34043 GROWTH, POVERTY, AND INEQUALITY Eastern Europe and the Former Soviet Union This report is part of a series undertaken by the Europe and Central Asia Region of the World Bank. The series covers the following countries: Albania Latvia Armenia Lithuania Azerbaijan Moldova Belarus Poland Bosnia and Herzegovina Romania Bulgaria Russian Federation Croatia Serbia and Montenegro Czech Republic Slovak Republic Estonia Slovenia FYR Macedonia Tajikistan Georgia Turkey Hungary Turkmenistan Kazakhstan Ukraine Kyrgyz Republic Uzbekistan GROWTH, POVERTY, AND INEQUALITY Eastern Europe and the Former Soviet Union GROWTH, POVERTY, AND INEQUALITY Eastern Europe and the Former Soviet Union Asad Alam, Mamta Murthi, Ruslan Yemtsov, Edmundo Murrugarra, Nora Dudwick, Ellen Hamilton, and Erwin Tiongson Europe and Central Asia Region ©2005 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 08 07 06 05 This book is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6193-7 ISBN-13: 978-0-8213-6193-1 e-ISBN: 0-8213-6194-5 DOI: 10.1596/978-0-8213-6193-1 Library of Congress Cataloging-in-Publication Data Asad, Alam, 1962­ Growth, poverty, and inequality : Eastern Europe and the former Soviet Union / Alam Asad, Mamta Murthi, Ruslan Yemtsov. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-8213-6193-1 ISBN-10: 0-8213-6193-7 1. Poverty--Europe, Eastern. 2. Poverty--Former Soviet republics. 3. Europe, Eastern--Eco- nomic conditions--1989­ 4. Former Soviet republics--Economic conditions. 5. Equality--Europe, Eastern. 6. Equality--Former Soviet republics. I. Murthi, Mamta, 1965­ II. Yemtsov, Ruslan. III. Title. HC244.Z9P6123 2005 339.4'6'0947--dc22 2005043452 Cover photo by: Anatoliy Rakhimbayev. Cover design by: Naylor Design, Inc. Contents Foreword xiii Acknowledgments xv Acronyms and Abbreviations xvii Overview 1 Trends in Poverty in the Region, 1998­2003 4 Factors Contributing to Poverty Reduction, 1998­2003 16 Nonincome Dimensions of Well-Being, 1998­2003 22 Prospects for Poverty Reduction 29 The Role for Public Policy 34 Conclusions 41 1. Nature and Evolution of Poverty, 1998­2003 47 Introduction 47 Consumption Poverty 51 Poverty in Nonincome Dimensions 67 Conclusions 76 2. How Has Poverty Responded to Growth? 79 Growth and Poverty Reduction 80 Growth Elasticities, or, How Responsive Is Poverty Reduction to Growth? 81 vi Contents Changes in Distribution, What Happened and Why 87 The Relative Shares of Growth and Changes in Distribution in Poverty Reduction 90 Rural-Urban and Other Subnational Differences in Poverty Reduction 92 Conclusions 96 3. The Role of Labor Markets and Safety Nets 107 How the Poor Can Connect to Growth 107 Economic Opportunities Have Expanded 110 The Poor Took Advantage of New Opportunities 119 Why Are Many Workers in the Region Still Poor? 129 Conclusions and Policy Recommendations 140 4. Affordable Access to Quality Services 147 Education 149 Access to, and Affordability of, Health Services 161 Energy and Other Utility Services 171 Conclusions 183 5. Prospects for Poverty Reduction 189 Alternative Scenarios for Growth, Poverty Reduction, and Inequality 189 Patterns of Growth: Implications for Growth and Inequality 195 The Role for Public Policy 203 Conclusions 213 Appendix 219 A. Data and Methodology 215 B. Key Poverty Indicators 236 Bibliography 279 Index 293 Boxes 1 Using Purchasing Power Parity to Measure Poverty 6 2 Nonincome Dimensions of Poverty and Millennium Development Goals in the Region 32 3 Data for This Report: The World Bank's ECA Household Survey Archive 42 Contents vii 1.1 What Is an Appropriate Poverty Line for the Region? 49 1.2 What Would Someone in the Region Living on Two Dollars a Day Consume? 52 1.3 National Poverty Assessments Confirm Poverty Trends Based on International Poverty Lines 53 1.4 Vulnerable Groups and Poverty: Roma, IDPs, and Institutionalized Populations 64 1.5 Life Satisfaction in the Region Remains Low 75 3.1 In Most Countries, Household Survey Data Report Higher Employment Figures than ILO Statistics 113 3.2 The Role of Agriculture in Transition 116 3.3 The Role of Remittances in the Region 120 3.4 Improvements in Targeting: Lessons from Recent Policy Reforms 130 3.5 Global Trends in the Number of Working Poor 132 3.6 Informal Employment in Transition Economies 134 3.7 Labor Market Study Discusses Ways to Enhance Job Opportunities in the Region 141 3.8 Raising Agricultural Productivity 144 4.1 Survey Data Provide Limited Information about Access to, and Quality and Affordability of, Utilities 172 4.2 What Has Happened with District Heating? 173 4.3 Electricity and Water Tariffs Remain below Benchmarks for Full-Cost Recovery 178 5.1 EU Accession and Poverty Reduction Objectives 192 5.2 Depleted Social Capital of the Poor Limits Opportunities 196 5.3 Nonincome Dimensions of Poverty and Achieving the MDGs in the Region 199 Figures 1 More than 40 Million People Moved out of Poverty during 1998­2003 3 2 At the Country Level, Absolute Poverty Has Declined Almost Everywhere 5 3 The Lowest National Poverty Line in the Region Is around Two Dollars a Day 8 viii Contents 4a Children Face a Greater Risk of Poverty than Other Population Groups; in Most Cases, This Risk Has Increased over Time 9 4b Poverty Reduction in Secondary Cities and Rural Areas Has Lagged behind Capital Cities 10 4c The Unemployed Face Higher Risks of Poverty than the Employed; This Risk Has Grown over Time in the CIS 12 4d There Are Marked, and in Some Cases Increasing, Differences in Poverty across Regions 12 5 Working Adults and Children Continue to Form the Bulk of the Poor in the Region 13 6 In Some Countries, Poverty Is Shallow; in Others, Deep 14 7 While Changes in Distribution Have Gone Either Way in the EU-8 and SEE, They Have Moved in Favor of the Poor in the CIS 15 8 Since 1999, Growth Rates in the Region Have Been High, with the CIS the Most Rapidly Growing Subregion 17 9 Employment-to-Population Ratios Are Well below Lisbon Targets (70 percent) in the EU-8 and SEE and Often Trending in the Wrong Direction 18 10 The Poor Have Benefited More than the Rich from the Growth Rebound in the CIS 19 11 The Share of Growth in Poverty Reduction Is Dominant across All Regional Subgroups 20 12 Access to Secondary Education Has Gone Up Virtually throughout the Region, but Some Countries Continue to Struggle to Arrest the Decline in Primary Enrollment Rates 24 13 Hospital Utilization Rates Have Recovered, but Remain at Low Levels in Parts of the Low Income CIS Group 26 14 The Poor Make Greater Use of So-Called Dirty Fuels for Heating 27 15 In the Low Income CIS Countries, the Reliability of Water Supply Is Low and Shows Little Improvement 28 16 Household Expenditures on Utilities Have Increased 29 Contents ix 17 Growth Will Move an Additional 21 Million People out of Poverty by 2007, but 40 Million Will Remain Absolutely Poor and More Than 100 Million Vulnerable to Poverty 31 1.1 More Than 40 Million People Moved out of Poverty during 1998­2003 51 1.2 Poverty Incidence Varies across Countries in the Region, around 2003 55 1.3 Poverty Depth in the Region, 1998­2003 56 1.4 Levels and Changes in Poverty by Employment Status, 1998 to 2003 58 1.5 Change in Poverty by Education for Representative Countries 59 1.6 Capital Cities Gained More than Other Cities and Rural Areas, 1998­2003 60 1.7 Variation of Poverty Risks by Regions, 1998/9­2002/3 61 1.8 Changes in Poverty by Age, Relative to National Average 62 1.9 The Poor in the Region around 2003 66 1.10 Most Nonworking Poor Live in Households Where Someone Works 66 1.11 Life Expectancy at Birth, 1990­2003 68 1.12 Incidence of Tuberculosis, 1990­2003 69 1.13 Poverty in the Dimensions of Consumption, Access to Water, and Health 71 2.1 Since 1999, Growth Rates in the Region Have Been Higher than the World Average 81 2.2 Growth Has Been Accompanied by Poverty Reduction 82 2.3 The Poor Have Benefited More than the Rich from the Growth Rebound in the CIS 84 2.4 Poverty Is More Responsive to Growth, the Higher the Level of Income and the Lower the Level of Inequality 86 2.5 Distribution Has Moved in Favor of the Poor in Most CIS Countries 88 2.6 "Decomposition" of Inequality Does Not Explain Declines in Most CIS Countries 89 2.7 Share of Growth in Poverty Reduction Is Dominant across All Regional Subgroups 91 2.8 Increase in the Ratio of Rural to Urban Poverty in Most Countries 93 x Contents 2.9 Urban Poverty Is More Responsive to Growth and Falling (or Rising) More Rapidly than Rural Poverty 94 2.10 Partial Elasticity of Poverty Reduction to Growth Is Lower in Rural Areas 95 3.1 Connecting the Growth to the Poor 108 3.2 Real Wage Growth Typically Outpaced Net Employment Growth in Transition Economies 111 3.3 The Structure of Employment Has Changed 112 3.4 Employment in Service Sector Expanding; in Agriculture, Mixed 114 3.5 Value Added per Worker Is Lowest in Agriculture 115 3.6 Real Wage Changes Correlate with Poverty Changes 121 3.7 Poor Gained from Real Wage Gains in SEE and the CIS 122 3.8 Changes in Employment Rate, 1998­2002, by Quintiles for Selected Countries 126 3.9 Safety Nets Cover Many Poor in the Region 128 3.10 Sectoral Wage Employment for the Poor and Nonpoor in Selected Countries 133 3.11 Large Wage Gap between Poor and Nonpoor Persists across the Region 138 3.12 Productivity Distribution of Old, Restructured, and New Enterprises 139 3.13 Wage Increases Outstripped Productivity Gains during the Economic Recovery in the Region 143 4.1 MDGs in the Region: Infant Mortality and TB Incidence 148 4.2 Regional Coverage of Education, Ages 7­14 150 4.3 Inequality in Access to Primary Education in the Region, 1998­2002 151 4.4 Regional Coverage of Education, Ages 15­17 152 4.5 Inequality in Access to Secondary Education, 1998­2002 153 4.6 Gender Inequality in Access to Secondary Education, 1998­2002 153 4.7 Recent Declining Trends in Regional Mathematics Performance (TIMSS) 156 4.8 Mathematics Performance in Selected Countries of the Region, 1999­2003 157 Contents xi 4.9 Role of Subnational Governments in Education, 1995­2002 158 4.10 Aging Teaching Force in the Region, 1995­2003 159 4.11 Cervical Cancer in the Region and Western Europe, 1970­2002 163 4.12 Hospital Utilization 164 4.13 Utilization Rates of Health Services by Quintiles 166 4.14 Ratio of Out-of-Pocket Health Spending to Household Total Consumption, 1998­2003 167 4.15 Health Insurance and Utilization in Armenia, 2001 171 4.16 Reliability of Electricity in the Region in the Early 2000s 174 4.17 The Deterioration in Water Provision in Tajikistan and Moldova 175 4.18 Household Expenditure Shares for Electricity, Heating, Water, and Sewerage Have Increased from 1998 to 2002/2003 176 4.19 Electricity Payments Are a Larger Share of Household Expenditures for Poor Households (Quintile 1) than for Rich Households (Quintile 5) 177 4.20 Poor Households Are Less Likely than Rich Ones to Use Clean Fuels 182 5.1 Population of the Region by Poverty Status, 1990­2002, and Outlook for 2007 191 5.2 Trends from Global Projections 195 Tables 1 Transfer Payments for Social Protection Have Had an Important Role to Play in Reducing Poverty outside of the Low Income CIS Countries 21 2 Achieving Subgroup-Appropriate Poverty Reduction Targets over the Long Term (2015) Will Require Significantly Higher GDP Growth Rates 33 2.1 Poverty Has Been More Responsive to Growth in the Middle Income CIS Countries and SEE than Elsewhere 83 3.1 Work Does Not Protect Families from Poverty in the Region 109 xii Contents 3.2 The Evolution of Pension Spending by Groups of Countries 119 3.3 Transfer Payments for Social Protection Have Had an Important Role in Reducing Poverty outside the Low Income CIS Countries 131 4.1 Mathematics Performance, 1995­2003 155 4.2 Impoverishing Effects of Catastrophic Health Expenditures 169 4.3 In Most Countries, Households in Secondary Cities Were More Likely to Heat with Dirty Fuels in 2003 than in 1998 181 5.1 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, MDG-Related Targets 194 5.2 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, Country-Specific Targets Focused on Economic Vulnerability 194 5.3 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, Country-Specific Targets with European Vision 194 Foreword Five years ago the World Bank report, Making Transition Work for Every- one, helped focus public attention and debate on the urgent challenge of reducing poverty and inequality in the countries of Eastern Europe and Central Asia. The report found that one out of five people sur- vived on less than $2.15 a day compared with fewer than one in twenty-five a decade earlier. The sharp increase in poverty was driven in large part by the collapse in incomes and the increase in inequality. The working poor, children, rural households, and specific groups (such as the Roma in Eastern Europe) were at greatest poverty risk. Capabilities of the poor were being endangered from falling access to education, corruption (particularly in public service delivery), nutri- tional deficiencies, and communicable diseases. And prospects for rapid poverty reduction seemed remote. Much has changed in these countries since then. Growth has rebounded in the Commonwealth of Independent States (CIS) from the depths of the financial crisis of 1998. This was driven by several factors, in particular, the boost to the Russian Federation's competi- tiveness provided by a large devaluation, the supply response from structural reforms that had been undertaken by many of the Region's countries, and the income windfall from the large and unexpected increases in the prices of oil and other energy exports. At the same time, the process of European Union integration has helped broaden xiii xiv Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union markets, lock in domestic reforms, and attract higher investment in acceding countries. The cessation of war in the western Balkans has also helped make the economic environment more conducive to investment and growth. Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union takes stock of the impact of growth on poverty and inequality during 1998­2003, provides systematic evidence-based analysis to understand the different outcomes, and suggests areas for further action. It is heartening to find that during this period growth has helped more than 40 million people move out of poverty. This has been aided by improved income distribution in many countries, espe- cially in the CIS. But poverty and vulnerability still remain a signifi- cant problem. More than 60 million are poor (that is, living on incomes less than $2.15 a day), and more than 150 million are vul- nerable (that is, living on incomes between $2.15 and $4.30 a day). Most of them are in the middle-income countries even as the low income countries have higher rates of poverty. Working families con- stitute the largest group among the poor. Many others face depriva- tions in access and quality of public services. Regional inequalities, both between and within countries, are large. The heterogeneity of countries in the Region suggests a differentiated approach across countries. But it is obvious that prospects for reducing poverty and vulnerability--and achieving the Millennium Development Goals-- will be crucially dependent on the ability of countries to accelerate growth and create well-paying jobs; improve the quality of educa- tion, health care, and public infrastructure; and strengthen social safety nets. Achieving these will require stepping up efforts to com- plete the institutional and policy reform agendas and sustaining them. This report--the first in a new series of regional studies--is an important contribution to our thinking about how the World Bank can work more effectively with clients and partners in the Region to reduce poverty and vulnerability in a rapidly changing world. Forth- coming reports on jobs, trade, infrastructure, migration, and demo- graphics will look at the key economic and social opportunities and challenges. I hope that these reports will stimulate debate, promote better understanding, and spur action to bring about prosperity for all. Shigeo Katsu Vice President Europe and Central Asia Region Acknowledgments This study was prepared by a core team led by Asad Alam, Mamta Murthi, and Ruslan Yemtsov and comprising Edmundo Murrugarra, Nora Dudwick, Ellen Hamilton, and Erwin Tiongson. Additional writ- ten contributions were provided by Jo Swinnen, Karen Macours, Vic- toria Levin, Signe Zeikate, Taras Pushak, Emil Daniel Tesliuc, Sheetal Rana, Cem Mete, Alan Wright, Marina Bakanova, and Lire Ersado. The analysis of data was carried out by Taras Pushak, Jossey Moies, Zurab Sajaia, Stefania Rodica Cnobloch, Diane Steele, Lucian Bucur Pop, Akshay Sethi, Victoria Levin, and Sayyora Umarova. The Regional Household Data Archive, established by the Office of the Chief Economist of the Europe and Central Asia (ECA) Region, was the main data source. Judy Wiltshire was responsible for processing the document and, together with Helena Makarenko, provided invaluable assistance to the team during the process of preparation. Comments, suggestions, and materials were provided by Arup Banerji, Radwan Shaban, Phillip Keefer, Peter Lanjouw, Berk Ozler, Marta Menendez, Ximena Del Carpio, Juan Carlos Guinarte, Taras Chernetsky, Shabih Mohib, Daniela Gressani, Bernard Funck, Benoit Blarel, Alan Wright, Pierella Paci, Louise Cord, Matthew Verghis, Kin- non Scott, Stefano Scarpetta, David Kennedy, Julian Lampietti, Jan Rutkowski, Merrell Tuck, Armin Fidler, Alexander Marc, Jean- Jacques Dethier, Shahrokh Fardoust, and Reema Nayar. Bruce Ross- xv xvi Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Larson assisted the team in conceptualizing the report. Many other colleagues helped in preparing the report. The report was prepared under the overall guidance of Pradeep Mitra. Valuable support and suggestions were provided by Cheryl W. Gray. Additional guidance was provided by the peer reviewers, John Page, Dennis de Tray, Anthony Atkinson, and by Shigeo Katsu, and Annette Dixon. The study benefited from close cooperation with the World Bank World Development Report 2006 team, including Michael Walton, Francisco Ferreira, and Tamar Manuelyan Atinc. The report has also benefited from discussions with Johannes Linn of the Brook- ings Institution. The study team acknowledges helpful comments received during consultations held in early 2005 with the European Commission (EC), the European Investment Bank, and the EuropeAid Co-opera- tion Office; with the German Ministry for Economic Cooperation, KfW Bankengruppe, and the German Agency for Technical Coopera- tion; with Jessica Irvine and other staff of the U.K. Department for International Development (DFID); with William Buiter and Steven Fries of the European Bank for Reconstruction and Development (EBRD); with Sorbonne's Research Center on Transition and Devel- opment Economics, IMPACT, and the Social Policy Division of the Organisation for Economic Co-operation and Development (OECD); with the French Ministry of Finance; and with Gerry Redmont of the United Nations Children's Fund (UNICEF). The study team has also received feedback from participants at the "Europe after the Enlarge- ment" conference, convened at the Center for Social and Economic Research in Warsaw in April 2005, especially Andrew Warner, Mikhail Dimitriev, and Anthony Shorrocks, and from participants at the annual conference of the Higher School of Economics in Moscow the same month. The report was also discussed at the Conference on Labor Markets, Growth, and Poverty Reduction Strategies for the Western Balkan Region in Thessaloniki, Greece, in May 2005, and the authors are grateful for comments received, especially from Antonin Braho, Richard Marshall, Elizabeth McKeon, and Egbert Holthuis. Book design, editing, and production were coordinated by the World Bank's Office of the Publisher. Dina Towbin was the produc- tion editor, and Francis Speltz edited the manuscript. Acronyms and Abbreviations AIDS acquired immune deficiency syndrome ALB Albania ARM Armenia ASR age-standardized rate AZE Azerbaijan BEL Belarus BiH Bosnia and Herzegovina BUL Bulgaria CEE Central and Eastern Europe CIS Commonwealth of Independent States CPI consumer price index CRO Croatia CZE Czech Republic DFID U.K. Department for International Development DHS Demographic and Health Survey EBRD European Bank for Reconstruction and Development EC European Commission ECA Europe and Central Asia ECHP European Community Household Panel ECSIE Infrastructure and Energy Services Department ECSSD Environmentally and Socially Sustainable Development Network xvii xviii Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union EI Expenditure and Income (survey) EST Estonia EU European Union EU-15 First 15 member states of the European Union EU-8 Eight new member states of the European Union Eurostat Statistical Office of the European Communities EU-SILC European Community Statistics on Income and Living Conditions FYR Former Yugoslav Republic of (Macedonia) GDP gross domestic product GEO Georgia HBS Household Budget Survey HIV human immunodeficiency virus HUN Hungary IBRD International Bank for Reconstruction and Development IDP internally displaced person ILO International Labour Organization IMF International Monetary Fund IN Integrated (survey) ISCED International Standard Classification of Education IZA Institute for the Study of Labor KAZ Kazakhstan KILM Key Indicators of the Labour Market KYR Kyrgyz Republic LAT Latvia LIS Luxembourg Income Study LIT Lithuania LPG liquefied petroleum gas LSMS Living Standards Measurement Study MAC Macedonia, former Yugoslav Republic of MDG Millennium Development Goal MOL Moldova NBER National Bureau of Economic Research OECD Organisation for Economic Co-operation and Development PISA Programme for International Student Assessment POL Poland PPP purchasing power parity RLMS Russian Longitudinal Monitoring Survey ROM Romania RUS Russian Federation SAM Serbia and Montenegro Acronyms and Abbreviations xix SEE Southeastern Europe SLK Slovak Republic SLN Slovenia SNA System of National Accounts SSA Sub-Saharan Africa TAJ Tajikistan TB tuberculosis TIMSS Trends in International Mathematics and Science Study UKR Ukraine UN United Nations UNICEF United Nations Children's Fund UNMIK United Nations Mission in Kosovo UNU United Nations University UZB Uzbekistan WBI World Bank Institute WDI World Development Indicators WHO World Health Organization WIDER World Institute for Development Economics Research Note: All dollar amounts are U.S. dollars ($) unless otherwise indicated. Overview This study examines the impact of growth on poverty and inequal- ity in Eastern Europe and the Former Soviet Union during 1998­2003. It updates the World Bank's previous study on poverty, entitled Making Transition Work for Everyone, which appeared in 2000. It asks three questions: What are the recent trends in poverty and inequality? Why do we see different outcomes across coun- tries? And how can public policy help maximize the impact of growth on poverty reduction? To measure poverty, an absolute poverty line of $2 a day1 is used, comparing it with household consumption per capita. This line is a closer approximation to basic material needs in the Region than the well-known global standard of $1 a day because of the additional expenditures on heating and warm clothing that are required by the cold climate. Using an absolute--as opposed to a relative--poverty line allows us to focus on those who are deprived of the most basic needs, rather than those who may be deprived relative to their better- off fellow citizens. It also allows us to determine trends over time and make comparisons across countries, both of which would be difficult if we were using a relative notion of poverty. In addition to the $2-a-day poverty line, a $4-a-day line is used to capture the notion of "eco- nomic vulnerability," that is, to measure the proportion of the popula- 1 2 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union tion that is not absolutely poor, but could become poor in the event of an economic downturn. In terms of poverty levels, the Region is best thought of in four dis- tinct subgroups of countries. The eight new member states of the Euro- pean Union (EU-8) have low poverty (less than 5 percent) confined to specific subgroups of the population.2 Countries in Southeastern Europe (SEE) have generally moderate levels of poverty (around 5­20 percent).3 The same is true of the middle-income countries in the Commonwealth of Independent States (CIS).4 The low-income coun- tries in the CIS, however, have extremely high levels of poverty (more than 40 percent).5 In addition to countries in these four subgroups, the Europe and Central Asia Region of the World Bank (ECA) also includes Turkey. Wherever possible, we treat Turkey as a "benchmark" against which to evaluate the performance of postsocialist countries in the Region. Turkey has moderate poverty. We also include two coun- tries from outside the Region as benchmarks: Colombia (a middle- income country) and Vietnam (a low-income country). Summary The resurgence of growth in the eastern half of the Region, particularly in the CIS, resulted in a significant decline in poverty in the Region dur- ing 1998­2003 (figure 1):6 more than 40 million people have moved out of poverty during this period. Where roughly 20 percent of the population (or one in five people) was living in poverty, today poverty affects only 12 percent (or one in eight people). While much of this poverty reduction has occurred in the populous middle-income coun- tries in the Region (Kazakhstan, the Russian Federation, and Ukraine), poverty has fallen almost everywhere. During 1998­2003, poverty fell in most countries of the Region, except for Poland and Lithuania in the EU-8 and Georgia in the low income CIS group.7 However, in the con- text of the EU-8, the low overall levels of poverty need to be borne in mind. Income (consumption) inequality showed no clear trend in the EU-8 and SEE; however, inequality declined in the CIS, with the notable exceptions of Georgia and Tajikistan. At the same time that 40 million people have moved out of poverty in ECA as a whole, more than 60 million people remain poor, and more than 150 million people are economically vulnerable.8 Progress on the nonincome dimensions of poverty--such as access to educa- tion, health care, safe water, and heating--is very mixed, with improvements in some cases and deterioration in others. In educa- tion, although access has improved, no subregion is free from coun- tries experiencing declining standards. In health, no subregion is free Overview 3 from the growing epidemic of human immunodeficiency virus (HIV) FIGURE 1 and acquired immune deficiency syndrome (AIDS). Quite apart from More than 40 Million HIV/AIDS and other communicable diseases, attaining the health People Moved out of Millennium Development Goals (MDGs) will prove difficult for many Poverty during countries in the CIS and SEE because of the failure of health services 1998­2003 to deliver adequate and timely services. Access to key infrastructure Distribution of Population by Poverty Status services--in particular lighting and heating--is actually deteriorating in some countries of the low income CIS group. 100 The single most important factor behind the significant decline in 90 poverty in the period in question is high growth in the CIS, where the 215.1 80 bulk of the poor reside. Combined with moderate levels of inequality, 264.2 economic growth has delivered significant poverty reduction. To some 70 extent, this rebound in growth rates in the CIS is unsurprising, 60 although at the height of the financial crisis in Russia and neighboring 50 contagion--which came at the end of a decade of difficult transition-- population it was hard to see the prospects for resumption in growth. A further of 40 % 160.7 factor influencing poverty reduction since 1999 is the reduction in 153.3 30 consumption inequality in some countries of the CIS, which, too, can 20 be viewed as a rebound from the levels observed in the 1990s. Because the substantial reduction in poverty is the result of a 10 102.0 61.2 unique constellation of factors--rapid "catch-up" growth in the CIS 0 accompanied by reductions in inequality in some countries--prospects 1998­9 2002­3 for poverty reduction going forward are less propitious. Very few coun- Nonpoor: tries, even those that have made the most progress in reducing above$4.30aday Vulnerable: poverty, have been successful in creating jobs to fully replace those above$2.15and below$4.30aday that have been destroyed. To some extent, reduction in overall Poor: below$2.15aday employment was only to be expected, given the socialist legacy of high employment-to-population ratios. However, the failure to generate a Source: World Bank staff estimates using ECA Household Surveys sufficient number of jobs means that employment-to-population Archive. ratios have been falling, except in a few of the rapidly growing coun- Note: In million persons and in per- tries of the CIS. In the EU-8 and SEE, the employment ratio is well cent to population. Poverty lines con- verted to local currencies using 2000 below what is found in Organisation for Economic Co-operation and PPP. Data refer to ECA Region as de- Development (OECD) countries. If it persists, this failure to expand fined by the World Bank, and Turkey is included in the aggregate figures. employment will fundamentally limit the poverty reduction impact of growth and act as a brake on further reduction of absolute poverty. This is an issue even in countries where poverty is relatively low (for example, Poland, where rising poverty is related to the growing divide between those with and without employment). In addition to concerns on the jobs front, there is a marked regional and spatial dimension to both the income and nonincome dimen- sions of poverty in the Region. The most rapid declines in poverty have been observed in capital cities, as opposed to secondary cities and rural areas. In parts of the CIS, poverty rates are just as high in 4 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union secondary cities as in rural areas. Throughout the Region, the quality of public services is also poorer outside of capital cities, and trends are mixed, with some differences between capital and noncapital areas getting smaller and others larger. Overall, there is a substantial agenda of reforms if countries wish to reduce poverty in all its dimensions over the coming years. While specific actions will vary from country to country, all countries need to focus on policies that will accelerate rates of growth and ensure that benefits are widely shared among the population. In addition, efficiency and equity concerns warrant strengthening delivery of edu- cation, health, and public utilities services, and enhancing social pro- tection. It is also essential to monitor progress on poverty reduction. Poverty and poverty reduction have a special significance in the Region that is different from that in other parts of the world. First, as mentioned previously, the cold climate means that the notion of "basic needs" has to be expanded to take into account the need for warmth. Only a small fraction of the population in the developing world would require a similar expansion of the basic needs set. Sec- ond, many countries in the Region completed the demographic and epidemiological transition a few decades ago. This graying of the pop- ulation poses significant challenges for public policy, particularly where there are trade-offs involved in relation to the working (or the young) versus the elderly. There is also a greater burden of noncom- municable diseases, with implications for costs and access to health care. Again, there are few countries at equivalent levels of income that face a similar challenge. Finally, the legacy of the former socialist systems of production means that huge inefficiencies exist in the way production is organized, how infrastructure is deployed, and where people are located. Breaking with the past represents not only an opportunity but also a challenge. Trends in Poverty in the Region, 1998­2003 Since 1998, absolute poverty at $2 a day (or, more accurately, $2.15 a day) has declined in most countries in the Region (figure 2). Two countries, Georgia and Poland, bucked the trend of declining poverty, and in another one, Lithuania, poverty was largely unchanged. These trends, which are based on comparable consumption aggre- gates specially constructed for the purposes of this report, reflect the use of the latest purchasing power parity (PPP) exchange rates (2000 PPP) available for the countries of the Region. The use of different PPP revisions affects the ranking of a few countries in the Region, Overview 5 especially those that experienced hyperinflation or continue to prac- tice administrative price setting, but leaves the overall extent of poverty and trends unchanged (see overview box 1). As it did five years ago, absolute deprivation varies enormously across the Region. At one end of the spectrum are countries in the low income CIS group such as Tajikistan, where the proportion of the population living on less than $2.15 per day is more than 70 percent, while at the other end are countries in the EU-8 such as Hungary, where absolute poverty, by this definition, is virtually absent. Coun- tries fall into three broad groups: those with high poverty (all low income CIS countries), those with low poverty (typically EU-8), and those in between with moderate poverty (typically SEE and middle income CIS countries). These groupings are not hard and fast, with some countries in SEE (for example, Bulgaria) and the middle income CIS group (for example, Belarus) having low levels of poverty. Even where incomes have grown and absolute material depriva- tion at $2.15 per day is low, the standard of living is not high, and FIGURE 2 At the Country Level, Absolute Poverty Has Declined Almost Everywhere Poverty Rates by Country 100 90 80 70 60 50 population 40 of % 30 20 10 0 1998 2002 1998 2003 1998 2002 1999 2003 2001 2003 1998 2002 1999 2002 2001 2003 1998 2002 2003 1999 2003 2003 2000 2003 1999 2003 2000/1 1998/9 Hungary Lithuania Poland Romania Bulgaria Belarus Russian Kazakh- Georgia Uzbeki- Moldova Armenia Kyrgyz Tajikistan Fed. stan stan Rep. EU-8 SEE Middle income CIS Low income CIS Below $2.15 a day Above $2.15 but below $4.30 Source: World Bank staff estimates using ECA Household Surveys Archive. Note: 2000 PPP. 6 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union OVERVIEW BOX 1 Using Purchasing Power Parity to Measure Poverty An absolute poverty line, as the name implies, attempts to establish the value of consumption that a person needs to stay out of poverty, regardless of time and place.The first widely accept- ed global poverty estimates, produced by the World Bank's World Development Report in 1990, chose a poverty line measured in 1985 PPP. Chen and Ravallion (2001) have since updated these numbers, using 1993 PPP exchange rates for consumption. The PPPs were again updated for the Region in 1996, and this updated set was used by Making Transition Work for Everyone (World Bank 2000a). This report uses the most recent PPP numbers from 2000 (OECD 2003). More recent data on PPP are more relevant for the transition economies of the Region because they reflect contem- porary (in many cases, liberalized) prices, as opposed to the administered prices of the past. For Turkey, a country without the legacy of an administratively directed economy, all PPP revisions produce approximately the same poverty counts (see figure). The economies of all formerly so- cialist countries exhibit significant changes, with more recent numbers being more plausible. For example, it is highly implausible that poverty in Uzbekistan is negligible (which is the impression that one gets using 1993 and 1996 PPPs and could be traced to widespread price controls in that country practiced during the 1990s). Errors can also go the other way (that is, overstate pover- ty), as appears to be the case when the 1993 PPP is used for Georgia. In addition to issues with relative prices, Georgia experienced hyperinflations around 1993, which would have made meas- urement of prices problematic. It would be incorrect to say that the 2000 PPP revision solves all comparability problems. Where interferences in market mechanisms continue, price surveys that form the basis for PPPs will deliver incorrect results (a factor that can be partly responsible for the low poverty headcount for Belarus). large shares of the population are found to consume between $2.15 and $4.30 per capita per day. This group, while not absolutely deprived, is likely to have relatively low savings and is vulnerable to poverty in the event of shocks that affect earning potential. Of course, an absolute poverty line of $2.15 a day (or some mul- tiple) is one of many potential lines that could be drawn. Often what is relevant from the perspective of the poor is the level of resources that may be needed in the country context to be free from hunger, cold, and other forms of deprivation. In this report, the authors have chosen to use an absolute concept of deprivation, not only to focus more on those who are deprived in some "fundamen- tal" sense but also to facilitate comparisons across countries and over time. The basic needs without which individuals would be Overview 7 Poverty Rates at $2.15 a Day with Different PPPs Countries Sorted by Poverty, Based on 2000 PPPs 90 80 poor 70 as 60 defined 50 40 30 population of % 20 10 0 Fed. Rep. HungaryBelarus PolandUkraineLithuaniaBulgaria Turkey Romania Albania GeorgiaArmenia Moldova Tajikistan Russian Kazakhstan Uzbekistan Kyrgyz 1993 PPP 1996 PPP 2000 PPP Sources: Staff estimates; OECD 2003. The total poverty headcount for the Region does not change much whether one uses the 1993 PPP or the 2000 PPP, although individual country assessments are affected. However, the 1996 PPP (with few exceptions: Bulgaria, Estonia, and Lithuania) produces a lower poverty count than the 2000 PPP does. It is important to note that only the ongoing global International Comparison Pro- gram (www.worldbank.org/data/icp/), expected to produce results by 2007, will address funda- mental problems of all existing sets of PPP in their application to poverty comparisons. Interna- tionally comparable poverty data produced for this study need to be interpreted with due caution. Sources: World Bank staff; World Bank 2000a; Chen and Ravallion 2001; Kakwani and Sajaia 2004; and OECD 2003. absolutely deprived are typically reflected in national poverty lines.9 As might be expected, standards of income required to ensure against material deprivation in richer countries are higher, so national poverty lines are positively related to income levels. National poverty lines in the Region suggest that a poverty line around $2 per capita per day might indeed be a relevant absolute floor (figure 3). When compared with national poverty lines from a random selection of non-Region countries, the Region's poverty lines are found to be higher on average, perhaps reflecting the higher cost of basic needs due to the extremely cold climate in cer- tain countries. However, high though they may seem, even the highest national poverty lines in the Region are substantially lower than poverty lines of two of the poorest EU-15 countries,10 Greece 8 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 3 The Lowest National Poverty Line in the Region Is around Two Dollars a Day 9$ Greece 8$ Portugal 7$ day/person Hungary 6$ per Bulgaria 5$ lines, 4$ Latvia Ukraine poverty 3$ Tajikistan 2$ National 1$ Burkina Faso Nigeria 0$ 1$ 10$ 100$ Level of consumption per capita, per day/per person, log Non-ECA countries ECA countries Sources: Region: World Bank staff estimates; non-Region: Kakwani and Sajaia 2004 and OECD 2003; EU: Dennis and Guio 2004. Note: Latest years of available data used, all values expressed in 2000 PPP $. and Portugal. Using measures of absolute deprivation that are more consistent with national poverty lines but still modest (such as $4.30 per capita per day), it is evident that absolute deprivation continues to exist even in relatively well-off countries such as EU-member- state Hungary (figure 2). How has poverty risk evolved in the past five years? Looking below the national averages on population subgroups, four characteristics stand out for raising poverty risk (that is, poverty incidence) above average: being young, living in a rural area or (in some cases) a sec- ondary city, being unemployed, and having low levels of education.11 Although not equally important in all subregions, these were the same groups identified as having a higher-than-average poverty inci- dence five years ago (World Bank 2000a). Outside the low income CIS countries, children face a substantially higher risk of poverty than other population groups do. Relative poverty risk for children has actually increased in the past five years because poverty incidence has fallen less rapidly among families with children than for other groups (figure 4a). Overview 9 Residents of rural areas face a higher risk of poverty than those in cities do (figure 4b). Among the rural dwellers, children usually face the highest poverty risk, often multiple times the national average. But in some countries of the CIS, poverty risks are as high in second- ary cities as in rural areas. Indeed, lumping capital cities together with other urban areas can be misleading because of the contrast between their positions. Over the past five years, with few exceptions, poverty has declined far more rapidly in capital cities than elsewhere. As a result, residents of rural areas and secondary cities face a far greater risk of poverty relative to capital city dwellers than previously. Outside the low income CIS countries, the unemployed face sig- nificantly higher risks of poverty than the employed do (figure 4c). With the resumption of sustained growth in the CIS and an improve- ment in the position of the employed, the relative risk of poverty faced by the unemployed has increased significantly. The risk of poverty falls with educational attainment. As shown in the report, over time, the risk of poverty of those with basic education FIGURE 4a Children Face a Greater Risk of Poverty than Other Population Groups; in Most Cases, This Risk Has Increased over Time 2.5 2 1.00 = 1.5 incidence 1 poverty Country 0.5 0 1998 2002 1998 2002 2000 2002 2001 2003 1999 2002 2001 2003 1999 2002 1998 2002 2001 2003 2000 2003 1999 2003 Hungary Poland Romania Bulgaria Russian Kazakh - Moldova Georgia Armenia Kyrgyz Rep. Tajiki- Fed. stan stan EU-8 SEE Middle income CIS Low Income CIS < 6 years 7­14 years 15­17 years between 18 and 65 years 66 years Source: World Bank staff estimates using ECA Household Surveys Archive. Note: 2000 PPP. Poverty line for the EU-8 and Bulgaria is $4.30 per day. Risk of 1 indicates that an age group is no more or less likely than the average to fall into poverty. 10 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4b Poverty Reduction in Secondary Cities and Rural Areas Has Lagged behind Capital Cities Poland Hungary EU-8 Lithuania Romania SEE Macedonia, FYR Russian Fed. CIS Middle income Kazakhstan Moldova Georgia CIS Kyrgyz Rep. income Armenia Low Tajikistan Uzbekistan ­ 0 + 0 50 100 Change in poverty, Percent of population % defined as poor 2003 capital cities 2003 secondary cities 2003 rural areas Change in poverty, Change in poverty, Change in poverty, capital cities secondary cities rural areas Source: World Bank staff estimates using ECA Household Surveys Archive. Note: 2002 data used instead of 2003 in Russia, Poland, and Hungary. The benchmark year to measure change is 1998, ex- cept in Kazakhstan, where it is 2001, Kyrgyz Republic and Uzbekistan (2000), and Tajikistan (1999). Poverty line for the EU- 8 and FYR Macedonia is $4.30 per day; everywhere else it is $2.15 per day in 2000 PPP. Overview 11 or less, relative to other groups, has increased, reflecting their handi- cap in benefiting from new economic opportunities. Ethnicity is also associated with higher-than-average poverty inci- dence in some cases. Data on ethnicity are sometimes not covered in surveys, and even where they are, sample size may preclude any robust conclusions. While the data do not allow trends to be inferred, relatively strong evidence exists that in more than one country, groups such as the Roma of Central and Eastern Europe (CEE) face a substantially higher incidence of poverty than the general population does (World Bank 2001c; World Bank 2002g; and World Bank 2005e). Available evidence on other minorities is mixed, with some faring worse than average, such as the Turkish minority in Bulgaria or the Russian minority in Latvia, while others do better, such as the Russian minority in the Kyrgyz Republic or the Hungarian minority in Romania (World Bank 2003i; World Bank 2003k; World Bank 2004g). The relative position of minorities is a function of human capital and other endowments relative to the population as a whole and of their position in relations of power, which may vary from group to group. Within countries, poverty incidence shows marked variation, and there is evidence that regional differences are growing over time in some countries (figure 4d). This is because poverty rates have typi- cally declined more sharply in capital cities and other prosperous areas of trade and tourism than in rural areas or secondary towns. In a related vein, many countries outside the low income CIS group (where information is more limited) show high and highly persistent differences in unemployment rates across regions. Composition of the poor. Most of the poor in the Region comprise work- ing adults and children, who between them account for 60­75 per- cent of the poor (figure 5). In most instances, poor children are children of working parents. This structure of poverty, with the pre- dominance of working families (that is, households with working adults), is no different from that of the past, although the share of working families has declined. The next largest group comprises those out of the labor force, followed by the unemployed and the elderly. With regard to location, urban and rural residents are evenly split, each constituting around 50 percent of the poor in the Region as a whole. This split is influenced by an interaction of higher-than-aver- age poverty risk for rural residents and their relatively low share in the population. In relation to subregions, rural residents form the bulk of the poor in the low income CIS group (70 percent of the poor), SEE (62 percent), and the EU-8 (51 percent). The only subregion that 12 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4c The Unemployed Face Higher Risks of Poverty than the Employed; This Risk Has Grown over Time in the CIS 3.50 3.00 2.50 index risk 2.00 1.50 poverty Unemployed/employed 1.00 0.50 Poland Romania Belarus Russian Kazakh- Moldova Georgia Tajikistan Fed. stan EU-8 SEE Middle income CIS Low income CIS 1998 poverty of unemployed/poverty of employed* 2002 poverty of unemployed/poverty of employed* Source: World Bank staff estimates using ECA Household Surveys Archive. Note: For Kazakhstan, 2001 is used instead of 1998, and 2003 instead of 2002; for Tajikistan, 1999 instead of 1998 and 2003 instead of 2002. For Romania and Moldova, 2003 is used instead of 2002. For the EU-8 and Belarus, the poverty line is $4.30; everywhere else it is $2.15 a day in 2000 PPP. FIGURE 4d There Are Marked, and in Some Cases Increasing, Differences in Poverty across Regions 3.0 2.5 1 = 2.0 rate 1.5 poverty 1.0 National .5 0 1999 2002 1998 2003 1998 2002 1998 2002 Poland Romania Russian Fed. Georgia Source: World Bank staff estimates using ECA Household Surveys Archive. Note: The box depicts the spread in regional poverty rates (boxes plot upper and lower boundaries, called interquartile range, of the interval where most of the regional poverty rates would fall, and the whiskers the extremes). Dots represent outlying regions in a statistical sense. Overview 13 is dominated by urban poor is the middle income CIS group (only 41 percent rural poor). Trends in poverty depth. At the end of the 1990s, the general under- standing was that poverty in the Region, while widespread, was rela- tively shallow. The pattern five years later appears more varied, with the Region's countries now spanning a wide range (figure 6). Using Turkey, Vietnam, and Colombia as the benchmarks shows that poverty in the low income CIS group is fairly deep, but in the middle income CIS countries and SEE, it is fairly shallow. In the EU-8, the picture (rel- ative to a $4.30 poverty line) is mixed, with examples of both shallow and deep poverty. Trends in inequality. While there is no clear trend in the EU-8 and SEE, consumption inequality in the CIS declined (with few exceptions) between 1998 and 2003 (figure 7).12 By 2003, consumption inequal- ity in the Region as a whole looked broadly comparable to that in OECD countries and East Asia. While inequality in consumption does not appear egregiously high, subjective data suggest that people in the Region continue to find inequality to be excessive. This may be FIGURE 5 Working Adults and Children Continue to Form the Bulk of the Poor in the Region 100 75 population 50 poor of % 25 0 1998 2002 1999 2003 1999 2002 1999 2003 1999 2003 Poland Romania Russian Fed. Moldova Tajikistan EU-8 SEE Middle Low income CIS income CIS Children (<16 years) Working (employed+self-employed) Unemployed Inactive Elderly (66 years) Source: World Bank staff estimates using ECA Household Surveys Archive. Note: For the EU-8 and Bulgaria, the poverty line is $4.30. 