantity of household waste per capita Share of municipal waste that is deposited on landfills Forest resources Removal in relation to increment Changes in standing volume Forest planting and sowing Fish resources Taxation Fishing mortality Spawning stock development Quotas Source: Statistics Norway 1995. Source:.van Esch, S. A. draft Source: Environment Canada 1996. 1996. growing wealth, does not necessarily guarantee The Materials Balance Approach sustainability-although the chances are better. Other examples of systemic indicators include Materials balance sheets have been developed in the materials balance approach and the nutrient industrial cowutries as a way to screen a country's balances approach. economy for pollution "leaks" that would lead to International Progress in Indicator Development 105 Box 8.3 The World Bank's 1997 World Development Indicators (WDI) The main sections of the WDI are the following: * Energy efficiency, dependency, and emissions: real GDP * People per unit of energy use; net energy import as a per- * Environment centage of commercial energy use; total carbon di- * Economy oxide emissions per capita and per unit of real GDP * States and markets . Urbanization: urban population; urban population * Global linkages. as a percentage of total population; average annual The tables and indicators presented in the environ- growth rate of urban population; population in ur- ment section are: ban agglomerations of a million or more; popula- * Land use and deforestation: land area; rural popu- tion in the largest city as a percentage of total lation density; cropland, permanent pasture, and population; percentage of urban population with other land as a percentage of total land area; to- access to sanitation tal forest area; annual deforestation * Traffic and congestion: number of vehicles per 1,000 * Protected areas and biodiversity: nationally pro- people; number of vehicles per kilometer of road; tected area; total number and threatened species road traffic volumes in million vehicle-kilometers; of mammals, birds, and higher plants number of people killed or injured per 1,000 vehicles * Freshwater use: per capita freshwater resources; * Air pollution: emissions and ambient concentrations annual freshwater withdrawals; freshwater with- of suspended particulate matter and sulfur dioxide drawals by agriculture, industry, and domestic * Government commitment: status of country environ- use; percentage of rural and urban population mental profile; national conservation strategy; with access to safe water biodiversity profile; compliance with Convention * Energy use: production and use of commercial en- on International Trade in Endangered Species of ergy; annual and per capita use of commercial Wild Flora and Fauna; participation in climate energy; use of traditional fuels; growth rate and change, ozone layer, chlorofluorocarbon control, per capita production of electricity and law of the sea treaties. Source: World Bank 1997. an accumulation of a dangerous pollutant in the how much "environment" goes into different environment. Typically this approach dealt with products and lifestyles. persistent groups of compounds, like lead com- A related effort is highlighted in box 8.4. The pounds. One-off studies were conducted to help World Resources Institute, in conjunction with the identify priorities for pollution control policies. Wuppertal Institute, the Netherlands Ministry of Statistics Netherlands built up considerable ex- Housing, Spatial Planning and the Environment perience with these studies during the late 1970s and the National Institute for Environmental and 1980s. Studies, Japan Environment Agency, have de- Later developments have widened the appli- tailed trends in material throughput in their re- cation of the materials balance approach. Most spective countries. significantly, as pollution management became The Nutrient Balances Approach more operational, it needed routine and timely measurement of its impact. And indeed, once the Relative to many other environmental indica- key factors in a country's balance of copper com- tors, more experience exists with the use of nu- pounds, for example, had been identified, esti- trient balances-although most of the literature mating the annual load of copper compounds to relates to land quality at the field level. Nutrient the county's soils became a matter of routine balances belong to the broader materials balances rather than a special study. family of indicators. They come in slightly dif- In addition, a much more aggregate form of ferent forms, depending on whether the issue is materials balances has been developed at the nutrient loss from soils (as in the case of Kenya) Wuppertal Institute in Germany. This approach or excess nutrients brought into the environment groups together all inputs that go into the pro- (as in the Netherlands). In both situations nitro- duction of a given commodity on a mass basis. gen is often taken as representative of all This provides a rough-and-ready way to compare nutrients. 