14 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 6 In Some Countries, Poverty Is Shallow; in Others, Deep Deficit in Consumption of an Average Poor Person as Percentage of Poverty Line Hungary Poland EU-8 Lithuania Estonia Serbia & Montenegro SEE Albania Romania Russian Fed. Middle CIS income Kazakhstan Armenia CIS Moldova Uzbekistan income Kyrgyz Rep. Low Georgia Tajikistan Vietnam Colombia Turkey Benchmarks ­5 ­15 ­25 ­35 ­45 Depth in % to poverty line Source: World Bank staff estimates using ECA Household Surveys Archive. Note: For the EU-8, the poverty line is $4.30; everywhere else it is $2.15 per day per capita in 2000 PPP, latest year of avail- able data used. related in part to the fact that, despite recent falls in the CIS, inequal- ity remains significantly higher than at the outset of the transition. The decline in inequality in most of the CIS (with the notable exceptions of Georgia and Tajikistan) also runs counter to widely held perceptions that the "bounce-back" in growth has gone hand in hand with widening income differentials. There are at least three reasons why subjective and objective measures may suggest different trends. First, changing relative positions of different population subgroups may leave overall inequality unchanged, but may lead to the impres- sion of growing inequality. For example, the rise in the position of capital city dwellers relative to residents of secondary cities may leave overall inequality unchanged (as, for example, would be the case if the two groups simply switched positions in the income distribution), but may contribute to the perception of a growing divide. Overview 15 FIGURE 7 While Changes in Distribution Have Gone Either Way in the EU-8 and SEE, They Have Moved in Favor of the Poor in the CIS (Georgia and Tajikistan Excepted) Estonia Hungary EU-8 Lithuania Poland Albania Bulgaria SEE Romania Serbia & Montenegro Belarus CIS Kazakhstan income Russian Fed. Middle Ukraine Armenia CIS Georgia Kyrgyz Rep. income Low Moldova Tajikistan Colombia Turkey Benchmarks Vietnam 0.100 0.200 0.300 0.400 0.500 Gini Index 1998 2003 Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Gini index for per capita consumption. Second, measures of inequality typically employed in the litera- ture (including in this report) are measures of relative inequality; however, subjective perceptions often relate to absolute differences in income, not relative ones. For example, if everyone's income increases by 10 percent, measures of inequality (such as the Gini 16 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union coefficient employed in this report) would show no change in the distribution of income.13 However, the absolute difference in the incomes of the rich and the poor would have increased (for example, 10 percent growth means a larger absolute increase in income for the rich than for the poor), contributing to a perception of growing inequality.14 Third, sampling and nonsampling errors mean that survey data may do a poor job of measuring income growth at the top of the dis- tribution and thus may end up underestimating upward drift in inequality. This is a worldwide problem; however, if there is confi- dence in the quality of the data at the lower end of the distribution, one can be reasonably confident of trends in poverty and inequality in the vicinity of the poverty line. In the Region, because of data improvements in all but a handful of countries, the measured trends in poverty and inequality are robust. Factors Contributing to Poverty Reduction, 1998­2003 Growth in output and wages. Since 1999, the growth of the gross domes- tic product (GDP) in the Region has been impressive (higher than the world average), with the CIS the most rapidly growing subregion. For the CIS, the recovery of growth in Russia has been an important fac- tor. The devaluation that accompanied the financial crisis in Russia was important for restoring the exchange rate to a more competitive level and spurring the recovery of exports and growth. Combined with high prices for oil and other natural resources, this gave a huge boost to the Russian economy, which has in turn become a regional locomotive for many neighboring countries. Structural reforms that had been undertaken by many of the CIS countries enabled an improved supply response when the opportunity presented itself. For the EU-8, the prospect of accession provided a strong impetus for both reforms and growth, while the restoration of peace and stability in SEE was an important factor in sustaining recovery (figure 8). Mirroring the growth in output, there has been a sharp upswing in average wages in all economies in the Region. For example, in the low income CIS group, real wages have almost doubled since 1997. Data on wages by profession or by position in the income distribution suggest that this upswing has been shared alike by unskilled and skilled, poor and nonpoor. In most cases, wage increases have been larger than increases in productivity, reflecting in part the bounce- back of wages from the very low levels observed in the late 1990s in the CIS and parts of SEE. Overview 17 While growth has resulted in the creation of new and more pro- ductive jobs, only the fast-growing economies of the CIS have been able to create jobs at a sufficient pace to replace ones that were lost. Employment-to-population ratios have therefore stayed steady or declined almost everywhere outside the CIS (figure 9).15 Because of the failure to engender sufficient job creation, the EU-8 and SEE are well below the Lisbon targets of 70 percent of the labor force in employment.16 The position is somewhat different in the CIS, where employment levels tend to be higher and, in many cases, are trending upward (with some notable exceptions such as Georgia and Tajik- istan). While some fast-growing countries have succeeded in increas- ing wage employment (for example, Russia), in many low-income countries the main source of employment growth has been through the expansion in self-employment (for example, Moldova). Even where employment ratios are stable, there has been a contin- uing reallocation of labor across sectors. In most of the EU-8, agricul- ture employment fell, and its relative share of employment is now close to the EU benchmark. In contrast, agriculture employment increased in most SEE and low income CIS countries. Expansion of employment in services was observed in almost all countries in the Region. FIGURE 8 Since 1999, Growth Rates in the Region Have Been High, with the CIS the Most Rapidly Growing Subregion 15 10 5 % 0 rate, 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Growth ­5 ­10 ­15 ­20 ­25 EU-8 SEE Middle income CIS Low income CIS Source: World Development Indicators (World Bank 2005i). 18 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union What have these developments in the labor market meant for the poor? Growth incidence curves, which plot the increase in household income (consumption) by percentile, show that in most--but not all--cases, the poor saw a substantial increase in incomes during this period (figure 10). This is not surprising, given that at the start of this period, most of the poor consisted of households with working adults (World Bank 2000a) who would no doubt have benefited from the growth in wages across a range of sectors and professions. Little infor- mation exists on trends in earnings of the self-employed. One coun- try where the income of the poor declined rather than increased, leading to an increase in poverty, is Poland.17 Interestingly, this reduc- tion in incomes for the poor coincided with significant employment reduction in the economy as a whole and for the poor. Decline in inequality. As the growth incidence curves suggest, growth in incomes was proportionately higher for the poor than the rich in the CIS. This fact underlies the fall in inequality discussed previously. In contrast, in the EU-8 and SEE, growth was either pro-rich or pro- poor, depending on the country; therefore, trends in inequality are mixed. In the CIS, poverty reduction was aided by the fact that incomes of the poor grew more rapidly than those of the rich (that is, FIGURE 9 Employment-to-Population Ratios Are Well below Lisbon Targets (70 percent) in the EU-8 and SEE and Often Trending in the Wrong Direction 80 70 60 50 40 Percentage 30 20 10 0 Poland Hungary Bulgaria Romania Belarus Russian Georgia Moldova Tajikistan Fed. EU-8 SEE Middle income CIS Low income CIS Wage employment Left bar corresponds to 1998 or earliest year Self-employment Right bar corresponds to 2003 or latest year Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Employment and self-employment levels are derived from household survey data and may differ from official statis- tics in some respects. Employment population ratio is a percentage of employed among the population of working age (16­64 years old). Overview 19 the distribution of income changed in favor of the poor). In contrast, in countries such as Poland or Romania, poverty reduction was ham- pered by the fact that the incomes of the poor grew more slowly than those of the rich. What factors account for these changes in distribution? While there is no common pattern, there are some common trends across countries in the Region. In the CIS, declining wage arrears have been a feature of the economic recovery. Wage arrears were regressive in impact, driving up inequality among wage recipients (Lehmann and Wadsworth 2001); therefore, arrears reduction has likely been bene- ficial to equality. In contrast, in Poland and Romania, upward pres- sure on inequality from falling participation rates has been reinforced by rising inequality among wage earners. The latter is no doubt related to the further decompression in wages in these countries (World Bank 2003k; World Bank 2004h; World Bank 2005g).18 What were the roles of growth and changes in distribution in poverty reduction? Figure 11 plots the shares of growth and changes FIGURE 10 The Poor Have Benefited More than the Rich from the Growth Rebound in the CIS Poland (EU-8) 1999­2002 Romania (SEE) 1999­2002 8 16 6 14 4 12 % 2 % 10 0 8 ­2 6 Growth, ­4 Growth, 4 ­6 2 ­8 0 ­10 ­2 1 10 20 30 40 50 60 70 80 90 100 1 10 20 30 40 50 60 70 80 90 100 Percentile Percentile Russian Federation (middle income CIS) 1999­2002 Moldova (low income CIS) 1999­2002 70 100 90 60 80 % 50 % 70 40 60 50 30 40 Growth, 20 Growth, 30 20 10 10 0 0 1 10 20 30 40 50 60 70 80 90 100 1 10 20 30 40 50 60 70 80 90 100 Percentile Percentile Percentile growth rate Growth rate in mean Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Percentiles measure position in the distribution of per capita consumption from the poorest to the richest (100th percentile). 20 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union in distribution in poverty reduction for selected growth periods since 1998. The "growth share" measures how much poverty reduction can be attributed to growth in mean incomes on its own (that is, assuming no changes in the distribution), while the "distribution share" measures how much can be attributed to changes in the distri- bution of incomes alone, assuming no change in mean income. Fig- ure 11 highlights the overwhelming importance of growth to poverty reduction over the period in question. Relative to growth, the contri- bution of changes in distribution to poverty reduction has been rela- tively small. But as small on average as they appear, changes in distribution have clearly been quite important in some countries. For example, in Poland in 1998­99, 40 percent of the increase in poverty is attributable to the increase in inequality, while 60 percent is attrib- utable to the decline in income. In a number of countries in the CIS, the share of changes in distribution to poverty reduction in the period since the end of the financial crisis is more than 20 percent. Role of public transfers. In addition to wages, public transfers are an important component of household income and play an important role in poverty reduction. In most countries, social protection benefits FIGURE 11 The Share of Growth in Poverty Reduction Is Dominant across All Regional Subgroups 20 10 points % 0 poverty, ­10 in ­20 Change ­30 Fed. Fed. Rep. Poland Poland Hungary Belarus Romania Romania Ukraine Armenia Georgia Georgia Moldova Moldova Tajikistan Kazakhstan Kazakhstan Russian Russian Kyrgyz 2000­1 1998­9 2001­2 1998­9 2000­1 2000­1 2001­2 2002­3 1998­9 1999­ 2002­3 2000­2 1998­9 2001­2 2000­1 1998­9 1999­ 1999­ 2002 2002 2003 EU-8 SEE Middle income CIS Low income CIS Overall change in poverty: Due to growth Due to inequality Source: World Bank staff estimates using ECA Household Surveys Archive. Overview 21 have increased in the past five years in real per capita terms, along with the growth of fiscal revenues. Where data are available, they suggest that benefits have also improved in both coverage and ade- quacy. The reduction in arrears, particularly in pensions but also in other benefits, has no doubt contributed to these improvements. As a result, social protection transfers have come to play an important role in reducing poverty. Indeed, poverty would have been significantly higher in a hypothetical "no-transfers" situation (overview table 1). While somewhat simplistic, particularly in assuming no behavioral response in the no-transfer scenario (except in a few instances), the data are nonetheless illustrative of the importance of public transfers to poverty reduction, especially outside the low income CIS group. Private transfers. In the low income CIS countries and parts of SEE, remittances and other private transfers by far exceed publicly pro- vided resources. In some cases, remittances accounted for more than 10 percent of GDP and boosted consumption levels, including among the poor, helping to reduce poverty; however, the size of the impact is difficult to estimate because of various data limitations (Chernetsky Forthcoming). OVERVIEW TABLE 1 Transfer Payments for Social Protection Have Had an Important Role to Play in Reducing Poverty outside of the Low Income CIS Countries Increase in poverty without Country Year all social transfers, % EU-8 Poland 2001 141 SEE Bosnia & Herzegovina 2001 68 Bulgaria 2001 156 Romania 2002 49 Serbia 2003 41 Montenegro 2002 34 Middle income CIS Belarus 2002 143 Kazakhstan 2002 100 Russian Fed. 2002 68 Low income CIS Armenia 2001 12 Kyrgyz Rep. 2001 10 Benchmark Countries Vietnam 1998 5 Sources: For ECA, World Bank, various poverty assessments; for Vietnam, Van De Walle (2002). Note: Simulations use national poverty lines. Some behavioral response is assumed in Romania (50 percent of transfer in- come is replaced) and Serbia (72 percent of transfer income is replaced in rural areas, 87 percent in urban areas). 22 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union The impact of public and private transfers on inequality is mixed. Parts of the public transfer system, such as well-targeted social assis- tance programs, can be fairly progressive. Others may be regressive. The largest program in most countries, however, is the public pension program, which tends to be distributionally either neutral or regres- sive. The overall impact thus varies from country to country, with examples of both fairly progressive and fairly regressive systems in the Region. Unfortunately, there has been little systematic study of the contribution of public and private transfers to changes in inequal- ity over this period. Nonincome Dimensions of Well-Being, 1998­2003 What are the trends in the nonincome dimensions of well-being? Although there has been a reduction in poverty, trends in the nonin- come dimensions of well-being, such as access to education, health care, safe water, and heating, are markedly variable. Inequalities in access, whether to good schooling or health care or reliable water and electricity, persist and in some cases have increased, particularly in the CIS. In these countries, many people have thus come to have more income in their pockets, but in access to services and quality of services they may be no better off. Education. The most acute form of education deprivation is illiteracy. Average literacy among the transition countries of the Region is high (more than 98 percent), and in the transition country with the low- est level of literacy (Tajikistan), 96 percent of adults are literate. In a benchmark country, Turkey, literacy was much lower to start with and, despite increases, stood at just 87.5 percent in 2002. Thus, the extreme form of education deprivation does not appear to be a major issue in the Region. Since 1998, many countries in the Region have maintained or improved their high levels of school enrollment. Most countries entered the 1990s with a widespread network of education services that enabled them to achieve almost universal coverage in compulsory education. However, some of these achievements were eroded during the 1990s, particularly among the low income CIS group and some countries in SEE, although, even with the decline in coverage, enroll- ment in the compulsory cycle was typically higher than in comparator countries. Since 1998, enrollment in the compulsory cycle has been maintained or improved, except for some poor countries such as Geor- gia, the Kyrgyz Republic, and Tajikistan, which have still not managed Overview 23 to arrest the decline (figure 12). While there is evidence of some income gradient in enrollment, with children from better-off households hav- ing better coverage, the gradient is not large. Continuing high coverage in most countries of the Region suggests that the prospects for meeting the MDG of universal primary enrollment are fairly good (World Bank 2005c). Gender inequality in compulsory education has not been an issue, except in Tajikistan, where it continues to warrant attention. Compared with the primary level, enrollments at the secondary level have increased throughout the Region. This increase has gener- ally been accompanied by a reduction in enrollment gaps across income groups, except in a few low income CIS countries. Urban- rural gaps have also been reduced in virtually all countries. Interest- ingly, gender inequalities at this stage of education favor girls. The exceptions to this are Bulgaria and Tajikistan. Although the ratio of female to male enrollments in Tajikistan has increased over the past five years, it continues to be low by the standards of the Region. Returns to education, which were highly variable during the 1990s, particularly in the CIS, have now stabilized at levels similar to those of market economies. This underlines the value of investment in educa- tion, not only for its own sake but also as a means for ensuring ade- quate standards of living, particularly for the poor. However, returns are a function not simply of access but also of quality of education. Compared with enrollments, trends in quality of education are less sanguine. Despite increases in spending on a real per capita basis almost everywhere in the Region, the failure to invest sufficiently in the quality of infrastructure or staff means that quality is not being maintained. For example, Trends in International Mathematics and Science Study (TIMSS) data suggest that although performance of eighth graders continues to remain good relative to those in other countries, including those in OECD countries at higher levels of income, scores are declining in all but a handful of countries in the EU- 8. Where the analysis is available, it suggests that the declines are in large part due to a sharp increase in the share of students who are seri- ously underperforming. Often, these students tend to be in schools where the quality of service provision is marginal, such as rural schools. Reading scores of 15-year-olds from the Programme for Inter- national Student Assessment (PISA) present no different a picture. Apart from a small handful of countries in the EU-8, scores are declin- ing or low. Health care. Trends in health status and health care utilization are mixed. Declines in male life expectancy (particularly in the successors of the Former Soviet Union), which had become one of the most 24 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 12 Access to Secondary Education Has Gone Up Virtually throughout the Region, but Some Countries Continue to Struggle to Arrest the Decline in Primary Enrollment Rates Enrollment in Primary Education (7­14 years) 100 98 96 school in 94 enrolled 92 % 90 88 Poland Hungary Bulgaria Romania Russian Kazakh- Kyrgyz Georgia Moldova Armenia Tajikistan Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998 2002­03 Enrollment in Secondary Education (15­17 years) 100 95 90 85 school 80 in 75 70 enrolled % 65 60 55 50 Hungary Poland Bulgaria Romania Russian Kazakh- Armenia Georgia Kyrgyz Moldova Tajikistan Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998­99 2002­03 Source: World Bank staff estimates using ECA Household Surveys Archive. Overview 25 widely documented negative health outcomes of the transition, have generally been arrested. However, many of the proximate causes of high male mortality, notably the high incidence of cardiovascular and circulatory disease and death from accidents and acts of violence, remain. As with male life expectancy, child and maternal mortality are also trending in the right direction. However, the very slow progress in achieving reductions in mortality and concerns about the delivery and quality of critical medical services imply that many coun- tries in the CIS appear unlikely to meet the child and maternal mor- tality­related Millennium Development Goals (MDGs) (World Bank 2005c).19 There is a growing threat to the health of the Region's population from HIV/AIDS and tuberculosis (TB), particularly in the CIS, but also to some extent in SEE and the EU-8 (the Baltic countries). The Region as a whole currently has one of the most rapidly growing infection rates of HIV/AIDS in the world, due to problems related to the increase in injecting drugs and commercial sex work, a concurrent increase in the incidence of sexually transmitted infections (STIs), high migration rates, limited capacity of governments and civil soci- ety to implement effective preventive responses, and low levels of awareness of HIV and STIs. Drug transit through the CIS and growth of local consumer markets for drugs are also contributing to the prob- lem. At current rates of infection and treatment, the HIV/AIDS MDG is unlikely to be attained by a broad swath of countries in the Region (World Bank 2005c). Countries in the Region have a large network of public health providers that distributed generous services and that suffered major fiscal restrictions during the 1990s. Between 1994 and 1999, coun- tries in the Region spent on average 4 percent of GDP on health, rang- ing from 1 percent in Georgia (low income CIS group) to 9 percent in Croatia (SEE). After 1999, some countries in the low income CIS group continued to experience reductions in public spending on health. Other poor countries were able to stem the decline in spend- ing, but only at very low levels of spending (for example, Armenia). Even where funding may be on the upswing, outside the EU-8 it is not close to levels experienced at the outset of the transition. While funding levels may have stabilized or even increased, this is not reflected in improving quality, particularly for the poor, because of three factors. First, the very large network of providers has largely been retained, resulting in an underfunded, and hence ineffective, network in many countries. Second, the lack of resources for basic interventions like public health activities has resulted in a repeated failure to stem communicable diseases. Third, the changing demographic composition 26 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union of the population (which is becoming older) has altered the morbidity profile and increased the costs of health provision. Vulnerable popula- tion groups (such as rural or small-town dwellers or the poor) are expected to have borne the brunt of not only the misallocation but also the failure of resources to keep pace with needs. The lack of public resources increased the use of fee-for-services in a mostly unregulated setting, reducing the demand for health care. Official statistics show the decline in utilization of health care during the 1990s, but after 1999, this remained stable or even recovered (fig- ure 13). Inpatient care in the low income CIS group declined more than 20 percent between the mid-1990s and 2000, only to stop after 2001. Countries in the middle income CIS group continue to have very high hospital utilization rates, higher than the EU average. Sur- vey data on utilization, which control for need, suggest a similar pic- ture, but point in many cases to persistent differentials across rural and urban areas, particularly in the CIS. Infrastructure. Turning to infrastructure, here too the picture is very mixed. Data problems make this a particularly difficult area to analyze. However, what is clear is that countries of the Region began transition reasonably well covered with basic utility services, but the economic shocks of the early reform years left providers strapped for funds, which meant that utilities deteriorated Regionwide for much of the 1990s. Since then, the decline in utility performance (as measured by access and quality) has been reversed or slowed. Electricity has shown FIGURE 13 Hospital Utilization Rates Have Recovered, but Remain at Low Levels in Parts of the Low Income CIS Group 33 people 28 23 100,000 per 18 13 8 Admissions 3 1994­8 1999 2000 2001 2002 Lithuania (EU-8) Romania (SEE) Belarus (middle income CIS) Georgia (low income CIS) Source: World Health Organization (WHO), based on official statistics of hospital admissions. Overview 27 the greatest improvement: providers have maintained near universal coverage while improving reliability in subregions where it was partic- ularly poor, such as in the low income CIS countries. Other recent gains include the expansion of gas supply networks to many house- holds affected by the collapse in district heating, and the improvement of water reliability in some countries. Despite these improvements, many households, including many urban households in the CIS and in some parts of SEE, continue to use dirty fuels such as coal, wood, and oil for heating because they lack access to gas and cannot afford, or are not reliably provided with, electricity. Available survey evidence suggests that the reliance on solid fuels for heating affects the poor more than the rich (figure 14). This could have a long-term impact on the health of the poor due to the negative impact of increased indoor pollution. Over the past five years, access to clean heat has become less FIGURE 14 The Poor Make Greater Use of So-Called Dirty Fuels for Heating Percentage of Population Using Clean Sources of Heating in the Poorest and the Richest Quintiles (Latest Available Year) Hungary EU-8 Bulgaria SEE Romania Kazakhstan CIS Middle income Armenia CIS Moldova income Low Tajikistan Turkey mark Bench- 0 10 20 30 40 50 60 70 80 90 100 % of population Poorest quintile Richest quintile Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Clean fuels are all sources of heating for a household except solid fuels such as wood and coal, which are classified as "dirty." 28 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union equal in some countries of the low income CIS group (for example, Armenia), although in others (such as Tajikistan), the revival of district heating has improved equity. Although time-series data tracking water availability are available for only a few countries, the evidence from the low income CIS coun- tries shows the influence of years of little maintenance and no invest- ment in water provision (figure 15). Although households officially still have water connections, little water flows through the pipes. On average, Tajik households receive water for less than six hours each day, and although not depicted here, households in smaller cities and rural areas have the least water. Are key public services affordable? Both catastrophic health spending and the additional cost of utilities pose the challenge of affordability for the poor. Household expenditure shares for utilities have continued to increase from the late 1990s to the present (figure 16). They are typically higher for poor households than for rich ones. On average, however, expenditure shares are highest in the EU-8, followed by SEE, the middle income CIS countries, and the low income CIS group, as shown in the figure. The increase in utility expenditure shares is largely driven by the increasing cost of electricity and, in the EU-8, the price of water. FIGURE 15 In the Low Income CIS Countries, the Reliability of Water Supply Is Low and Shows Little Improvement Access to Uninterrupted Water Supply 24 20 day per 15 supply of 10 Hours 5 0 1997 2001 1999 2003 2002 1999 2003 Bulgaria Armenia Georgia Tajikistan SEE Low income CIS Poorest quintile Richest quintile Source: World Bank staff estimates using ECA Household Surveys Archive. Overview 29 FIGURE 16 Household Expenditures on Utilities Have Increased Expenditure Shares on Electricity, Heating, Water, and Sewerage 24 22 20 18 16 14 expenditures 12 10 total of 8 % 6 4 2 0 FYR Fed. Rep. Latvia Estonia Poland Hungary Albania Turkey Lithuania Bulgaria Belarus Romania Ukraine Armenia Georgia Moldova Vietnam Azerbaijan Tajikistan Colombia Montenegro Kazakhstan Russian Kyrgyz Uzbekistan & Macedonia, Serbia EU-8 SEE Middle income Low income CIS Benchmark CIS 1998 or earliest 2003 or latest Source: World Bank staff estimates using ECA Household Surveys Archive. Note: For Albania, Latvia, Ukraine, Serbia and Montenegro, Azerbaijan, Turkey, and Colombia, data before 2002 are not available. For Estonia, Armenia, Kyrgyz Re- public, and Uzbekistan, 2000 is used instead of 1998; for Tajikistan, 1999; and for Kazakhstan, 2001. Catastrophic health expenditures run the danger of impoverishing households in parts of the Region such as the low income CIS coun- tries, where the health system relies heavily on household contribu- tions and households are relatively poor. Outside the EU-8, where the impact is more limited because of higher incomes, simulations undertaken for the purposes of this report suggest that catastrophic health spending can increase the fraction of the poor population by 3­9 percent. Prospects for Poverty Reduction Given what has been achieved, what are the prospects for poverty reduction? In drawing lessons, it is worth reminding ourselves of the main concerns five years ago when Making Transition Work for Every- one (World Bank 2000a) was published. Then, although growth had 30 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union recovered in the EU-8 and parts of SEE, recovery was slow at best in the CIS, where most countries suffered an additional blow because of the financial crisis in Russia. The collapse in output, particularly in the CIS, combined with the increase in inequality, meant a sharp increase in poverty. Prospects for poverty reduction were felt to be unclear even in the event of the resumption of growth because a core group of the very poor--the long-term unemployed and socially excluded--were likely to be bypassed by growth. Despite notable achievement in the pretransition period, education and health sectors were under strain and working to the detriment of poor families and the economic mobility of their children. And to top it all, data issues clouded researchers' understanding of poverty. The picture looks different five years later, and one's understanding of the challenge of poverty needs to be suitably nuanced. Economic growth has firmly returned to the Region, and all countries are expe- riencing positive growth. In addition, changes in inequality have been modest over this period. Moreover, in the CIS, which had previously seen the sharpest increases, inequality has (with the few exceptions noted previously) abated. The rise in output and the moderation in inequality have together meant a substantial reduction in poverty.20 The long-term unemployed and the socially excluded remain a con- cern, particularly because very few countries outside of the rapidly growing economies of the CIS have succeeded in raising the share of the population that is employed. However, the failure to raise overall employment levels has not acted as a brake on poverty reduction to date (except perhaps in a few countries such as Poland) because, in most countries, the bulk of the poor retain some attachment to the labor market and have benefited from the bounce-back in real wages. The health and education sectors in most countries have benefited from increased levels of funding, but because of remaining inefficien- cies in delivery, this has not translated into uniform improvements in quality. Access is possibly less of a concern, with poor quality and high cost (or reduced affordability) becoming key dimensions of depriva- tion. The quality of data has improved, encouraging more confidence about observed trends. At the same time, shortcomings remain, espe- cially with regard to data on the quality dimension of public services. However, while the picture looks somewhat different, and notwith- standing the impressive reduction in poverty in the Region over the past five years, it is clear that there is a long road ahead, not just for the low income CIS countries but also for many middle-income countries in the Region (where poverty rates are lower, but where most of the poor live). This is for a number of reasons. First, despite the recovery, poverty rates remain significant: for many countries, it will be some Overview 31 time before absolute poverty is eradicated (on this, see further below). Second, the recovery is still recent for many, if not most, countries, and large numbers remain vulnerable to poverty in the event of an eco- nomic downturn. Third, despite recent improvements, morale (as revealed in self-reported assessments of well-being) remains low com- pared with that of countries at similar levels of income. This may be related to uncertain prospects for the future. Low morale may also be related to greater inequality and erosion in access and quality of public services compared with the past. Projecting poverty rates over the medium term, using available economic growth forecasts from the World Bank's Global Economic Prospects 2005, indicates that poverty will not disappear altogether and, together with economic vulnerability, will affect 30 percent of the population by 2007. This is not to say that there will be little poverty reduction. In fact, poverty will fall by 7 percent a year, or 21 million fewer people will be in poverty in the five years covered by the forecast period (figure 17). Impressive though this reduction may FIGURE 17 Growth Will Move an Additional 21 Million People out of Poverty by 2007, but 40 Million Will Remain Absolutely Poor and More Than 100 Million Vulnerable to Poverty 100 90 212.1 80 258.2 70 329.2 60 50 population of 40 160.7 % 30 153.3 108.8 20 10 102.0 61.2 40.0 0 1998­9 2002­3 By 2007* Nonpoor: above $4.30 a day Vulnerable: above $2.15 and below $4.30 a day Poor: below $2.15 a day Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Growth rates for 2002­2007 are from the World Bank's Global Economic Prospects. * = simulations. 32 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union OVERVIEW BOX 2 Nonincome Dimensions of Poverty and Millennium Development Goals in the Region Of the Region's subregions, the EU-8 is perhaps the least challenged by the nonincome di- mensions of poverty. Most countries have met, or are likely to meet, all of the nonincome MDGs. However, for the Baltics, current trends suggest that the spread of HIV/AIDS may not be effectively combated. The nonincome dimensions of poverty are likely to challenge some countries in SEE. It is not clear that countries such as Bulgaria and Romania will be able to combat the spread of HIV/AIDS. Romania may also struggle to meet the water access MDG because only 16 percent of the sizable rural population is assessed to have access to an improved water source. The nonincome dimensions of poverty, particularly related to health, are likely to chal- lenge the middle income CIS countries. None of these countries is assessed as likely to be able to combat the spread of HIV/AIDS. The targets for reductions in child mortality and mater- nal mortality may also not be achieved in some countries. It should be pointed out, however, that because of the age and epidemiological profile of these countries, proportionately higher gains in life expectancy would accrue from reducing adult mortality through the control of noncom- municable diseases than from achieving targets related to the MDGs. The low income CIS countries are most severely challenged on nonincome dimensions of poverty. Most MDGs are unlikely to be met in the low income CIS group. Indeed, only the MDG regarding attaining gender equity in schooling is on track. In benchmark Turkey, the nonincome dimensions also represent a challenge. In particular, even though the gender gap has been closing, girls are significantly underrepresented in primary and secondary schools. The MDG for maternal mortality also appears unlikely to be achieved, with maternal mortality rates unusually high for a middle-income country. Source: World Bank 2005c. be, some 40 million people are projected to remain absolutely poor in the Region by 2007. Naturally, faster growth could lead to faster reduction of poverty rates. Sustained economic growth is hence a crucial component of any poverty alleviation strategy. These projec- tions do not incorporate any worsening of the income distribution, which would undermine the impact of growth. Given that inequality levels in the Region are, broadly speaking, at the low end by world standards, some worsening of inequality over the medium term would not be surprising. These projections should therefore be under- stood as a best-case scenario. Overview 33 Looking behind the regional aggregates and using subgroups' spe- cific targets indicate that all country subgroups face challenges in poverty reduction over the longer term. A forward-looking agenda could, for example, aim to meet the MDG on poverty--halving absolute poverty by 2015 compared with 1990 levels--which is most relevant for the low income CIS countries. For the middle income CIS group and SEE, which have moderate poverty, the MDG is not suffi- ciently ambitious. Eliminating economic vulnerability, which this report takes as having income (consumption) levels that are above twice the poverty line, is an appropriate "modified" MDG for this group. For the EU-8, who have minimal absolute poverty but are sig- nificantly poorer than the 15 EU member states they have joined, a modified MDG could be to reduce poverty by half, assuming the poverty line to be the lowest among the EU-15.21 Overview table 2 shows that, despite recent progress, all subgroups face a real chal- lenge in poverty reduction. The required growth rates are signifi- cantly higher than the rates at which these countries are expected to grow. Moreover, as was the case in figure 17, no change in income distribution is assumed. With worsening inequality, either the goals would not be achieved, or the required growth rates would be even higher. As with income, countries in the Region will continue to face challenges on nonincome dimensions of well-being. This is most simply characterized through the prospects of attaining the MDGs on education, health, and the environment, which in many ways rep- resent a subset of aspirations on the nonincome side. The World Bank (2005) projects the likelihood that countries will attain the MDGs, based on recent trends in the indicators. It concludes that the OVERVIEW TABLE 2 Achieving Subgroup-Appropriate Poverty Reduction Targets over the Long Term (2015) Will Require Significantly Higher GDP Growth Rates Long-term growth rates Medium-term Baseline required to growth forecast poverty rate Groupings meet targets * (2002­7)** Specific target (2002)*** Low income CIS 5.6% 3.9% Reduce poverty by half relative to 1990 (at $2.15 a day) 52.3% Middle income CIS 9.7% 6.8% Eliminate economic vulnerability ( at $4.30 a day) 39.6% SEE 10.8% 5.4% Eliminate economic vulnerability (at $4.30 a day) 55.3% EU-8 6.6% 4.3% Reduce poverty by half, taking as poverty line the lowest line in EU-15 today 36.6% Source: World Bank staff estimates; growth rates are from GEP. Note: * Subregional country averages weighted by GDP. ** GEP, Global Economic Prospects (World Bank 2005b). *** Population weighted. 34 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union attainment of the nonincome MDGs in the Region is not assured. The health MDGs, which require reductions in child and maternal mortality and effectively combating the spread of communicable dis- eases such as HIV/AIDS, represent a challenge in many, if not most, parts of the Region (see overview box 2). For the low income CIS group, most MDGs--including improvements not only in health but also in education and water supply--remain a challenge. Thus, tak- ing both the income and nonincome dimensions into account, the picture is one of a recovering Region, but one where much still remains to be done. The Role for Public Policy It is clear from the above that accelerated and shared growth, along with reform of public service delivery and better targeting of social programs, will be key to making progress on both income and nonin- come dimensions of poverty. It is also important to be able to monitor progress in poverty reduction. Within these four areas, what are the priority actions for public policy? Accelerating Shared Growth It is difficult--based on the experience of the past five years--to overemphasize the importance of raising and sustaining high rates of growth for poverty reduction. As the simulations in overview table 1 suggest, accelerated growth is essential for poverty reduction. The EU-8 is already well placed to take advantage of the new economic opportunities and market integration provided by EU accession. Enhanced competition and the mobility of both products and factors of production that EU accession provides will likely become a dynamic source of growth in the future. This is also true, but perhaps to a more limited extent, for countries with the prospect of accession. But for low and middle income CIS countries that do not yet have such an external driver for change, domestic catalysts remain crucial. Good economic governance and responsible leadership must take advantage of the relatively good economic times to put into place policies and institutions that would enhance growth. Understanding the policies and institutions that lead to strong and sustained rates of growth is therefore a first step in reducing poverty. While this report has less to say on factors that drive growth--which is not the focus of this study--the pursuit of sound economic policies is a necessary precondition. These include sound monetary and fiscal Overview 35 policies (reflected in, for example, moderate-size government and low inflation), a climate conducive for investment, a relatively well-devel- oped financial system, and trade openness. Countries of the Region are, with few exceptions, relatively well integrated in world markets, although more can be done (World Bank 2005 Forthcoming c). However, beyond these broad issues around promoting growth, the diagnosis in this report points to a number of areas where more could be done either to increase the assets of the poor or to create greater returns to their assets. These relate to (a) promoting enter- prise reform, (b) boosting growth and productivity in agriculture, and (c) promoting opportunities for those in lagging towns and regions. The report considers each in turn. Promoting enterprise sector reform. Encouraging the growth of new, more productive firms and strengthening the financial discipline for exist- ing enterprises continue to be important for both poverty reduction and accelerated GDP growth. The typical economy of former socialist countries continues to face significant productivity differences across old, restructured, and new firms within the same sector (World Bank 2002h). New firms are typically the most productive, reflecting not just the more efficient use of resources but also the relative dynamics of different kinds of firms and the very different policy environment in which they function. This historic legacy of transition is reflected in the large earnings gap between the poor and the nonpoor as many of the workers belonging to poor households are trapped in the old, unrestructured, low-productivity firms. Accelerating reform of the enterprise sector and of the business climate as a whole to create a level playing field across all firms and--in particular--to encourage the entry and growth of new firms is thus an important factor for equalizing the returns to labor and reducing poverty. This is particu- larly important for CIS countries, which despite recent progress con- tinue to lag behind other countries in ease of doing business (World Bank Forthcoming d). Boosting growth and productivity in agriculture. Many of the poor in the Region are in rural areas, where poverty is proving more resistant to growth than in urban areas. Agriculture is the main activity in rural areas; thus, stimulating agricultural growth is crucial for poverty reduction. Where land reforms have been implemented, especially where initial conditions favor labor-intensive cultivation (low income CIS), land distribution resulted in significant productivity and income gains to rural households. Where land reforms are incomplete (for example, some middle income CIS) or where land market operations 36 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union need to be improved to facilitate land restructuring (for example, in SEE), significant income gains can be attained from accelerated reforms. In all countries, future gains in reducing poverty in rural areas would hinge on eliminating key market imperfections in input and output markets essential for enabling self-employed farmers to lift themselves out of poverty. In particular, the integration of rural areas into national credit markets is critical for further investments and productivity growth in agriculture. More broadly, improving the investment climate in rural areas is very important. Increasing evi- dence shows that investments in food processing, agribusiness, trade, and retail companies play a crucial role in helping small farmers over- come input and output market imperfections, in helping them upgrade the quality of their products, and in accessing markets (World Bank 2005a). Beyond measures related to agriculture, inte- gration of the rural poor into national labor markets--either through rural off-farm employment or by improving access to urban labor markets--and adequate social safety nets will be crucial for sustained income growth and poverty reduction, particularly in the middle income CIS and SEE countries. Emphasis on rural service delivery and infrastructure is also critical, especially in the low income CIS, not only for its instrumental role in raising rural incomes, but also as an aspect of poverty that warrants attention in its own right. Promoting opportunities in lagging regions. Countries in the Region face substantial differences in poverty rates between urban and rural areas and between capital cities and smaller towns that, if severe, risk per- petuating intergenerational poverty and inequality traps and act as a drag on economic growth. Most countries seek to address regional inequalities through the maintenance of a stable macroeconomic envi- ronment, the creation of a level playing field for businesses, and fiscal transfers for targeted programs in lagging regions. But more can be done. First, countries need to enhance labor mobility. When people move to economic nodes that promise a higher expected income, it helps to reduce spatial income disparities. Adoption of appropriate poli- cies to encourage movement, supported by the development of urban housing markets and policies, credit markets, and entitlement reform can provide a strong stimulus to inter- and intraregional mobility and help improve income levels in relatively poorer areas while also boost- ing competition, productivity, and growth in destination areas. Second, in countries with decentralized fiscal systems, there is a strong role for equalizing resource transfers to address regional inequalities. In partic- ular, social and economic reforms in the lagging regions can be encour- aged through market-based incentives, including the use of competitive Overview 37 allocation mechanisms for fiscal transfers. Third, education and health service delivery should be strengthened in lagging regions to ensure adequate human capital formation as a route out of poverty. In partic- ular, existing inequalities in access to public services and quality of ser- vices provided need to be addressed as a priority. Strengthening Public Service Delivery Ensuring access and improving quality of education and health care require strengthened accountability arrangements. Although low lev- els of spending are an issue, more so in education than in health care, only a few countries spend less than is warranted, given levels of income. Thus, going forward, most countries will need to operate within the available resource envelope. Reforms will therefore have to focus on improving the quality and efficiency of public spending. Enhancing quality and equity of education services. In education, the low income CIS group needs to stem the decline in primary enrollments and quality of education, in particular by ending the situation in which staff are underpaid and complementary expenditures (on textbooks, heating, and repairs) are underfinanced, while at the same time employment and, in some cases, facilities remain well above standards common in much richer countries. In addition, some countries may need to ensure greater equity in education spending across subna- tional regions (for example, the Kyrgyz Republic). Ensuring access to primary education is much less of an issue outside the low income CIS countries. Here the main issue is secondary education, where quality and relevance to market demand are often in question. Governance reforms that both strengthen government accountability for outcomes as well as increase participation and voice will be essential to improv- ing outcomes. Lessons from the experience of the EU-8 in raising qual- ity certainly point in this direction. In particular, decentralization of services to allow for a greater role for both school administrators and parents has an important role to play in stemming declines in quality. Strengthening access to, and quality of, health care. In health care, low- income countries suffer from having to provide for a range of ser- vices when budget resources are limited, but even the available allocations are not spent wisely. This is reflected in the large share of household contributions in total health spending. Improving utiliza- tion among the poor is closely linked to financing and quality issues. Tough decisions are required on the size of the basic package and a major reallocation of expenditure--and greater accountability for its 38 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union use--implemented to improve access to, and quality of, care (World Bank 2005d). To improve matters, accountability relationships between politicians and citizens need to become more effective (through such means as more organized voice power of citizens, cit- izens' report cards, and informed voting), and the accountability relationships between politicians and providers need to be strength- ened (through such means as clarifying responsibility, aligning incentives between policy maker as principal and provider as agent, and better enforcement of contracts between organizational and front-line providers). Countries such as Armenia have shown that, even with limited resources and high poverty rates, improvements in key dimensions such as affordability can be made, albeit on a mod- erate scale. At the other end of the spectrum, the EU-8 is struggling to maintain the easy access to a wide range of health services in a context of rising costs. Clearly, further efficiency-enhancing mecha- nisms as well as private financing will need to be found to control expenditures. Managing reform of utilities. Service quality in many of the infrastruc- ture services is extremely poor in the low income CIS countries, and even in richer countries there are large disparities between service quality for the poor and the nonpoor. The infrastructure needs of the poor are unlikely to be met without reform of the utilities sector to bring it to a financially self-sustaining basis, which would encourage much-needed upkeep and maintenance of viable infrastructure and improvements in service quality. Improving financial performance will involve raising tariffs, which--except for power, where there has been some movement toward cost recovery, and possibly water in the EU-8--are well below cost recovery levels. Further movement toward full cost recovery in power is expected to have a limited impact on poverty, except in the poorest countries. However, across- the-board increases in the full range of utilities is expected to have a more serious impact on poverty (World Bank 2005 Forthcoming b). The social impact of tariff increases will need to be factored into the sequencing and pace of reforms in the event of across-the-board reforms in a range of utilities. Where the social safety net is adequate, it can be expected to mitigate the impact on the poor. However, where social safety nets are relatively thin, as for example in the low income CIS group, other options are worthy of consideration. For utilities that can be metered, lifeline tariffs can serve as a useful temporary cushion; however, where lifelines are not practical (as, for example, in sectors where consumption cannot be measured), reforms would need to be calibrated to affordability. Overview 39 Enhancing Social Protection Strengthening the social safety net. Given their importance for poverty reduction and the broad improvements over time, it should be clear that countries need to maintain ongoing social insurance and social assistance reforms, which are largely designed to improve sustainabil- ity, and to enhance coverage and targeting of the poor within the available resource envelope. In the low income CIS group, the main constraint will continue to be the fiscal means to cover the population adequately. In the middle income CIS group and SEE, although there is more fiscal space for social protection, there is also greater resist- ance to reforms, as suggested (for example) by the difficulties with the monetization of privileges in Russia. While the objective of the reforms is not in question, the difficulties in implementation serve as a useful reminder of the importance of sequencing with other social and economic reforms, the need to protect the most vulnerable groups, and an appropriate communications strategy to explain the benefits of reforms. Where systems are more generous, as for exam- ple in parts of SEE and the EU-8, a balance will need to be struck between the need for social protection and labor market incentives. Strengthening targeted interventions for marginalized groups and minorities. This may be in the form of assistance in cash or in-kind (such as edu- cation, health, or housing), depending on the nature of the group. For the long-term unemployed or nonparticipants, active labor mar- ket programs can be particularly relevant. But evidence from success- ful training programs suggests that these should be targeted, offered on a selective basis, with clear links to potential employers, and in collaboration with the private sector. It is important to bear in mind that there is limited evidence of successful retraining programs from the low income CIS group. In some cases, interventions may need to be integrated across many fronts. For example, for the Roma minor- ity of the EU-8 and SEE, governments are taking a holistic approach to ending persistent exclusion by setting goals in four areas--educa- tion, employment, health, and housing. Other minorities may require a different approach. The elderly, especially those who are very old or living alone, may also require special interventions such as supple- mentary cash benefits or provision of assisted living services. For most marginalized groups, however, additional assistance, whether in cash or kind, need not be provided by the public sector alone. Civil society organizations, community-based groups, and other organizations could also be encouraged to come into the sector under the overall direction of the government. 40 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Ensuring adequate minimum wages. Minimum wages are an important policy instrument for enhancing the income security of the poor. These can help provide a floor to income, but need to be kept at a rea- sonable level. In this context, the large real increases in the CIS in recent years have brought minimum wages closer to subsistence lev- els. However, future increases in minimum wages need to be consid- ered carefully, so as not to become so high that they have negative effects on growth and employment, with adverse impacts on poverty. Where variations in regional income and labor market profiles are large, governments may need to consider setting region-specific min- imum wages, which may help improve the employability of certain groups of workers (such as younger workers and those in lagging regions). This is an issue particularly in the EU-8 and SEE, where minimum wages represent a relatively high proportion of the average wage and the adverse impact of common minimum wages is particu- larly noticeable. For example, the relatively high minimum wage is found to constrain employment opportunities for the low-skilled in countries such as Lithuania and Poland. Monitoring Progress on Poverty Reduction Huge progress has been made in recent years in improving the qual- ity and accessibility of poverty data. Countries need good survey data to monitor changes in poverty and to evaluate the impact of specific policy actions on the poor. This report documents significant progress in collecting up-to-date high-quality data across the Region. The pre- vious report on poverty (World Bank 2000a) relied on a single survey for many countries and could produce an estimate of poverty over time for only three countries. With full data sets closed to users out- side statistical offices, the report also had to rely on partial data. Since then, many countries have started implementing regular surveys that periodically collect representative data on income and nonincome dimensions of living standards. In addition, data are provided openly to researchers for the purposes of study and policy evaluation. These improvements are not only confined to EU-8 countries (for example, Hungary) but also cover SEE (for example, Romania), middle income CIS countries (for example, Kazakhstan), and low income CIS coun- tries (for example, Georgia and Moldova). But many challenges remain. First, improvements to data quality and availability are very recent, and for many countries in the Region reliable data on poverty changes can be obtained for only a few recent years. The efforts in collecting data need to be maintained. Second, survey coverage and response rates have fallen over time in all coun- Overview 41 tries, and there is a need to strengthen the technical capacity of sta- tistical offices to curb this trend and deal with it appropriately. Third, wide gaps exist in data collection on the nonincome dimensions of poverty: there are practically no attempts to gauge trends in the qual- ity of health care and infrastructure services, and even indicators of access are not consistently collected. Given the changing nature of poverty with the increasing role of nonincome components, this gap is the most worrying. Fourth, not all countries have opened their data sets to researchers, undermining the effective use of public funds spent on data collection and monitoring. These areas--keeping up with periodic surveys to provide comparable data, collecting informa- tion on nonincome dimensions, and opening up access to survey data--are priorities for action to ensure adequate information sup- port for poverty reduction efforts. (See overview box 3 for a discus- sion of the data used for this report.) Conclusions The countries in the Region have made significant progress in reducing poverty. More than 40 million people moved out of poverty during 1998­2003. Much of this poverty reduction derives from the growth rebound in the CIS countries. But poverty and vulnerability still remain a significant problem: more than 60 million are poor, and more than 150 million are vulnerable. Most of the poor are the working poor. Many others face deprivations in access and quality of public services. Regional inequalities both between and within countries are large. The highest levels of absolute poverty are found in poor countries of Cen- tral Asia and the South Caucasus, but most of the Region's poor and vulnerable are in middle-income countries. Notwithstanding the tremendous heterogeneity among countries in the Region, reducing poverty and vulnerability requires an accel- eration of shared growth, strengthening of public service delivery, better targeting of social protection, and regular monitoring of progress in poverty reduction across the Region. In promoting accel- erated shared growth, the report emphasizes (a) further reform of the enterprise sector to encourage the release of resources from the old, less productive sectors to the new, more productive sectors; (b) further reforms to promote agriculture and rural growth by integrat- ing rural areas into the rest of the economy with regard to labor and capital markets, access to credit, trade and services; and (c) policies to promote greater opportunity in lagging regions. Public service deliv- ery needs to be improved through increasing the accountability of 42 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union OVERVIEW BOX 3 Data forThis Report:TheWorld Bank's ECA Household Survey Archive To arrive at the internationally comparable assessment of poverty, this report uses primary unit record data from recent household surveys to construct a comparable indicator of living stan- dards across all countries in the Region. Income data, an alternative to consumption for meas- uring living standards, remain particularly difficult to collect in transition countries. In contrast, practice has shown that consumption data can be gathered with a great degree of precision. Survey consumption modules have become more detailed over time and better capture various dimensions of consumption. In relying on consumption of goods and services by a household as the measure of living standards, a number of conceptual and practical issues needed to be addressed. First, unlike food, consumer durables and housing are consumed over a long time. It is customary, there- fore, to include the imputed value of the consumption flow associated with the possession of consumer durables (including housing), but exclude the expenditure on the purchase of these goods. However, for the Region, data availability limits the application of this approach to all countries. This report does not, therefore, include estimates of flow of services of durables, nor have researchers added in durable purchases or rents. Second, when con- sumption is used as a measure of well-being, higher consumption should indicate a higher level of well-being. For most consumption items, this correspondence is reasonable; howev- er, for some categories, such as health expenditures, this correspondence is questionable. As a result, health expenditures were not included as a part of consumption (Deaton and Zai- di 2002). Third, given the significance of spatial differences, the authors adjusted for spatial price differences, using the same set of information in all countries. In the cases where data government and the voice and participation of citizens. This is essen- tial to improving access and quality of social services, which are not only important in their own right but also of instrumental value in helping the poor move out of poverty. The report emphasizes the need to further strengthen the social safety net to meet the chal- lenges of restructuring economies. Finally, monitoring progress on poverty reduction on a regular basis needs good-quality household survey data sets that are publicly available for research and analysis. Endnotes 1. All monetary amounts are in U.S. dollars ($) unless otherwise indicated. 2. The new member states of the European Union are the Czech Republic, Overview 43 were collected over a long time, it was also necessary to adjust for changes in prices over time. Fourth, households in the Region cope with poverty by relying on an array of nonmar- ket strategies, including producing their own food and engaging in reciprocal exchange with other households and institutions. A consistent approach was used in assigning a monetary value to these components of consumption. Fifth, to adjust for differences in household composition, researchers took the simplest approach and used the per capita scale. Sixth, the procedure, which conforms to methods used in other international household survey data depositories (such as the Luxembourg Income Study; see Gottschalk and Smeeding [1997] and www.lisproject.org) was used to clean the data of outliers across all data sets. Following a consistent approach across all data sets gives reasonable confidence that differ- ences across countries in the final consumption measure are due to differences in the primary data and are not due to the method of aggregation. This is the first time (to the authors' knowl- edge) that comparable consumption aggregates have been constructed for countries in the Re- gion. (Full details are provided in appendix A, Data and Methodology.) The constructed estimate of real per capita consumption has several shortcomings that reflect some persistent data problems in the Region. First, it ignores the differential impact of price in- creases on the poor and nonpoor. No price indices for low-income groups that would allow this issue to be addressed are routinely available in the Region. Second, over time, there has been considerable deterioration in response rates in many countries. Countries deal with this problem in different ways, which may have (as yet unknown) implications for survey-based poverty and inequality measures. Notwithstanding these limitations, consumption indicators constructed for this report produce a reasonably reliable anchor to assess changes in poverty and distribution during 1998­2003. Sources: Deaton and Zaidi 2002, Gottschalk and Smeeding 1997, and www.lisproject.org. Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic, and Slovenia. 3. Countries in SEE consist of Albania, Bosnia and Herzegovina, Bulgaria, Croatia, the former Yugoslav Republic (FYR) of Macedonia, Romania, and Serbia and Montenegro; the subregion also includes a territory of Kosovo, now UNMIK. 4. Middle-income countries in the CIS are Belarus, Kazakhstan, Russia, and Ukraine. 5. Low-income countries in the CIS are Armenia, Azerbaijan, Georgia, the Kyrgyz Republic, Moldova, Tajikistan, Turkmenistan, and Uzbekistan. 6. The Commonwealth of Independent States includes all countries that were part of the Former Soviet Union, except for the three Baltic republics, Estonia, Latvia, and Lithuania, which acceded to the European Union in May 2004. 7. The analysis of poverty trends uses data from 15 countries in the Region for which comparable representative surveys are available over the entire 44 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union 1998­2003 period or its significant part: Armenia, Belarus, Bosnia and Herzegovina, Bulgaria, Estonia, Georgia, Kazakhstan, the Kyrgyz Repub- lic, Lithuania, Moldova, Poland, Romania, Russia, Tajikistan, and Uzbek- istan. The appendix to the study discusses country coverage in detail. 8. Aggregate poverty figures include Turkey in addition to the Region's poverty headcount, and thus refer to the ECA Region as defined by the World Bank. 9. Many of these poverty lines have been drawn in-country from house- hold survey data and represent a level of consumption that would allow a typical household to meet international minimum caloric require- ments, with an additional allowance for nonfood basic needs. They are a far cry from the outdated and questionable poverty lines used in many countries in the early years of the transition. 10. The first 15 member states of the European Union are Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden, and the United Kingdom. 11. For the purposes of discussing poverty risks in the EU-8 countries and selected countries in SEE and the middle income CIS group, we use a higher poverty line of $4.30. The $2.15 line catches a fairly small pro- portion of the population in these countries and thus may not provide robust results. 12. For the reasons explained in the appendix, we use consumption as an indi- cator of welfare at the household level. The other often applied measure--income--tends to show much higher levels of inequality, but it suffers from severe problems of underreporting in many transition economies; it is unclear whether high levels of income inequality are driven by the noisy data. To deal with this problem while comparing lev- els of inequality in the Region with other countries, we consistently rely on consumption-based indices. 13. The Gini coefficient is a standard measure of inequality: it takes values between 0 (complete equality) to 1 (extreme inequality, when all income is appropriated by the richest person) (see details in the appendix). 14. The point here is not whether relative or absolute notions of inequality are "right," but simply that people's perceptions of inequality are differ- ent from how economists have chosen to measure it. 15. Employment rates among women do not compare unfavorably with those of men, except in a few countries such as Bosnia and Herzegovina and Tajikistan. 16. Targets agreed to at the Lisbon Summit in March 2000, where heads of state and governments of the European Union decided that the EU should adopt the strategic goals for the next decade of becoming "the most competitive and dynamic knowledge-based economy [with] greater social cohesion." The 70 percent employment rate for the work- ing-age population is one of such targets. 17. Georgia (not depicted here) is another country where the incomes of the poor have declined in real terms. 18. Thus, trends in wage inequality are mixed, with inequality falling in some countries and rising in others. 19. The MDGs represent an international effort to promote poverty reduc- tion and human development by establishing measurable yardsticks for Overview 45 eight goals to be achieved by 2015. To achieve the child and maternal mortality MDGs, countries would have to reduce child mortality by two- thirds and maternal mortality by three-quarters by 2015. The 1990 benchmark against which progress is assessed is often not suitable for the Region's countries because many social indicators reached their nadir in the mid-1990s or later. The projections made by the World Bank are based on recent trends in the indicators (see World Bank 2005c for fur- ther details). 20. In technical terms, the elasticity of poverty reduction to growth has proved to be high for the most part. 21. Portugal has the lowest poverty line: around PPP $7.50 per capita per day. CHAPTER 1 Nature and Evolution of Poverty, 1998­2003 Poverty has declined significantly in Eastern Europe and the Former Soviet Union (the Region) since the financial crisis in Russia (1998­99). Then, one in five people (or 20 percent of the population) were living in poverty. Five years on, this figure is close to one in eight people (or 12 percent).1 This report documents the substantial reduction in poverty that has been achieved since 1998 and discusses why poverty has been more responsive to growth in some countries than in others. It explores the main channel--the labor market-- through which resurgent growth has contributed to poverty reduc- tion. It examines whether, and to what extent, nonincome dimensions of welfare have improved alongside improvements in household income. Finally, it discusses prospects for the reduction of both poverty and overall economic vulnerability and what role pub- lic policy can play. This chapter discusses the nature and evolution of poverty in the period since 1998. Introduction How many people are living in absolute deprivation in Eastern Europe and the Former Soviet Union? To answer this question, one needs to measure material well-being and establish a poverty line. 47 48 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Data on well-being for this report are derived directly from represen- tative household surveys to ensure comparability across countries and over time (see overview box 3 and the appendix, A. Data and Methodology). Survey data are available for 23 countries from the Region, although the data do not span the entire period 1998­2003 in all countries.2 To deal with the diversity of the Region, these coun- tries are grouped into four subregional clusters: · EU-8, the group of countries that recently acceded the European Union (EU) (and have the lowest poverty) · Southeastern Europe (SEE), the group of countries in Europe that are either formal candidates or have some prospect of acceding to the EU--Albania, Bosnia and Herzegovina, Bulgaria, Croatia, FYR Mace- donia, Romania, and Serbia and Montenegro (moderate poverty) · Middle income CIS countries, consisting of Belarus, Kazakhstan, Russia, and Ukraine (moderate poverty) · Low income CIS countries, comprising Armenia, Azerbaijan, Geor- gia, the Kyrgyz Republic, Moldova, Tajikistan, Turkmenistan, and Uzbekistan (high poverty)3 Turkey, included in the Europe and Central Asia (ECA) Region in the World Bank classification, but which does not have the same his- torical legacy as the other countries, is treated as a benchmark or comparator country to reveal differences between poverty in a tran- sition economy context in the Region and poverty in a developing country. Two other benchmarks from outside the Region are also used: Colombia (survey covers 2003) and Vietnam (latest available survey is 1998). One dollar a day is not enough in the Region. In many parts of the world, the $1-a-day line is used to measure absolute deprivation. However, because of the cold climate and other features of countries in the Region, this line is too low (see box 1.1). As a result, this report uses the $2-a-day line (or, more accurately, $2.15 per person per day) to measure the extent of the absolute material poverty. A higher poverty line, the $4-a-day line (or, $4.30 per person per day) is used to measure "economic vulnerability," by which is meant those who are not absolutely poor, but are nonetheless vulnerable to poverty (see box 1.1 and box 1.2). These lines are converted into local currencies, using 2000 pur- chasing power parity (PPPs) (see overview box 1 and the appendix for comparison of PPPs used in this report to other PPP revisions of 1993 and 1996) and country-specific consumer price indexes (CPIs).4 Nature and Evolution of Poverty, 1998­2003 49 BOX 1.1 What Is an Appropriate Poverty Line for the Region? This report uses an absolute concept of poverty, which is consistent with a large body of literature in which poverty is seen as the inability to meet basic material needs (Ravallion 1994). Although the notion of basic needs differs across countries, it can be reasonably well defined as the current cost of the subsistence consumption basket (see box 1.2). In practically all countries in the Re- gion, one finds groups of the population unable to meet such basic needs. This group and the group who is "nearby" in income terms are the focus of this report. The alternative measure of deprivation--relative poverty--has also been used in the literature (Atkinson, Marlier, and Nolan 2004). However, the difficulties that it creates for monitoring differences across countries and changes over time within countries make the authors favor the absolute poverty approach. What would be an appropriate absolute poverty line for countries in the Region?The World Bank often uses $1 a day for cross-country comparisons, which has since 1990 come to be regarded as providing the absolute minimum standard of living. The $1-a-day poverty line (in 1985 PPPs) was chosen because it was the most typical poverty line among the low-income countries (later updated, using 1993 PPPs, to $1.075 a day). None of the countries in the Region was considered when coming up with this estimate. National Poverty Lines and Consumption in 2000 PPPs 9$ Greece 8$ Portugal 7$ day/person Hungary 6$ per Bulgaria 5$ lines, 4$ Latvia Ukraine poverty 3$ Tajikistan 2$ National 1$ Burkina Faso Nigeria 0$ 1$ 10$ 100$ Level of consumption per capita, per day/per person, log Non-ECA countries ECA countries Sources: ECA: World Bank staff estimates; non-ECA: Kakwani and Sajaia (2004) and OECD (2003) for 2000 PPP; EU: Eurostat (2003) recalculated from per equivalent to per capita, using formula by Deaton and Zaidi (1999); and Dennis and Guio (2004). Note: Dashed gridlines correspond to $1.075, $2.15, and $4.30 a day per capita. (Box continues on the following page.) 50 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 1.1 (continued) Comparing national poverty lines for groups of countries (see figure) reveals that, as elsewhere in the world, there is a close correlation in the Region between the average standard of living and the national minimum needs definition. However, no country in the Region has a poverty line close to $1 a day. On the contrary, the lowest poverty lines cluster around the $2 mark, which fortuitously is twice the $1-a-day line. The vertical distance between fitted lines for countries in the Region and those outside the Region, which translates into a higher national poverty line for the same level of consumption, is suggestive of higher costs of basic needs in the Region.This is not surprising in a part of the world where climatic conditions mean that warm clothing and heating, both of which can be expensive, are essential for survival. Two dollars a day (or, more accurately, $2.15, which is exactly double $1.075) is therefore used as an absolute poverty line. A higher poverty line ($4.30 a day) is also used as a proximate vulnerability threshold to identify households who are not suffering absolute material deprivation, but are vulnerable to poverty. Al- though it seems somewhat arbitrary, it does bear some relation to empirically observed vulner- ability to poverty. Analysis of panel data suggests that households with per capita consumption at least twice the poverty line face considerably reduced risk of becoming poor (World Bank 2002c). Sources: World Bank staff estimates; Ravallion 1994, Atkinson, Marlier, and Nolan 2004; Kakwani and Sajaia 2004; OECD 2003; and World Bank 2002c. Material and nonmaterial poverty are closely linked. Poor people in the Region have much in common with people in other parts of the world; namely, an inability to buy basic material needs. However, the socialist legacy of high access to social services (for example, health care) and infrastructure (for example, heating), which have since been eroded, means that people feel an acute sense of depriva- tion relative to the past. For this reason, both income and nonincome dimensions of well-being are considered when trying to understand the evolution of living standards in the period since 1998. Material poverty remains in the center, because many of its nonmaterial aspects--such as the psychological pain of being poor, low achieve- ments in education and health, vulnerability to shocks, and a sense of powerlessness--are in fact closely linked to material poverty. At the same time, nonmaterial poverty does not entirely overlap with mate- rial poverty and deserves distinct consideration. The rest of this chapter is organized as follows: the first section presents the profile of material poverty and its changes over time, the second section discusses the nonincome dimensions of well-bring, and the third section presents conclusions. Nature and Evolution of Poverty, 1998­2003 51 Consumption Poverty More than 40 million people moved out of poverty in the Region between 1998 and 2003. The resurgence of growth in the Region during 1998­2003--particularly in Russia and other CIS countries-- has substantially reduced the number of people in poverty. Poverty incidence fell from around 20 percent at the start of the period to around 12 percent by the end (figure 1.1). In addition to 60 million poor in 2003, more than 150 million remained "vulnerable" to poverty (between $2.15 and $4.30 per day).5 Thus, close to half of the FIGURE 1.1 population is either poor or under threat of poverty. Although some More Than 40 Million countries in the Region are among the poorest in the world--Tajik- People Moved out of istan has a poverty headcount comparable to Cameroon, Côte Poverty during d'Ivoire, the Republic of Yemen, and Honduras (see World Bank 1998­2003 2005i)--two-thirds of the poor in the Region live in middle-income Distribution of Population by countries, with Kazakhstan, Poland, Romania, Russia, and Ukraine Poverty Status jointly accounting for more than half of all poor.6 100 90 Country-LevelTrends in Incidence and Depth of Poverty 215.1 80 264.2 Poverty reduction varied across countries: some countries did not 70 experience any poverty reduction, while poverty fell significantly in others. Although most countries in the Region have seen a reduction 60 in poverty from 1998­99 to 2002­3, the degree to which countries 50 have succeeded in reducing poverty has varied a great deal (see population of 40 overview figure 2 and appendix table 2 for country-level data on % 160.7 153.3 poverty over time). The largest reduction in poverty was achieved in 30 Moldova, where 22 percent of the population moved out of absolute 20 poverty.7 Poverty reduction in Tajikistan was equally impressive, with 10 102.0 more than 15 percent of the population moving out of absolute 61.2 poverty. In contrast, neither Georgia nor Poland experienced any 0 poverty reduction--in fact, poverty increased in both countries-- 1998­9 2002­3 while poverty incidence in Lithuania was unchanged. The trends in Nonpoor: above$4.30aday poverty derived from the comparable consumption measure pro- Vulnerable: above$2.15and duced for this study correspond closely to poverty changes monitored below$4.30aday Poor: with national definitions of poverty (see box 1.3). below$2.15aday In relation to country groupings, the most rapid poverty reduction Source: World Bank staff estimates occurred in the middle income CIS group, followed by the low income using ECA Household Surveys Archive. CIS group and SEE. There was no change in poverty in the EU-8. It is important to note here, however, that EU-8 countries had already Note: In million persons and in per- cent to population. Poverty lines con- grown out of poverty in its absolute sense by 1998, and further progress verted to local currencies using 2000 PPP. Data refer to ECA Region as de- in poverty reduction for this group is more appropriately assessed using fined by the World Bank, and Turkey the concept of economic vulnerability. is included in the aggregate figures. 52 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 1.2 What Would Someone in the Region Living onTwo Dollars a Day Consume? Based on prevailing market prices, the average food expenditure needed to meet basic caloric requirements with the cheapest products available on the market is around $1.18 a day at 2000 PPP (selected countries in CIS, population-weighted). Interestingly, it is found to be in a relative- ly narrow range from the cheapest basket of $1.15 a day in Tajikistan to around $1.22 in Kaza- khstan. National data show that such allowances cover only very meager baskets (composed predominantly of wheat, beans, milk, and oil). Based on this evidence, taking an international line of $1.075 a day would simply violate the presumption of measuring basic needs, even in the poorest countries of the Region. A person living at the poverty line of $2.15 a day would have been able to spend about $1 a day toward other needs. Such needs in the Region primarily consist of heating and lighting. The ap- proximate monthly electricity needed to light an apartment with 3 bulbs and run basic appliances (for example, a refrigerator) is 150 kilowatt-hours. At prevailing prices of around 2­5 cents per kilowatt-hour, when converted into PPP (PPP exchange rates are typically 3­4 times market lev- els), and adjusting for family size (3­4 people per household), this amounts to $0.07­$0.17 per day. Heating would require significantly more. For example, Eurelectric's (2003) "typical con- sumer" on average requires an additional 350 kilowatt-hours per month around the year, or an additional $0.17­$0.42 per day in PPP. Wood and gas are cheaper sources of heating, but esti- mates from poor countries in the Region suggest wood for heating and cooking would cost at a minimum $80­$100 per year at the current exchange rate, which amounts to around $0.30 per day.Thus, essential energy can eat up a quarter to a half of the dollar that remains after the pur- chase of the minimum food basket. After purchasing food and energy, the person would have little to put toward miscellaneous es- sential nonfood items, such as clothing and transport.These needs are not negligible and do not represent a luxury. Warm clothing is essential in cold climates and, in groups such as children, requires replacement on an annual basis as children grow. In the Balkans, households are known to enter into complex reciprocal exchange arrangements to economize on children's clothing. Ex- penditure on warm clothing could easily amount to a minimum of $50 per child per year. Trans- port costs are also important because maintaining access to the labor market and basic social capital requires some minimum mobility. A sizable share of the population in the Region cannot afford even this frugal bundle. Sources: World Bank staff; Kakwani and Sajaia 2003; Wu, Lampietti, and Meyer 2004; and Euroelectric (2003). Countries varied even more with regard to changes in economic vulnerability or the number of near poor. Often the reduction in the number of people below the absolute poverty line was accompanied by mixed outcomes in economic vulnerability (between $2.15 and Nature and Evolution of Poverty, 1998­2003 53 BOX 1.3 National Poverty Assessments Confirm PovertyTrends Based on International Poverty Lines The World Bank carried out a number of Poverty Assessments in the countries of the Region over the recent years. These reports focused on the evolution of poverty using national (official) definitions of poverty and living standards suited to the circumstances of each county. Although levels of poverty as defined in this study according to international poverty lines may well differ from national assessments of poverty, trends in poverty universally point in the same direction. In Moldova, the report entitled Recession, Recovery, and Poverty in Moldova (World Bank 2004i) shows that growth and income poverty are closely tied--poverty rose sharply and be- came deeper and more severe during the recession that followed the Russian crisis, but the rate, depth, and severity of poverty began to recede with the subsequent recovery. By 1999, 71 percent of the Moldovan population was poor according to national definition of poverty, but by 2002, the poverty rate had receded to 48 percent. In Russia, the Reducing Poverty through Growth and Social Policy Reform (World Bank 2005g) doc- uments that by 1999, because of the collapse in incomes and jump in inequality, poverty levels reached an all-time high for the transition period. Luckily, economic rebound after the crisis was both impressive and broad based--albeit uneven--across both sectors and regions. All this led to a dra- matic reduction in poverty. Russia succeeded in cutting poverty in half between 1999 and 2002, from about 42 percent in 1999 to 20 percent in 2002, using a consistent national poverty standard. In Belarus, the Poverty Assessment (World Bank 2004b) reports that according to national measures of poverty, the headcount ratio has fallen from 39 percent of population in 1996 to 27 percent in 2002, and further to 18 percent in 2004, implying that about 2 million people moved out of poverty.This can be fully accounted for by broad-based economic growth beneficial to la- bor. Strong growth in labor-intensive sectors (such as services, food processing, and machinery), backed by government wage and income policies, helped to ensure that the growth benefits were broadly shared by the population. The poverty reduction of Belarus is impressive but vul- nerable because Belarus's significant comparative advantages at its main export markets, the main source of its impressive growth, are eroding quickly. In the Kyrgyz Republic, as argued in Enhancing Pro-Poor Growth (World Bank 2003i), major strides in the past few years toward macroeconomic stability and economic growth started pay- ing off in poverty reduction. With increased productivity and a shift toward higher-valued prod- ucts, the agricultural sector has led economic growth since 1996, although gold and trade have also contributed to the recovery. Based on the full nationally representative survey (only compa- rable since 2000), the number of poor people was reduced by an estimated 300,000 individuals between 2000 and 2001 (poverty declined from 63 percent to 56 percent, using the national def- inition of poverty). Analysis of a panel subset of the survey shows that out of every 100 poor peo- ple in 1998, 23 people are estimated to have escaped poverty by 2001. (Box continues on the following page.) 54 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 1.3 (continued) In Poland, the trends point in the opposite direction. The report Growth, Employment and Liv- ing Standards in Pre-Accession Poland (World Bank 2004h) shows that the growth slowdown during 1998­2002 led to a reversal of the previous trend toward poverty reduction. Low growth rates and growing inequalities combined to increase poverty after 1998--from around 13 per- cent in 1998 to 15 percent in 2001, according to national definition. Moreover, poverty has be- come increasingly a "permanent" phenomenon associated with lack of skills, long-term unem- ployment, and residence in small towns or in particular regions. In Georgia, the Poverty Update (World Bank 2002c) and subsequent analysis focused on a key puzzle: why, despite positive growth, was there no poverty reduction in the country? The ex- treme poverty incidence edged up during 1998­2003 from around 14 percent to 17 percent of the population. Poverty (officially defined) stood at around 50 percent. The report argued that growing poverty between 1997 and 2000 was due to increased inequality in the distribution. Rural poverty worsened considerably. Growth was too weak and too concentrated in a narrow set of sectors, and there were no effective mechanisms to redistribute its benefits. Sources: World Bank 2002c; World Bank 2003i; World Bank 2004b; World Bank 2004h;World Bank 2004i; and World Bank 2005g. $4.30 per day). In some cases, the number of vulnerable increased (for example, Armenia, Kazakhstan, the Kyrgyz Republic, and Moldova). In others, such as Russia, there was little change. By contrast, Hungary and Romania reduced economic vulnerability significantly. In underlying poverty incidence by country, countries form three distinct clusters: low, moderate, and high poverty (figure 1.2). At one end of the spectrum, Armenia, Georgia, the Kyrgyz Republic, Moldova, Tajikistan, and Uzbekistan are characterized by high inci- dence of poverty. At the other end, countries such as Belarus, Bul- garia, Estonia, Hungary, Latvia, Lithuania, Poland, and Ukraine have negligible poverty at $2.15 a day. A much higher poverty (or vulner- ability) line of $4.30 a day still yields relatively limited poverty in these countries. Countries in the middle--Kazakhstan, Romania, and Russia--have moderate poverty (in the range of 10­20 percent) at $2.15 a day, but very sizable poverty at $4.30 a day. High preva- lence of poverty at $4.30 in these countries is a reflection of high vulnerability to economic downturns, which, as the financial crisis in Russia has shown, could easily lead to doubling of absolute poverty counts in a space of a year. Comparing poverty in the Region to benchmarks shows that each of these comparator countries fits in a particular cluster mentioned above: Colombia in the low-poverty Nature and Evolution of Poverty, 1998­2003 55 cluster, Turkey in the moderate-poverty cluster, and Vietnam in the high-poverty cluster. In some countries in the Region, poverty remains shallow, while in others, it is deep, and deepening. For countries for which data are available over time, poverty has become deeper in some countries with initially low depth, such as Poland and Romania. On the con- trary, poverty depth fell in some poorest countries and approached more moderate levels, for example, in Armenia, Moldova, and Tajik- istan. In other countries, it remained stable or increased. Despite some convergence, differences between countries in poverty depth remain very large (figure 1.3) and warrant different approaches to poverty reduction. Profile of the Poor Looking beneath national aggregates, which groups face a higher- than-average risk of poverty, and which groups constitute most of the poor? The two are not the same. Certain subgroups can have an FIGURE 1.2 Poverty Incidence Varies across Countries in the Region, around 2003 100 75 50 population of % 25 0 FYR Fed. Rep. Latvia Hungary Estonia Poland Bosnia Turkey Lithuania Bulgaria Belarus Romania Albania Ukraine Georgia Moldova Armenia Vietnam Azerbaijan Tajikistan Colombia Montenegro Russian Kazakhstan Uzbekistan Kyrgyz & Macedonia, Serbia EU-8 SEE Middle income Low income CIS Benchmarks CIS Below $2.15 a day Above $2.15 but below $4.30 a day Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 2 for latest year. Note: In 2000 PPPs. 56 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 1.3 Poverty Depth in the Region, 1998­2003 Hungary Poland EU-8 Lithuania Bulgaria SEE Romania CIS Russian Fed. Middle income Kazakhstan Armenia Moldova CIS Kyrgyz Rep. income Low Georgia Tajikistan ­5 ­15 ­25 ­35 ­45 Poverty deficit, % 1998 or earliest 2002 or latest Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 2 for data and years used by country. Note: Poverty depth is the distance between consumption per capita of an average poor and poverty line, expressed as per- cent of poverty line. For the EU-8 poverty depth is assessed using $ 4.30 a day, for other countries, $ 2.15 a day. extremely high incidence of poverty, but may not form most of the poor because of their small share in the population. The discussion focuses first on groups with high risk of poverty and then on groups that constitute a large share of the poor. Although the poverty profile is changing, the same groups are found to be at high risk of poverty as in 1998. Five years ago, the World Bank identified four subgroups of the population as having a higher incidence of poverty than others: in particular, the unem- ployed, the less well educated, the rural population, and children (or large families). However, because these groups made up a relatively small share of the total population, they rarely constituted the largest group among the poor, except in a few countries. In most of the Nature and Evolution of Poverty, 1998­2003 57 Region's countries, the largest group of the poor was found to be employed, with a secondary education, often living in urban areas, and of working age. Children formed the second largest group. Taken together, working adults and their dependent children (or, working families) accounted for an overwhelmingly large proportion of the poor. Five years later, this picture remains largely true. The unemployed face high and increasing poverty risk. As might be expected, the unemployed face a higher-than-average incidence of poverty (right panel in figure 1.4). Moreover, there has been an ongo- ing deterioration in their position relative to the employed (left panel in figure 1.4). Indeed, in the middle- and low-income countries in the CIS, such as Belarus, Moldova, Russia, and Tajikistan, there has been a sharp increase in relative poverty risk of the unemployed (that is, poverty risk relative to the employed), basically reflecting the sub- stantial improvement in the living standards of those in employment. By contrast, five years ago, the difference between the poverty risks of the two groups (that is, employed and unemployed) was relatively small. In the poorer CIS group, Georgia represents the case of the dif- ferences that remain as subdued as they were five years ago. Low education is strongly associated with poverty. As reported pre- viously, in the Region, poverty incidence falls with level of schooling. Figure 1.5 shows absolute poverty risk by levels of education for four representative countries and its evolution during 1998­2003. It shows that in countries with significant poverty reduction (mostly in the CIS), all educational categories shared equally. A different outcome emerges in countries with slow poverty reduction (mostly in EU-8 and SEE). In these countries, those with the lowest and the highest educa- tional attainment experienced no change, while poverty incidence for the middle groups changed in parallel. Clearly, the low-growth envi- ronment has done little to reduce poverty among the least educated. Capital cities gained the most from growth and rural areas the least. Rural areas face the highest poverty risk, followed by secondary cities. Rural poverty is typically higher than urban poverty (except for Armenia and Belarus), but the gap has increased in the past five years. This is because economic growth has resulted in more rapid poverty reduction in urban than in rural areas and, in a few instances (Arme- nia, Georgia, and Poland), rural poverty has actually increased (left panel in figure 1.6). Capital cities, and the poor residing in capital cities, gained most, to the point that in some countries, poverty in capital cities has been practically eliminated. Barring some exceptions driven by peculiar historic circumstances (for example, FYR Macedo- nia), capitals have much lower poverty than other cities, sometimes strikingly so (Kazakhstan, Russia, and Uzbekistan). 58 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 1.4 Levels and Changes in Poverty by Employment Status, 1998 to 2003 Poland EU-8 Romania SEE Belarus CIS Russian Fed. income Middle Kazakhstan Moldova CIS Georgia income Low Tajikistan ­ 0 + 0 50 100 Change in poverty, Percent of population % defined as poor 1998 poverty of unemployed 2002 poverty of unemployed 1998 poverty of employed 2002 poverty of employed Poverty change for unemployed Poverty change for employed Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 6 for country-level data and years used. Note: 1998 or earliest available year is used as starting point and 2002 or latest available as end point; for Poland and Belarus, poverty line is $ 4.30 a day, for other countries, $ 2.15; in 2000 PPPs. Nature and Evolution of Poverty, 1998­2003 59 FIGURE 1.5 Change in Poverty by Education for Representative Countries Poland 1998­2002 Romania 1999­2002 35 25 30 20 25 15 population 20 population of of % 15 % 10 rate, rate, 10 5 Poverty Poverty 5 0 0 None or Basic General Vocational University None or Basic General Vocational University less than secondary secondary less than secondary secondary primary primary Russian Federation 1998­2002 Moldova 1999­2002 20 90 80 15 70 60 population of 10 population % of 50 % rate, rate, 40 5 Poverty 30 Poverty 0 20 None or Basic General Vocational University None or Basic General Vocational University less than secondary secondary less than secondary secondary primary primary Poverty rate, 2002 Poverty rate, 1998 Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 5 for country data and years used. Note: For Poland, poverty line is $ 4.30 a day, for other countries, $ 2.15, in 2000 PPPs. Changes over time show that because of concentration of economic opportunities, capital cities have reduced poverty faster than any other areas have, and often secondary cities did no better on average than rural areas. As a result, a large gap is now observed in rural and urban poverty rates in middle income CIS countries, the EU-8, and SEE (right panel in figure 1.6). In countries where rural populations are large or the rural-urban gap is particularly high, the rural poor are in a major- ity. Thus, rural poor represent only between 25 and 40 percent of all poor in Bulgaria, Estonia, and Russia, but close to 70 percent of all poor in Kazakhstan, the Kyrgyz Republic, Moldova, Romania, and Tajik- 60 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union istan. Because of the rising relative incidence of rural poverty, the Region as a whole has seen an increase in the share of poor living in rural areas. At the end of the past decade, 45 percent of all poor in the Region lived in rural areas. This share has since risen to 50 percent. FIGURE 1.6 Capital Cities Gained More than Other Cities and Rural Areas, 1998­2003 Poland Hungary EU-8 Lithuania Romania SEE Macedonia, FYR Russian Fed. CIS Middle income Kazakhstan Moldova Georgia CIS Kyrgyz Rep. income Armenia Low Tajikistan Uzbekistan ­ 0 + 0 50 100 Change in poverty, Percent of population % defined as poor 2003 capital cities 2003 secondary cities 2003 rural areas Change in poverty, Change in poverty, Change in poverty, capital cities secondary cities rural areas Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 3 for country data and years used. Note: For the EU-8 and FYR Macedonia poverty line is $ 4.30 a day, for other countries, $ 2.15; in 2000PPPs. Nature and Evolution of Poverty, 1998­2003 61 Poor regions are noticeably lagging behind, and there is increasing differentiation between regions with regard to poverty. With further disaggregation of poverty rates by subregions within countries, a simi- lar picture of increasing concentration of poverty in relatively disad- vantaged areas emerges. Figure 1.7 shows that the gap between poorer and richer regions has increased in the past five years. In part, this is due to rising relative poverty in rural areas. It is also because poverty reduction in secondary cities has lagged behind the progress in capital cities and large population centers, where most of the new economic opportunities are concentrated. The trend toward rising regional dis- parities in poverty rate for the Czech Republic, Hungary, and Poland is also documented by Förster, Jesuit, and Smeeding (2005), using differ- ent data sources and different definitions of poverty. The comparison with benchmark countries presented in figure 1.7 offers interesting insights into the specifics of poverty in the Region. The extent of varia- tion in poverty risks within countries of the Region spans the whole range from relatively moderate to extreme. But the location of box dia- grams in figure 1.7 suggests that concentration of poverty within coun- tries of the Region remains quite distinct from that of other world FIGURE 1.7 Variation of Poverty Risks by Regions, 1998/9­2002/3 3.0 2.5 1 = 2.0 rate 1.5 poverty 1.0 National .5 0 1999 2002 1998 2003 1998 2002 1998 2002 2003 2002 1998 Poland Romania Russian Fed. Georgia Colombia Turkey Vietnam Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 3 for data used. Note: The boxes represent range for variation of the poverty rates. The line in the middle of the box represents the ratio of the median to the national poverty rate. The box extends from the 25th percentile to the 75th percentile of the distribution of regional poverty rates, the so-called interquartile range (IQR). Whiskers extend the box by 50 percent. Regional-level poverty rates are for the level at which surveys are representative, and rural and urban areas for each region were treat- ed separately--samples were designed to provide urban-rural breakdown within each region; the number of regions varies between 176 in Russia and 16 in Georgia. Dots refer to outlying regions with poverty rates (relative to the median) more than 1.5 times the interquartile range. 