106 Expanding the Measure of Wealth Box 8.4 Indicators of material flows Many of the materials removed, moved, and processed and consumed in a third. So a significant portion of to support industrial economies are not captured in the natural resource use that supports a nation's eco- conventional economic accounting. Those accounts do nomic activity c:an occur outside its borders. Of these not make explicit many activities that require envi- four countries, the United States is largely self-suffi- ronmental modifications or uses of natural resources cient in natural resources due to its size, but in the that have potential environmental impacts. For ex- other three countries the foreign proportion of the TMR ample, measures such as the Gross Domestic Product is between 35 to 70 percent. These industrial countries (GDP) do not include the movement or processing of gain the benefits of resources while the environmen- large quantities of materials that have no (or even tal cost of their production falls on others, often on negative) economic value. An understanding of the developing countries. environmental consequences of economic behavior re- The TMR takes into account both hidden flows quires physical accounts of the material basis of that and foreign components of natural resource use, as behavior. In a unique collaboration the Wuppertal In- well as direct inputs of natural resources into the stitute in Germany, the National Institute for Environ- economy. Even when normalized to a per capita ba- mental Studies in Japan, the Netherlands Ministry of sis, the TMRs of modem industrial economies are Housing, Spatial Planning, and Environment, and the enormous. In 1991 the TMR per capita of these four World Resources Institute in the United States have countries ranged from 45 to 85 metric tons of natural developed new physical accounts of the material ba- resources per person. Over the 20 years of the time sis of their industrial economies that parallel tradi- series (box figure 8.4.1) there was a surprising trend tional economic accounting (World Resources Insti- toward convergence of these quantities even though tute 1997). the details of those material flows differed significantly They propose a new summary measure, the To- (box figure 8.4.2). tal Material Requirement (TMR) of an industrial A parallel set of physical accounts provides the economy. The TMR is the sum of all the material that basis for the construction of new indicators that com- is moved or extracted from the environment in sup- bine both physical and economic information. In par- port of the economy. Some of this material enters the ticular it allows the construction of measures of an economy as a commodity, but much of it is never seen economy's material intensity in a way that is more in economic accounts. These are called "hidden flows." comprehensive than traditional measures, incorporat- These hidden flows, which are associated with extrac- ing as it does both hidden and foreign components tive activities, harvesting of crops, and infrastructure (box figure 8.4.'3). The results for these four countries development (for example, soil erosion, overburden, shows a clearly declining pattern of materials inten- dredging, and excavation) are immense. In the four sity, supporting the conclusion that economic activity countries studied, from 55 to 75 percent of the TMR is growing somewhat more rapidly than natural re- arises from these hidden flows. National accounts in source use. physical terms are required to routinely document Ultimately, sustainable development will require such uses of natural resources and their potential en- a closer understanding of how the economic and en- vironmental effects. vironmental aspects of human activity interact, as well But natural resources are frequently extracted in as actions based on such understanding. Indicators of one country, transformed into products in another, physical flows, such as the TMR, could guide progress Box figure 8.4.1 Total material requirement, annual Box figure 8.4.2 Primary contributions to TMR, flows per capita 1991 Tons per capita Tons per capita 40 m Metals and industrial minerals i20 ma Fossil fuels 35 t m Construction minerals 100 ,> ,,,, United States | sg Renewables = 30 X i lrnfrastructure excavation Xn 8 Erosion X f 2 4 D | t hw0. ",5 60~~~~~~~~~~~~~~ Germany Japan Netherlands United States International Progress in Indicator Development 107 Box 8.4 Indicators of material flows (coninued) Box figure 8.4.3 Overall material intensity (TMR/ toward more efficient use of natural resources. Since GDP) Index what leaves the industrial system as wastes is closely related to the volume of material inputs, policies that TMR/GDP Index (1975 = I00) reduce the use of primary natural resources not only 1 10 diminish extraction pressures, but also wastes and pol- lution. Similarly, policies that make natural resource 100 > use more efficient or increase recycling lower environ- %;s>\ mental impacts over the entire materials cycle. 90 As8,3¢ Germany Further work is anticipated that will provide I_5,, Netherlands ' adding outputs, including products and waste; con- 80 J > \ ' tinuing to harmonize definitions and methods; devel- oping detailed indicators of the material cycle within 70 Jaa economic sectors-that would also be useful to policy Un6te0 St tsmakers; and expanding the number of countries, both 60 75 1980 1985 1990 1994 industrial and developing, using these indicators- 1975 1980 1985 1990 1994 adding a global reach to the understanding of mate- Five-year average rial flows. In most parts of the world, including Africa The history of nutrient balances in the Neth- and some regions of Latin America, the major erlands illustrates some important, general concern is with nutrient depletion: more nitro- points. To begin with, a ten-year-long standoff gen is carried off with crops than is brought into delayed official publication of nutrient balances the soil ("fixed") by natural processes and supple- -the obstacle being determining the standards mented through chemical fertilizer and manure. that would represent an acceptable nutrient load. This reduces soil fertility and leads to declining Eventually, publication of a range of balances re- yields (see the discussion on land quality indica- vealed that there was a problem by any standard. tors in Honduras in chapter 5). The appropriate Furthermore, once public debate broadened, indicator in this situation is the fraction of nitro- the original calculation scheme was significantly gen in the soil that is lost annually. An example simplified-accepting some information loss in for Kenya is given in box 8.5. The beauty of this return for timeliness and transparency. The sim- indicator is its transparency, which makes it easy plification made it possible to use a single indi- to compare with predetermined acceptable nu- cator f6r reporting to parliament on the trient levels. In addition, it can be applied to scales performance of this aspect of environmental other than the national scale, and it is an easy policy during the previous year. way to evaluate economic and environmental sce- The lesson from this review of experience with narios by their effect on the nutrient balance. nutrient balance accounts is that the approach can The Netherlands has one of the best known be used in situations of nutrient deficiencies as (and most extreme) cases of nutrient excess. The well as nutrient surplus. More generally, it illus- core of the problem is industrial-scale cattle and trates how indicator definitions can evolve with pig farming that has developed rapidly since the the changing needs of policymaking. 1960s, driven by European Union price guaran- tees for meat, and the ample availability of ani- Conclusion mal feed from outside the European Union (mainly from Latin America, Thailand and the Many organizations around the world are devel- United States). Nutrient balances have been used oping indicators to provide quantitative measures to identify the resulting pollution problems of progress toward environmentally sustainable caused by intensive husbandry, and this informa- development. The geographical distribution of tion has helped place the issue on the political publications shows that the origin of indicator agenda (see box 8.5). work lies in the consensus-based styles of gov- 108 Expanding the Measure of Wealth Box 8.5 Nutrient balances Operationalizing a nitrogen balance for Kenya The fraction of nitrogen in the soil that is lost annually can be estimated using the following simplified scheme from Smaling (1993 and 1996) and Stoorvogel and Smaling (1990). Fraction lost = (out-in)/stock where "ouf' is nitrogen in harvested product, "in" is nitrogen in mineral fertilizers, and "stock" is nitrogen in soil. a. Average maize yield 1,460 kg/ha (FAO 1993) b. Nitrogen content in maize (out) 25 kg/ha (Stoorvogel and Smaling 1990) c. Total fertilizer consumption 95,000 tons (FAO) d. Average nitrogen content of fertilizer 20 percent e. Total nitrogen consumption 19,000 tons (c*d) f. 25 percent of fertilizer to maize 4,750 tons (0.25*e) g. Area under maize 1,200,000 ha (FAO data) h. N fertilizer to maize (in) 4 kg/ha (f*g) i. Average stock of nitrogen in soil 3,750 kg/ha in maize-growing areas' The fraction lost can therefore be calculated as follows: (25- 4) /3,750 = 0.0056. Based on the following tentative classification