62 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union regions. The poor in the Region are still spread evenly throughout the countries, while in benchmark countries representing the developing world, they are concentrated in a few poor regions. Children face much higher poverty risk than the elderly, which is increasing over time relative to the average. Figure 1.8 shows poverty risks by age and its evolution over time (relative risks to fall below the $2.15 poverty line, where a risk of 1 indicates that an age group is no more or less likely than the average to fall into poverty). High poverty risk, by which is meant higher-than-average incidence of poverty, among families with children remains a major concern in all coun- tries of the Region. This figure provides evidence that small children (under 6 years) face elevated poverty risks. The same applies to older children (under 17 years). The elderly, in contrast, are characterized by somewhat lower poverty risk (except for Georgia), especially in the EU-8 and Kazakhstan. Indeed, in contrast to the situation five years ago, no country has the elderly as the group facing the highest risk--a possible reflection of the regularization of pension payments and reduction of pension arrears. FIGURE 1.8 Changes in Poverty by Age, Relative to National Average 2.5 2 1.00 = 1.5 incidence 1 poverty Country 0.5 0 1998 2002 1998 2002 2000 2002 2001 2003 1999 2002 2001 2003 1999 2002 1998 2002 2001 2003 2000 2003 1999 2003 Hungary Poland Romania Bulgaria Russian Kazakh - Moldova Georgia Armenia Kyrgyz Rep. Tajiki- Fed. stan stan EU-8 SEE Middle income CIS Low Income CIS < 6 years 7­14 years 15­17 years between 18 and 65 years 66 years Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 4 for data used. Note: For the EU-8 and Bulgaria poverty line is $ 4.30 a day, for other countries, $ 2.15. Nature and Evolution of Poverty, 1998­2003 63 The diagnosis of a worsening situation for families with children does not hinge on using a per capita scale to assess welfare. A key question when considering relative risks of different demographic groups is whether findings are sensitive to assumptions about economies of scale. The per capita standard used to construct the poverty profile in this study assumes no economies of scale. Children tend to live in large families--where there may well be scale economies--while the elderly tend to live in small families. With suf- ficiently large economies of scale, it is possible for relative risk rankings for children and elderly to be reversed. However, in countries where one would expect economies of scale to be significant (for example, middle-income countries), the disadvantage of families with children is so pronounced that the ranking does not change with changing val- ues of the economies-of-scale parameter. So, for example, children are consistently poorer than the elderly in Bulgaria, Hungary, Poland, and Romania. In low-income countries, this is not always the case (espe- cially in Moldova, as documented in Mencini and Redmond [2005]), but given the dominance of food consumption in the household con- sumption basket (see appendix, part A, chart 2), it is not clear that other than per capita scales are particularly relevant. On balance, chil- dren tended to be the poorest group in the Region during 1998­2003. The changes in the poverty profile by age have been most pro- nounced in countries with moderate poverty. In most countries, in addition to facing higher-than-average risk of poverty, children, espe- cially small children, have seen noticeable increase in their poverty risk relative to other groups. This is particularly evident in middle- income countries such as Hungary, Kazakhstan, and Russia. In Poland, the poverty risk of small children has improved somewhat, but it is undone by the increasing poverty of youth. The poorest coun- tries, in contrast, again with the notable exception of Moldova, have not seen major shifts in poverty risks faced by children. The elderly as a group has, in general, held a steady position vis-ŕ-vis the rest of the population or has improved its well-being. One important exception is Georgia, where poverty risk increased somewhat for the elderly. Vulnerable groups. Certain social groups (internally displaced persons [IDPs], Roma, and other socially excluded groups) suffer extremely high poverty, even controlling for their individual demographic and labor market characteristics (see box 1.4); however, these groups often account for only a minority among the poor. Which groups constitute most of the poor? Looking at the compo- sition of the poor by labor market status (figure 1.9), children and working adults together constitute most of the poor. Because most 64 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 1.4 Vulnerable Groups and Poverty: Roma, IDPs, and Institutionalized Populations A number of studies carried out in the past five years in countries of the Region document ex- tremely high poverty risks for some social groups: ethnic minorities, refugees, institutionalized persons, and disabled. These risks persist even when one controls for household and individual characteristics.The combination of income poverty and low social capital for these groups often reflects a "totality of exclusion" (Szalai 2002). Roma minority. Poverty rates among Roma are a multiple of poverty rates of the general popu- lation and other vulnerable groups. For example, a staggering 60.5 percent of the Roma are poor in Serbia, compared with 6 percent among the general population (World Bank 2005e). As a mi- nority that has experienced centuries of discrimination, the Roma have high intragroup interac- tions, but their networks do not extend beyond their ethnic group. A cross-country comparison found that the bulk of poor Roma households were headed by someone with primary or less than primary education, a factor that in turn limited employment opportunities and increased de- pendence on social assistance (Revenga, Ringold, andTracy 2002;World Bank 2004d). Roma fre- quently live in settlements where unclear property ownership and inadequate documentation prevent them from claiming social assistance or enrolling their children in school. Their high prevalence in informal sector employment further limits their access to health care and unem- ployment benefits. Social and cultural factors also affect access and interactions with service providers. Exclusion has been furthered by overrepresentation of Roma children in "special schools" for disabled children, and language barriers create difficulties in communicating with teachers, doctors, and local welfare officials (Ringold, Orenstein, and Wilkens 2003). poor children are in families with at least one working adult, working families form the largest group of the poor. The next largest group, which could be the elderly, the unemployed, or the inactive, depends on the country in question. The poverty profile in the countries of the Region is strikingly different from the composition of the poor in benchmark countries. Working poor are the majority among the poor in the Region. It is worth pointing out that in the EU-15, exclusion from employment is a typical correlate of poverty; and the share of the working families, or working poor, is typically small, while the share of jobless among the poor is large. In contrast, in the Region, except for Hungary, there are more working adults among the poor than there are nonworking adults. Thus nonemployment--while closely associated with Nature and Evolution of Poverty, 1998­2003 65 Internally displaced persons (IDPs). In all countries where data are available, IDPs consistently are found among the poorest group (see, for example, World Bank 2003l). Migrants lose social capi- tal through displacement, where the activities and structures that supported social relationships in their previous environment are missing. This is particularly true for people displaced by war. A study of displacement in the Region shows that poverty from loss of assets and unemployment of IDPs leads to the "hollowing out" of communities of displaced persons because potential en- trepreneurs and leaders migrate to more fertile environments (Holtzman and Nezam 2004). Formerly institutionalized populations (such as young people leaving residential facilities) start life with a combination of poor skills and very low social capital that puts them at risk of poverty and exploitation. Many lack functioning family ties, and their remaining social networks consist of people as isolated and disadvantaged as themselves. Throughout the Region, many children are still excluded from mainstream schools because of restricted mobility, sensory impairment, learning difficulties, or minority ethnicity.They spend their school years in institutions that isolate them and drastically reduce opportunities for mainstream social engagement (Grammenos 2003; Clert and Gomart 2004). The disabled, especially the severely disabled, is a group that often faces discrimination and se- vere constraints in engaging with society.The sheltered workshops that once provided incomes, a place in society, and regular social contact outside the home no longer function. Often, dis- ability pensions effectively exclude disabled people from the labor market. Current Kyrgyz legis- lation, for example, prohibits some disabled people from earning a living. At the same time, dis- ability benefits are insufficient (Toralieva and Maslova 2004). Sources: Grammenos 2003; Clert and Gomart 2004; Holtzman and Nezam 2004; Szalai 2002; Toralieva and Maslova 2004; Ringold, Orenstein, and Wilkens 2003; Revenga, Ringold, and Tracy 2002; and World Bank 2003l, World Bank 2004d, and 2005e. poverty--does not represent in itself a major poverty dimension. Some countries in the EU-8 such as Hungary seem to have moved closer to the EU-15 typical pattern, but still harbor significant num- bers of working poor. This point is brought home by figure 1.10, which shows for four representative countries the distribution of poor between households where someone is employed (working families) and households where no one is employed (jobless families). Hun- gary has the largest share of poor without any connection to employ- ment (approximately 37 percent of the poor live in households where no one has a job), but such households account for less than one quarter of all poor elsewhere. The predominance of working poor in the Region remains in stark contrast with the rest of Europe, where poverty is concentrated among the jobless. 66 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 1.9 The Poor in the Region around 2003 Composition of the Poor by Age and Employment in Selected Countries in the Region and Benchmarks 100 75 population 50 poor of % 25 0 Poland Hungary Romania Bulgaria Kazakh- Russian Moldova Georgia Tajiki- Turkey Colombia Vietnam stan Fed. stan EU-8 SEE Middle Low income CIS Benchmarks income CIS Children (<16 years) Working (employed+self-employed) Unemployed Inactive Elderly (66 years) Source: World Bank staff estimates using ECA Household Surveys Archive; see appendix tables 4 and 6 for data and years used. Note: For the EU-8 and Bulgaria $ 4.30 a day used as a poverty line; other countries, $2.15. FIGURE 1.10 Most Nonworking Poor Live in Households Where Someone Works Share of All Poor Accounted for by Nonworking Individuals Differentiated between Those Living with Someone Who Works (Working Families) and Jobless Families 100 Jobless families 90 80 Working families 70 100 = 60 50 Poor 40 All 30 20 10 0 Hungary, Romania, Russian Fed., Tajikistan, 2002 2002 2002 2003 Children Unemployed Retired & elderly Students Inactive Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Hungary $4.30, other countries $2.15, at 2000 PPP used as poverty line; jobless families are defined as household where no one is working for wage or is self-employed. Nature and Evolution of Poverty, 1998­2003 67 Poverty in Nonincome Dimensions Although poverty may no longer be a growing problem in most coun- tries in the Region, deprivation in nonincome dimensions remains a source of concern. Across nonincome dimensions, health status in most countries of the Region is a major factor of deprivation and, in many respects, is deteriorating. Progress in other nonincome dimen- sions is also mixed: better financing of public services has helped to maintain access, but erosion of quality, combined with deterioration of affordability, has excluded many poor households. The relative worsening of service quality--both in access (poor versus nonpoor, urban versus rural) and with regard to historical parameters (where subsidized services were universally available)--affect subjective per- ceptions of poverty. Health Life expectancy losses have been difficult to reverse. All countries in the Region suffered declines in life expectancy during the 1990s. For some countries, this decline was brief, for others fairly protracted. In the latter group--although the declines have bottomed out or have begun to be reversed--many countries have not recovered to pre- transition levels (figure 1.11). For example, in Russia and Ukraine, life expectancy is five years below pretransition levels. Most of the decline in life expectancy is accounted for by premature (or avoid- able) deaths in the most productive age groups, and it affected males particularly strongly. In Russia alone, the total number of male pre- mature deaths was 2 million during 1992­2000, which, when com- pared with all deaths, suggests that one out of five deaths during the period was preventable (Nolte, McKee, and Gilmore 2004). Interpret- ing premature death as an extreme manifestation of health poverty, the risk for Russian males to suffer for this form of deprivation was about 3 percent. Although it is recovering, life expectancy in the Baltics, Belarus, Kazakhstan, Russia, and Ukraine is significantly lower than in much poorer Armenia and Georgia.8 In transition, premature mortality affects both the poor and non- poor. Evidence on linkages between premature mortality and poverty is more limited. Available evidence from panel surveys suggests that premature mortality is not concentrated among the poor, but affects all groups of the population to the same extent (Brainerd and Cutler 2005). At a more aggregate level, Ivashenko (2005) finds that varia- tion in regional poverty rates accounts for less than one-half of regional excess male mortality. 68 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Maternal and child mortality remain high in some countries. Mor- tality among children and women is also an issue, although more so in some countries than in others. Infant and under-five mortality rates are declining in most countries of the Region; however, in some parts of the Region, particularly in the low income CIS countries, progress in reducing child mortality is too slow.9 In addition to the need for better pre- and postnatal care, lagging countries need to develop better case management techniques for the treatment of infant and early childhood diseases, both at home and in the com- munity. Maternal mortality (and maternal health) is also an issue in a number of countries in the Region. A number of low income CIS countries and some middle income CIS countries, as well as bench- mark Turkey, may well not achieve the Millennium Development Goal (MDG) targets related to maternal health.10 Key to reducing maternal mortality is access to emergency obstetric care and a refer- ral system that enables women to reach life-saving treatment in time. Despite the high proportion of births that are attended to by medical professionals, health systems in many parts of the Region are not able to deliver the timely services that are essential to averting maternal death. Related to maternal health more broadly is the issue of exces- FIGURE 1.11 Life Expectancy at Birth, 1990­2003 73.0 72.0 71.0 years 70.0 birth, at 69.0 68.0 expectancy 67.0 Life 66.0 65.0 64.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 EU-8 SEE Middle income CIS Low income CIS Turkey Source: WHO. Note: EU-8: Hungary, Poland, Estonia, and Latvia; SEE: Romania and Bulgaria; middle income CIS: Belarus, Kazakhstan, Russia, and Ukraine; low income CIS: Arme- nia, Georgia, the Kyrgyz Republic, Moldova, and Uzbekistan. Nature and Evolution of Poverty, 1998­2003 69 sive reliance on abortion for birth control in large parts of the Former Soviet Union. This is thus partly a legacy issue but also a reflection of the fact that contraceptives remain in short supply and are relatively expensive. Despite declines in recent years, abortion rates remain among the highest in the world, with negative consequences for maternal health. Public health indicators show worrying trends outside the EU-8. Of course, preventable mortality is the most extreme form of health deprivation. Other forms of health deprivation include increased risk of disease and reduced access to medical help when in need. An increase in the incidence of noncommunicable diseases is to some extent to be expected because most countries have completed the epidemiological transition (from communicable diseases as a major source of morbidity to other health risks). However, coun- tries (especially in the CIS) are facing growing epidemics of com- municable diseases such as TB and HIV/AIDS (figure 1.12). The increase in injecting drugs and commercial sex work throughout the Region, a concurrent increase in the incidence of sexually transmitted infections (STIs), high migration rates, limited capacity of governments and civil society to implement effective preventive FIGURE 1.12 Incidence of Tuberculosis, 1990­2003 120 100 people 80 100,000 per 60 incidence 40 TB 20 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 EU-8 SEE Middle income CIS Low income CIS Turkey Source: WHO. Note: EU-8: Hungary, Poland, Estonia, and Latvia; SEE: Romania and Bulgaria; Middle income CIS: Belarus, Kazakhstan, Russia, and Ukraine; low income CIS: Ar- menia, Georgia, the Kyrgyz Republic, Moldova, and Uzbekistan. 70 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union responses, and low levels of awareness of HIV and STIs have all contributed to the growing epidemic. At current rates of infection and treatment, the HIV/AIDS MDG is unlikely to be attained by the Region. Perceptions of health status show mixed trends. Although suffer- ing from a number of drawbacks, subjective perceptions of health sta- tus reflect many of these objective trends. Between 20 and 25 percent of the population in the Region report their health status being "bad" or "very bad"--significantly more than in any country in the EU. As is the typical pattern with these data, the poor report less bad health than the rich (controlling for age), pointing to the influence of income and possibly education in such self-assessments. This difference is reflected at the national level in reported acute morbidity rates that vary from fairly low levels in poor countries (for example, 3 percent of the population per month in Armenia, Tajikistan, and Uzbekistan) to fairly high levels in richer countries (more than 20 percent in Belarus and Russia). Over the past five years, morbidity rates have declined, perhaps reflecting rising incomes. Chronic health conditions that limit daily activities provide per- haps a better assessment of underlying health conditions. Reported incidence of chronic conditions has been increasing. In the CIS, between 25 and 55 percent of the population report such conditions (with the highest proportion observed in Moldova, the lowest in Georgia and Kazakhstan, and Russia falling in between). The pro- portion of the population reporting chronic conditions falls to around 10­20 percent in SEE and the EU-8 (with the lowest in Romania and the highest in Hungary), but remains higher than in most EU countries.11 Utilization of health services in low-income countries has declined to very low levels. Utilization of health services, defined as the frac- tion of sick individuals who use health services, represents another dimension of deprivation, namely deprivation of health care. Here the picture is mixed. Utilization has gone up in most countries, but in a number of poor countries remains at very low levels, reflecting both supply-side (poor quality) and demand-side issues (high cost) in accessing health care. There is clear evidence that the poor experience greater financial barriers to accessing health care than the rich. Figure 1.13 examines the degree of overlap between different notions of poverty as they relate to income and health. There is clearly a degree of overlap, but it is less than complete. This reinforces the need to think of health and other indicators (such as access to education or good-quality housing) as dimensions of well-being distinct from income (or poverty). Nature and Evolution of Poverty, 1998­2003 71 FIGURE 1.13 Poverty in the Dimensions of Consumption, Access to Water, and Health Russian Federation, 2002 No access when ill = 39 Consumption poor = 8 28 2 3 1 2 7 9 No piped water = 19 All population = 100 Georgia, 2002 Consumption poor = 49 31 4 No access when ill = 9 2 1 2 13 No piped water = 19 4 All population = 100 Romania, 2002 No access when ill = 7 3 Consumption poor = 16 1 3 3 No piped water = 42 11 27 All population = 100 Source: Staff estimates from survey data. Note: Percentages to population. Size and overlap exaggerated in some cases for clarity of exposition. 72 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Education Inherited literacy rates are high in the Region. The most acute form of education deprivation is illiteracy. Average literacy among the transi- tion economies of the Region is high (more than 98 percent), and in the transition country with the lowest levels of literacy (Tajikistan), 96 percent of adults are literate. Thus, this extreme form of education deprivation does not appear to be a major issue in the Region. Few studies in the Region focus on actual ability to read and understand text (as opposed to self-declared literacy), but the ones that do show that functional illiteracy is not uncommon. However, data deficien- cies preclude an examination of whether this is more of an issue than in other parts of the world, or any examination of trends. The lack of education and poverty are closely correlated, but a majority among the poor in the Region have completed secondary education. Although there is no information on functional literacy and poverty, data on number of years of completed education show that there is a concentration of adults with incomplete primary edu- cation among the poor. However, in contrast to most developing countries, including benchmarks, a clear majority among the poor in the Region are those who completed secondary education.12 But the quality of education is not improving and is ill suited to the needs of the labor market. Although more extreme forms of educa- tion poverty are not an issue, concerns remain about the overall qual- ity of education imparted by inherited systems. With few exceptions, there has been an increase in the proportion of students who under- perform at the secondary school level, according to international assessments of educational quality.13 If this trend is not stemmed, the fraction of children and youth with poor education will continue to grow. The value of the skills imparted by education systems in a large number of countries is also questionable and bears little relation to what may be required by the market (see, for example, Yemtsov, Mete, and Cnobloch 2005). Infrastructure and Housing The steady erosion of infrastructure networks due to neglected main- tenance (especially in the CIS countries) has taken a toll on access to, and quality of, infrastructure services. Even though the ability to pay has increased since the resumption of growth in all parts of the Region, the lagged effect of postponed maintenance has resulted in falling access and increased deprivation, particularly in the CIS. How- ever, as with education, the inherited networks still offer access far Nature and Evolution of Poverty, 1998­2003 73 greater than in countries at comparable income levels, and depriva- tion generally does not take extreme forms. Water. Lack of access to safe water is the most acute form of depriva- tion in this area and is an important proximate cause of poor health from waterborne diseases. Interpreting access to safe water in the very narrow sense of connection to a working tap (or faucet), one does not find significant water access "poverty" in the Region, except for some specific groups (for example, some rural areas) and some poor countries. Connection rates are impressive and, in some instances, have expanded over the past five years (Tajikistan). Con- nection rates in rural areas are obviously lower than in urban areas and, in some instances, have actually deteriorated in the past five years (for example, the Kyrgyz Republic and Uzbekistan).14 Connec- tion rates, however, overestimate access to water. Water is often not available for more than few hours a day, particularly in the low income CIS group (see chapter 4 for more evidence). There is grow- ing evidence that tap water is not meeting basic quality requirements in many instances. It is also possibly the case that water supply dis- ruptions take a heavier toll on urban dwellers in apartment buildings with limited alternatives (widespread in the Region) than for rural residents. Although there are concerns about quality both in uninter- rupted supply and bacteriological content, there is limited evidence of significant health impacts of limited access to safe water. There have been few significant outbreaks of waterborne diseases in the Region. Heating. The ability to maintain ambient temperature at home through the winter is a basic need in the colder parts of the Region. In contrast to water (where even cost recovery tariffs will not under- mine the budgets of the poor), "energy" poverty is a source of wide- spread concern. Although few studies suggest extreme deprivation-- such as death from exposure to cold--there is abundant evidence of less extreme forms of deprivation, such as the increase in the use of so called "dirty" fuels for heating indoors and reduced activity levels due to inadequate heating (see, for example, Wu, Lampietti, and Meyer 2004). Because of high connection rates to electricity (virtu- ally 100 percent in both rural and urban areas) and availability of other fuels on the market, energy "poverty" is more often than not related to the inability of the poor to afford clean sources of fuel and is thus one of the consequences of poverty. Sanitation services and housing. Most countries in the Region are highly urbanized, and provision of safe sewerage and solid waste disposal are 74 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union important aspects of household well-being. Among countries of the Region, Uzbekistan has the lowest connection rates to sewerage (50 percent of urban households). In other CIS countries, connection rates vary between 70 and 90 percent for urban areas, while those in the EU-8 and SEE are generally above 90 percent. Connection rates are significantly lower among the poor, again with Uzbekistan having the most acute problem (80 percent of urban poor are not connected to sewerage). In other countries, nonconnection rates among the poor are also high, ranging between 50 percent in the low income CIS countries to 25 percent in the middle income CIS group and SEE. There is little information on other forms of sanitation. Substandard urban housing in many parts of the Region is con- centrated in informal settlements, often parts of large cities where temporary housing is built without permits. Dwellers of slums often belong to poor marginalized groups, such as Roma and IDPs. How- ever, the overall prevalence of slums is low compared with other regions (UN Millennium Project 2005). There is growing anecdotal evidence of the rise in the number of slum dwellers in the Region, but accurate estimates are elusive. Overcrowded housing is a significant problem in the poorest countries of the Region, affecting primarily capital cities. Subjective Poverty The effect of the transition shock on people's morale is not over: the Region remains the one with the most pessimistic perceptions and lowest self-rated satisfaction with life (see box 1.5). Subjective per- ceptions of well-being provide a measure of a population's own assessment of poverty. Although material deprivation is the most important correlate of subjective poverty, nonincome deprivation also plays a role. Despite recent improvements in the self-assessed welfare mirroring income gains, the formerly socialist countries are characterized by lower satisfaction with life for their level of income than any other region of the world. Some authors suggest that inequality aversion is an important fac- tor behind the lower subjective perceptions of well-being in the Region (see Sanfey and Teksoz 2005;15 also Ravallion and Lokshin 1999). Others, however, find limited evidence of inequality aversion in the CIS relative to other countries (see Murthi and Tiongson forth- coming). Decline in access and quality of public services may also have a role to play in explaining perceptions of well-being, particu- larly given the legacy of fairly uniform access. Nature and Evolution of Poverty, 1998­2003 75 BOX 1.5 Life Satisfaction in the Region Remains Low Subjective data from the World Values Survey, spanning 1990­2000, suggest that after declining from the early to the late 1990s, subjective valuations of welfare in the transition economies of the Region improved in the most recent period. In other words, subjective evaluations appear to follow the broad trend in material deprivation. At the same time, these data reveal that, com- pared with other countries, the formerly socialist countries are characterized by lower satisfac- tion with life for their level of income (see figure below). Five countries in the Region--Armenia, Belarus, Moldova, Russia, and Ukraine--are in the bottom decile in the world distribution of sat- isfaction scores and fare much worse than the benchmark countries represented on the graph. Two of the Baltic States--Latvia and Lithuania--are in the next-to-bottom category, along with Albania, FYR Macedonia, and Romania. In general, the new EU members score much better, with Slovenia's score of 7.23 (above France's, 7.01, and not far off from Great Britain's and Ger- many's, 7.40 and 7.42, respectively) in life satisfaction. Life Satisfaction in the Region, While Improved, Is Low Vis-ŕ-Vis Other Countries, around 2000 8.5 8.0 7.5 score 7.0 6.5 6.0 satisfaction Life 5.5 5.0 4.5 4.0 3.5 0 $10,000 $20,000 $30,000 $40,000 $50,000 GDP per capita, 2000 PPPs World, 1999­2000 ECA, 1995­7 ECA, 1999­2000 Source: World Values Survey. Sources: Sanfey andTeksoz 2005; Murthi andTiongson forthcoming; andWorldValues Survey (see www.worldvaluessurvey.org). 76 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Conclusions Poverty has declined significantly in the Region since 1998 because of the resumption of growth. At the same time, progress in the non- income dimensions of well-being are mixed with improvements recorded in some instances and stagnation or declines in others. All in all, this has resulted in a change in the nature of poverty in the Region. Although the share of population below the poverty thresh- old has shrunk, the proportion who are "poor" in dimensions such as health, good-quality schooling, or housing has not declined in tandem across all dimensions. This has changed the nature of poverty, with a growing share or "weight" of nonincome dimensions in overall poverty. What is specific about poverty in the Region? Poor people in the Region have much in common with the poor in other parts of the world; namely, an inability to meet basic material needs. But the Region's his- torical legacy, challenges of transition, and unique geographical factors, have clear impact on defining these basic needs. Climatic conditions mean that warm clothing and heating are essential for survival. Demo- graphic factors have their effects too. On the one hand, the rapid aging of the population in many countries means that there are fewer children for the poor to support, boosting the impact of earnings on poverty. But on the other hand, reductions in the family size mean that households in the Region are exposed to shocks that otherwise could be managed within the intrafamily solidarity networks. "Graying" of the population, which brings with it a burden of "rich country" epidemiological pat- terns, also means additional health care costs. Transition itself affects the nature of poverty in the Region. The costs of moving large masses of population, which the previous sys- tem concentrated around nonviable production units, to new loca- tions and sectors where they can rely on more sustainable livelihoods are enormous, and the process of reallocation has proven to be slow. It is further slowed down by infrastructure bottlenecks (persistent due to neglect of maintenance and low and inefficient investment) and housing market rigidities, which limit population mobility. The inherited production systems also result in persistent differences of productivity across firms with different histories (new, restructured, and old), which in turn result in persistent differences in earnings and pockets of high concentration of poverty among the working popula- tion. All of these factors not only determine the nature of poverty in the Region in the present, they are also powerful drivers of future progress in poverty reduction. A discussion of these issues follows in the next chapters. Nature and Evolution of Poverty, 1998­2003 77 Notes 1. Throughout this report, $2.15 per capita per day (in 2000 PPP) is used as the absolute poverty line. 2. The countries not covered by the poverty data for 1998­2003 are Croa- tia, the Czech Republic, the Slovak Republic, Slovenia, and Turk- menistan. Data from UNMIK were also not used. Moreover, several countries are represented by a single survey (year) or by surveys that are not sufficiently comparable to assess trends in poverty over time: Alba- nia, Azerbaijan, and Serbia and Montenegro. 3. Though no recent data are available for Turkmenistan, it is classified as part of the low income CIS group, with a high level of poverty, based on the data (LSMS 1996) reported in Making Transition Work for Everyone (World Bank 2000a). 4. The World Bank's first regional poverty report, Making Transition Work for Everyone, used the 1996 set of PPPs. For differences in poverty rates from using other estimates of PPPs, see overview box 1. 5. Turkey is included in the aggregate figures, thus the referral to the ECA Region in the World Bank classification. Poverty for countries with miss- ing data was extrapolated based on their population size and average subregional poverty incidence. 6. Here the distribution of the Region's poor by countries is assessed with- out Turkey. 7. Because survey data are used to obtain poverty estimates for the popula- tion as a whole, they have only a certain degree of precision, which dif- fers across countries, but normally falls in a ± 2 percentage-point confidence interval around the point estimate. 8. Several explanations are advanced to rationalize this apparent puzzle, including dietary patterns, social environment (that is, family structure, the educational system, social networks, and so forth), physical environ- ment (for example, exposure to toxic substances, safety at home and work, housing conditions and degree of overcrowding, and urban-rural differences), genetic endowment, and behavior (for example, adherence to treatment regimens, lifestyle choices like smoking). 9. The child mortality MDG calls for a two-thirds reduction in child mor- tality during 1990­2015. This goal is not meaningful for countries in the Region with low child mortality because a two-thirds reduction would require attaining child mortality levels that are below what is found in high-income countries. For others with higher mortality, the goal is still relevant. One complicating factor in measuring progress is that many countries use the old Soviet concept of infant mortality, which is contrary to WHO best practices and underestimates infant (and child) mortality. Alternative survey- and model-based methods help address this problem, but are not available with the same fre- quency as official data. 10. As with child mortality, the maternal mortality MDG--which calls for a reduction by three-fourths in maternal mortality rates--is not realistic in a number of countries. In countries where the goal is relevant, signifi- cant improvements are needed if the goal is to be met. 78 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union 11. Data for the Region are staff estimates based on household surveys. EU figures are reported in EC 2002. 12. See appendix table 5. Note that because of a low poverty headcount with the $2.15-a-day poverty line in EU-8 countries, data on poverty by edu- cation level from this subregion should be used with caution. 13. See chapter 4 for the assessments of secondary school students according to the evaluations of PISA and TIMSS. 14. Refer to appendix table 9 for country-level data on water connection. 15. Sanfey and Tukoz find a strong negative association between country- level inequality and overall life satisfaction in the Region's countries, in contrast to a positive association in the nontransition case. CHAPTER 2 How Has Poverty Responded to Growth? Since the end of the financial crisis in the Russian Federation, most countries in the Region have experienced sustained growth. As a result, poverty has declined substantially, although by a greater margin in some countries than in others. This chapter seeks to explain why differences in poverty reduction have been observed across countries in the Region. In part, these variations occurred simply because growth rates have differed. In general, where growth has been stronger, poverty reduction has been greater. How- ever, even when allowing for differences in growth rates, the response of poverty to growth has varied across countries. This dif- ference in responsiveness relates not only to differences in initial conditions but also to changes in the distribution during the period in question. Indeed, a number of fast-growing countries--for exam- ple, those in the CIS--have also seen shifts in the distribution of income toward the poor. As a result, poverty has declined more rap- idly than might be expected. In contrast, where the distribution was unchanged or moved against the poor, poverty reduction was atten- uated. In the extreme, with low growth and adverse movements in inequality--as was the case, for example, in parts of the EU-8-- poverty actually increased. These issues are discussed in further detail in this chapter. 79 80 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Growth and Poverty Reduction Virtually all countries in the Region experienced positive rates of growth (as measured by GDP) during 1999­2003, some at rates unprecedented in the past quarter of a century.1 Since 1999, regional output has increased by more than 25 per cent. Not only was growth widespread but it also consistently exceeded the world average (fig- ure 2.1). Growth rates, however, remain below those achieved in East Asia, largely because of the extremely strong performance of China. Within the Region, the highest growth rates were recorded in the middle income CIS countries (where the bulk of the poor reside), followed by the low income CIS countries, then SEE, and finally the EU-8. This strong growth performance resulted from several factors, key among which was the ability to take advantage of a favorable external environment. Countries were helped, in varying degrees, by the strength of the domestic policy environment. Despite the strong growth performance, it was not until about 2004 that the Region as a whole returned to the level of GDP recorded in 1990 (see Ĺslund 2001). Also, although the Region as a whole may have resumed earlier levels of output, the GDP in some countries remains significantly below its pretransition level. For example, Georgia and Moldova are struggling to rise above half the level of GDP they recorded in 1990, whereas Ukraine is at 60 per- cent of 1990 levels.2 Moreover, during the 15 years that the Region has taken to recover from the transition shock, world output increased by 43 percent. Despite the recovery of output, poverty in the Region has more than doubled compared with the late 1980s (see chapter 1), largely because of the rise in inequality since the onset of the transition. What factors account for the strong growth performance of the Region in the past five years? Factors vary by regional subgroup. For the CIS, the recovery of growth in Russia has been an important fac- tor. The devaluation that accompanied the financial crisis in Russia was important for restoring the exchange rate to a more competitive level and spurring the recovery of exports and growth. Combined with high prices for oil and other natural resources, the devaluation gave a huge boost to the Russian economy, which has, in turn, become a regional locomotive for many neighboring countries.3 Structural reforms that many of the CIS countries had undertaken enabled an improved supply response when the opportunity pre- sented itself. For the EU-8, the prospect of accession provided a strong impetus for both reforms and growth, while the restoration of peace and stability in SEE was an important factor in sustaining recovery. How Has Poverty Responded to Growth? 81 FIGURE 2.1 Since 1999, Growth Rates in the Region Have Been Higher than the World Average 15 10 5 % rate, 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Growth ­5 ­10 ­15 ECA LAC EAP World Source: World Bank staff estimates, using World Bank (2005i). Note: All averages are population weighted. Poverty has, in general, responded to growth in all subgroups (see fig- ure 2.2).4 During 1998­99, when a number of countries experienced contraction--in particular in the CIS, because of the impact of the finan- cial crisis in Russia--poverty increased. However, with the resumption of growth, there have been significant declines in poverty throughout the Region, with few exceptions. To illustrate this fact, very few observations appear outside the lower right and upper left quadrants of figure 2.2, showing that where growth has been positive, poverty has declined, and where it has been negative, poverty has increased. The one observation in the upper right quadrant relates to Poland, where positive growth has gone hand in hand with an increase in poverty because of changes in dis- tribution. Comparison with benchmark countries suggests that the amplitude of changes in the Region has been historically remarkable. Growth Elasticities, or, How Responsive Is Poverty Reduction to Growth? Although it has declined, the degree to which poverty has responded to growth varies across countries. As figure 2.2 suggests, even where 82 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 2.2 Growth Has Been Accompanied by Poverty Reduction +40 +20 %, poverty +0 in Change ­20 ­40 ­20 ­10 +0 +10 +20 +30 +40 Growth rate in real consumption per capita (survey), % EU-8 SEE Middle income CIS Low income CIS Benchmarks Source: World Bank staff estimates using ECA Household Surveys Archive and "Pro-Poor Growth in the 1990s" (World Bank 2005f). Note: Selected periods, for countries with comparable data series over time, see appendix for detailed country-level data. For EU-8 $4.30 a day in 2000 PPP used as a poverty line, $2.15 otherwise. All data are expressed as annual changes. Benchmark countries include data spanning 1990s and early 2000s from: Vietnam, El Salvador, Uganda, Ghana, India, Tunisia, Bangladesh, Senegal, Brazil, Burkina Faso, Bolivia, Indonesia, and Zambia. growth (as measured by change in real consumption) is the same, the change in poverty can vary by a substantial margin. The degree to which poverty responds to growth is encapsulated in the notion of elasticity, which measures the change in poverty for 1 percent change in growth. In general, the elasticity would be negative, because growth and poverty tend to move in opposite directions: positive growth typically means a decline in poverty, and negative growth normally indicates an increase. Table 2.1 presents simple averages of the elasticity of poverty reduction to growth for the four main subregions in the Region. These averages are based on data from countries for which comparable time-series data are available over the period in question. They should be treated as indicative, rather than as fully representative, of the nature of poverty response in the subregion in question. With these caveats, poverty has been the most responsive to growth in the mid- dle income CIS countries and in SEE. Indeed, in these two subre- gions, an additional 1 percent of growth has lowered poverty by more than 2 percent over the past five years. By contrast, in the low income CIS countries and in the EU-8, every 1 percent of growth has lowered poverty by 1.3­1.4 percent. How Has Poverty Responded to Growth? 83 TABLE 2.1 Poverty Has Been More Responsive to Growth in the Middle Income CIS Countries and SEE than Elsewhere Average elasticity (total) of poverty to growth in consumption per capita, Subregion Countries 1998­2003 EU-8 Hungary, Poland ­1.3 SEE Romania, Bosnia and Herzegovina ­2.5 Middle income CIS Belarus, Kazakhstan, Russian Federation, Ukraine ­3.1 Low income CIS Armenia, Kyrgyz Republic, Moldova, Tajikistan, Uzbekistan ­1.4 Sources: World Bank staff estimates using ECA Household Surveys Archive. Country-level data derived from information reported in table 2 in the Appendix. Note: Averages are simple cross-country means. They should be treated as indicative, rather than as representative, of typical values found in the subregion. Poverty line is $4.30 a day per person in 2000 PPP for the EU-8, and $2.15 elsewhere. What factors explain why poverty has been more responsive to growth in some countries than in others? Before turning to explana- tions, it is worth addressing the issue of data quality. The Region was characterized by problematic data, especially in the years of the tran- sition, when statistical systems were suffering from both declining budgets and limited capacity to deal with the changing nature of the economy. Since then, however, data quality has improved, both on the survey side and on the national accounts side, not only because of economic recovery but also because of the investment of growing resources in statistical capacity building. As a result, data quality is broadly comparable to that in other regions, and indeed, with few exceptions, national accounts and survey data give a broadly consis- tent picture of consumption growth (see annex 1 for further details.) Thus, one can have a reasonable degree of confidence in the patterns indicated by the data. Now to the factors that explain the diverse response of poverty reduction to growth. Quite obviously, what matters for poverty reduction is not growth per se, but growth in incomes (or consump- tion) of the poor. The impact of economic growth on household income is most simply represented by growth incidence curves, which describe how growth affects income, not just on average but also across the range of the income distribution. Figure 2.3 plots the growth in incomes (consumption) across the percentiles of the income distribution, using survey data from selected countries. The way in which growth affects the incomes of the poor obviously varies by country and by period. Figure 2.3 highlights the importance of the distribution of growth to the overall responsiveness of poverty to growth. Although there is 84 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union considerable diversity within each subregion, the discussion focuses here on one country per subgroup. Starting with Poland in the EU-8, although there was a modest growth in income5 on average since 1999, growth was concentrated in the upper 40 percent of the distri- bution. The lower 60 percent of individuals experienced a contraction in incomes. As a result, poverty increased. By way of contrast, in Romania in SEE, although the poor benefited less than the rich from growth, there was a positive growth in income for all households. As a result, poverty declined. Russia (in the middle income CIS group) and Moldova (in the low income CIS group) represent another inter- esting contrast to both Poland and Romania. In Russia and Moldova, the poor benefited proportionately more than the rich from the growth rebound. As a result, the decline in poverty was greater than would have been the case had growth been distributed more evenly.6 Putting aside changes in distribution, what other factors explain why poverty is more responsive to growth in some countries than in others? The simple arithmetic of poverty reduction shows that the change in poverty can be decomposed into a "growth effect" (defined as the change in poverty in response to changes in average income, FIGURE 2.3 The Poor Have Benefited More than the Rich from the Growth Rebound in the CIS Poland (EU-8) 1999­2002 Romania (SEE) 1999­2002 8 16 6 14 4 12 % 2 % 10 0 8 ­2 6 Growth, ­4 Growth, 4 ­6 2 ­8 0 ­10 ­2 1 10 20 30 40 50 60 70 80 90 100 1 10 20 30 40 50 60 70 80 90 100 Percentile Percentile Russian Federation (middle income CIS) 1999­2002 Moldova (low income CIS) 1999­2002 70 100 90 60 80 % 50 % 70 40 60 50 30 40 Growth, 20 Growth, 30 20 10 10 0 0 1 10 20 30 40 50 60 70 80 90 100 1 10 20 30 40 50 60 70 80 90 100 Percentile Percentile Percentile growth rate Growth rate in mean Source: World Bank staff estimates using data from ECA Household Surveys Archive; see appendix, chart 1, group A for detailed description of the surveys. How Has Poverty Responded to Growth? 85 holding distribution constant) and a "distribution effect" (defined as the change in poverty in response to changes in distribution, holding average income constant). The growth effect or, in elasticity terms, the growth elasticity, which measures the percentage of change in poverty for a 1 percent change in mean income, holding constant the distribution, gives one measure of the responsiveness of poverty to growth. Note that this measure is different from the measures reported in table 2.1, in which the distribution of income was not being held constant. For this reason, the elasticity with constant dis- tribution is referred to as the partial elasticity. The other, reported in table 2.1, is the total elasticity. Although it can be estimated empiri- cally, the partial elasticity has no simple analytical form. However, if the distribution of income is assumed to be log normal, then it can be written as an explicit function of the initial level of inequality and the initial level of income (Bourguignon 2003). In particular, the higher the initial level of inequality, the lower the (partial) elasticity of poverty reduction to growth, and the higher the initial level of income, the higher the (partial) elasticity of poverty reduction to growth. The intuition behind the first proposition is simply that the higher the initial level of inequality, the less the poor benefit from any inequality preserving growth in average income. Hence, poverty reduction is lower. The second proposition is less intuitive and is related to the shape of the income distribution. Figure 2.4 illustrates the relationship between the (partial) elastic- ity of poverty reduction, the initial level of income (measured relative to the poverty line), and the initial level of inequality, using examples from various countries in the Region.7 For countries in the low income CIS group, such as the Kyrgyz Republic, Moldova, and Tajik- istan, poverty is not very responsive to growth: the (partial) elasticity is below 1 for a range of inequality levels. As incomes increase, the elasticity increases. Thus, for countries in the middle income CIS group, SEE, and the EU-8, elasticities are higher. The elasticity is lower where initial inequality is high: therefore, the higher the initial level of inequality, the more "shallow" the curve. Figure 2.4 also shows that the elasticity is more sensitive to inequality at high incomes. For example, at income levels where average consumption equals the poverty line, lowering inequality from 0.45 (the upper range in the Region) to 0.3 (the median) would increase the (partial) elasticity from around ­0.75 to around ­1.25 (66 percent). At three times the level of income (consumption to poverty line equals 3), it would raise the (partial) elasticity from around ­1.5 to around ­3.5 (133 percent). Conversely, an increase in inequality reduces the (par- tial) elasticity of poverty reduction more sharply at high incomes. 