suggested in Smaling (1993), nitrogen is being lost from Kenyan soils but not at an alarming rate. Fraction lost < 0 Enrichment 0 < Fraction lost 0.005 Equilibrium 0.005 < Fraction lost < 0.01 Slightly alarming 0.01 < Fraction lost < 0.05 Moderately alarming 0.05 < Fraction lost < 0.2 Severely alarming 0.2 < Fraction lost Very severely alarming Nitrogen balancefor soil and groundwater in the Netherlands, 1990 (van Eerdt and others 1996). In contrast nitrogen balance for soil and groundwater in the Netherlands, as illustrated below, features a consider- able accumulation of nitrogen. This can be traced to import of feed and use of artificial fertilizers. Input (million kgs of N) Output (million kgs of N) Animal manure 486 Withdrawal by harvested crops (including grass) 456 Artificial fertilizers 412 Volatilization of ammonia 53 Sewage sludge 5 Denitrification 330 Atmospheric deposition 161 Leaching and surface runoff toward inland waters 124 Wastes 31 Accumulation 173 Other 41 Total 1,136 Total 1,136 Notes: 1. For this illustration the following soil types and nitrogen contents have been assumed for area under maize: 50% ferralsols, acrisols 0.1 % N = 2,500 kg. N/ha 50% luvisols, nitisols, phaseozems 0.2% N = 5,000 kg. N/ha 2. Excluding the amount of ammonia that volatilizes during manure-spreading andi finds its way back to the soil via atmo- spheric depositions (43 miRlion kilograms of nitrogen). ernment of Canada, the Netherlands, and examples of implicit or explicit indicator work are Scandinavia (Porter 1995). These pioneering coun- developing in regions with different backgrounds tries continue to refine their indicator methods (Australia, the Baltic states, Japan, LatinAmerica, and adapt them to the evolving needs of policy the United Kingdom, and the United States). In and environment management. Meanwhile, many other countries the uses of indicators are begin- International Progress in Indicator Development 109 ning to be studied (southern Africa and some European Commission. 1996. "Enviromnental ndi- Mediterranean countries). cators and Green Accounting: Practical Steps International programs like that of the CSD towards the Implementation of the Communi- of the United Nations, and of the OECD, with its cation from the Commission to the Council and expa,are contributing to this the European Parliament on Directions for the eproliferanion mfembership, methods to underpin European Union on Environmental Indicators proliferation of empirical methods to underpin and Green National Accounting." (COM (94) environmentally sustainable development poli- 670 final). Brussels. cies. The resulting indicator sets clearly exhibit a European Environment Agency. 1995. Environment "common gene" (Hammond and others 1995) that in the European Union 1995: Reportfor the Review is manifest in a number of dimensions: the use of of the Fifth Environmental Action Programme. K. a pressure-state-response framework, sometimes Wieringa, ed. Copenhagen. expanded to pressure-state-impact-response; Hammond, A., A. Adriaanse, E. Rodenburg, D. policy issues as a structuring element, with link- Bryant, and R. Woodward. 1995. Environmental eto targets represented in all sorts of variants; Indicators: A Systemic Approach. Washington, agd emphasis on the provisional character of the D. C.: World Resources Institute. and emphasis on the provisional character of the Hettelingh, J.-P., M. Posch, P. A. M. de Smet, and R. first-edition indicator sets. The UNEP is now un- J. Downing. 1995. "The Use of Critical Loads in dertaking an ambitious effort to apply indicators Emission Reduction Agreements in Europe." for environmentally sustainable development in Water, Air, and Soil Pollution 85: 2381. global and regional scenario analysis (UNEP Pieri, C., J. Dumanski, A. Hamblin, and A. Young. forthcoming 1997) and others may follow suit 1995. Land Quality Indicators. World Bank Dis- (DPCSD forthcoming). cussion Paper 315. Washington, D. C. The various programs also feature common Rump, P. 1996. State of the Environment Reporting: methodological problems. Almost all programs Source Book of Methods and Approaches. Nairobi: report difficulties with response indicators. This has United Nations Environment Programme/ prompted fresh initiatives to expand the devel- Smaling,E.M.A. 193. "AnAgro-EcologicalFrame- opment of indicators (and the underlying data- work for Integrated Nutrient Management with base) and to extend field testing to a wider Special Reference to Kenya." Ph.D. dissertation. number of countries. Agricultural University, Wageningen, The Neth- erlands. Notes -. 1996. Personal communication. Statistics Norway. 