86 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 2.4 Poverty Is More Responsive to Growth, the Higher the Level of Income and the Lower the Level of Inequality 0.00 TAJ 99 MOL 99 ­0.50 MOL 00 MOL 98 KYR 00 GEO 00 GEO 01 GEO 99 Gini > 0.45 ­1.00 KYR 01 MOL 01 MOL 02 GEO 97 ARM 00 GEO 98 KAZ 01 ­1.50 ARM 01 KAZ 02 RUS 99 RUS 00 RUS 98 ­2.00 RUS 97 ROM 00 ROM 99 RUS 01 ­2.50 ROM 01 0.45 > Gini > 0.3 BEL 99 Elasticity ROM 98 BEL 98 ­3.00 BEL 00 SER 02 POL 01 Gini < 0.2 POL 00 BEL 01 POL 99 ­3.50 POL 98 UKR 02 0.2 > Gini > 0.3 ­4.00 HUN 99 HUN 00 ­4.50 HUN 98 HUN 01 ­5.00 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25 3.5 3.75 Consumption to poverty line (1 if average consumption = poverty line) EU-8 SEE Middle income CIS Low income CIS Source: World Bank staff simulations based on country-level data from ECA Household Surveys Archive. Note: The same poverty line ($2.15 a day at 2000 PPP) was applied across all counties. This framework helps draw attention to the fact that changes in dis- tribution essentially play two roles in poverty reduction. The first is a direct or "one-time" effect on poverty because of an increase or decrease in inequality (the distribution effect discussed previously). The other is indirect and acts through the (partial) elasticity of poverty to growth. A permanent reduction in inequality not only directly reduces poverty in the same period but also contributes to poverty reduction by increasing the (partial) elasticity of poverty reduction to growth. What explains the differences in the response of poverty reduction to growth, as described in table 2.1? Initial conditions play a large role; however, so do changes in distribution. In the low income CIS countries, low income is an important factor behind the low (total) elasticity reported for this group in table 2.1. Indeed, as figure 2.4 shows, the (partial) elasticities for this group of countries are clus- tered around the ­1 mark. Moreover, as is discussed more closely in the next section, except for Tajikistan, countries in this group experi- enced a decline in inequality over the period in question. This decline resulted in greater poverty reduction than might be expected based on initial conditions alone. As a result, an average (total) elasticity of ­1.4 is observed. How Has Poverty Responded to Growth? 87 In the middle income CIS countries, a wide range of elasticities are observed, from ­1.5 in Kazakhstan to more than ­3.5 in Ukraine. The average (partial) elasticity for this group is around ­2.5. However, as in the low income CIS group, most countries experienced a decline in inequality over the period in question. As a result, poverty was more responsive to growth than initial conditions would suggest, and an average (total) elasticity of ­3.1 is observed. In SEE, (partial) elasticities are comparable to those found in the middle income CIS group, clustering around ­2.5. However, there is no clear trend in inequality for countries in this group for the period studied (see the next section). Average (total) elasticity is thus similar to the average (partial) elasticity. For the EU-8, the low reported (total) elasticity in table 2.1 seems at odds--at first glance--with the (partial) elasticities in figure 2.4, which are the highest in the Region.8 However, in countries such as Poland, slow growth and rising inequality means that, in many instances, income growth among the poor was negative, and poverty increased despite positive growth. Averaging across periods in which poverty increased (and thus the elasticity was positive) and those in which it decreased (and thus the elasticity was negative) results in a low (total) elasticity for this group of countries as a whole. Changes in Distribution, What Happened and Why As the previous section suggests, changes in distribution have an important role to play in enhancing or reducing the poverty impact of growth. It is therefore useful to understand what changes have occurred and, to the extent possible, why. With few exceptions, changes in inequality during 1998­2003 have been relatively modest.9 At the same time, distinct patterns of change are discernable at the subregional level. In the CIS, the overall impres- sion is of stable or declining Gini coefficients (see appendix A, Poverty Indexes), except for Georgia and Tajikistan. In SEE and the EU-8, the picture is mixed, with both increases and decreases (figure 2.5). Are there any common factors underlying these trends, particu- larly in the CIS countries, where there has been a tendency for inequality to fall (again, with the notable exception of Georgia and Tajikistan)? To examine this further, the study decomposes inequality into the contribution of inequality "between" groups and inequality "within" groups. The focus is on the market for labor in which the bulk of incomes are earned, dividing up households into groups char- acterized by wage employment, entrepreneurial activities, subsis- 88 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 2.5 Distribution Has Moved in Favor of the Poor in Most CIS Countries Estonia Hungary EU-8 Lithuania Poland Albania Bulgaria SEE Romania Serbia & Montenegro Belarus CIS Kazakhstan income Russian Fed. Middle Ukraine Armenia CIS Georgia Kyrgyz Rep. income Low Moldova Tajikistan Colombia Turkey Benchmarks Vietnam 0.100 0.200 0.300 0.400 0.500 Gini Index 1998 2003 Sources: World Bank staff estimates using ECA Household Surveys Archive; see appendix table 2 for country-level data and years. tence activities, and nonemployment (retirement, unemployment, and so on).10 The study uses the Theil entropy measure of inequality, which can be conveniently decomposed into contributions of inequality within and between groups. The share of within-group inequality is the product of inequality within the group and the share of the group. Thus, the share of a particular group to overall inequal- ity may change either because inequality within the group has changed or because the share of the group in the total population has How Has Poverty Responded to Growth? 89 changed. The share of between-group inequality is the inequality that remains if all households in a group are given the group's average income (that is, there is no within-group inequality). The sum of the within-group and between-group contributions equals 1. Figure 2.6 plots the shares of the within-group and between-group elements for eight countries, treating inequality in the initial year as equal to 100. The changing pattern of inequality in the Region does not offer any simple explanation. A few broad generalizations, however, do emerge. First, between-group inequality has had little, if any, role to play in explaining changes in inequality. Second, the growth of entre- preneurship has been a factor pushing up inequality in most coun- tries. This is because as a group, it is associated with higher inequality in outcomes than wage employment or subsistence activities are, and its share in the total population has been rising. There are, however, exceptions to this finding: notably Georgia and Romania, where a decline in the share of households characterized by entrepreneurial activity has resulted in a falling contribution of this group. Third, the rise in the "contribution" of the nonemployed is an important factor behind rising inequality, particularly in the EU-8 and SEE. The rise is due to growing inequality within this group, accompanied, in some cases, by the rising share of this group. Growing inequality among the nonemployed may be a reflection of the increasingly poor opportuni- FIGURE 2.6 "Decomposition" of Inequality Does Not Explain Declines in Most CIS Countries Share of Between- and Within-Group Inequality in Theil Index 120 100 80 index period 60 of Theil 40 of start % at 20 0 Poland Hungary Romania Kazakhstan Russian Georgia Moldova Tajikistan Fed. EU-8 SEE Middle income CIS Low income CIS Between groups Within wage earners Within entrepreneurs Within subsistence Within nonworking Note: Left bar corresponds to 1998 or earliest year. Right bar corresponds to 2003 or latest year. Source: World Bank staff estimates using ECA Household Surveys Archive. Note: Theil entropy measure, see World Bank 2005j for detailed technical discussion of decomposition techniques. 90 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ties for those who are unemployed or out of the labor force to sustain their standard of living relative to pensioners and can be related to the failure to raise the share of the employed in the total population. (The issue of jobless growth in the EU-8 and SEE is addressed in chapter 3.) Beyond these generalizations, how different factors come together is very much a country-specific matter. In Russia, for example, where overall inequality declined, the main factor is the shift from self- employment (whether entrepreneurial or subsistence) to wage employment, accompanied by a decline in inequality among wage earners. One factor explaining this decline is the reduction in arrears, which has been a feature of the economic recovery in the CIS. Wage arrears were regressive in impact, driving up inequality among wage recipients (Lehmann and Wadsworth 2001). It is therefore likely that arrears reduction has been beneficial to equality. In Moldova, too, overall inequality declined, not because of changing shares of differ- ent groups, but because of a decline in within-group inequality for all major groups (that is, wage employees, entrepreneurs, and subsis- tence farmers). The reduction in wage inequality may be because of arrears reduction; however, the changes in inequality among entre- preneurs and those engaged in subsistence farming require further investigation. In contrast, in Poland and Romania, upward pressure from nonworkers has been reinforced by rising inequality among wage earners. This is no doubt related to the further decompression in wages in these countries (World Bank 2003k; World Bank 2004h; World Bank Forthcoming-a). The Relative Shares of Growth and Changes in Distribution in Poverty Reduction Given the importance of both growth and changes in distribution to poverty reduction, it is useful to understand the relative importance of the two factors. Figure 2.7 plots the shares of growth and changes in distribution to poverty reduction for selected growth periods since 1998. The "growth share" measures how much poverty reduction can be attributed to growth in average incomes alone (that is, assum- ing no change in the distribution), while the "distribution share" measures how much poverty reduction can be attributed to changes in the distribution of incomes alone, assuming no change in average income. The growth share is a function of not only the rate of growth but also the (partial) elasticity of poverty reduction to growth. Where both growth and changes in distribution have been favorable or unfa- vorable to poverty reduction, both contributions can be expected to How Has Poverty Responded to Growth? 91 go the same way. Where poverty has increased despite positive growth (for example, Poland during 2001­2), the share of growth to poverty reduction is negative. Figure 2.7 highlights the overwhelming importance of growth to poverty reduction over the period in question. With few exceptions, the contribution of growth to poverty reduction is more than 75 per- cent. Relative to growth, the contribution of changes in distribution to poverty reduction has been relatively small. This is perhaps not that surprising, given the modest changes in inequality over the period. Thus, while inequality is an important part of the story, out- side of a few countries and periods, it is the less important partner. Although small on average, changes in distribution have clearly been quite important in some countries. For example, in Poland dur- ing 1998­99, 40 percent of the increase in poverty is attributable to the increase in inequality and 60 percent to the decline in income. In 2000­2001, the impact of changes in inequality is even greater. In fact, it explains all of the increase in poverty. Indeed, as poverty increased, despite growth in household incomes on average, the contribution of growth to poverty reduction is negative. In a number of countries in the CIS, the share of changes in distribution to poverty reduction in the period since the end of the financial crisis is more than 20 percent. FIGURE 2.7 Share of Growth in Poverty Reduction Is Dominant across All Regional Subgroups 20 10 points % 0 poverty, ­10 in ­20 Change ­30 Fed. Fed. Rep. Poland Poland Hungary Belarus Romania Romania Ukraine Armenia Georgia Georgia Moldova Moldova Tajikistan Kazakhstan Kazakhstan Russian Russian Kyrgyz 2000­1 1998­9 2001­2 1998­9 2000­1 2000­1 2001­2 2002­3 1998­9 1999­ 2002­3 2000­2 1998­9 2001­2 2000­1 1998­9 1999­ 1999­ 2002 2002 2003 EU-8 SEE Middle income CIS Low income CIS Overall change in poverty: Due to growth Due to inequality Source: World Bank staff estimates using data from ECA Household Surveys Archive; see appendix, chart 1, group A for detailed description of data sets. 92 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Rural-Urban and Other Subnational Differences in Poverty Reduction The discussion now turns from the national aggregates to examine the extent to which the response of poverty to growth varies at the subnational level. Chapter 1 drew attention to the relatively large dif- ferentials in rural and urban rates of poverty in the Region outside of the low income CIS countries. With few exceptions, this relativity (by which is meant the risk of poverty in rural areas relative to urban areas) appears to have increased over time in SEE, the middle income CIS countries, and the low income CIS countries, although in the lat- ter group the increase is modest. The gap in poverty between urban and rural areas in the Region remains modest compared with what is observed in developing countries (figure 2.8). The rising relative risk of poverty in rural areas is because poverty headcounts have declined more sharply in urban than in rural areas. By contrast, there is no clear trend in the relative poverty risk in the EU-8. It is worth trying to understand fully the exact nature of these changing relativities. Figure 2.9 presents the trends in poverty head- counts for one country in each of the four regional groupings. Because poverty rates vary substantially across the four countries, dif- ferent scales were deliberately chosen to highlight the changes that are relevant for each country. Although the trends in rural and urban poverty broadly track each other, where poverty has declined or increased, it has declined or increased somewhat more rapidly in urban than in rural areas. In other words, poverty has responded more strongly to growth (whether positive or negative) in urban than in rural areas. For countries where poverty has declined, the stronger urban trend is easier to observe in Romania (as representative of SEE) and Russia (as representative of the middle income CIS countries) than in Moldova (as representative of the low income CIS countries), where rural areas experienced a particularly strong decline in poverty between 1999 and 2001. The figure presents only one country, Lithuania (EU-8), where poverty has increased for most of the period, and here the sharper increase in urban poverty is noticeable. It should be clear from the discussion in the previous chapter that if poverty in urban areas were to be broken down further, the contrast between trends in capital cities (which would tend to lead the pack of all urban areas) and rural areas would be even more striking. What factors underlie the lower responsiveness of rural poverty to growth? The framework developed earlier in this chapter provides some insight into the issue. As discussed previously, the (partial) elas- ticity of poverty reduction in relation to growth is a function of initial How Has Poverty Responded to Growth? 93 FIGURE 2.8 Increase in the Ratio of Rural to Urban Poverty in Most Countries Hungary Lithuania EU-8 Poland Bulgaria Romania SEE Serbia Belarus CIS Kazakhstan income RussianFed. Middle Ukraine Armenia Georgia CIS KyrgyzRep. income Moldova Low Tajikistan Turkey Colombia Benchmarks Vietnam 0 1 2 3 4 5 6 Ratioofruraltourbanpoverty 1998orearliest 2002orlatest Source: World Bank staff estimates using data from ECA Household Surveys Archive; see appendix table 3 for country-lev- el data and years used. Note: For EU-8, Belarus, and Bulgaria $ 4.30 a day at 2000 PPP is used as poverty line, otherwise $2.15. levels of income and initial levels of inequality. In general, income in rural areas is lower than income in urban areas in the Region. Where this is combined with higher inequality, the responsiveness of poverty to growth in rural areas is lowered further, making poverty reduction "doubly" difficult. Capital cities have even higher incomes and thus are expected to have more favorable conditions for poverty reduc- 94 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 2.9 Urban Poverty Is More Responsive to Growth and Falling (or Rising) More Rapidly than Rural Poverty Lithuania Romania (EU-8) (SEE) 50 30 poor poor 45 as as 25 40 20 defined 35 defined 30 15 25 population population 10 of 20 of % % 15 5 1998 1999 2000 2001 2002 2003 1998 1999 2000 2001 2002 2003 Russian Federation Moldova (Middle income CIS) (Low income CIS) 25 85 poor poor as as 75 20 defined defined 65 15 55 10 population population 45 of of % % 5 35 1997 1998 1999 2000 2001 2002 1998 1999 2000 2001 2002 2003 Urban Rural Source: World Bank staff estimates using data from ECA Household Surveys Archive; see appendix table 3 for data. Note: Poverty line is $ 4.30 at 2000 PPP used as poverty line for Lithuania, $2.15 a day in other countries. tion, although higher inequality there may diminish their potential in reducing poverty with a given growth rate. How exactly all these fac- tors interplay can be studied with survey data. (Partial) elasticities (based on distributionally neutral growth) are presented for the Region and benchmark countries in figure 2.10 (note that because elasticity is negative, in the figure the smaller elas- ticity is closer to the top). In general, one observes very similar rural and urban secondary cities' (partial) elasticities in the low income CIS countries, reflecting similar levels of consumption and inequality in rural and urban areas. However, capital cities clearly stand out, with much higher elasticities than observed in either rural or secondary cities. Outside the low income CIS countries, rural incomes are sig- How Has Poverty Responded to Growth? 95 nificantly lower, and the gap in the elasticity between urban and rural areas the largest (although here too there are exceptions: see, for example, Belarus). Where lower incomes in rural areas are combined with higher levels of inequality, it results in a lower (partial) elasticity of poverty reduction in rural areas. Even where inequality in rural areas may be no higher than in urban areas, the impact of lower incomes typically results in a lower rural elasticity (see figure 2.10). Thus, for most countries outside of the low income CIS group, initial conditions with regard to not only income but also (in some cases) inequality are such as to make poverty in rural areas less responsive to growth. Intuitively, a lower elasticity of poverty to growth in rural areas is not difficult to understand. Rural households have access to land and the means to produce their own food, which is a very important item of consumption for the poor. One would therefore expect during a recession or a macroeconomic downturn that rural poverty will rise less sharply than urban poverty. Conversely, in an upswing, one would expect urban poverty to fall more sharply because of the bet- ter integration of the urban poor into labor markets. FIGURE 2.10 Partial Elasticity of Poverty Reduction to Growth Is Lower in Rural Areas 0 ­1 ­2 ­3 Elasticity ­4 ­5 ­6 ­7 Hungary Poland Romania Bulgaria Russian Kazakh- Belarus Armenia Moldova Georgia Kyrgyz Tajiki- Vietnam Colombia Turkey Fed. stan Rep. stan EU-8 SEE Middle income CIS Low income CIS Benchmarks Rural/national per capita consumption Rural areas Other urban areas (secondary cities) Capital cities Source: World Bank staff estimates using data from ECA Household Surveys Archive; see appendix table 3 for the latest years available by country. Note: Simulations using the assumptions of distributional neutral growth and the latest year of survey data. For EU-8, Belarus, and Bulgaria $ 4.30 a day at 2000 PPP is used as poverty line, otherwise $2.15. 96 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Lower (partial) elasticity of poverty reduction in rural areas implies that without higher rates of growth in rural areas or signifi- cant improvements in distribution, reductions in rural poverty can be expected to lag behind reductions in urban poverty. Over time, this can lead to an increasing relative risk of poverty in rural areas and a concentration of poor in rural areas, unless mitigated by migra- tion from rural to urban areas. Conversely, where growth is nega- tive, the increase in poverty among rural residents can be expected to be lower than among urban residents. These expected changes in the risk of poverty in rural areas relative to urban areas appear to have been borne out since 1998. Outside of the low income CIS countries, where changes were marginal, there has been a growing relative risk of rural poverty in subregions (and countries) where poverty has declined. Where poverty has increased (for example, Poland and, to some extent, Lithuania), the relative risk of poverty in rural areas has declined. As a result, the relative risk of poverty in rural areas has grown in SEE and the middle income CIS countries, while in the EU-8 (with its mix of poverty outcomes), there is no clear trend. As with rural-urban differences, the framework developed earlier in this chapter is also useful for understanding growing regional dif- ferences in poverty. Like rural areas, many poor regions, especially those with high levels of inequality, could face a poverty "trap" (that is, face poverty rates that respond very slowly to growth). Over time, such regions could become a "pocket of poverty" in an otherwise growing national economy. This is likely to be important in countries where there are large regional differences, such as Russia.11 From a poverty perspective, differences in the response of poverty to growth suggest that without concerted efforts to raise growth rates in rural areas and lagging regions, persistent differentials in poverty are likely to be maintained for some time, or they may even grow over time. Thus, a key issue for public policy is: what should be done about these persistent or growing differentials? Chapter 5 discusses some key measures to address regional and spatial inequality. Conclusions This chapter elaborated on the role of growth and changes in distri- bution in explaining poverty reduction in the Region since the end of the financial crisis in Russia. In general, high rates of growth and the overall responsiveness of poverty to growth have meant that growth has been the most important factor in explaining changes in poverty. How Has Poverty Responded to Growth? 97 In the CIS countries, growth has gone hand in hand with improve- ments in distribution (with the notable exception of Tajikistan), which has enhanced the impact of growth on poverty reduction. Some of the reduction in inequality may simply be a feature of "catch- up" growth and may be due to factors such as the elimination of wage arrears. For this reason, it is not clear that further improvements in distribution can be expected in the future. As a result, poverty reduc- tion may become more difficult, particularly if distribution shifts away from the poor. Outside the CIS countries, trends in inequality have been mixed. However, in the EU-8, growing unemployment and polarization of employment opportunities away from the poor are an important and growing source of inequality in some countries. Where this is not addressed, it will act as a significant brake on poverty reduc- tion and social inclusion. This chapter also elaborated on how poverty is less responsive to growth in rural areas than in urban areas. As a result, there is a per- sistent and, in many cases, growing gap in poverty incidence between rural and urban areas, which can be expected to grow over time. Addressing this gap will require a concerted effort to raise rates of growth in rural areas if subnational gaps are not to persist and mili- tate against attempts to reduce overall poverty. Much the same applies to lagging or poor regions, which also face disadvantages sim- ilar to those of rural areas. In the chapters that follow, the discussion turns to the impact of growth on household income. The most important source of house- hold income is the market for labor (chapter 3). Following a discus- sion of labor markets and poverty, the focus turns to the role of service delivery in influencing poverty, both in the short term, through impacts on nonincome dimensions or capabilities, and in the medium to long term, through impacts on human capital (chapter 4). Finally, the study examines prospects for poverty reduction in the future and public policy priorities for reducing poverty (chapter 5). Annex 1: How Accurately Do Survey Data Record Changes in Consumption in the Region? There are two main sources of information on consumption: house- hold sample survey data ("survey data") and national accounts data. Survey data, collected from households, form the basis for analysis of poverty and distribution. National accounts are the basis for measur- ing economic growth and include an estimate of private consumption 98 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union (or private consumption expenditure), which is typically computed as a residual from national income after backing out other sources of final expenditure. Although related in principle, the concept of consumption meas- ured by the two sources is different. For example, survey data gen- erally exclude consumption of goods and services by unincorporated businesses and nonprofit organizations; however, these are included in the national accounts measure of private consumption expendi- ture. Another difference is imputed rent from owner-occupied hous- ing, which is covered in national accounts, but typically not constructed for surveys. The two sources are also subject to inde- pendent measurement error, and the price deflators typically used to bring expenditures to real terms also differ. As a result, the two sources yield different estimates of consumption and consumption growth. Although the two sources provide different estimates of consump- tion, they are reasonably well correlated across most regions and data sets. In the past, however, data from the Region have been considered somewhat suspect because of the perception that household-survey- based private consumption data are below acceptable quality stan- dards and contain numerous oddities. Ravallion (2001b) finds no correlation between the private consumption growth rates from sur- veys and those from the national accounts. The data covered 27 growth episodes in 19 countries, mostly in the period from 1988 to 1996. For the same reason, Ravallion (2001a) drops data for the Region's countries in a cross-country study on the MDGs. Adams (2002) documents growth and inequality changes among the various regions of the world and argues that the data from the Region should not influence the debate on global patterns of poverty and inequality change because of their poor quality. Although this may have been true in the early years of transition, which were characterized by economic and institutional decline and crisis, the data work undertaken for this report suggests a different picture. To conduct the analysis, this report has re-created consump- tion aggregates from recent surveys in a comparable manner to enable good cross-country comparisons. The data suggest a strong positive correlation in growth rates as measured by the two sources (see annex figure 1). In fact, when growth rates of survey means are regressed on growth in private consumption (national accounts), a ß- coefficient of 0.95 (t = 6.37) is obtained, which is comparable to the estimate of 0.84 (t = 5.74) reported by Ravallion (2001b) for coun- tries of the world, excluding the Region. Annex figure 1 plots the 38 regional "periods" used in the report alongside the graph used by How Has Poverty Responded to Growth? 99 ANNEX FIGURE 1 Growth of Household Consumption Measured by SNA vs. Surveys a. World 30 year 20 per % 10 mean, 0 survey in ­10 rate ­20 Growth ­30 ­15 ­10 ­5 0 5 10 15 Consumption growth rate from National Accounts, % per year between survey data Source: Ravallion (2001b), figure 3. Note: Data cover 115 growth "spells" from the 1980s and 1990s in Africa, East and South Asia, and Latin America. b. ECA 30 year per 20 % 10 mean, 0 survey in rate ­10 Growth ­20 ­30 ­15 ­10 ­5 0 5 10 15 20 Consumption growth rate from National Accounts, % per year between survey data Source: Staff estimates. Note: Data cover 38 growth spells in the ECA region over 1998­2003. 100 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Ravallion (2001b). As might be expected from the above regression results, the pictures look very similar. Within the picture of broad consistency between the two meas- ures, there are some notable outliers. In a few instances, consump- tion growth as recorded by the two sources goes in opposite directions; for example, Georgia (1998­99, 1999­2000) and Romania (1999­2000). In others, although the growth rates go in the same direction, they differ quite substantially; for example, Moldova (1998­99, 2001­2) and Russia (1998­99). At the same time as these deviations are relevant for the measurement of poverty in the indi- vidual countries, they do not detract from the picture of overall con- vergence between data patterns in this Region and other regions. This impression of broad convergence is reinforced when the ratio of consumption measured by the two sources is examined. As argued by Deaton (2004), consumption as measured by survey data is typi- cally found to be lower than consumption as measured by national accounts. Moreover, the ratio of the two is often lower in richer coun- tries than in poorer ones, which is attributable, at least in part, to the difficulty of sampling well-off households. As a result, many coun- tries have seen a decline in the ratio between survey consumption and national accounts consumption over time. Following Deaton's approach, one finds a similar relationship in the Region, with the poor countries tending to have the higher ratios (see annex figure 2). The average ratio of survey consumption to national accounts consumption is 0.69, which is lower than the aver- age reported by Deaton for the Region's countries in his sample (0.84).12 As with growth rates, within the broad picture of consis- tency with patterns from other regions, there are some notable out- liers. Belarus and Ukraine stand out as having relatively high ratios among the group of middle-income countries. As noted by Deaton and various other authors (Adams and Raval- lion; but also Bloem, Cotterell, and Gigantes 1998), it would be incorrect to presume in favor of national accounts: this alternative can also be subject to many errors and may well be overstating growth in consumption. Bloem, Cotterell, and Gigantes (1998) argue that the introduction of the 1993 System of National Accounts (SNA) in most of the formerly centrally planned economies led to the emer- gence of a number of problems. Some of these problems probably cause overestimates of national accounts variables, while others cause underestimates, and it would be purely coincidental if these effects cancel each other out. Most researchers now would agree that a unit ratio between survey and national accounts is rarely a sign of data quality. How Has Poverty Responded to Growth? 101 ANNEX FIGURE 2 Level of Household Consumption in SNA vs. Surveys Consumption from Surveys to SNA consumption in ECA, 1998­2003 1.20 1.00 Accounts, ratio 0.80 National of consumption, 0.60 Survey 0.40 6 7 8 9 10 GDP per capita, log Source: World Bank ECA Household Survey Data Archive. Note: Data cover 51 estimates of consumption from 13 countries in ECA over 1997­2003. See tables 1 and 2 in the ap- pendix. Deaton (2004) uses variance in the ratio of survey to national accounts consumption as a crude indicator of combined survey and national accounts quality. He argues that problems with national accounts notwithstanding, survey measures are more likely to vary from year to year because of changes in sample and survey design, and from country to country because survey protocols are less stan- dardized than national accounts. Applying this measure to data from 59 surveys from the Region, Deaton does not find evidence of partic- ularly high variance in the Region. Analysis with the more recent and standardized data used in this report suggests an even closer match. Overall, the emerging evidence of consistency of data quality in the Region with that from other regions does not come as a surprise and can be related both to the economic rebound in the Region and to investments in improving statistical capacity (in both national accounts and sample surveys) that have occurred in the Region. As observed earlier, this does not mean that there are no problems in using survey data in the Region. Overall, however, there is a weak case for regarding the Region as something of an outlier for the pur- poses of global analysis. Within the Region, further efforts can be made to improve overall data quality, especially in countries where there are inexplicable inconsistencies or trends. However, the initial efforts in improving data quality appear to have paid off. 102 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ANNEX TABLE 1 Ratio of Private Consumption from Household Surveys and National Accounts Unweighted Population weighted No of Mean Standard Standard Mean Standard Standard surveys Ratio error deviation ratio error deviation Estimate (ECA) according to this report 57 0.632 0.025 0.190 0.646 0.021 0.159 Estimate according to Deaton (2004) All 277 0.860 0.029 0.306 0.779 0.072 0.191 EAP 42 0.819 0.069 0.224 0.863 0.031 0.110 ECA 59 0.847 0.038 0.230 0.796 0.040 0.184 LAC 26 0.767 0.094 0.329 0.585 0.078 0.193 MENA 20 0.955 0.104 0.300 0.867 0.111 0.270 OECD 33 0.781 0.052 0.097 0.726 0.032 0.076 SA 23 0.649 0.063 0.122 0.569 0.036 0.103 SSA 74 1.000 0.061 0.415 1.089 0.089 0.459 Sources: Deaton (2004) and staff estimates based on ECA Household Survey Archive. Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America and Caribbean; MENA = Middle East and North Africa; SA = South Asia; SSA = Sub-Saharan Africa. Annex 2: Inequality in the Region in Comparative Perspective How equal or unequal is consumption in the Region relative to other parts of the world? The answer to this question is not straightfor- ward because comparison of inequality across countries and regions is confounded by the noncomparability of the measures reported by different sources (see World Bank 2005j). This study nonetheless makes a first attempt to place inequality in the Region in interna- tional perspective (see annex figure 3). For the countries in the Region, the authors use the data developed for this study, for which a concerted effort has been made to ensure comparability. However, the same degree of comparability is difficult to achieve for the wider international sample. To increase comparability, alongside the data from this study, measures are reported that pertain to per capita con- sumption from other countries (inequality in per capita income is generally higher). However, the authors have not been able to con- trol for other factors such as differences in the definition of con- sumption across surveys. Therefore, the comparisons should be treated as an approximate, rather than exact, picture of differences across countries. As the accompanying figure suggests, the Region shows the full spectrum of inequality outcomes, from fairly unequal to fairly equal. Median inequality in the Region is lower than in the rest of the devel- oping world; however, it is broadly comparable to the median inequality in rich countries. How Has Poverty Responded to Growth? 103 ANNEX FIGURE 3 Inequality in the Region in an International Perspective 0.15 0.30 0.45 0.60 Consumptionpercapita,Ginicoefficients AFR EAP ECA HIC LAC MENA SAR Sources: World Bank 2005j; World Bank 2003g; Bazen and Moyes 2003; and Sieminska and Garner 2002. Note: HIC stands for "high income countries," which include France, Greece, Israel, Italy, Spain, Taiwan, United Kingdom, and United States. What factors "account" for inequality in the Region? In other words, to what extent can inequality be explained by inequality between groups, such as rural residents versus city dwellers, or high school graduates versus those with less than high school education? Looking at several types of partition (including education, age and gender of the household head, rural versus urban residence, and 104 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union region), the World Development Report 2006 (World Bank 2005j) finds that the two factors that explain the highest share of total inequality in the Region are education and region (each explaining around 10 per- cent of total inequality in the median country). However, these factors are no more nor less important than in other parts of the world. There- fore, the Region does not stand out as one where educational or regional differences are exceptionally important as drivers of inequality. The World Development Report data do not include Russia, a country with vast regional differences. It is an open question whether includ- ing Russia would give a different impression of inequality in the Region (or at least in Russia). One respect in which the Region does appear to be somewhat dif- ferent is the smaller role of rural-urban inequality. Unlike in most developing regions, differences between rural and urban areas do not explain a significant share of the overall national inequality (the share in the median country is about 5 percent, in the Region less). In this respect, the Region appears more like the high-income countries in the sample, where rural-urban differences play a relatively small role in explaining total inequality compared with other factors. Endnotes 1. The only countries to have experienced negative growth in GDP since 1999 are FYR Macedonia (2001) and Turkey (2001). The Kyrgyz Repub- lic experienced zero growth in 2002. 2. Data from the early years of the transition may not be fully comparable in some cases; therefore, comparisons should be treated as illustrative, rather than fully indicative. 3. In Russia, it is estimated that higher oil prices accounted for growth in excess of 4­5 percent per year. 4. For all analyses in this chapter, this study uses data from 15 countries (10 countries from the A cluster on chart 1 in the appendix, plus Estonia, Kazakhstan, the Kyrgyz Republic, Ukraine, and Uzbekistan) spread across the four subgroups (see chapter 1 for groupings for the Region). For the quality of association between household survey and national accounts data in the Region, see annex 1 to this chapter. 5. Term "income" here is used for explanation purposes, but all data refer to consumption. This is justified by poor reporting of income data in household surveys across the Region, making recorded consumption level a more accurate reflection of living standards at the household level (see also appendix, A. Data and Methodology, for a detailed discussion of the use of income versus consumption in the Region). 6. The time of the financial crisis in Russia (1998­99) was also one of declining inequality. Incomes contracted across the board; however, the contraction was sharper in the uppermost deciles. The fact that the rich How Has Poverty Responded to Growth? 105 suffered proportionately more than the poor (or, what is the same thing, that the distribution of income shifted in favor of the poor) moderated the impact of the crisis on poverty. 7. Elasticities are computed using the growth rate in the year in question. The distribution of income is assumed to be standard log normal, with parameters estimated from survey data. For the relationship between the Gini coefficient and dispersion parameter s, see Dikhanov (1996). The distributionally neutral change in poverty is taken from Epaulard (2003, 12). 8. However, the poverty line used to produce figure 2.4 is $2.15 a day, while $4.30 was applied for estimates reported in table 2.1; the lower poverty line automatically implies higher elasticities. Use of $2.15 in many EU-8 countries produces a headcount that is not statistically sig- nificant based on actual survey data, and thus any empirical analysis of elasticity is meaningless. For the theoretical distribution simulated to produce figure 2.4, this is not an issue. 9. For a discussion of inequality in the Region in an international context, see annex 2 to this chapter. 10. The definitions used are as follows: dependence on (a) wage employ- ment: no working members who are self-employed and minimal income from self-production (<5 percent); (b) entrepreneurial activities: at least one adult in self-employment, but minimal income from self-production (<5 percent); (c) subsistence activities: at least one adult in self-employ- ment and significant income from self-production (>5 percent); and (d) nonemployment: no adult in employment or self-employment. 11. See Kolenikov and Shorrocks (2003) for an analysis of factors underly- ing differences in poverty rates across Russian regions. Large regional disparities are not confined to big countries: Hungary is also known for large dispersion across regions. See, for example, Förster, Jesuit, and Smeeding (2005). 12. For the report, this study uses a fairly parsimonious definition of con- sumption to enhance comparability across countries. In particular, the survey-based consumption measures do not include flow of services from durables, rent (where paid), and catastrophic health care. CHAPTER 3 The Role of Labor Markets and Safety Nets In 2003, more than two-thirds of the poor in the Region (or around 40 million poor people) belonged to families where someone worked. Although economic growth has served the poor well (par- ticularly the working poor), they remain the largest group among the poor. This chapter analyzes the main channels through which growth affected the well-being of the poor during 1998­2003. It shows that alongside higher wages, increased transfers were instru- mental in reducing poverty. But neither higher wages nor increased transfers can be expected to sustain poverty reduction in the Region. The chapter concludes that higher productivity and enhanced employment generation are needed to sustain poverty reduction. To achieve this, policy makers need to push for the continuation of structural reforms to bring market discipline to old enterprises and encourage entry by new firms. How the Poor Can Connect to Growth The poor connect to growth processes in various ways, direct and indirect. This chapter adopts a simple framework (used by "Pro-Poor Growth in the 1990s," World Bank 2005f) to analyze how economic growth shapes the opportunities available to the poor in the Region. 107 108 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union There are three main channels that affect different groups among the poor. The unemployed poor directly benefit from increased employ- ment resulting from growing demand for their labor. The working poor gain from rising real wages or higher productivity of their self- employment. Growth can also trickle down to the nonworking or eco- nomically inactive poor through increased public and private transfers (figure 3.1). Policies affect the scale of opportunities open to the poor. Geographic location, gender, or membership in a specific group (eth- nic, political, and so forth) influence access to these opportunities. Other individual circumstances (dependency rates and so forth) determine whether a given growth in earnings or transfer income is sufficient to move a household out of poverty. Different patterns of growth have different effects on the poor, depending on where they are. Table 3.1 presents labor market profiles of the poor (defined as employment status of the household head) in four representative countries of the Region: Poland (EU-8), Romania (SEE), Russia (middle income CIS group), and Moldova (low income CIS group). This table complements the data on poverty by individual labor market status discussed earlier (chapter 1, figures 1.4 and 1.9). This study adopts the definition of the working poor in line with the one developed by the Indicators Subgroup of the EU Social Protection Committee.1 It defines the working poor based on the work intensity FIGURE 3.1 Connecting the Growth to the Poor New employment Country policies and conditions · Education · Social policies Country policies and · Labor market policies conditions Growth · Education Productivity Outcomes opportunities · Infrastructure and wages for the · Political economy poor · Redistribution of Household and group assets characteristics · Health · Education · Gender Public/private · Dependency transfers Source: Adapted from World Bank 2005f. The Role of Labor Markets and Safety Nets 109 TABLE 3.1 Work Does Not Protect Families from Poverty in the Region Poverty Profile by Sector and Type of Employment of the Household Head, Selected Countries, around 2002 Poland Romania Russian Federation Moldova Household head Poverty Share of Poverty Share of Poverty Share of Poverty Share of employment rate Poor rate Poor rate Poor rate Poor Sector of employment Agriculture 35.8 15.1 26.3 33.7 20.1 24.8 67.1 35.7 Industry 35.3 24.2 9.1 9.8 13.4 14.3 51.3 9.1 Services 19.2 29.5 10.7 21.4 3.2 27.9 48.3 26.9 Type of employment Public employee 18.8 16.7 5.0 5.0 5.0 35.4 48.9 11.6 Private employee 31.8 35.4 7.6 11.6 5.5 28.2 56.3 35.6 Self-employed 25.7 16.8 26.3 48.3 9.9 3.4 61.9 24.3 Employed 25.6 69.0 14.8 64.9 6.6 67.1 56.7 71.4 Not employed 30.4 31.0 16.3 35.1 12.7 32.9 57.4 28.6 Total 27.1 15.3 8.6 56.9 Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: For Poland $4.30 a day at 2000 PPP is used as a poverty line; for other countries in this table $2.15 is used. of the household as a whole. If no member of a household with work- capable members worked for even a single day in the reference period, such a household is classified as "jobless"; all other households with employed work-capable members are classified as "working." Work does not protect families from poverty in the Region. Table 3.1 shows that the working poor (in a broad sense) constitute two- thirds of the poor. Their risk as a group is noticeably lower than the average poverty incidence, especially in Poland and Russia, but it is definitely above zero everywhere. There are also clear differences across sectors, with agriculture characterized by an elevated poverty risk and services by a significantly lower risk. The growth of the ser- vice sector, therefore, can be expected to have different conse- quences, depending on whether it is translated into the increase in employment (in which case, it will strongly contribute to poverty reduction) or the rise in earnings (in the latter case, the impact on poverty will be minimal because workers in the sector are above the poverty line already). With regard to ownership structure, the public sector has the lowest incidence of poverty, while self-employment is characterized by the highest incidence. These channels may reinforce each other, but they can also inter- act negatively. There are complex interactions between the three channels presented in figure 3.1. For example, if wage growth out- paces productivity improvements, it may depress the demand for labor. Excessive and poorly designed transfers may create depend- ency traps and discourage the poor from taking advantage of new 110 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union employment opportunities. Underdeveloped safety nets may, on the other hand, prevent the poor from changing jobs, thereby locking them into low-productive activities. This chapter is organized around the three channels presented in figure 3.1. The next (second) section presents trends in wages, employment, and public and private transfers during 1998­2003. The third section assesses not only how well the poor in the Region were able to take advantage of new opportunities by entering the labor market or moving to more productive and remunerative occupations but also how well they were served by the transfer system. The fourth section explains why the poor were able to connect to growth in dif- ferent degrees. The fifth section concludes by reviewing the main findings and discussing implications for policy. Economic Opportunities Have Expanded Rapid real wage growth outpaced employment generation in the Region. Figure 3.2 presents the evidence on broad economywide trends in employment and wages. It shows that between 1998 and 2003, real wages increased in all countries in the Region. Successes in generating new employment were less impressive, especially in SEE and the EU-8, where job destruction during the period exceeded job creation (see box 3.1). In the poorest CIS countries, average real wages have almost dou- bled since 1997. This rebound should be put in an historic context. In the early 1990s, real wages in transition economies fell sharply. In CIS countries, where enterprises adjusted to the fall in output by delaying salaries rather than by shedding redundant labor, real wages gener- ally fell more than in the EU-8, where enterprise restructuring was carried out through labor retrenchment. Since the mid-1990s, real wages have recovered everywhere, rising faster than output and pro- ductivity in most cases. Comparison with Vietnam in figure 3.2 sug- gests that rapid increases in wages (from a low base) are not unique to the Region, but issues of sustainability are key (discussed in the last section of this chapter). Average wages are above the poverty threshold for most countries in the Region. The increase in real wages translated directly into raising consumption of workers and pulled many out of poverty. Although around 1998, practically all countries in the low income CIS group had average wages below the poverty standard ($2.15 a day per capita),2 in 2002, wages would put an average worker in poverty only in Tajik- istan. In the EU-8 and SEE (except for Serbia and Montenegro), aver- age wages also exceeded the economic vulnerability threshold ($4.30). The Role of Labor Markets and Safety Nets 111 FIGURE 3.2 Real Wage Growth Typically Outpaced Net Employment Growth in Transition Economies Real Wages and Employment by Country Groups and Benchmark Countries; 1998 = 1.00 EU-8 SEE 1.60 1.60 1.40 1.40 1.00 1.00 = 1.20 = 1.20 1998 1998 1.00 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 2003 Middle income CIS Low income CIS 1.60 1.60 1.40 1.40 1.00 1.00 = 1.20 = 1.20 1998 1998 1.00 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 2003 Turkey Vietnam 1.60 1.60 1.40 1.40 1.00 1.20 1.00 1.20 = = 1998 1.00 1998 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 2003 Employment index Real wage index Sources: ILO Key Indicators of the Labour Market (KILM); LABORSTA (ILO); General Statistics Office of Vietnam; and Turkey's State Planning Organization. Note: The following countries were used to compute averages. Wage data are average wages of full-time workers in all sectors of economy deflated using CPI in- dexes. Simple averages are used. The Consumer Price ondex was used to deflate current wages. For Vietnam: Data on wages includes only state sector employees. For Turkey: Wage data is for private sector employees only. Employment data before 2000 included persons 12 years or older, and later--persons 15 years or older. Employment data is for civilian employment only. The structure of employment has changed, even though employ- ment levels might have been stable or declining. Figure 3.3, based on household data (which is typically different from official unemploy- ment data, as discussed in box 3.1) reports dynamics of employment for selected countries in the Region and for benchmark countries, breaking 112 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 3.3 The Structure of Employment Has Changed Wage- and Self-Employment Rates Over Time for the Region and Benchmark Countries 80 70 60 population 50 age 40 30 20 working of 10 % 0 Poland Estonia Hungary Bulgaria Romania Belarus Russian Georgia Moldova Tajikistan Turkey Colom-Viet- Fed. bia nam EU-8 SEE Middle income Low income CIS Benchmarks CIS Wage employment Left bar corresponds to 1998 or earliest year. Self-employment Right bar corresponds to 2003 or latest year. Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: Employment and self-employment levels are derived from household survey data and may differ from official statistics; includes full-time and part-time em- ployment with at least one hour of gainful work in the reference period of the survey. The age brackets are 16-64 (inclusive) for all countries. it down by wage employment and self-employment. In Bulgaria, Geor- gia, Moldova, Romania, and Russia, noticeable structural shifts between self-employment and wage employment occurred during the period. But even in 2002, all transition economies in the Region (except for Georgia) had significantly lower rates of self-employment, com- pared with benchmark countries. It suggests that there exists a poten- tial for further shifts in the employment structure (box 3.1). The major source of new employment in low income CIS countries has been the growing sector of self-employed (own-account) work- ers. Self-employment was the main source of employment growth in the low income CIS group. In these countries, self-employment accounts for about half of total employment, compared with 17 per- cent and 20 percent for middle income CIS countries and the EU-8, respectively. Figure 3.3 shows significant reallocation between two main forms of employment, wage employment and self-employment, suggesting the existence of large flows in the labor market. The process of labor reallocation between sectors is far from over. The transition process has disproportionately affected the manufac- turing sector, resulting in a reduction in the share of employment in industry, while the service sector share of total employment has The Role of Labor Markets and Safety Nets 113 BOX 3.1 In Most Countries, Household Survey Data Report Higher Employment Figures than ILO Statistics This report relies on information from the Household Budget Survey (HBS) or variations on it, such as the Integrated Survey or the Living Standards Measurement Study (LSMS). Even though this type of data is not primarily intended to measure employment, it provides representative coverage and collects information on earnings alongside information on characteristics and ac- tivities of household members. The HBS programs across transition countries have benefited from international technical assistance, with a fair degree of useful unification and standardiza- tion. As a result, their quality is sound. HBS data have been used extensively in empirical stud- ies that explore the relationships between employment, earnings, and poverty. However, HBS- based figures may differ from official labor data, which rely on different sources. As the table in this box suggests, trends in employment, as reflected in different sources, point in the same di- rection (except for Bulgaria, Georgia, and Moldova). Such discrepancy is not unique to the Re- gion, with HBS data generally capturing higher employment ratios than specialized labor market surveys. Such a discrepancy may come from a different reference period to qualify respondents as employed (in the HBS, typically a longer survey period) or to better capture some informal ac- tivities in the HBS/LSMS integrated surveys. Employment Rates from HBS/LSMS and Official Labor Data, Percentage of Population EU-8 SEE Poland Estonia Hungary Bulgaria Romania 1998 2002 2000 2003 2001 2002 1995 2003 1998 2002 Survey 60.9 54.0 61.4 62.3 63.1 61.4 61.6 62.7 63.9 61.9 ILO data 59.0 51.5 60.3 62.0 56.5 55.5 42.2 41.5 64.2 57.6 Middle income CIS Low income CIS Benchmarks Belarus Russian Fed. Georgia Moldova Turkey Colombia 1998 2002 1999 2002 1999 2002 1998 2002 2002 2002 Survey 69.7 67.7 64.0 67.3 69.2 59.7 64.2 71.0 55.0 62.8 ILO data 65.4 64.3 57.6 65.0 49.5 53.2 52.9 52.4 50.8 51.6 Sources: World Bank staff estimates using data from ECA Household Surveys Archive, ILO Key Indicators of the Labour Market (KILM), and Labor Sta- tistics (LABSTAT), http://laborsta.ilo.org/. Note: Age brackets in HBS/LSMS are chosen to be consistent with the official definition. generally expanded in response to rising demand for services from both consumers and enterprises (World Bank Forthcoming-a). The shifts of employment between sectors of the economy contin- ued between 1998 and 2003. Figure 3.4 reports shifts in employment between agriculture and services over the past five years. In most coun- tries, even where aggregate employment ratios have been stagnant, the shift of employment between sectors continued. These shifts imply that 114 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union in a typical country, at least 5 percent of workers change their jobs on a net basis.3 Data for two benchmark countries (Colombia and Turkey) suggest that these developments are in line with sectoral reallocation observed in rapidly developing economies with structural transforma- tion. Expansion of services, a general pattern observed in all countries in the Region, is reflecting a global trend and has generally meant good news for the poor in the Region because this sector is characterized by the lowest poverty risk (see table 3.1). In contrast, the evolution of agri- cultural employment varies by subregion. In most EU-8 countries, agri- cultural employment has fallen, and its share is now close to the EU benchmark; however, agriculture employment has increased slightly in most SEE and low income CIS countries. Employment shifts have generally enhanced productivity. The implications of the reallocation of labor across sectors are better understood by taking into account the productivity differentials between sectors. Data on sectoral productivity (measured as value added per worker) reported in figure 3.5 are fully consistent with the poverty profile by sectors of employment (table 3.1): higher produc- tivity implies lower poverty. Figure 3.5 also illustrates differences in productivity levels across sec- tors and across country groups. In EU-8 and middle income CIS coun- FIGURE 3.4 Employment in Service Sector Expanding; in Agriculture, Mixed Structure of Employment by Sectors for Subregional Groups and Benchmarks, 1998­2003 Agricultural employment Service employment 70 60 50 40 employment 30 total of % 20 10 0 EU-8 SEE Middle Low Turkey Colombia EU-8 SEE Middle Low Turkey Colombia income income income income CIS CIS CIS CIS 1998 2002 or latest Sources: ILO Key Indicators of the Labour Market (KILM) database and World Bank staff estimates. Note: In 2000 US$. For this figure CEE countries include the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovak Republic, and Slovenia. SEE countries include Bulgaria, Croatia, and Romania. Low income CIS countries include Armenia, Azerbaijan, Georgia, the Kyrgyz Republic, Moldova, Tajikistan, and Uzbekistan. Middle income CIS countries include the Russian Federation. The Role of Labor Markets and Safety Nets 115 tries, the recent shift in employment away from agriculture and into ser- vices can be seen as movement into more productive sectors. It also reflects general trends in productivity-enhancing job reallocation within and across industries, or resource flows from less productive to more productive firms within industries (Brown and Earle 2004a and b). While EU-8 countries approached or exceeded levels of productivity typically observed in middle-income developing economies, the middle income CIS group and SEE significantly lagged behind. At the same time, low income CIS economies exhibit levels and patterns of produc- tivity typical of low-income developing countries. The increase of agricultural employment in the low income CIS group also potentially enhances productivity. In low income CIS countries, agricultural self-employment has underpinned the growth of aggregate employment and the reduction of poverty. Because it generally reflects the movement of people from unem- ployment (or potential unemployment) into agricultural self- employment, the growth of agricultural employment has improved household welfare. In fact, the evidence presented in this report (see box 3.2 and appendix B. Key Poverty Indicators, table 3) sug- gests that where labor intensity is highest (that is, where the ratio of labor to land is high, as is typical in the poorest CIS countries), FIGURE 3.5 Value Added per Worker Is Lowest in Agriculture Sectoral Value Added per Worker, Subregional Groups and Benchmarks, in 2000 PPP Dollars 12 10 8 6 thousands $ 4 2 0 EU-8 SEE Middle Low Colombia Turkey Vietnam income income CIS CIS Benchmarks Agriculture Industry Service Sources: World Bank World Development Indicators (WDI) database; ILO Key Indicators of the Labour Market (KILM) data- base; and World Bank staff estimates. Note: In 2000 US$, 2002 or latest available year. For this figure CEE countries include the Czech Republic, Estonia, Hun- gary, Latvia, Lithuania, Poland, Slovak Republic, and Slovenia. SEE countries include Bulgaria, Croatia, and Romania. Low income CIS countries include Armenia, Azerbaijan, Georgia, the Kyrgyz Republic, Moldova, Tajikistan, and Uzbekistan. Middle income CIS countries include the Russian Federation. 116 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.2 The Role of Agriculture inTransition Transition to market has affected agriculture through several channels.The liberalization of prices and the trade and subsidy cuts caused a dramatic decline in farm profitability and rural incomes in the Region at the onset of the transition. Reduced domestic demand, with falling incomes and subsidy cuts, was reinforced by falling foreign demand and increased import competition with trade liberalization. Land reforms, farm restructuring, privatization of agrifood companies, and lib- eralization of markets have initially caused important disruptions and sometimes reinforced out- put declines and income falls. More recently, these reforms have been sources of growth. The process of reforms, the implementation, and the effects, both initially and more recently, have varied tremendously between countries. There are differences between groups of coun- tries in the Region in the levels of agricultural productivity and implications for labor markets and poverty. Productivity in agriculture remains lowest in low and middle income CIS countries, but it is also relatively low compared with other sectors in national economies elsewhere. The em- ployment levels are high in the low income CIS group and SEE. In some EU-8 countries (Hun- gary), agricultural productivity was increasing very fast and converged to average levels, while in others (Poland), the gap remains wide open. But employment levels in agriculture are declining very rapidly there.Thus, the challenges and policy implications are different. In the EU-8, large-scale agriculture led productivity growth, with major labor shedding from farms. Effects on poverty have been mitigated by increased social transfers. In those countries, large-scale privatized farms have laid off surplus workers, who have either found jobs in other sectors, become unemployed, or gone into early retirement. The countries could manage asso- ciated fiscal costs because initial levels of agricultural employment were low. Productivity growth came mainly from major gains in labor productivity on large farms, rather than from yield in- creases, which started only as of the mid-1990s. SEE faced different challenges. Rural households in Bulgaria, Romania, and countries in the Balkans possessed relatively developed and capital-intensive agricultural production systems, and productivity gains from shifting to small farms were less than those in poor, labor-intensive agricultural countries, while the costs of losses of scale economies and technology disruptions were larger. With limited access to credit, inputs, and technology and few off-farm employment opportunities in rural areas, there were many constraints limiting growth or investments. The safety net systems were less developed than in the richer Central European countries, and households that were laid off by large farms had to fall back to farming to complement their in- come. As a consequence, relatively few people left farming, and there was even some inflow of labor into agriculture as people laid off in industry fell back to semisubsistence agriculture. The problems were further complicated because land restitution concentrated land ownership in the group of older households.Young and dynamic people left the rural areas in search of better op- portunities in the urban areas, and often abroad, pushing up the share of older, low-skilled, and The Role of Labor Markets and Safety Nets 117 less-educated people in rural areas and dampening productivity improvements. In the middle income CIS group (for example, Kazakhstan and Russia), large farms have re- mained, and the restructuring of these large-scale farms did not lead to significant income gains in rural areas, but it has not resulted in open unemployment either.This can be explained by the more capital-intensive nature of agriculture in these countries. Indeed, raising productivity of large farms depends to a much greater extent on access to other inputs and factors, which col- lapsed in the early years of transition. Low agricultural productivity in these countries is further exacerbated by relatively poor human capital stock.The recent improvements in rural poverty in those countries are likely due to an improvement in agricultural prices and overall liquidity in the economy, which have pushed up wages and improved services and wage payments, but not through significant gains in productivity. In the low income CIS countries, although it is true that observed productivity is extremely low, one should take into account the extremely low capital intensity of this production, which often developed as a coping strategy. An important factor mitigating against even lower agricultural productivity was land reform. Land distribution to poor rural households during transition in- duced important productivity gains and enabled self-employment in agriculture, which helped mitigate the negative shock of transition. Figure 3.6 shows the high degree of correlation be- tween labor intensity in agriculture and the growth of household farming. In labor-intensive rural economies, access to land through distribution of land plots to rural households induced impor- tant growth in productivity and income in rural areas. This is what happened in China in the late 1970s, in Vietnam in the mid-1980s, in Albania in the early 1990s, in the Kyrgyz Republic in the mid-1990s, in Azerbaijan in the late 1990s, and in Moldova after 1999 (which may explain why Moldova is the only country where rural poverty declined much more strongly than urban pover- ty during 1998­2003). Factor Intensity and the Growth of Household Farms Individual farming 5 years after reform 100 China 80 Balkans use 60 land Caucasus of 40 Share 20 Central Europe Core CIS countries 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Persons per 100 hectares Sources (figure): Rozelle and Swinnen 2004; (box): Macours and Swinnen 2004. 118 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union the ratio of rural to urban poverty rate also tends to be lower--an association consistent with the critical role of labor-intensive agri- culture in mitigating poverty. Nevertheless, the expansion of agri- cultural employment in low income CIS countries also reflects the reallocation of labor toward a sector where productivity is lowest and where further increases in productivity may not be possible without addressing market imperfections in factor and output mar- kets. By contrast, reallocation of labor to services has been slow (see box 3.2). Changes in employment and earnings discussed so far represent only a part of the channels that connect the poor to growth. Changes in the real value and direction of transfers represent another major channel. The Region is characterized by a significant amount of redistribution that takes place both through the govern- ment budget in the form of taxes, social contributions, and transfers and through private channels in the form of remittances, gifts, and in-kind intrafamily reallocations (see World Bank 2000a). At the onset of transition, the severe recession led to large declines in gov- ernment revenues throughout the Region and consequently large declines in social safety-net spending. However, better macro and fiscal performance after 1998­99 helped to address major gaps in financing safety nets and to increase the real value of transfers. This has facilitated the regularization of pensions and social benefit pay- ments and relieved poverty pressures on the working poor and on the elderly. There are substantial differences between countries in the Region in the role of public transfers, with richer countries spending two times more in GDP shares than poorer countries do. By 2000, spend- ing on social security and welfare in the low income CIS group con- stituted (on average) 6.5 percent of GDP, while in the EU-8, it represented 13.8 percent of GDP. Fiscal management (with some exceptions) has been careful over the period and avoided major spikes in transfer payments, but even with roughly stable shares of GDP, economic growth resulted in increases in real transfer payments. The expansion of real transfers is reflected in the real value of pen- sions, which constituted between 70 percent of all public transfer spending in the EU-8 and 50 percent in the poorest CIS countries. Table 3.2 shows that there was an increase in the real value of pen- sions in all country groups, particularly strong in SEE and the middle income CIS countries. Public transfers (at least pensions) played a much more limited role in the low income CIS countries, and an average public pension remains well below the poverty threshold. However, very high levels of transfer payments in the EU-8 and SEE The Role of Labor Markets and Safety Nets 119 TABLE 3.2 The Evolution of Pension Spending by Groups of Countries Middle income Low income EU-8 SEE CIS CIS 1998 2002 1998 2002 1998 2002 1999 2002 Pension spending/GDP 9.72 9.47 7.80 8.45 6.85 6.90 4.50 3.34 Real Pension Index 1.00 1.05 1.00 1.30 1.00 1.40 1.00 1.04 Pension Spending, $* 871 917 434 565 241 339 59 61 Source: Regional Fiscal Database * Annual per capita in 2000 PPP Note: EU-8 = Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, and Slovak Republic; SEE = Albania, Bulgaria, and Croatia; middle income CIS = Belarus, Kazakhstan, the Russian Federation, and Ukraine; and low income CIS = Armenia, Azerbaijan, Georgia, the Kyrgyz Republic, and Moldova. are causes for concern because they may introduce perverse incen- tives to the nonworking poor and contribute to fiscal imbalances that hamper prospects for growth. Private transfers compensated limitations of public transfers in low-income countries. Data on private transfers are extremely scarce in the Region, but they are believed to exceed by several times what is available as public social welfare, especially in countries where remittances play a major role (see box 3.3). As economic growth in several large countries in the Region attracted migration of workers from poorer economies, remittances have expanded too. The evidence presented in this section suggests that the economic growth in the Region's countries translated into an expanded set of opportunities open to people. Were the poor able to take advantage of these opportunities? The next section assesses how the gains from economic growth were distributed among various groups. The Poor Took Advantage of New Opportunities Opportunities have expanded everywhere in the Region, in the form of either increased earnings or new employment. Parallel increases in public and private transfers made it easier for the poor to connect to growth. Among these factors, increasing real wages were of para- mount importance so far. Figure 3.6 presents evidence on how the gains in real wages were associated with poverty reduction in the Region between 1998 and 2003, as well as in several benchmark countries. Each symbol on the graph represents a period of poverty changes as measured by the survey data and a corresponding change in the economywide average real wage. The Region's countries clearly stand out in the size of the changes in real wages and in their impact on poverty. 120 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.3 The Role of Remittances in the Region During the past 15 years, there were huge migration flows within the Region, as well as from coun- tries in the Region to the rest of the world. By 2003, close to 30 million citizens of the Region's countries were residing abroad. It is evident that such size of migration, most of which is due to economic reasons, has had an effect on the economies of both the receiving and sending coun- tries. Migrants` transfers to relatives and friends at home have recently gained prominent impor- tance in several countries of the Region. In 2003, remittances to the Region from relatives and friends working or living abroad amounted to about 10 billion U.S. dollars.The Balkans and Eastern Europe received the major bulk of total migrants' transfers into the Region. In Albania, Bosnia and Herzegovina, Moldova, andTajikistan, they account for more than 10 percent of GDP (in Moldova, a quarter of GDP in 2002).These numbers, based on the balance-of-payments statistics, are likely to underestimate the scale of remittances because of the predominance of informal channels to remit. Remittances can play an important role in poverty alleviation.They spur domestic consump- tion, investments, and human capital accumulation and thus contribute to growth. The study of their effects on the poor is made difficult by relatively poor reporting in the household surveys, but available statistics suggest that the poor rely more on remittances than the nonpoor. Source: Chernetsky Forthcoming. The differences across countries highlight the different roles of channels connecting the poor to growth. Some growth periods in the EU-8 resulted in a sizable wage growth in highly productive sectors, with initially zero poverty accompanied by fixed or falling overall employment levels. Thus, there are some periods over which wage and poverty changes were not correlated. In SEE, employment has fallen; gains in real wages have been moderate, while transfers have expanded considerably; hence, the potential for dissociation between poverty and wages. In the middle income CIS group, employment opportunities, earnings, and transfers have all increased at a simi- larly high pace, and the relationship between wages and poverty was the tightest. Finally, in the low income CIS group, there was a grow- ing divide between expanding job opportunities in less-productive agriculture and stagnant employment in increasingly better-paid, more-productive sectors. These differences had important implica- tions for the poor. Higher average economywide wages were good for the poor. Fig- ure 3.6 suggests that despite some dissociation between wage and poverty movements in the Region, there remains a strong link between changes in real wages and poverty. Earnings growth helps The Role of Labor Markets and Safety Nets 121 FIGURE 3.6 Real Wage Changes Correlate with Poverty Changes Change in Poverty and Change in Real Wages in Percentage by Country Groups, Annual Periods, 1998­2003 +40 +20 %, +0 poverty in Change ­20 ­40 ­20 ­10 +0 +10 +20 +30 Change in real wage, % EU-8 SEE Middle income CIS Low income CIS Benchmarks Sources: World Bank staff estimates using data from ECA Household Surveys Archive for poverty and ILO for data on wages. Data on benchmark countries are from World Bank 2005f and include Bangladesh, Bolivia, El Salvador, Tunisia, and Vietnam during 1990­2002. Note:. The line represents a trend derived with the OLS regression across points representing the Region's countries. the working poor. Because many poor in middle income CIS coun- tries were working poor, employed for wages, the relationship between real wage growth and poverty reduction is the strongest for this group. But if the working poor represent a smaller fraction of all poor (in the EU-8 and SEE), or the working poor consist predomi- nately of the self-employed (low income CIS countries), the real wage gains will have less-consistent effects on poverty. Moreover, if the growth in wages is not equitable, increases in inequality may under- mine the impact of wage gains on poverty reduction. The poor gained from wage increases in many countries. Figure 3.7 uses household survey data to examine the relationship between average economy-wage changes and poverty by tracing the evolution of real wages across the spectrum of distribution. It shows that real wages have typically improved for both rich and poor workers, but by different degrees. In Poland, only the well-off have gained, while in Moldova, Romania, and Russia, there was a similar rate of increase for the top and bottom quintile. 122 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Occupational wage data also suggest broad-based wage growth across the Region. The International Labour Organization (ILO) data on monthly earnings for selected occupations for a sample of CEE and CIS countries during 1998­2003 also suggest that real earnings have generally increased across occupations and across countries. Most important, real earnings of blue-collar workers have increased, in some countries more rapidly than in the others. Wages of construc- tion laborers and welders have increased by about 30­50 percent FIGURE 3.7 Poor Gained from Real Wage Gains in SEE and the CIS Real Wage by Quintiles of Consumption: 1998­2003 ($ per month in 2000 PPP Poland Romania (EU-8) (SEE) 700 300 600 month 500 month 200 per per $ $ 400 300 100 Poorest Second Third Fourth Richest Poorest Second Third Fourth Richest Quintile of consumption Quintile of consumption Russian Federation Moldova (middle income CIS) (low income CIS) 400 200 300 month 200 month 100 per per $ $ 100 0 0 Poorest Second Third Fourth Richest Poorest Second Third Fourth Richest Quintile of consumption Quintile of consumption Average monthly wage, 1998 Average monthly wage, 2002 Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: Per capita consumption, reported monthly wages in 2000 PPP US$. The Role of Labor Markets and Safety Nets 123 between 1999 and 2003 in the Czech Republic, Latvia, and Moldova, but only by 5­10 percent in the Kyrgyz Republic, Poland, and Roma- nia (ILO 2004). Wage inequalities have recently declined, especially in countries with initially high levels. Chapter 2 provided decompositions of inequality changes and highlighted the role of wages as a driver of the overall inequality outcomes. Direct estimates of the Gini coefficient for wages based on survey data show stability or declines for most countries in the Region. Declines in earnings inequality were also more pronounced than the increases.4 Improvements of wage distribution in CIS can be traced to a reduc- tion of wage arrears. As economies have recovered, the incidence of wage arrears has fallen in all CIS countries. In Russia, for example, the proportion of workers with arrears rose steadily from 1994 through 1998, when it reached 63 percent, then fell sharply in 2000 to 29 percent. Similarly, the average number of overdue monthly wages fell from 3 to 1 between 1998 and 2000. At its peak, wage arrears were regressive in impact, driving up inequality among wage recipients (Lehmann and Wadsworth 2001). Not surprisingly, arrears reduction has been beneficial to equality (World Bank 2005g). Returns to education have stabilized. Extensive literature has emerged in recent years documenting the rapid increases in returns to education during the early stage of the transition. Few country case studies cover the pretransition period in the EU-8 and middle income CIS countries all the way through the late 1990s using comparable data; those that do suggest that the sharpest increases in returns to skills happened in the early transition and that returns seem to have largely stabilized by 2000 at around 8 percent per year of schooling, a level observed in most market economies.5 Nonworking poor had problems connecting to growth in countries with little job creation. In many EU-8 and SEE countries, notably in FYR Macedonia and Poland, unemployment rates have been increas- ing, despite growth. Over the past five years, unemployment rates have fallen only in a few countries, particularly in the middle income CIS group and the Baltic States. Where major industrial restructuring was postponed (for example, in Bulgaria and Romania in SEE), unemployment exploded in the aftermath of renewed reform efforts. Another concern is long-term unemployment. The incidence of long-term unemployment in transition economies is higher than that in their advanced-economy counterparts, and some countries have experienced rapidly increasing shares of long-term joblessness among the unemployed. Among advanced market economies, the incidence of long-term unemployment is below 40 percent, and some countries 124 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union have successfully reduced long-term joblessness in recent years. In contrast, in the EU-8, some 40 to 50 percent of the unemployed have been jobless for at least a year; the incidence of long-term unemploy- ment has been generally stagnant, although in some countries it has risen rapidly since 1997. For example, long-term joblessness among the unemployed in Poland rose from 38 percent in 1997 to 48 percent in 2002. Over this same period, long-term unemployment increased from 28 percent to 50 percent of the unemployed in the Czech Republic. In low income CIS countries, long-term unemployment is also remarkably high. Some 70 percent in Armenia have been unem- ployed for at least a year. In the Kyrgyz Republic, the incidence of long-term unemployment among the registered unemployed increased from less than 10 percent in 1995 to nearly 30 percent in 2001 (Babetskii, Kolev, and Maurel 2003). Youth unemployment also remains high throughout the Region, about two to three times the average unemployment rate. This has discouraged many young workers and, in some cases, has led to increasing inactivity. Women took advantage of new opportunities. Gender inequality in employment over the transition does not point to a specific disadvan- tage of females (Paci 2002). Transition affected men and women dif- ferently; but in nearly two-thirds of the countries, the ratio of female to male activity rates has increased slightly, indicating that women are more likely to be employed than men (only Bosnia and Herzegovina, the Kyrgyz Republic, and Tajikistan show worsening outcomes for female employment). There is limited evidence to suggest a reversal of these trends in the recent past. However, there are also emerging gen- der differentials in the extent to which formal employment has been replaced by informal economic activity, but with ambiguous effects on poverty. The poor benefited from expanded job opportunities and lost out where they shrank. Figure 3.8 uses household survey data to see how the employment rate has moved over time for the poor as compared with the rich. It shows that in countries where employment has increased noticeably (Moldova and Russia), all income groups have benefited from this increase, and the poor benefited approximately to the same extent as wealthier households. In Poland, where employ- ment fell, all groups suffered, but the poor suffered disproportion- ately. In Romania, the employment rate has not moved for any quintile of the distribution, and the poor have an employment rate well below that of the rich. Where job opportunities have been shrinking, it has hurt youth and workers with poor human capital endowments the most. As mentioned previously, long-term unemployment is high and, in The Role of Labor Markets and Safety Nets 125 many instances, growing. Across all countries, the bulk of long-term unemployed are those with lower education attainment, inhabitants of rural and remote regions, and representatives of ethnic minorities (for example, Roma in the EU-8 and SEE). Some groups among the poor literally face handicaps to participat- ing in the growth process through employment. Rising health inequal- ities could well be an important obstacle to equitable (and sustainable) economic growth and poverty reduction because the relationship between health status and employment is extremely strong. For those individuals who are between ages 40 and 59, Mete and Liu (2005) find that poor health status leads to a 56.7 percentage-point decline in the probability of employment in Romania. Reflecting on the past, deteri- orating population health status during transition might have con- tributed to the decline in employment rates. As for the upcoming challenges in the future, deterioration of health status could emerge as a serious obstacle in achieving higher employment rates, productivity, and (by extension) a more-equal distribution of income. Unemployment in depressed regions has proven to be stubborn. Bornhorst and Commander (2004) find that regional disparities in unemployment rates are large and grew from 1991 to 2001. In recent years, the gaps between regions with the lowest unemployment rates and the regions with the highest unemployment rates have remained large, about 10 percentage points for most countries for which data are available and more than 50 percentage points in Russia. In Russia, the gap narrowed between 1997 and 2000, but has steadily increased since then. Compared with selected OCED counterparts, transition economies have experienced a higher degree of variation in regional unemployment rates; the dispersion in Russia, in particular, is higher than that in most comparators. Bornhorst and Commander argue that this labor market phenomenon is consistent with rising long-run unemployment. It is also associated with rising inactivity, because high unemployment rates have tended to discourage workers. As noted earlier, the real value of transfers increased in most coun- tries, and quite substantially in the middle income CIS countries and in SEE. This increase occurred largely through the existing social safety nets, which on aggregate and in most countries in this group are distributionally neutral. This implies that increases in transfers during 1998­2002 were passed on to the poor and nonpoor alike. The story in the low income CIS group is more complex. Although there was a small gain in the overall amount of public transfers, social assis- tance programs aimed at improving their targeting to the poor, so it is possible that income gains for this group are greater than the averages would suggest. 126 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 3.8 Changes in Employment Rate, 1998­2002, by Quintiles for Selected Countries Poland Romania (EU-8) (SEE) % % 80 80 population, 70 population, 70 age age 60 60 working working to to 50 50 employed 40 employed 40 of of Ratio Ratio 30 30 Poorest Second Third Fourth Richest Poorest Second Third Fourth Richest Consumption quintile Consumption quintile Russian Federation Moldova (middle income CIS) (low income CIS) % % 80 80 70 population, population, 70 age age 60 60 working working to 50 to 50 40 40 employed employed of of 30 30 Ratio Ratio Poorest Second Third Fourth Richest Poorest Second Third Fourth Richest Consumption quintile Consumption quintile Employment rate 1998 Employment rate 2002 Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: Per capita real consumption used to rank households. Employment rate is the share of wage and self-employed among 16- to 64-year-olds. Private transfers, especially remittances, grew during this period and came to play a major role as a source of income growth for the poor, especially in the low income CIS countries. The employment opportunities that expanded in the middle income CIS countries and easier access to labor markets in developed countries set off migration flows. The ensuing flow of workers' remittances has been a sizable source of hard currency for many economies (see box 3.3). Massive outflow of migrant labor from low-income countries of the CIS The Role of Labor Markets and Safety Nets 127 toward middle-income countries may have also had effects on local labor markets, pushing real wages up. Public and private transfers not only expanded but their effect on poverty also seems to have become stronger. Where data are avail- able, they suggest that social benefits have also improved in targeting, coverage, and adequacy. The reduction in arrears, particularly in pen- sions but also in other benefits, has no doubt contributed to these improvements. As a result, social protection transfers have come to play an important role in reducing poverty. Figure 3.9 puts together available evidence on the coverage of social protection programs in general, and pension systems in partic- ular, across the Region's countries. It is derived from household sur- vey data. To make an assessment of coverage, the poor are defined based on ex ante consumption (before the receipt of transfers). As the figure suggests, there is some overlap between social insurance (pen- sions) and other forms of social protection. Figure 3.9 shows that social protection programs generally cover the poor quite well. In Azerbaijan, Belarus, Bulgaria, Georgia, Hun- gary, Poland, and Romania, nearly 100 percent of all poor (assessed based on consumption before transfers) receives some form of social transfer. But even in the poorest CIS countries, coverage rates are high, exceeding 50 percent (except for Tajikistan). But there are stark differences across countries in coverage by social assistance programs targeted at the poor. In EU-8 and SEE countries, close to 80 percent of the poor are captured by these programs. The low income CIS countries show coverage rates around 20 percent (except for Azer- baijan and Uzbekistan). Although the size of transfers in the EU-8 and SEE makes them relatively efficient in reducing poverty, con- strained funding of social programs in low income CIS countries, combined with low coverage, translates into relatively small effects on poverty (see table 3.3). In part, the efficiency of programs in many countries is due to pro- grams targeting the poor increasingly well, although the improve- ments are slower than anticipated (see box 3.4). In Romania, for example, 50 percent of all social protection spending is going to the poor, compared with 47 percent five years ago, a rather limited suc- cess for significant reform efforts. Efficiency also varies enormously between countries. In the Kyrgyz Republic, only 20 percent of funds go to the poor, as opposed to 70 percent in Poland. There is strong evidence that the targeted part of social protection is operating increasingly well (in Kazakhstan, the share of social assistance spend- ing received by the poor increased from 6 percent five years ago to 56 percent in 2003). 128 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 3.9 Safety Nets Cover Many Poor in the Region Coverage of Social Protection by Country Groups, around 2003 Hungary Lithuania EU-8 Poland Albania Bosnia & Herzegovina Bulgaria SEE Macedonia, FYR Romania Serbia & Montenegro CIS Belarus Middle Kazakhstan income Armenia Azerbaijan Georgia CIS Kyrgyz Rep. income Low Moldova Tajikistan Uzbekistan 0 20 40 60 80 100 % of poor households Social assistance Pensions and social assistance Pensions only No public transfers Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: Poor are defined based on ex ante consumption levels, $2.15 is used; see also box 3.4 for detailed references to the ongoing study. The Role of Labor Markets and Safety Nets 129 Social protection transfers were helping to reduce poverty, and poverty would have been significantly higher in a hypothetical "no- transfers" situation (table 3.3). Although somewhat simplistic (partic- ularly in assuming no behavioral response in the no-transfer scenario, except in a few instances), the data are nonetheless illustrative of the importance of public transfers to poverty reduction, especially outside the low income CIS group. Data on private transfers are scarce and do not allow any system- atic assessment of trends. The limited data on remittances (see box 8) suggest that private transfers play a much more important role than public transfers in low income CIS and some SEE countries. Unfortu- nately, it is unclear what happened to the distribution of private trans- fers over time (Chernetsky Forthcoming). Government transfer policies sometimes conflict with labor sup- ply incentives among the poor, although this is more likely to be an issue in the EU-8 and (to some extent) SEE. The design of public transfer systems often implies high marginal tax rates (withdrawal of benefits) for earnings if the poor move from inactivity and unem- ployment to jobs. This creates powerful disincentives for the poor for entering the labor market (Poland: World Bank 2004h; the Slovak Republic: World Bank 2002g). Why Are Many Workers in the Region Still Poor? Despite the evident progress to the benefit of the working poor, every third worker in countries of the Region remains poor or economically vulnerable. Although this is a lower incidence than observed globally because of the low workers-to-population ratio in the Region (see box 3.5), the employed nonetheless represent the largest group among the poor. This section provides answers to a question: why, despite gains to the employed across all countries of the Region in the past five years, do many workers remain poor? There are three groups of explanatory factors. First, the level of productivity (real wages) in the economy or in a particular sector determines whether an average worker is productive enough to earn incomes above the poverty threshold. Second, group and individual characteristics of workers, their human capital, and depend- ency rates may explain why their living standards are lower than those of nonpoor workers. Third, policy, institutional, and structural causes inher- ent in the process of transition explain why, despite similar characteris- tics, some workers remain in poverty despite the nationwide productivity gains. These three sets of factors are discussed in turn. 130 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.4 Improvements inTargeting: Lessons from Recent Policy Reforms Most countries of the Region target social benefits to the poorest.The range of benefits encom- passes social assistance (cash or in-kind), scholarships or free school supplies, health-fee waivers, subsidized medicines, and utility services (heating, electricity, transport, and so forth). Although the overall targeting performance of these programs is very heterogeneous, a few of the Region's programs are among the best performers in the world (Coady, Grosh, and Hoddinott 2004). To study targeting performance further, theWorld Bank has studied key design and implementation arrangements of six well-targeted programs in the Region: Family Poverty Benefit (Armenia) and the Unified Monthly Benefit (the Kyrgyz Republic) in the low income CIS group, Ndihme Ekonomika (Al- bania) and Guaranteed Minimum Income (Bulgaria and Romania) in SEE, and the Social Benefit (Lithuania) in the EU-8. All programs under review transfer a larger share of their benefits to the poor- est quintile, compared with other social assistance programs (see figure that follows). Share of Social Protection Benefit to Lowest Quintile Low-income countries (LICs) Ndihme Ekonomika, Albania 41 Family poverty benefit, Armenia 30 Unified monthly benefit, Kyrgyz Rep. 37 Average nonpension program, LIC 27 Medium-income countries (MICs) Guaranteed minimum income, Bulgaria 58 Social benefit, Lithuania 60 Guaranteed minimum income, Romania 64 Average nonpension program, MIC 27 0 10 20 30 40 50 60 70 % of funds going to the poorest quintile Source: World Bank staff estimates using data from ECA Household Surveys Archive. Although the overall cost of these programs is rather modest, they provide an effective shield against poverty and destitution. Two models seem to emerge. In the EU-8 and SEE, targeted programs are residual programs that serve those households not assisted by other programs, such as social pensions, unemployment benefits, and child allowances.Targeted programs with budgets between 0.25 and 1 percent of GDP are found to reduce the extreme poverty in such countries. In contrast, in the low income CIS countries, fiscal constraints have led to the re- placement of the diversified system of social assistance inherited from the socialist regime by one large, targeted program. These programs transfer about 1­2 percent of GDP and cover 25­35 percent of the bottom decile of the population. The Role of Labor Markets and Safety Nets 131 Programs that use more than one targeting method to select their beneficiaries have better tar- geting performance. Although all programs use means or proxy means tests, these are often combined with work requirements or categorical filters to strengthen targeting. All successful programs seek to balance the need to protect the poorest with maintaining adequate work in- centives. All programs use a comprehensive indicator of household means, which include both formal and informal income, as well as tests of assets--an important feature for reducing inclu- sion and exclusion error. An important finding of the World Bank's evaluation is that administra- tive costs are moderate, and the marginal costs associated with targeting are only a small com- ponent of the total cost. Source: Coady, Grosh, and Hoddinott 2004. TABLE 3.3 Transfer Payments for Social Protection Have Had an Important Role in Reducing Poverty outside the Low Income CIS Countries Increase in poverty without Country Year all social transfers, % EU-8 Poland 2001 141 SEE Bosnia & Herzegovina 2001 68 Bulgaria 2001 156 Romania 2002 49 Serbia 2003 41 Montenegro 2002 34 Middle income CIS Belarus 2002 143 Kazakhstan 2002 100 Russian Fed. 2002 68 Low income CIS Armenia 2001 12 Kyrgyz Rep. 2001 10 Benchmark Countries Guatemala 2000 9 Vietnam 1998 5 Sources: For ECA, World Bank, various poverty assessments; for Guatemala, World Bank 2003; for Vietnam, Van De Walle 2002. Note: Simulations use national poverty lines. Some behavioral response is assumed in Romania (50 percent of transfer in- come is replaced) and Serbia (72 percent of transfer income is replaced in rural areas, 87 percent in urban areas). For Guatemala, transfers include both public and private transfers. Sectoral Profile Low levels of productivity in the low income CIS countries explains why most workers are poor there. Figure 3.5 showed average levels of productivity (in US$ value added per worker) by groups of countries and sectors in the Region. Clearly, when an average agricultural worker 132 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.5 GlobalTrends in the Number of Working Poor A recent ILO report (2004) finds that some 1.39 billion workers and their families are living below the US$2-a-day level in the world. It estimates that one-third of all employed in transition economies were poor around 2003. This aggregate estimate is quite close to the assessment presented in this report, based on actual country-level data. There are important differences in the way ILO and this report estimate the number of working poor. The difference in PPP is one source of these differences (ILO used poverty estimates based on 1993 PPP). This report uses different poverty lines ($2.15 in CIS countries, but $4.30 in the EU-8). The ILO estimates the number of working poor based on the overall poverty rate multiplied by the total labor force.The formulation assumes that the poverty rate of working-age people is equal to that of the popula- tion as a whole and that the labor force participation rate of the poor is equal to that of the pop- ulation as a whole. Both assumptions, according to evidence presented in appendix tables, may not be true for individual countries of the Region, but they reflect the average levels well. Direct evidence on consumption levels of those in employment used to establish whether they are poor or not poor provides much richer insights. Source: Kapsos 2004. in the low income CIS group adds less than 2 dollars per day of value to output, it is impossible to expect that he or she will not be poor. Low agricultural productivity in the middle income CIS countries and in SEE drives agricultural workers into poverty. Low productivity of agriculture in these countries keeps workers there mired into poverty. It is important to understand country variations and patterns of change (see box 3.2), which reflect not only the initial conditions of these countries but also their policies. The analysis suggests that low agricultural productivity in these countries reflects a failure to address the main constraints to rural growth. The poor have not moved to more-productive occupations to the same extent as the nonpoor have. Comparison of changes in sectoral employment shares over time shows that the differences in employ- ment patterns between poor and nonpoor remain large. Figure 3.10 reports results for four representative countries (similar data for other countries yield similar conclusions). The poor are overrepresented in agriculture, the least-productive and shrinking sector, but in some countries, employment in agriculture has even expanded for the poor. The poor are overrepresented among the self-employed, especially in low income CIS countries. In principle, self-employment covers a wide range of occupations, from aspiring entrepreneurs to subsis- The Role of Labor Markets and Safety Nets 133 FIGURE 3.10 Sectoral Wage Employment for the Poor and Nonpoor in Selected Countries Poland Romania Russian Federation Moldova (EU-8) (SEE) (middle income CIS) (low income CIS) 100 75 50 employment of % 25 0 1998 2002 1998 2002 1999 2002 1999 2002 1999 2002 1999 2002 1999 2002 1999 2002 Poor Nonpoor Poor Nonpoor Poor Nonpoor Poor Nonpoor Agriculture Industry Service Source: World Bank staff estimates using data from ECA Household Surveys Archive . Note: Sectoral employment data are derived from household survey data and may differ from official statistics; includes full-time and part-time employment with at least one hour of gainful work in the reference period of the survey. tence farming. Subsistence agriculture, however, is normally charac- terized by high poverty risks. Applying the criteria for defining sub- sistence farming discussed in chapter 2 (analysis of inequality), 20 percent of the population in Georgia, 24 percent in Kazakhstan, and 40 percent in Moldova rely on subsistence farming as the main source of their livelihood. Only 2 percent in Hungary and Poland, 11 percent in Romania, and 14 percent in Russia do the same (see figure 3.10). Informal employment plays an ambivalent role for the poor. Infor- mality is often discussed as a synonym of self-employment, although this is not entirely correct (box 3.6). The informal sector represents a conglomerate of different activities, some of which result in low pro- ductivity and poverty and some of which do not. But many among the working poor are found to be employed in the informal economy. A large informal sector may also have indirect effects on poverty by reducing the tax collections and thus limiting resources available for social programs and services. However, the definition of the informal sector remains one that is difficult to implement in empirical studies of poverty. It is even harder to analyze distinct types of informal employment and their implications for poverty. So far, there is no solid evidence that informality in itself drives workers to poverty in the Region. 