1995. "Nordic Environmental In- 1. These include Belgium, Bolivia, Brazil, Canada, dicators: Draft Document." Prepared for the Nor- China, Costa Rica, Czech Republic, Finland, Ger- dic Environmental Monitoring and Data Group many, Hungary, Malaysia, Morocco, South Af- by the Nordic Indicator Group, Oslo. rica, the Netherlands, Ukraine, and Venezuela. Stoorvogel, J. J., and E. M. A. Smaling. 1990. Assess- 2. To the Dutch parliament and to the Nordic min- ment of Soil Nutrient Depletion in Sub-Saharan isters of the environment, respectively. Af 1 (2d ede). Wugeningenn 3. Urban air quality does figure in official reporting AfTica: 198s2000. Vol. 1 (2d ed.). Wageningen, that encompasses the formal indicators. The Netherlands: The Winand Staring Center. Swart, R. J., and J. Bakkes, eds. 1995. Scanning the References Global Environment: A Framework and Methodol- ogyfor Integrated Environmental Reporting and As- DPCSD (United Nations Department for Policy Co- sessment. Nairobi: United Nations Environment ordination and Sustainable Development). 1996. Programme/RIVM. Indicators of Sustainable Development: Framework U. K Department of the Environment. 1996. Indica- and Methodologies. New York: United Nations. tors of Sustainable Developmentfor the United King- . Forthcoming. Major Trends and Sustainable De- dom. London: HMSO. velopment. New York: United Nations. UNEP (United Nations Environment Programme). Environment Canada. 1996. National Environmental Forthcoming. Global Environment Outlook: A New Indicator Series. Ottawa. State of the Environment Report. Nairobi. 110 Expanding the Measure of Wealth USAID (United States Agency for International De- Institut Fran~ais de l'Environnement. 1995. Physical velopment). 1995. Annual Performance Report Environmental Accounting: Land Use/Land Cover; 1995: Environment Chapter. Washington, D. C. Nutrients and the Environment. Orl6ans. - 1996. Managing USAID's Environmental Port- 1996 Indicateurs de performance environne- folio for Results: Proceedings of the March 1995 Environmental Workshop on Performance Measure- mentale de la France. Orl6ans. ment. Washington, D. C.: USAID. Ministryfor Environrnent and Regional Policy. 1994. van Eerdt, M. M., C. S. M. Olsthoorn, and P. K. N. Environmental Indicators of Hungary. Budapest. Fong. 1996. "Nutrient Balance Sheets inAgricul- Ministry of the Environment. 1996. Trends in the Finn- ture." In Environment Related-Agricultural Statis- ish Environment-Indicatorsfor the 1997 OECD En- tics in the Nordic Countries. Technical Report 62. vironmenital Performance Review of Finland. Copenhagen: Nordisk Statistisk Sekretariat. Helsinki van Esch, S. A. Draft. "Performance Indicators of O elsinkio Environmental Policy." Background document Orgaimsahton for Economic Co-operation and De- to the ".dctr ithNaonl Eniomn velopment 1993. Indicators for the Integration of to the 'Indicators in the National Environmen-r, tal Program 1997-2000." RIVM, Bilthoven. Environmental Concerns into Energy Policies. Paris. World Bank. 1997. World Development Indicators. _. 1993. Indicatorsfor the Integration of Environmen- Washington, D.C. tal Concerns into Transport Policies. Paris. World Resources Institute. 1997. Resource Flows: The Ma- . 1994. Environmental Indicators-OECD Core Set. terial Basis of Industrial Economies. Washington, D. C. Paris. Selected indicator publications -. Various years. OECD Environmental Perfonrance Reviews. Paris. (Already published: Austria, Bul- Danish Ministry of the Environment. Various years. garia, Canada, Germany, Iceland, Italy, Japan, the Environmental Indicators. Copenhagen. (Pub- Netherlands, New Zealand, Norway, Poland, Por- lished annually since 1991.) tugal, Sweden, United Kingdom, United States.) 10 Enabling the Safe Use of Biotechnology: Principles and Practice 11 Biodiversity and Agricultural Intensification: Partnersfor Development and Conservation 12 Rural Development: From Vision to Action (forthcoming) 13 Integrated Pest Management: Strategies and Policiesfor Effective Implementation 14 Rural Finance: Issues, Design, and Best Practices (forthcoming) 15 The Economics of Involuntary Resettlement (forthcoming) 16 Social Assessmentsfor Better Development: Case Studies in Russia and Central Asia 17 Expanding the Measure of Wealth: Indicators of Environmentally Sustainable Development Related ESD Publications Monitoring Environmental Progress: A Report on Work in Progress Nurturing Development: Aid and Cooperation in Today's Changing World Toward Sustainable Management of Water Resources Water Supply, Sanitation, and Environmental Sustainability: The Financing Challenge The World Bank Participation Sourcebook THE WORLD BANK IS1818 1 I St,-cctt .NA.\ WLshinlAton, 1).(. 201433. l S.\ I'cicpllonc: 202-477-1234 F icsim1il: 202-477-6391 ICIe.lx: .\1I 6414. WO(II .I)IAANK I( 1 248423 W RI AM).ANK1 (ia,hilc Vidress: INT1'[.\I C\I) NV\ :SII I1Nt (;Y) NI)C \Norld WXidc Web: hittp://NN-w%.% Xorldhaill-.ol-0 F- n.ail: hookos(i %torl'l)bonis.or4, ISBN 0-8213-3956-7