134 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.6 Informal Employment inTransition Economies Informal sector jobs are usually defined as value-adding activities outside the tax net and regu- lations. These activities may be unregistered and untaxed by their nature (household subsis- tence economy) or emerge because of purposeful evasion and noncompliance. According to the ILO, the informal economy is a sum of production units operating as unincorporated enterpris- es.This definition emphasizes that the direct relationship between operating revenue of the pro- duction unit and workers' well-being is a constituent feature of informal sector employment. There are four distinct types of informal employment: (a) subsistence-type activities; (b) small- scale entrepreneurial activities; (c) informal wage labor, usually of a casual nature; and (d) em- ployment in formal businesses, with part or all of payment consisting of undeclared wages (for tax reasons or to avoid the withdrawal of social payments and other benefits they are entitled to as unemployed). Evidence shows that some informal sector activities, especially of types (a) and (c), are charac- terized by lower productivity and by higher inequality of outcomes (risk).That would normally im- ply that poverty of workers in the informal sector is higher than in the formal sector, which is in- deed the case in a typical developing country. In transition countries, there is evidence that some small-scale firms chose to remain informal, with resulting undercapitalization and lower productivity than comparable firms in the formal sector. Several studies have documented that many working poor in the Region are employed as wage laborers in such "classical" informal sector activities. But this is only a part of the informal sector employment in transition countries. Individual and Group Characteristics Differences in consumption between the poor and nonpoor can be decomposed into several factors: measures of dependency, labor mar- ket participation, labor earnings per income earner, and other factors (such as transfers). Workers from the lowest consumption quintile in all countries in the Region for which such analysis has been under- taken have higher dependency ratios (usually about 10­20 percent more dependents for each employed) and lower employment rates (sometimes significantly so, especially in SEE and the EU-8, where poor households have a third less employed than nonpoor households). But the importance of these factors as drivers of poverty dimin- ishes compared with the role of earnings. Results presented five years ago and updated in various country studies (Bosnia and Herzegovina, see World Bank 2003d; Serbia and Montenegro, see World Bank 2003l) eloquently show that between 60 and 75 percent of the gap in The Role of Labor Markets and Safety Nets 135 In addition, formal employment in transition also often takes characteristics that make it indis- tinguishable from informal employment. For example, workers of old unrestructured enterprises subject to wage arrears often are characterized by high uncertainty of earnings and low capital endowments, exactly as their informal sector counterparts. On the other hand, workers choos- ing to hide their jobs in fear of losing entitlements to social benefits are generally not poor.Thus, a formal sector job in transition does not automatically imply incomes above the poverty line, and not all informal workers are poor (because of the risk premium, informal activity can be highly re- warding). In addition, a buoyant informal sector can generate additional dynamism in the econo- my and have positive spillover effects on the poor in general. Although the size of the informal sector is notoriously difficult to estimate, informality is report- edly large in low income CIS countries (close to one-half of all employment) and smaller in CEE countries (about a quarter of employment). In relative importance of various types of informali- ty, there are significant differences between the low income CIS group (predominantly subsis- tence farming), middle income CIS countries and SEE (undeclared wage employment in manu- facturing and services, with extensive subsistence), and the EU-8 (household entrepreneurship and undeclared paid jobs in the service sector). Informal sector activities also follow different dy- namics in the EU-8, SEE, and CIS.The EU-8 countries are characterized by a "flat" trend line rep- resenting the size of the informal sector in the recent period. The SEE countries seem to have the size of the informal sector peaking around 1996­99 and have recently seen some reduction. Little is known with certainty about the role and dynamics of the informal sector in low income CIS countries. Sources: ILO 2004;Yoon and others 2003; Schneider 2002; and Commander and Rodionov 2005. consumption levels between poor and nonpoor workers can be explained by this single factor: differences in earnings per each employed. The education profile of poor workers helps to explain some part of this earnings gap. Lower education results in less pay and thus cor- relates with higher levels of poverty. The empirical analysis of earning functions undertaken for the purposes of this report shows that by 2003, a worker with primary education faced a wage disadvantage of 20­40 percent across countries of the Region, compared with a worker with secondary education (Yemtsov, Mete, and Cnobloch 2005). A recent analysis of detailed firm surveys in several countries reveals the presence of education-specific wage differentials within blue-collar skill grades. In particular, blue-collar workers with rela- tively less education have seen their wages fall in Hungary, Romania, and Russia (Commander and Köllő 2004), reflecting relative produc- tivity developments. 136 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Workers with low skills have lower chances to find employment, and once they do, their wages are lower. Thus, differences in labor market prospects between skill groups are translated into different poverty outcomes. But what is interesting is that there are large dif- ferences in wages within the same educational clusters between poor and nonpoor workers, suggesting that skills are not a full explanation of why certain workers are receiving low wages. Gender is also a major source of wage differentials in the Region. The crude gender gap is significant--but declining--in most countries. Controlling for other characteristics, women currently face 3 percent (Bosnia and Herzegovina, 2002) to 25 percent (Tajikistan, 1999) lower wages compared with those of males, with most of the differential not accounted for by their human capital characteristics (Paci and Reilly 2004).6 Although being an important indicator in itself, the wage gen- der gap cannot explain why certain workers are poor. Low-paid female workers often come from nonpoor families and provide secondary sources of income; therefore, their low pay is not a critical factor in determining the poverty status of the household. Standard human capital characteristics explain only a small fraction of the wage gap between poor and nonpoor workers. In-depth studies of wage gap determination for EU-8 countries (Poland, Newell and Socha 2005; the Czech Republic, Munich 2003) and middle income CIS countries (Russia and Ukraine, Gorodnichenko and Peter 2004) find that less than half of the gap in wages between the bottom decile and the median worker can be accounted for by standard Mincerian human capital characteristics (age, education, gender, and so forth). In addition, this gap is quite persistent. Although the explained share of the gap has increased remarkably compared with that of the mid- 1990s in all countries studied, it remains much less determined by human capital characteristics than similar gaps in market economies. Institutional Factors To understand the nature of wage differentials between poor and nonpoor workers requires going beyond individual and household characteristics and connecting earnings of workers to performance of their enterprises. Labor markets in the Region are still functioning less than optimally, with serious barriers to competition and free movement of workers, which are required to equalize wage rates across firms (World Bank Forthcoming-a). This results in persistent differences in pay across firms within the same sector and close links between firm-level productivity and wages. Unfortunately, there are practically no data in the Region that would connect household infor- The Role of Labor Markets and Safety Nets 137 mation on workers to firm characteristics; therefore, indirect evi- dence must be relied on. Two observationally similar workers may have very different pay rates in the Region. Figure 3.11 presents household survey data on average monthly wages for male urban prime-age full-time workers in private manufacturing enterprises by levels of skill and poverty sta- tus (expressed in PPP terms). The figure shows that there is a signifi- cant gap between poor and nonpoor workers with the same characteristics across all country groups. Expressed not as absolute differences in monthly pay rates, but as a percentage gap, it is great- est in Moldova and Russia (more than 50 percent). Remarkably, the gap is significantly narrower in Colombia than in any country in the Region, and there are no manufacturing full-time workers with voca- tional education who are poor in Turkey. These differences can be traced to several factors: differences in wages across firm sizes, differences in wages across subsectors of man- ufacturing, and differences in wages between the formal and informal sectors. Most likely, all these factors coexist and are not entirely unique to transition. But on top of these common factors is also an additional unique source: persistent differences in productivity that are observed between old, restructured, and new enterprises of the same sector in transition economies. As argued elsewhere (see World Bank 2002h), interaction between old enterprises, restructured enterprises, and new entrants is key to economic growth in transi- tion. These differences in productivity are illustrated in figure 3.12. The historical legacy in the form of persistent differences in produc- tivity between old, restructured, and new enterprises can persist, reflecting political economy factors blocking or facilitating reforms that bring about discipline over the old and restructured firms and encourage the entry of new firms. Several sources based on survey data suggest that dispersion of productivities and the resulting dispersion of earnings are indeed larger in transition economies compared with market economies, reflecting the historical legacy described above. First, a detailed study of wage dispersion by narrowly defined subsectors of manufacturing (European Commission 2003) reveals that new member states and accession countries are characterized by a much lower explanatory power of regressions that predict average wages (by subsectors) based on a set of common industrial structure characteristics than are the EU-15. The highest share of "unexplained" dispersion in sectoral wages is found in Bulgaria, the Czech Republic, and the Slovak Republic; the lowest in Hungary and Latvia. But even in the most advanced transition countries, the unexplained portion of dispersion 138 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 3.11 Large Wage Gap between Poor and Nonpoor Persists across the Region Average Monthly Earnings in PPP Dollars for Full-Time Urban Prime-Working-Age Males Employed in Manufacturing Private Firms 700 600 $ 500 wage, 400 monthly 300 Average 200 100 0 Poland Hungary Romania Bulgaria Russian Fed. Belarus Moldova Georgia Tajikistan Turkey* Colombia EU-8 SEE Middle income CIS Low income CIS Benchmarks Primary schooling, nonpoor Primary schooling, poor Vocational secondary education, nonpoor Vocational secondary education, poor Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: US$ in 2000 PPP, prime age is between 35 and 45 y.o. Poor defined as per capita consumption levels below $2.15, and $4.30 in Poland, Hungary, Bulgaria, Colombia, and Belarus. * In the Survey no worker with vocational education is poor in Turkey. in wages is far greater than in Germany, Ireland, Italy, or the United Kingdom. Such differences are reflected in excessively large regional variation in wages within subsectors. Second, a large literature on wage determination continues to show that the explanatory power of the Mincerian earning function7 remains significantly lower in transition countries of the Region than in market economies. It suggests that standard factors used to explain wage dis- persion, such as human capital characteristics, location, and standard job and firm characteristics combined (including sector and occupa- tion), explain around 30­40 percent of hourly wage dispersion in tran- sition economies (more in the EU-8 and less in the middle income CIS group); in market economies, 55­65 percent of variation is fully accounted for by these characteristics (Ukraine and Russia, Gorod- nichenko and Peter 2004; Hungary, Delteil, Pailhé, and Redor 2004). It is important to note here that compared with the mid-1990s, the explanatory power of standard earning functions in transition has increased by at least 10 percentage points (as documented in Newell and Reilly 1997; also see Fleisher, Peter, and Wang 2004), reflecting that restructuring that took place led to clear association between indi- The Role of Labor Markets and Safety Nets 139 FIGURE 3.12 Productivity Distribution of Old, Restructured, and New Enterprises Range of 0 old enterprises Productivity of old enterprises Range of 0 restructured enterprises Productivity of restructured enterprises Range of 0 new enterprises Productivity of new enterprises Source: World Bank 2002h. vidual productivity, job characteristics, and wages. The gap in how well an earning function predicts wages between markets and even the most advanced transition economies suggests that transition is not over. Finally, Munich (2003) directly compares wages and factors of wage determination between new private firms, restructured firms, and state firms in the Czech Republic. He finds not only persistent wage gaps between these groups of firms but also differences in returns to human capital. These studies suggest that the presence of a wage gap between poor and nonpoor workers can be traced to the productivity differen- tials at the micro level and that such differences can possibly be a fur- ther explanation for the working-poor phenomenon. These productivity differentials can be traced in turn to a number of institu- tional factors, but they are normally reflected in the degree of wage dis- persion across the enterprises. Two observationally similar workers may earn very different wages for the same types of job within the same sector. These differences are unique to the transition process and are persistent because there are barriers to competition to equal- ize productivity across historic types of firm and because workers are not taking full advantage of mobility. Lack of discipline and encour- agement in economic policies may thus be a factor behind poverty. Workers' mobility, in principle, could mitigate some of disparities in productivities, but transition economies are characterized by a legacy of low labor mobility springing from a high concentration of enterprise production and a history of low voluntary migration. These, in turn, have been sustained by institutional factors. First, compensation in Russia and many CIS countries tends toward nonmonetary benefits such as housing and childcare, thus sustaining worker attachment to firms. Second, moving costs have been high, commuting costs have 140 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union increased, and information about job opportunities has been poor. Third, institutional incentives discourage mobility. Subsidies have dis- proportionately favored home ownership over rentals. Privatization of cooperative houses and flats has promoted home ownership. Andrienko and Guriev (2004) find complementary evidence that migration is constrained by low liquidity and poor asset value of workers. High wages have encouraged outward migration, while high unemployment has tended to discourage it. Rising income thus increases, rather than decreases, labor outflow. Their estimates sug- gest that up to a third of Russian regions could be locked in poverty traps. Conclusions and Policy Recommendations The analysis up to this point has not addressed this question: what are the implications for policy? Trends described in this chapter are fully exploited in the companion study on labor markets in the Region (see box 3.7). The study also focuses on policy recommendations, varying by country groups, providing a comprehensive analysis of policy actions needed to spur job creation. The discussion here is limited to the linkages between the labor market and poverty. This discussion begins by considering three major concerns regard- ing poverty and the labor market highlighted thus far: · Wage increases have outpaced productivity growth. Wage gains in many countries have outpaced productivity improvements, squeezing out profit margins (figure 3.13). These developments underpin jobless growth and constrain further scope for employ- ment generation. To sustain poverty reduction and to permanently improve the welfare of working families, wage increases need to be driven by real improvements in productivity. · There is stubborn unemployment in the EU-8 and SEE. Employ- ment opportunities have been stagnant or declining, thus leading to a missed opportunity to reduce poverty further. In these coun- tries, transfer systems have been instrumental to smooth social costs of transition to the market. They have cushioned the poor, but by their design created perverse incentives and discouraged employment. · There is a persistence of low-productive employment in the low and middle income CIS countries. In many poor countries, the expansion of employment reflects an increase in low-productive The Role of Labor Markets and Safety Nets 141 BOX 3.7 Labor Market Study DiscussesWays to Enhance Job Opportunities in the Region The forthcoming labor market study for the Region, Enhancing Job Opportunities in Transition Countries: Eastern Europe and the Former Soviet Union, finds that the economies' growth in the 1990s in the Region had not resulted in an equivalent improvement in employment: only Slove- nia had (barely) exceeded the employment rate of the early 1990s.The study diagnoses the key causes of the Region's disappointing labor market outcomes.The dominant role has been played by the demand-side factors, while the supply-side effects were relatively limited. The study claims that the crux of labor market problems in the Region is insufficient rates of job creation. Thus, widespread defensive restructuring is an important factor behind relatively low rates of job creation, despite often significant economic growth. Another important reason is the high cost of doing business. Investment climate constraints to job creation range from the high risks as- sociated with operating a business in low income CIS countries (for example, policy unpre- dictability, insecure property rights, weak contract enforcement, and unreliable infrastructure) to the considerable administrative barriers in middle income CIS countries (for example, numerous permits, inefficient regulations, and red tape) to the high monetary costs in the EU-8 and SEE countries (for example, high taxation). The study examines the set of policies that are still needed to create more and better jobs in the Region's countries. It argues that improving labor market outcomes in the Region as a first pri- ority requires removing key constraints to firm entry and growth--that is, reducing the costs of doing business. In most countries of the Region, higher investment rates are necessary to ac- celerate economic growth and job creation. A second priority is to enhance labor market adapt- ability through less-stringent employment-protection regulations and decentralized bargaining between employers and workers. Active labor market programs, if properly designed and im- plemented, can also contribute by providing greater incentives for the unemployed to go back to work and for employers to employ disadvantaged individuals. The specific mix of policies that each country needs to adopt varies and depends upon its par- ticular economic and labor market situation, and thus priorities for reform are country-specific.To spur job creation, measures to improve the business climate can be identified by comparing the regulatory and institutional settings in the Region's countries with those of other economies with a vibrant private sector. As to improving labor market adaptability, countries where em- ployment protection legislation is stringent, but enforcement capacity is weak (mainly in the CIS and SEE), need to simultaneously liberalize the law and promote compliance so that core work- er rights are protected. Countries where the enforcement capacity is strong and legislation rela- tively flexible (mainly in the EU-8) need to focus on addressing specific constraints to labor mar- ket flexibility. For example, countries where there are few regulations on temporary employment, but regular employment is highly protected, need to reduce protection granted to regular employees to avoid creating labor market duality. Source: World Bank Forthcoming-a. 142 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union agricultural employment. In richer countries, a stagnant pool of workers, often characterized by poor endowments, remains employed in low-productive occupations. This casts doubt on whether poverty reduction will be sustained over time. Employ- ment, particularly when associated with low productivity, does not offer protection from poverty. What are some of the policy actions that would help address these concerns? Raising productivity and employment. With increasing productivity levels and growing wages, all or most of those who are already employed will be out of poverty. For those who are not employed, success in reducing poverty will critically depend on the ability to expand employment because small or negative overall employment change makes the growth process less favorable to the unemployed or new entrants. In the Region, the main job flows consist of mobility between jobs (World Bank forthcoming-a). Thus the "extra" pull of labor demand is needed to bring nonworking poor to the labor market. Whether the issue is to generate more employment or to raise pro- ductivity, it would seem that countries need to focus on improving the climate for investment, which will reduce obstacles for job cre- ation. World Bank (forthcoming-a) provides a detailed set of recom- mendations tailored to the Region's subregions, spanning a wide range of policy action and instruments. For the CIS countries in par- ticular, but also the SEE countries, balancing discipline over the old enterprises and encouragement of new entries is critical. This requires a policy environment that disciplines low-productivity old enterprises into releasing resources and encourages high-productivity new enter- prises to absorb those resources and to undertake new investment. There are also the well-known elements that create a better invest- ment climate. In low income CIS economies, special attention is needed to ensure productivity gains in agriculture (box 3.8). Resisting noncompetitive pressures to boost wages, for example, through unsustainable increases of minimum wages, will also be important. This is more of an issue in the EU-8, where minimum wages represent a high proportion of the average wage, around 40 or 50 per- cent. This binding minimum wage constrains wage flexibility at the bottom end of the distribution. In Poland, for example, a binding min- imum wage appears to be linked to the experience of downward wage rigidity (World Bank 2004h). In recent years, the real value of the min- imum wage has been somewhat eroded, but it remains a barrier to wage flexibility. Similarly, a relatively high minimum wage is found to The Role of Labor Markets and Safety Nets 143 FIGURE 3.13 Wage Increases Outstripped Productivity Gains during the Economic Recovery in the Region Real Wages and Value Added per Worker in Manufacturing in Selected Countries, 1997­2003; Indexes: 1997 = 1.00 Poland Romania 1.80 1.80 1.60 1.60 1.00 1.40 1.00 1.40 = = 1997 1.20 1997 1.20 1.00 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 2003 Russian Federation Moldova 1.80 1.80 1.60 1.60 1.00 1.40 1.00 1.40 = = 1997 1.20 1997 1.20 1.00 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 2003 Vietnam Turkey 1.80 1.80 1.60 1.60 1.00 1.40 1.00 1.40 = = 19971.20 19971.20 1.00 1.00 0.80 0.80 1997 1998 1999 2000 2001 2002 2003 1997 1998 1999 2000 2001 2002 Real gross wage Value added per worker in constant prices Sources: ILO, Key Indicators of the Labour Market (KILM) and WDI 2005; LABORSTA (ILO), and General Statistics Office of Vietnam. Note: Constant international 2000 US$ used (WDI 2005) to deflate current local currency units figures. For Vietnam, data on wages includes only state sector em- ployees. constrain employment opportunities for low-skilled workers in Lithua- nia (World Bank 2002h). In Estonia and Hungary, experience with minimum wage increases in recent years indicates such hikes reduce the employment prospects of selected groups of workers, in particular 144 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 3.8 Raising Agricultural Productivity Agricultural growth is crucial for poverty reduction, in particular in the poorest countries, where agriculture remains a major source of income and employment. Several policy actions can raise the productivity of agriculture.These include promoting land reforms where they are lagging (for instance, in middle income CIS) and improving land markets (for instance, in several SEEs) to fa- cilitate land restructuring. Improving the investment climate in general, and in rural areas specif- ically, is very important. Investments in food processing, agribusiness, trade, and retail compa- nies play a crucial role in helping small farmers to overcome input and output market imperfections, to upgrade the quality of their products, and to access markets.Integration of the rural poor in the labor markets (either through rural off-farm employment generation or by im- proving access to urban labor markets) will be crucial for sustained income growth, in particular in middle-income countries. Further increasing rural-urban mobility might help compensate for the human capital disadvantage of rural areas (for example, through private transfers). The inte- gration of rural credit markets is crucial for investments and productivity growth in rural activities, including agriculture. Source: Macours and Swinnen 2004. those directly affected by the increase, those in small firms, and those earning low wages (Hinnosaar and Rőőm 2003; Kertesi and Köllő 2003). In countries where the minimum wage does not represent a large proportion of the median or average wage, it may still be relatively high, particularly in depressed regions with large pools of unskilled workers. This is the case in Poland (World Bank 2004h). In Hungary, large disemployment effects of the minimum wage have been docu- mented, for the low skilled and those in depressed regions. However, in other countries such as the Slovak Republic, the minimum wage is low and does not appear to be binding even in poorer areas (World Bank 2002g). Reshaping social protection to aid the restructuring of the economy and employment growth will also be important. Countries need to maintain the momentum in the ongoing social insurance and social assistance reforms designed to improve sustainability and enhance coverage and targeting of the poor. In low-income countries, the main constraint will continue to be the fiscal means to cover the population adequately. Better alignment between public and private resources (which could start with improving the understanding of, and collect- The Role of Labor Markets and Safety Nets 145 ing better data on, private transfers) is needed to raise the efficiency of public funds. Although more fiscal space for social protection exists in the middle income CIS countries, there may also be greater resistance to reforms, as suggested by the difficulties with the monetization of privileges in Russia. Although the objective of the reforms is not in question, the difficulties in implementation are a useful reminder as to the importance of sequencing with other social and economic reforms, the need to protect the most vulnerable groups, and an appropriate communications strategy to explain the benefits of reforms. In coun- tries in SEE and the low-poverty countries of the EU-8, which have the most extensive social protection, a balance will need to be struck between the need for protection and labor market incentives. Endnotes 1. Dennis and Guio 2003. See also Atkinson and others (2002) and the Web site of the Directorate General: Employment and Social Affairs, of the European Commission, www.europa.eu.int, for details. 2. "Poverty wage" is defined as the level of wage in 2000 US$ PPP sufficient to keep a worker and 1.5 dependents (the typical ratio of workers to dependents in the Region) above the poverty line of $2.15 a day (if it is fully spent on consumption and the household does not have any other sources of income). 3. This is the net change; gross changes required to bring about such sizes of net changes are much higher. 4. Between 1999 and 2002 in Russia, the Gini coefficient for wages fell from 0.47 to 0.42; in Tajikistan, from 0.55 to 0.53. In contrast, in Poland, the Gini index for wages rose from 0.30 to 0.32; in Moldova, from 0.44 to 0.45. (Data on individual wages are from the Regional Household archive and represent all wage earners reporting positive wages in the reference period of the survey.) 5. For example, see Fleisher (2005); also Kertesi and Köllő (2001) for Hun- gary; Vodopivec (2004) for Slovenia; and Newell and Socha (2005) for Poland. 6. Countries included in the study of gender wage gap are Albania, Bosnia and Herzegovina, Bulgaria, Serbia, Poland, Tajikistan, and Uzbekistan. 7. Mincerian earning functions are based on the seminal research by Jacob Mincer (1974), which estimated the rate of return to education using log of earnings as the dependent variable and education (as well as experi- ence and experience squared) as independent variables. CHAPTER 4 Affordable Access to Quality Services The Region is characterized by a legacy of relatively high levels of human development achievement. Millennium Development Goal (MDG) indicator targets such as enrollment rates, infant or maternal mortality, and access to piped water show that most countries of the Region are generally better-off than other countries at equivalent levels of income. Although there was some worsening of trends in the early years of the transition, infant mortality rates (shown in figure 4.1), as well as mater- nal and child mortality, have been declining in recent years.1 The early transition years of overall economic decline and reduced fiscal resources were paralleled by an increase in private resources as households began to contribute toward the cost of services, not only in the social sectors but also in basic infrastructure. Because the delivery networks were not adjusted to reflect the different level and composi- tion of demand, public resources were spread across a large number of providers, reducing the effectiveness of services. This mismatch between available resources and the funds needed to maintain the existing networks was reflected in increasing wage arrears for staff in health and education and the inability to repair basic infrastructure such as gas and electricity. These shortages played a significant role in the difficult early years of transition and may have been partially solved as economic recovery provided additional, although limited, funds. Although quantitative evidence is hard to find, it is reasonable 147 148 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4.1 MDGs in the Region: Infant Mortality and TB Incidence Infant Mortality Rate, 1990­2002 TB Incidence 70 200 births 60 live 150 50 1,000 40 people 100 per 30 100,000 20 50 deaths per 10 Infant 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 EU-8 SEE Middle income CIS Low income CIS Turkey Sources: WHO-Health For All database (the Region and Turkey); PAHO (Colombia); and WDI (Vietnam). Note: Infant mortality is defined as yearly rate of deaths in children less than one year old. Missing data points have been interpolated. Country groups for this fig- ure : EU-8: Hungary, Poland, Estonia, and Latvia; SEE: Romania and Bulgaria; Middle income CIS: Belarus, Kazakhstan, Russia, and Ukraine; Low income CIS: Arme- nia, Georgia, the Kyrgyz Republic, Moldova, and Uzbekistan. to assume that the inability to maintain and renovate the service deliv- ery networks has affected the quality of services provided. Other dimensions of well-being, such as life expectancy, quality of education, or the incidence of communicable diseases such as TB or HIV/AIDS, reflect some of the outstanding challenges in service deliv- ery. For example, TB incidence in selected countries has been increas- ing. In Russia, it increased from 80 cases (per 100,000 population) in the mid-1990s to 113 cases in 2003, much higher than the average for the European Union (about 83). The HIV/AIDS epidemic in the Region shows alarming indicators. In Central Asia, the number of reported cases increased from 500 in 2000 to more than 12,000 in 2004 (Godinho and others 2005). In Ukraine, the country with the fastest-growing HIV epidemic, more than 12,000 new HIV-infected individuals were reported in 2004 (including 2,300 children), totaling more than 134,000 HIV-infected individuals (Lekhan, Rudiy, and Nolte 2004).2 Although other factors such as lifestyle or educational attainment are also important, the emergence of these new risks also reflects the mismatch between the level and quality of services (including information) and the actual needs of the population. These mixed social and human indicators reflect persistent chal- lenges in the processes underlying the outcomes in health, education, and other living conditions. Good health status, educational achieve- ment, or living conditions jointly reflect the nature of the public inter- ventions in these sectors and the ability of households to invest in Affordable Access to Quality Services 149 human and physical capital. Household characteristics are at the cen- ter of this process, but household inputs need to be combined with an effective network of services to produce desired outcomes. In this context, poverty in social services is understood as the deprivation from such services or, to be more precise, deprivation from affordable access to quality services. This chapter discusses three closely interrelated dimensions of poverty in service delivery: access and utilization, quality, and afford- ability. A household may be deemed poor if children do not have access to a school or if--even in the presence of a school--children cannot attend because of reasons beyond the household's control (lack of income to pay for fees, ethnic or language discrimination, and so forth). For some social services in the Region, such as education, these two dimensions (access and utilization) are almost the same because enrollment rates are very high. In other services like health care, access and utilization are different. Although access reflects the existence of a provider within the reach of a household, utilization captures the need to seek care and the ability to pay for those services. These dimensions of well-being must be read together; for exam- ple, the decline in quality or the increased cost of some services are associated with reduced utilization of these services. The main argu- ment in this chapter is that poverty as lack of access to services is per- haps not a major issue because of the inherited legacy of broad network coverage; however, the inability to actually use these ser- vices has gained relevance in recent times. In addition, the delivery network has persistent weaknesses that are reflected in the declining quality of services. Finally, the affordability dimension has acquired importance as formerly subsidized public services have come to rely increasingly on households' contributions. These arguments are illus- trated by using three sectors: education, health, and utilities. Education This section reviews access to education and its quality. It starts with primary education, but primarily focuses on the secondary level, where problems in the Region are most pressing. Coverage of Education Most countries in the Region inherited a wide network of education services that enabled them to achieve almost universal coverage in compulsory education. Although fiscal resources and enrollments 150 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union came under some pressure during the 1990s, particularly in the low income CIS group, enrollments continued to be high relative to levels of income. Primary Education During the period since 1998, most countries of the Region maintained, and some even improved, enrollment in primary school: enrollment at the primary level is more than 90 percent in all the countries (figure 4.2). However, not all low income CIS countries saw improvements-- Georgia, the Kyrgyz Republic, and Tajikistan experienced some reduc- tions in the proportion of children ages 7 to 14 years enrolled in school. On the other hand, in Moldova and Uzbekistan, the coverage of pri- mary-school-age children was increased. Most of the Region's countries are characterized by relatively equal coverage of education across income quintiles, and in many countries differences were further reduced. Figure 4.3 displays the ratio of enrollment rates between the richest and the poorest quintiles (or, the income gradient in primary education coverage). Bars close to the shaded area (= 1) indicate that the top and bottom quintiles have sim- ilar levels of coverage. Although children in better-off households FIGURE 4.2 Regional Coverage of Education, Ages 7­14 100 98 96 school in 94 enrolled % 92 90 88 Poland Hungary Bulgaria Romania Russian Kazakh- Kyrgyz Georgia Moldova Armenia Tajikistan Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998 2002­3 Source: World Bank staff estimates using data from ECA Household Surveys Archive. Affordable Access to Quality Services 151 FIGURE 4.3 Inequality in Access to Primary Education in the Region, 1998­2002 1.25 quintile 1.20 1.15 poorest to 1.10 1.05 richest of 1.00 Ratio 0.95 Poland Hungary Bulgaria Romania Russian Kazakh- Kyrgyz Georgia Moldova Armenia Tajikistan Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998 2002 Source: World Bank staff estimates using data from ECA Household Surveys Archive. have slightly better coverage than those in the poorest quintile (except in the Kyrgyz Republic and Russia), these differences are gen- erally not greater than 5 percent (Bulgaria stands out as a country with a relatively large gap of about 10 percent in 2002). In fact, coun- tries with the steepest income gradients (enrollment gaps of more then 5 percent) also showed the largest declines in the ratio between top and bottom because of improving coverage for the bottom quin- tile between 1998 and 2002. However, there are exceptions to this declining trend. For exam- ple, in Moldova, enrollment rates were very similar across the income distribution in 1999, but by 2003, enrollment rates in the richest quintile were more than 5 percent higher than those in the poorest quintile. This is partly because the post-1999 recovery benefited those in urban areas and the better-off, increasing socioeconomic differ- ences in education. Equally, enrollment rates in urban areas fell more significantly during the crisis (in fact, enrollment in rural areas were temporarily higher than in urban areas in 1998/99) because urban areas and the better-off suffered the largest consumption losses (Sig- noret and Murrugarra 2003). Enrollment in urban areas thus appears particularly vulnerable to crises. Secondary Education For children ages 15 to 17 years, the Region experienced a dramatic increase in coverage, reaching more than 85 percent. The improve- 152 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ments were evident for all countries, even for those with relatively low coverage, such as Moldova or Uzbekistan (figure 4.4). The over- all improvements in secondary education coverage may reflect the more attractive wages for better-educated individuals in both local and external labor markets (chapter 3). The overall improvement in coverage was paralleled by a reduc- tion in the large enrollment gaps across income groups (figure 4.5). The ratio of coverage rates for children in the richest and the poorest income quintiles is larger than at the primary level. For SEE countries like Bulgaria and Romania, coverage of the children in the top income quintile is more than 50 percent higher than that of the poorest quin- tile, compared with only 10 to 20 percent at the primary level. These differences, however, were reduced in most countries outside the low income CIS group, where gradients increased (except for the Kyrgyz Republic). The reduction of enrollment gaps across income groups is also observed across gender and geographic dimensions. At the secondary education level, there are small gender differences in enrollment, and these appear to be continuing to shrink (figure 4.6). In countries like Armenia and Moldova, where formal labor market opportunities are FIGURE 4.4 Regional Coverage of Education, Ages 15­17 100 95 90 85 school in 80 75 enrolled 70 % 65 60 55 50 Poland Hungary Bulgaria Romania Russian Kazakh- Kyrgyz Georgia Moldova Armenia Tajikistan Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998­9 2002­3 Source: World Bank staff estimates using data from ECA Household Surveys Archive. Affordable Access to Quality Services 153 FIGURE 4.5 Inequality in Access to Secondary Education, 1998­2002 2.0 1.9 quintile 1.8 1.7 1.6 poorest 1.5 to 1.4 1.3 richest 1.2 of 1.1 1.0 Ratio 0.9 Hungary Poland Romania Bulgaria Russian Kazakh- Armenia Georgia Tajikistan Kyrgyz Moldova Colombia Turkey Vietnam Fed. stan Rep. EU-8 SEE Middle Low income CIS Benchmarks income CIS 1998­9 2002­3 Source: World Bank staff estimates using data from ECA Household Surveys Archive. FIGURE 4.6 Gender Inequality in Access to Secondary Education, 1998­2002 1.3 1.2 males to 1.1 1.0 females 0.9 of 0.8 Ratio 0.7 Hungary Poland Romania Bulgaria Russian Kazakh- Armenia Georgia Tajikistan Kyrgyz Moldova Fed. stan Rep. EU-8 SEE Middle Low income CIS income CIS First year Last year Source: World Bank staff estimates using data from ECA Household Surveys Archive. limited and migration is common (especially among young men), young women tend to stay longer in school. These differences, how- ever, were reduced as enrollment rates among young men increased. Tajikistan stands out as the country with the largest gender gap at the secondary level: enrollment for girls is three-quarters that of boys, despite recent improvements. Urban-rural inequalities in access to education have also declined in most countries, including Tajikistan, the only country where coverage is higher in rural areas. 154 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Other sources of deprivation in access to education services include those related to ethnicity and language. Although ethnic dimensions do not represent a major preoccupation in many countries in the Region, the increasing availability and transparency of information has shed light on the deprivation of certain groups. One of them is the Roma population in Central and Eastern Europe. Across countries, 70 to 80 percent of Roma populations have less than primary education, while very few have completed both primary and secondary educa- tion. Moreover, most Roma children are enrolled in remedial "special schools," which are physically separate from other schools: between 75 and 85 percent of Roma children in the Czech Republic, Montene- gro, and the Slovak Republic and between 60 to 70 percent in FYR Macedonia and Serbia are enrolled in special schools. The combined effect of poverty, isolation, and education in a nonmaternal language only underscores the outstanding challenge of providing quality ser- vices to groups that face other sources of exclusion and vulnerability. In sum, the evidence during the past years shows a varied range of outcomes in poverty of education services. The biggest concern is in the low income CIS countries, where although some countries have made significant progress, others are still lagging behind. Coverage of secondary education, on the other hand, seems to have increased everywhere, even in the poorest countries. Still, public intervention to improve access to education services seems to be facing challenges in reaching certain minority groups. Quality of Education The deterioration of infrastructure during the early transition years had important impacts on the delivery of social services such as health and education.3 Schools came to be without adequate heating during the winter, and other services like electricity and water were irregu- lar. This section provides some evidence on the evolution of quality of inputs and its impact on education performance. Although education outcomes between 1995 and 2003 (such as achievement tests) are presented, linkages with specific inputs are discussed only for the period between 1995 and 1999. Detailed school surveys that included teachers' and students' characteristics, as well as data on mathemat- ics and science tests, are available only until 1999, precluding analy- sis for the later period.4 Test Scores as a (Partial) Reflection of Quality Assessing the quality of education requires examining the different elements in the education process, such as schools, teachers, and Affordable Access to Quality Services 155 (obviously) households. A positive interaction of these different ele- ments could result in students with higher educational achievements in the short run and increased productivity (and employment) in the long run. Countries of the Region have been characterized by very good per- formance levels in international tests. The results in mathematics from the Trends in International Mathematics and Science Study (TIMSS) show that the selected countries of the Region were per- forming at about the same level as some OECD countries like England and the United States (table 4.1). Hungary, Latvia, Lithuania, the Slo- vak Republic, and Russia have average scores better than England's. Despite the good performance (with a few exceptions), the Region's countries show worrying trends of declining performance over time. Although some countries performed well and may have even improved over time (such as Hungary or Lithuania), others have shown major losses in average scores (figure 4.7). Russia maintained its performance between 1995 and 1999, but by 2003, significant declines in average scores were observed. EU-8 countries (except for Estonia and Poland, not covered by TIMSS), have maintained a stable performance; but this hides important country heterogeneity. While the Baltic States and Hungary have maintained or slightly improved their performance, the Czech Republic, the Slovak Republic, and Slovenia experienced dramatic declines that are similar to those observed by some SEE countries. TABLE 4.1 Mathematics Performance, 1995­2003 (TIMSS mean scores for eighth grade students, ranked by 2003 score) 1995 1999 2003 Hungary 527 532 529 Slovak Rep. 534 534 508 Russian Fed. 524 526 508 Latvia 488 505 505 United States 492 502 504 Lithuania 472 482 502 United Kingdom 498 496 498 Slovenia 531 530 493 Armenia -- -- 478 Czech Rep. 546 520 -- Bulgaria 527 511 476 Romania 474 472 475 Moldova -- 469 460 Macedonia, FYR -- 447 435 Source: TIMSS Web site (www.timss.org). -- = did not participate. 156 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4.7 Recent Declining Trends in Regional Mathematics Performance (TIMSS) 540 520 score 500 test 480 Average 460 440 1995 1999 2003 Hungary (EU-8) Bulgaria (SEE) Russian Fed. (middle income CIS) Moldova (low income CIS) United States Source: World Bank staff estimates based on TIMSS data. The changes in average scores were uniformly observed across the achievement distribution. A drop in average scores for these high-per- forming countries may not have represented a major weakness in their education systems if they could have kept most of the students per- forming above the minimum required standards. The evidence from TIMSS, however, indicates that there are major increases in the fraction of students below minimum educational skills for their age.5 Between 1995 and 1999 (not shown here), two of the very best performers, the Czech Republic and Slovenia, multiplied their fraction of underper- forming students between two and three times to reach between 3 and 4 percent of their student populations. FYR Macedonia, Moldova, and Romania have more than one-fifth of their eighth graders not reaching minimum levels in mathematics by 2003. Between 1999 and 2003, Bulgaria almost doubled its fraction of underperformers, increasing from 10 to 18 percent, and the Slovak Republic increased from 4 to 10 percent. The only exceptions were Latvia and Lithuania, where the fraction of underperformers was cut by 40 to 50 percent, reflecting major gains in their average scores (figure 4.8). What Was the Role of Quality of School Inputs in the Changes in Performance? To assess the role of inputs and their quality in performance, four country case studies are used to examine schools, teachers, and house- hold characteristics. These countries are Latvia (with major progress), Affordable Access to Quality Services 157 FIGURE 4.8 Mathematics Performance in Selected Countries of the Region, 1999­2003 8 6 Lithuania 400 4 USA children than Hungary 2 England improvement in Slovenia Latvia less 0 no ­8 ­6 ­4 ­2 0 Romania 2 4 6 change Russian ­2 score Fed. ­4 reduction % ­6 who Cyprus Slovak Rep. ­8 Bulgaria deterioration ­10 deterioration no improvement change % change in average mathematics score Source: World Bank staff estimates based on TIMSS data. the Czech Republic (good performer, but losing significant ground between 1995 and 1999), Romania (stable at a low level), and Russia (good performer and stable). Although that analysis uses data between 1995 and 1999, the evidence described before suggests that the deteri- oration of quality is an ongoing phenomenon after 1999. Functioning of schools: infrastructure and governance. The limited fiscal resources and the large network of providers crowded out expendi- tures on maintenance of basic infrastructure. Between 1995 and 1999, the fraction of students in schools facing energy shortages increased. Even in Latvia, a country with major progress in education achievement, this proportion increased from 55 to 65 percent, and in Russia, this fraction jumped from 50 to 63 percent. The deterioration in infrastructure was also observed in other physical inputs: in Roma- nia, the fraction of students in schools lacking some instructional materials increased from 40 to 54 percent between 1995 and 1999. The transition process also brought other changes that affected the delivery of education. One important change was the decentraliza- tion process, through which most countries transferred some respon- sibilities for the provision of education services to lower levels of government.6 In all countries in figure 4.9, except for Croatia, the Slovak Republic, and Slovenia, subnational governments hold more than 50 percent of all public education sector responsibilities. The highest subnational government involvement in the sector is in Azer- 158 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union baijan, Belarus, and Kazakhstan. This process of decentralization in the delivery of education services took many different forms. In Alba- nia, Armenia, the Czech Republic, and Romania, basic education ser- vices were provided through deconcentrated regional branches of the central government, with varying (but mostly minor) roles for local governments. A different approach was followed in Hungary, Latvia, and Poland, where subnational governments bear most of the respon- sibilities for the provision of education services.7 The evidence on school autonomy obtained from TIMSS surveys corroborates this finding. The role of parents in setting school policy increased across all countries between 1995 and 1999. In some coun- tries, there is an increase in the role of the school in certain aspects of the service delivery. In the Czech Republic, the fraction of students in schools responsible for teachers' salaries increased from 58 to 80 per- cent in this period. In Romania, the share of students attending schools fully responsible for teachers' salaries increased from 7 to 16 percent, and those enrolled in schools with key roles in hiring teachers from 17 to 27 percent. Most schools in the Region were responsible for pur- chasing supplies in 1995, and this role increased even more by 1999. FIGURE 4.9 Role of Subnational Governments in Education, 1995­2002 100 90 80 70 total education in 60 on 50 shares 40 spending 30 Subnational 20 government 10 0 Rep. Rep. Fed. Rep. Latvia Poland Estonia Hungary Slovenia Albania Bulgaria Croatia Belarus Ukraine Georgia Czech Lithuania Romania Moldova Slovak Kazakhstan Azerbaijan Tajikistan Russian Kyrgyz EU-8 SEE Middle income CIS Low income CIS 1995 2001­2 Source: Zeikate 2004. Affordable Access to Quality Services 159 An aging teaching force. Lack of incentives (low salaries and the persist- ent arrears of the 1990s) led to an aging of the teaching force (figure 4.10). The fraction of eighth grade children with teachers more than 50 years of age almost doubled between 1995 and 2003, except for Lithuania. In the Czech Republic and Romania, the fraction of stu- dents with teachers more than 50 years of age increased from 30 to 40 percent by 1999, and in Romania it had risen to close to 50 percent by 2003. In Russia, the increase was from 21 to 41 percent between 1995 and 2003. Although the aging of the teaching force does not necessar- ily indicate a worsening of quality of education, the lack of funding for training (and retraining) in most countries suggests that teachers have not been adequately equipped with new pedagogical tools and, hence, students may not have benefited from new education approaches. Even in Latvia, a country with a strong education reform, this aging indicator, although constant at about 24 percent during 1995­99, increased to 33 percent by 2003. Besides the negative effect of a stag- nating teaching force, the lack of incentives for teachers to improve their skills and worsening school conditions may have affected the school environment and teaching practices in the Region. FIGURE 4.10 Aging Teaching Force in the Region, 1995­2003 50 old 45 years 40 50 35 over 30 25 teachers 20 with 15 10 children of 5 % 0 Hungary Latvia Lithuania Slovak Slovenia Bulgaria Romania Russian Rep. Fed. EU-8 SEE Middle income CIS 1995 1999 2003 Source: TIMSS 2005. 160 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Increased household heterogeneity. Quality can also be measured as the match between attributes of the students and those of the education services, such as language of instruction or examination. The increased presence of minorities and the inability to provide adequate educational programs to such groups may be interpreted as a quality deficit and can be reflected in educational achievement. This feature may have worsened since the Region witnessed a slow (but steady) increase of minorities in several countries. In fact, the fraction of stu- dents taking a math test in a language that they "speak only some- times at home" rose from 2 percent to more than 5 percent in Russia, and in Latvia this rose even faster, from less than 2 to more than 6 percent. Despite the emerging heterogeneity among students, chil- dren were less likely to skip classes in 1999 compared with 1995. In Romania, the fraction of students that skipped one class or more decreased from 66 to 50 percent. Similar improvements in attendance were observed in Latvia and Russia. Lessons on Performance and Quality Using microdata for the four countries' cases, changes in achievement scores between 1995 and 1999 were decomposed into factors related to student, teacher, school, and household background characteristics. The decomposition exercise was applied at different points of the achievement distribution to assess the role of these factors for explain- ing the relative share of students with lower and higher achievement.8 The deterioration of school characteristics played a negative role in education performance, but those effects were offset by house- hold attributes and, in some cases, by specific policies. Russia and Latvia constitute the examples of systems with major worsening of observable measures of school quality (such as school heating), which had a strong negative effect on scores. In these countries, however, the deterioration of school infrastructure was offset by household factors such as parents' education, and these compensat- ing effects were particularly important among those with lower scores. In Russia, teachers played a complementary role in compen- sating for the deterioration of schools. In Latvia, however, systemic factors such as the education reform and decentralization of services compensated for the negative school conditions. In other countries, the effects of the worsening education infrastructure were accompa- nied by detrimental effects of household conditions or teaching force. The drop in performance in the Czech Republic was driven by both school and household attributes, and these linkages were stronger among those with low scores. In Romania, teachers' attrib- utes also had negative effects, particularly for the low-scoring stu- Affordable Access to Quality Services 161 dents. These patterns show an even worse picture if specific popula- tion groups are addressed. Given the financial constraints from central governments, school managements are playing an increasing role in delivering education. Other governance factors have played a significant role in keeping up educational outcomes. Countries increasingly relied on local govern- ments for service delivery. The fraction of students in schools where hiring and firing, school budget, and salary decisions are made at the school level increased during the period. Not only did the role of prin- cipals and school governing bodies increase but they also had a positive impact on performance. In Russia, students had higher scores when they attended schools where the principal or the school governing body played an important role in staffing, wage setting, and formulating school budgets. In Latvia, the effects are not that marked; because of its broader education decentralization reform, it showed little variation in school responsibilities, compared with the Russian case. In sum, between 1995 and 2003, performance outcomes of education have remained relatively high in transition countries. These outcomes, which partly reflect the quality of education services, have been main- tained because of the skilled stock of human capital in both households and teaching force and, in some cases, by policy interventions such as an effective decentralization in service delivery. Although the decomposi- tion analysis uses data for 1995 and 1999, the continuous declines in per- formance across the Region suggest that while the Region may be able to live off its previous investments, these are eroding rapidly. The failure to maintain human and physical capital is resulting in environments that are not appropriate for effective education services and that are reflected in declining performance. This is particularly observable in rural areas and among poor households, which typically face the worst conditions. Policy interventions that improve the quality of education services are essential if the decline is to be brought to an end. Access to, and Affordability of, Health Services The inability to recover from adverse health shocks and maintain one's human capital is an important dimension of poverty (Narayan and others 2000). Utilization of health services is, in this sense, an interest- ing object because it reflects several linkages with poverty. First, it reflects the effects of poverty on human capital because impoverished households are more likely to face malnutrition, are more exposed to certain contagious diseases, or are less equipped to identify certain chronic diseases. Second, it also reflects the differential ability of 162 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union households to seek health care when needed. Even in health systems with entirely subsidized services, households have to incur other costs to receive treatment, such as transportation or the opportunity cost of waiting. Differences across households in financial capacity or the existence of institutional thresholds, such as ethnic barriers, define other forms of poverty in access to services. All of these forms of dep- rivation are evident in the Region. Although the Region's countries have a large public network of health providers who distribute generous services, they suffered major fiscal restrictions during the early 1990s. Between 1994 and 1999, the Region spent on average only 4 percent of GDP on health care, ranging from 9 percent in Croatia to close to 1 percent in Geor- gia. After 1999, some of the poorest countries (like Azerbaijan, Moldova, Turkmenistan, and Uzbekistan) continued to experience reductions in public expenditures on health, so much so that public health spending fell below 3 percent of GDP. Other poor countries like Armenia and Bosnia and Herzegovina have managed to stem the decline, but at very low levels of spending (around 2 percent of GDP). See annex table 1 for country-level data. The overall decline in public spending is paralleled by three fea- tures that have implications for health and poverty linkages. First, the very large network of providers has not been significantly adjusted for lower fiscal resources. This has resulted in an underfunded and, hence, ineffective network of providers. Second, the lack of resources for public health activities has resulted in repeated episodes of com- municable diseases that are easy to prevent. Third, the changing age composition of the population, which is becoming older, has changed the morbidity profile and increased the costs of health provision. In this overall constrained environment, in which health services have not adjusted to changed circumstances or demand, it should come as no surprise if quality declines, particularly for the poorest groups. However, assessing trends in quality of services is very com- plex because it requires information about specific failures in the pro- vision of health care. In countries with well-established information systems, quality of care is measured as hospital mortality or infections for specific types of patient and morbidity (Geweke, Gowrisankaran, and Town 2003). Efforts to standardize the collection of information in OECD countries have just begun with the identification of core indicators in different types of treatment (Marshall and others 2004). In the Region's countries, although this information is seldom recorded, anecdotal evidence corroborates the poor quality of health services because of outdated protocols, lack of basic materials and drugs, and the need to retrain personnel (Davidow 1996). Affordable Access to Quality Services 163 Despite the inability to observe direct measures of quality of ser- vice, some morbidities can be partially attributed to the quality of health services. The number of cases and mortality from certain dis- eases can be associated with deficiencies in service delivery when the prevention, identification, and treatment of those diseases can effec- tively reduce deaths. One such indicator is the number of cases and deaths due to cervical cancer. Although a number of factors affect cer- vical cancer, an effective primary health care system should be able to educate populations at risk and identify the morbidities, and hospital services should be able to treat and lessen mortality. Figure 4.11 shows cervical cancer rates for females between 15 and 44 years for selected countries of the Region and for Germany and France for comparison.9 Although Germany has been able to significantly reduce its morbidity in the past 20 years, most of the Region's coun- tries show an increasing morbidity profile. Only Poland has a declin- ing trend, but it is still at a very high level. FIGURE 4.11 Cervical Cancer in the Region and Western Europe, 1970­2002 5 4 3 women 100,000 2 per 1 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Poland (EU-8) Bulgaria (SEE) Russian Fed. (middle income CIS) Moldova (low income CIS) France (benchmark) Germany (benchmark) Source: Parkin, Whelan, Ferlay, and Storm 2005. Note: Age-standardized rate, females 15­44. 164 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Anecdotal evidence on quality of care suggests that quality of health services deteriorated even faster in rural areas, and that poor households are more prone to communicable diseases such as TB because of poor living environment. It also suggests that the percep- tion and management of chronic diseases is more difficult among the less educated. What Happened to Health Care Utilization? The reduction in public resources in health care in the Region increased the use of fee-for-services in a mostly unregulated setting, reducing the demand for health care. Official statistics show the decline in utilization of health care during the 1990s, but after 1999 these remained stable or even recovered (see annex table 2). Inpa- tient care in the Caucasus countries (Armenia, Azerbaijan, and Geor- gia), Moldova, and Tajikistan declined more than 20 percent between the mid-1990s and 2000. This decline stopped after 2001 and in some cases (Armenia) even recovered. Other countries like Belarus, Russia, and Ukraine continue to have very high hospital utilization rates that are higher than the average in the EU (less than 19)--in Russia, they are still increasing. Thus, official utilization data suggest that the low income CIS group has converged to very low levels of utilization, while countries with higher incomes have not adjusted their oversupplied network and still show very high levels of utilization (figure 4.12). FIGURE 4.12 Hospital Utilization (Inpatient Care) 33 people 28 23 100,000 per 18 13 8 Admissions 3 1994­8 1999 2000 2001 2002 Lithuania (EU-8) Romania (SEE) Belarus (middle income CIS) Georgia (low income CIS) Source: World Bank staff estimates based on official health statistics (see annex table 2). Affordable Access to Quality Services 165 Evidence from household surveys. Household survey data provide income-related inequalities in utilization. "Utilization" in the house- hold survey is defined as the fraction of those sick individuals who sought health care, reflecting the notion of need in health care. This is different from official statistics that record the number of inpatient admissions and outpatient contacts at the point of service.10 Figure 4.13 shows the survey-based utilization rates for selected countries by quintile. Armenia stands out as a country with one of the lowest uti- lization rates in the Region, which declined further between 1999 and 2003. Figure 4.13 (based on survey data) confirms official statis- tics reported in annex table 3. Russia comes close to Armenia, and other countries like Bulgaria and Romania (not shown) experienced large increases in utilization. The changes in utilization rates have been uneven across socio- economic groups. Figure 4.12 shows that generally there has been an overall recovery in utilization. Figure 4.13 helps one to hypothesize that when such gains have been made, they went hand in hand with maintaining relatively low income-related inequality. In countries such as Bulgaria, not only did utilization increase for all income quin- tiles but also the increase for the poorest was the greatest. Armenia is a particularly interesting case because of the large utiliza- tion gap between the poorest and richest quintiles. Utilization decreased between 1998 and 2003, despite a recovery in utilization between 2001 and 2003 (in both official and survey data). The reduction in utilization during the 1990s significantly increased the income gradient, but after 2001 a modest decline occurred because of a recovery in utilization among the poor. Improved coverage of the poor in the post-2001 period is explained by a recovery in public expenditures in health and the expansion of health insurance for families in poverty. Still, Armenia has one of the largest gaps across income groups in the Region. Utilization of health services among the poor suggests that poor coun- tries faced the worst declines in utilization during the 1990s and that the poor fared the worst. This pattern seems to be partially reversed after 1999 through changes that touched two aspects of utilization inequality. First, the improved economic conditions may have enabled households to bear the costs of seeking care. Second, the introduction of policies to provide health insurance for the poor and better funding of such policies may have resulted in better--or more affordable--utilization. How Are Households Paying for Health Care? Out-of-Pocket Payments and Catastrophic Expenditures Health status and health care utilization are important manifestations of economic status, but adverse health episodes can also cause house- 166 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4.13 Utilization Rates of Health Services by Quintiles Armenia (low income CIS) Bulgaria (SEE) 95 95 use 85 use 85 75 75 who who services 65 services 65 55 55 care care persons persons 45 45 ill ill of health 35 of health 35 % 25 % 25 15 15 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average Russian Federation (middle income CIS) Uzbekistan (low income CIS) 95 95 85 85 use use 75 75 who who services 65 services 65 55 55 care care persons 45 persons 45 ill ill 35 35 of health of health % 25 % 25 15 15 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average Colombia (benchmark) Vietnam (benchmark) 95 95 use 85 use 85 75 75 who who services 65 services 65 55 55 care care persons 45 persons 45 ill ill of 35 35 health of health % 25 % 25 15 15 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average 1998, 1999 2002, 2003 Sources: World Bank staff estimates using data from ECA Household Surveys Archive. See also appendix table 7 for country-level data by years. Note: Utilization rate shows percentage of respondents who used health services when sick over the reporting period; quintiles are based on consumption per capita. holds to fall into poverty because of large expenditures on drugs or treatment, or forgone income that reduces consumption. Because public expenditures were reduced during the 1990s, private expendi- tures played a more important role in the financing of the sector. This was reflected in an increasing share of health expenditures in house- hold budgets across different households (figure 4.14). In Belarus, a Affordable Access to Quality Services 167 FIGURE 4.14 Ratio of Out-of-Pocket Health Spending to Household Total Consumption, 1998­2003 Hungary (EU-8) Romania (SEE) 20 20 15 15 consumption consumption 10 10 total total of of 5 5 % % 0 0 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average Belarus (middle income CIS) Georgia (low income CIS) 20 20 15 15 consumption consumption 10 10 total total of of 5 5 % % 0 0 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average Armenia (low income CIS) Colombia (benchmark) 20 20 15 15 consumption consumption 10 10 total total of 5 of 5 % % 0 0 Poorest 2nd 3rd 4th Richest National Poorest 2nd 3rd 4th Richest National Quintiles average Quintiles average 1998, 1999 2002, 2003 Source: World Bank staff estimates using data from ECA Household Surveys Archive. country with broad and generous coverage of publicly provided health services, the share of households' budgets devoted to health doubled between 1998 and 2002. In Hungary, the increase was more than 30 percent, reaching almost 5 percent of the household budgets. But nowhere in the Region was the increase so fast and the burden so high as in low income CIS countries Armenia and Georgia. 168 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union The increase in households' contributions to health financing had an impact on poorer households. Although health care expenditures reflect the nature of the illness, they also capture some dimension of quality of treatment. This results in better-off households with higher out-of-pocket expenditures compared with those of poorer ones. However, relative to their incomes, the poor spend a larger fraction, and this share of health expenditures on their incomes has been increasing over time. The increasing costs of health care pose serious affordability con- cerns in some countries, and they partially explain the decline in uti- lization in the past decade. Treatment may be postponed, but often only to increase the severity of the illness and the cost of treatment. Then, when treatment (and its expenditures) cannot be avoided, large health expenditures may reduce the resources available for non- health-related spending. This is referred to in the literature as the impoverishing effects of catastrophic health expenditures. How Impoverishing Are Catastrophic Health Expenditures? Catastrophic health expenditures are usually defined as those extreme expenses that affect households' ability to maintain their consump- tion of basic items (Wagstaff and van Doorslaer 2001). This is differ- ent from simply examining the incidence of total health expenditures among the poor and the nonpoor (discussed earlier) because the impoverishing concept involves those nonpoor but vulnerable house- holds that may fall below the poverty line because of these unusually large unpredictable expenses. These effects may be underestimated in a country with low utilization of health services because many of those not using the network simply cannot sacrifice any additional consumption and may postpone care. Hence, there are potential impoverishing effects that are not observed because households sim- ply decline or postpone care. Across countries, simulations undertaken for the purposes of this report suggest that catastrophic health expenditures can increase the fraction of poor population by between 3 and 9 percent.11 Countries with vastly different funding and organization of their health sectors (such as Belarus and Armenia) experience similar impacts (table 4.2). Belarus, on one hand, has a health system that has changed very little from the previous model under the Former Soviet Union: public expenditures on health represent about 5 percent of GDP and sustain a large network of facilities and personnel (close to 120 beds and 45 doctors per 10,000 people). Although most health status indicators show low infant and maternal mortality, adult life expectancy is Affordable Access to Quality Services 169 declining because of adult male mortality (life expectancy for males is 62, one of the lowest in the Region). Households spend a small frac- tion of their budget on health, still reflecting strong public funding. However, the impoverishing effect is high because most households that need health care do seek care, even if they pay a small amount. The poverty impact, then, occurs through the broad number of pop- ulation affected. Armenia, on the other hand, represents a different picture. Although spending only 3 percent of GDP, Armenia still has a large network of providers and personnel (more than 43 beds and more than 34 doctors per 10,000 people). The public underfunding of the system has been temporarily covered by active participation of inter- national donors and households in the system (World Bank 2003b). Expenditures on health represent about 5 percent of the average household budget, and about half of those payments are made infor- mally.12 The levels of utilization, however, are much lower than in Belarus, and those who pay contribute a significant fraction of their incomes. The impoverishing effect of catastrophic expenditure is high in this case because of high expenditures among the house- holds seeking care. Mechanisms to Protect the Poor Health systems need to include cost recovery mechanisms not only to generate their own resources but also to introduce some incentives for the rational use of the network among consumers. Fully subsi- TABLE 4.2 Impoverishing Effects of Catastrophic Health Expenditures Poverty Indicators before and after Catastrophic Health Expenditures Impoverishing Impoverishing effect, Before After effect % impact Bulgaria (SEE) 4.8 6.3 1.5 31.9% Romania (SEE) 9.4 10.1 0.7 7.6% Belarus (middle income CIS) 19.0 20.6 1.6 8.4% Kazakhstan (middle income CIS) 15.8 16.3 0.5 3.2% Armenia (low income CIS) 41.2 44.6 3.4 8.4% Georgia (low income CIS) 40.3 43.9 3.6 9.0% Kyrgyz Rep. (low income CIS) 61.0 62.4 1.5 2.4% Moldova (low income CIS) 32.5 35.4 2.9 8.8% Tajikistan (low income CIS) 64.4 67.7 3.3 5.1% Uzbekistan (low income CIS) 39.0 40.7 1.6 4.2% Source: World Bank staff estimates. Note: Data used from most recent available household survey; poverty line used is $2.15 at 2000 PPP. 170 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union dized systems may generate excessive use of health resources and hence inefficient expenditures on unnecessary activities. Health sys- tems that depend fully on private expenditures are challenged by the households' ability to contribute, particularly when those expenses are large. There are, then, two areas that require public interventions to address market incompleteness or equity concerns. One is the con- struction of a functioning health insurance system that can provide households with a mechanism to cope with adverse health events. The second and most important issue for the poor is the introduction of health programs that provide financial coverage for those who can- not afford the cost of health services. One country where consecutive reforms have enhanced the provi- sion of health insurance for the poor is Armenia. First, in 1999, the government reformed the social assistance system, replacing more than 26 categorical benefits by a poverty-targeted benefit, improving the incidence of social assistance among the poor. Then, in 2001, the Ministry of Health made those beneficiaries of social assistance eligi- ble to receive a basic package of services with no charge. Although health insurance is still limited and faced lack of funding in the initial years, the evidence in 2001 and 2003 suggests that utilization has recovered and has improved more for the poorest households. Figure 4.15 shows the increases in utilization in 2001 compared with the poorest quintile in 1999. The figure shows that compared with the rates observed in 1999 for the poorest, most of the population expe- rienced an increase in the utilization of health services by 2001. But the increase was particularly pronounced for the poor who benefited from targeted health insurance. The effects of this expansion of health insurance are also associated with a better funding of the health care network by 2001. The positive effects of the eligibility expansion on utilization, or take-up, are also evident in higher-income countries with programs for the poor, such as Medicaid in the United States (Shore-Sheppard 2005). During the recent years while public funding for health services has improved, these services have come to rely much more on private expenditures. The negative implications of this financing structure for equity were evident during the 1990s, but better funding and better programs targeted to the poor have made services more affordable. The experience of Armenia suggests that even countries with limited resources and high poverty rates can make improvements in afford- ability and access among the poor, albeit on a moderate scale; how- ever, very few poor countries in the Region have followed in Armenia's footsteps. Affordable Access to Quality Services 171 FIGURE 4.15 Health Insurance and Utilization in Armenia, 2001 points) 30 relative 2001 and (percentage 20 1999 1999 in between quintile 10 utilization poorest in the Gain to 0 Poorest 2nd 3rd 4th Richest Quintiles Health insurance No health insurance Source: World Bank staff estimates based on household surveys. Energy and Other Utility Services The Region's countries entered the 1990s well covered with basic utility services, although rapid economic change meant that this infrastructure was not always the right kind or in the right locations. The economic shocks of the 1990s--which lasted longer in the CIS than elsewhere in the Region--meant that utilities deteriorated across the Region for much of the 1990s. Since then, the decline in utility performance (as measured by access, quality, and affordability) has been reversed or slowed in most countries. Electricity has shown the greatest improvement because providers have maintained near- universal coverage while improving reliability in the low-income countries of the CIS. Other recent gains include the expansion of gas to many households affected by the collapse in district heating, and the improvement of water reliability in some countries. Despite these improvements, many households, including many urban households, continue to use dirty fuels such as coal and wood for heating because they lack access to gas and cannot afford (or are not reliably provided with) electricity. In secondary cities, the increasing reliance of house- holds on dirty fuels in some countries represents an especially worri- some trend. 172 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union The Coverage Rates for Most Utilities Remain High Access to electricity and piped water remains quite high. Close to 100 percent of all households are connected to the electrical grid, while from 80­100 percent of urban households are connected to piped water. It is difficult to evaluate access to clean water for rural house- holds because of the limitations of the survey data (see box 4.1). Although electricity and network water access rates remain high, household access to district heating and hot water--provided in paral- lel by district heating companies--dropped precipitously in most low income CIS countries (see box 4.2). In 1989, 78 percent of urban resi- BOX 4.1 Survey Data Provide Limited Information about Access to, and Quality and Affordability of, Utilities Most surveys ask the wrong questions to evaluate whether rural households have access to clean water, improved sanitation, and clean heat. In urban areas, it is reasonable to as- sume that water, sewerage, and energy for heating should be provided through utility networks; however, this is not the case in rural areas. Even in much wealthier countries, rural households commonly depend on wells for water supply, septic tanks for sanitation, and liquefied petroleum gas (LPG) for heating and cooking. District heating and other network services make economic sense only in densely populated areas such as cities. Surveys ask only about connections, not service provisions. Power, water, gas, and district heating outages are quite common in some countries in the Region, especially the low-income countries. Surveys that ask households whether they are connected do not answer the question whether services are provided. Survey data about payments are weak. It is difficult to interpret payment data in the house- hold surveys for several reasons: (a) Households living in apartment buildings often receive a sin- gle bill for maintenance and state rent plus all utilities other than electricity (that is, the so-called communal services). As a result, they often do not know how much they pay for each utility. (b) Households often pay for communal services on a less-than-monthly basis so utility payments may be very uneven. Heating is inherently uneven because it is needed only in the winter. (c)The surveys provide information only about the amount paid, not the amount billed, so the amount of arrears is not known. (d) Some households are not asked to pay for electricity or other utilities at all. (e) Many households choose not to pay for reasons ranging from lack of cash to knowing the utilities cannot enforce payment to not paying because the household did not receive the service (without metering, households are typically billed normative amounts regardless of how much of the service is actually provided and consumed). Sources: World Bank staff; see appendix, A. Data and Methodology, for a detailed discussion. Affordable Access to Quality Services 173 dents in Armenia, 42 percent in Azerbaijan, 64 percent in Georgia, and 52 percent in Tajikistan were connected to district heating. By 2002, the percentage of urban residents connected to district heating had dropped to 6 percent in Armenia, 24 percent in Azerbaijan, 1 percent in Geor- gia, and 21 percent in Tajikistan, with an even smaller number of households actually receiving heat through the connections.13 As dis- trict heating networks shrank and electricity costs soared, affected countries increasingly emphasized gas. The result has been an increase in the number of households connected to the gas network in the low income CIS countries. Although this has allowed some households to switch from dirty fuels or costly electricity to a cleaner and less costly source of heating, many households continue to use dirty fuels for heating, cooking, and light, even if only sporadically. BOX 4.2 What Has Happened with District Heating? Apartments built during Soviet times were heated by district heating at little or no cost (the av- erage family in the Soviet Union spent less than 3 percent of income on all utilities, including heating). However, deep financial crises in the early transition years, often coupled with political and social unrest and the loss of deeply subsidized energy sources in Russia, meant that district heating has disappeared in a number of countries. Even in countries where district heating continues to function, such as Russia, the systems are at increasing risk of breakdown and stoppages (as the chart below shows). Disruptions of heat- ing supply have increased from the 1998­99 heating season. Although the reasons for the break- downs vary from year to year, those due to dilapidated equipment have increased somewhat steadily.The deficit of heating fuel has also grown.The failure of heating systems in a country as cold as Russia can severely damage health status, even causing death. Breakdown and Stoppages in District Heating in Russia 110 120 98 102 season 100 67 79 per 80 60 events 40 of 20 No. 0 1998­99 1999­00 2000­01 2001­02 2002­03 Dilapidation of networks and equipment Lack of preparation for winter (i.e., no fuel) Improper operation (lack of qualified workers) Natural disasters Other causes Source: Russian Federation Construction Agency (Rosstroi). 174 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union The Quality of Utility Services Is a Greater Challenge than Access Access measures only the presence of a connection, but it fails to cap- ture the larger challenge in the Region: service reliability or quality. The quality of electricity, water, gas, heat, and other infrastructure and energy services deteriorated during the early transition years, especially in the low income CIS countries. Although anecdotal evi- dence suggests that service reliability for electricity and water have improved in at least some countries, in others the reliability of ser- vices remains a serious challenge. Even in electricity, where reforms are most advanced and invest- ment has been the greatest, reliability cannot be assumed. As figure 4.16 shows, in countries such as Albania, Georgia, Tajikistan, and Turkmenistan, fewer than half of the households are supplied with electricity around the clock. The lack of reliable service is generally worse in marginal settlements, such as smaller cities and towns and rural areas, where poverty is also more prevalent. Although time-series data tracking water availability are available for only a few countries, the evidence from two of the poorer coun- tries, Tajikistan and Moldova, shows the influence of years of low maintenance and little investment in water provision. As figure 4.17 shows, households may be connected to water officially, but little water flows through the pipes. On average, Tajik households receive water for less than six hours each day, with the households in smaller FIGURE 4.16 Reliability of Electricity in the Region in the Early 2000s 100 80 electricity day 60 with a hours 40 24 households 20 of % 0 Albania Bosnia Bulgaria Kazakhstan Georgia Kyrgyz Rep. Tajikistan Turkmenistan SEE Middle Low income CIS income CIS Capital Other urban Rural Source: Hamilton and others 2004. Affordable Access to Quality Services 175 FIGURE 4.17 The Deterioration in Water Provision in Tajikistan and Moldova Running water in Tajikistan in 1999 and 2003, by location Running water in Moldova, 1996 to 2001 24 24 day 20 day 20 per per 16 16 hours 12 hours 12 8 8 Average 4 Average 4 0 0 Total Capital Other urban Rural 1996 1997 1998 1999 2000 2001 1999 2003 Sources: Tajikistan, staff calculations; Moldova, Regional Infrastructure Database. cities and rural areas having the least water. The dramatic drop from 1999 to 2003 suggests that the water utilities are bordering on com- plete collapse. In Moldova, utilities provide water for less than half the day, down substantially from about 20 hours per day in 1996­97. The average figure probably hides differences between regions, some of which are better-off and some of which are worse-off. The erosion of access to district heating networks has been rein- forced by continuing deterioration in service quality. In Russia, for example, district heating stoppages have increased overall since the late 1990s, as shown in box 4.2. Service failures have potentially dis- astrous consequences in this country because of its cold climate. Ser- vice deterioration and collapse also compel households to switch to other heating sources, including dirty fuels (which will be discussed further below). Poor-Quality Services Cost Households More As household expenditure shares for energy and other utilities con- tinue to increase, affordability may be a growing concern in some countries. As shown in figure 4.18, utility expenditure shares are highest in the EU-8, followed by those in SEE, the middle income CIS group, and the low income CIS group. Bulgaria, Moldova, and Roma- nia have seen especially large jumps in household expenditure shares in recent years. The increases in overall expenditures on utilities have been largely driven by increased payments for electricity as tariffs have gone up, payments have been enforced, and household reliance on electricity for heating has increased (because of loss of other heat- ing sources) in a number of countries. 176 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union FIGURE 4.18 Household Expenditure Shares for Electricity, Heating, Water, and Sewerage Have Increased from 1998 to 2002/2003 24 22 20 18 16 14 expenditures 12 10 total of 8 % 6 4 2 0 FYR Fed. Rep. Latvia Estonia Poland Hungary Albania Turkey Lithuania Bulgaria Belarus Romania Ukraine Armenia Georgia Moldova Vietnam Azerbaijan Tajikistan Colombia Montenegro Kazakhstan Russian Kyrgyz Uzbekistan & Macedonia, Serbia EU-8 SEE Middle income Low income CIS Benchmark CIS 1998 or earliest 2003 or latest Source: World Bank staff estimates, using data from ECA Household Surveys Archive. Note: Bulgaria--earliest data are from 1995; Tajikistan--earliest data are from 1999; Estonia, the Kyrgyz Republic, and Uzbekistan--earliest data are from 2000; Kazakhstan--earliest data are from 2001; Albania, Ukraine, Azerbaijan, and Turkey--data before 2002 are not available. Poor households devote a larger share of expenditures to paying for electricity than better-off households. Figure 4.19 compares the share of household expenditures for electricity in poorest income (quintile 1) and richest (quintile 5) households across the Region's countries. In every country for which data are available, electricity is a greater burden on poorer households than on richer ones. At the same time, there are significant regional differences. Households in the middle income CIS countries, for example, spend relatively little on electricity, whereas those in the EU, SEE, and some of the low income CIS countries where substantial tariff reform has occurred (Armenia, Georgia, and Moldova) spend relatively more. Despite tariff increases, tariffs for utilities remain well below cost recovery levels, assuming that investment needs are taken into account. Box 4.3 summarizes current electricity and water tariffs as compared with projected cost-recovery benchmarks. It is important to note that tariffs for both electricity and water will need to increase Affordable Access to Quality Services 177 FIGURE 4.19 Electricity Payments Are a Larger Share of Household Expenditures for Poor Households (Quintile 1) than for Rich Households (Quintile 5) Hungary EU-8 Poland Albania Bulgaria SEE Romania Serbia Belarus CIS Kazakhstan income Russian Fed. Middle Ukraine Armenia Azerbaijan CIS Georgia income Kyrgyz Rep. Low Moldova Tajikistan Turkey Bench- mark 0 2 4 6 8 10 12 Electricity payment as % of total household expenditure Poorest quintile Richest quintile Source: World Bank staff estimates using data from ECA Household Surveys Archive. Note: Nonpaying households have been excluded from calculations. substantially in nearly all countries before cost recovery will be achieved. To a large extent, the relatively lower water tariffs reflect less reform in the water sector because governments have prioritized elec- tricity, where the fiscal and quasi-fiscal losses were enormous, the interest of the private sector greater, and the ability to enforce pay- 178 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 4.3 Electricity and Water Tariffs Remain below Benchmarks for Full-Cost Recovery Residential electricity tariffs in most of the Region's countries have increased to levels sufficient to cover short-term operating costs (3 cents per kilowatt-hour). In only a few countries, howev- er, have residential tariffs increased to the rough benchmark of 7.5­8.5 cents per kilowatt-hour necessary for full-cost recovery, including coverage of capital investment.The chart below com- pares current residential tariffs with the lower boundary of an indicative regional benchmark (7.5 cents per kilowatt-hour, shown by a broken line). Residential tariffs in the CIS countries fall the furthest below this level, followed by those in SEE and the EU-8. Residential Electricity Tariffs Remain Well below the Benchmark of 7.5­8.5 Cents per Kilowatt-Hour 10 9 8 ECA benchmark 7 kWh/ 6 5 cents 4 US 3 2 1 0 Rep. Rep. FYR Rep. Latvia Estonia Poland Serbia Hungary Albania Turkey Lithuania Bulgaria Croatia Belarus Romania Ukraine Armenia Georgia Moldova Tajikistan Colombia Czech Slovak Montenegro Kazakhstan Azerbaijan Kyrgyz Uzbekistan Macedonia, EU-8 SEE Middle Low income CIS Bench- income marks CIS ment much simpler. In contrast to electricity, where metering has always been common and where disconnection is straightforward, water is not usually metered, and disconnecting residential customers in apartment buildings is technically difficult. The limited evidence available about gas and district heating tar- iffs suggests that both remain well below cost recovery levels and lag even further behind in overall sectoral reform. This may create addi- tional pressures in the future: late-reforming sectors may not be able to increase tariffs to needed levels because households, especially low-income households, will not be able to absorb additional Affordable Access to Quality Services 179 Residential water tariffs lag even further behind full cost recovery levels of $1 per cubic meter, as shown in the chart below. Only two countries (the Czech Republic and Lithuania) have reached the benchmark level ($1 per cubic meter, shown by a broken line).The low income CIS countries have the lowest residential tariffs, followed by the middle income CIS group, SEE, and the EU- 8. Residential Water Tariffs Remain Well below the Benchmark of $1 per Cubic Meter 1.4 1.2 World benchmark 1.0 meter 0.8 0.6 $/cubic US 0.4 0.2 0 & Rep. Rep. FYR Fed. Rep. Latvia Estonia Poland Hungary Croatia Serbia Lithuania Slovenia Albania Bosnia Romania Ukraine Armenia Georgia Moldova Tajikistan Colombia Vietnam Czech Slovak Herzegovina Kazakhstan Azerbaijan Russian Kyrgyz Uzbekistan Turkmenistan Macedonia, EU-8 SEE Middle Low income CIS Bench- income marks CIS Sources: World Bank Forthcoming-b. Data on Colombia: Latin American Energy Association (2004); Energy-economic infor- mation system. Energy Statistics; http://www.olade.org.ec/php/index.php?arb = ARB0000006; and staff calculations from tariff information provided at http://www.superservicios.gov.co/. Data for Vietnam are from http://www.ib-net.org/. increases. In some sense, increased electricity expenditures resulting from improved enforcement and increased tariffs may crowd out the ability of many households to absorb additional tariff increases in district heating and gas, where reforms have been implemented more slowly. Household Coping Options forTariff Increases Households have relatively few options for coping with tariff increases: (a) reducing utility consumption, (b) reducing nonutility 180 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union consumption, (c) incurring arrears, (d) relying on the social protec- tion system, and (e) fuel switching. All options present difficulties. The first option that households have to cope with increasing util- ity bills is to reduce consumption. Except for electricity, most utilities in most countries are not metered; hence, households are billed based on normative consumption. Without metering, households cannot economize to reduce bills. For electricity, poor households in low income CIS countries such as Armenia, Georgia, and Moldova have already reduced consumption to the absolute minimum. The second option that households have is to absorb tariff increases by maintaining utility consumption at the expense of other basic con- sumption items such as education, health, or food. Because house- holds cannot reduce consumption of utilities (which are normatively billed), this scenario may well be commonplace. The third option is not to pay, or not to pay in full. Review of non- payment patterns finds that a high proportion of households do not pay for utility services. Low-income households in the low income CIS countries are especially likely not to pay, or not to pay in full. Although weaknesses in the survey data (discussed in box 4.1) pre- clude development of a comprehensive picture of nonpayments and arrears for the Region, available evidence suggests that these prob- lems are widespread, especially in the poorer countries. In Serbia, for example, 1 in 5 people live in households with overdue electricity bills, and 1 in 10 in households with overdue bills for communal ser- vices (World Bank 2005h). In Azerbaijan, about one in five urban households report arrears for electricity, although fewer than 5 per- cent of households do not pay at all (World Bank 2004a). In the Russ- ian city of Norilsk in 2002, an average household paid only 82 percent of all amounts billed (Bashmakov 2004). Some evidence suggests that poor households are more likely to not pay or to have arrears. In Armenia, poor households report being dis- connected for nonpayment of electricity bills at a much higher rate than better-off families. In 2001, about two-thirds of the lowest-quin- tile households in the capital city of Yerevan were disconnected for nonpayments, while fewer than one-third of the highest-quintile Yerevan households reported disconnections for nonpayment. The pattern was broadly similar in other urban and rural areas in Armenia, although the disconnection rates were lower (World Bank 2003b). The fourth option that households have is to rely on the social pro- tection system. The two main mechanisms to protect the poor from tariff increases are lifeline (block) tariffs (where consumption can be metered) or targeted subsidies. Lifeline tariffs provide for consumption of a minimal or basic amount by all consumers at a subsidized price. Affordable Access to Quality Services 181 Higher rates are charged for amounts consumed above the basic block to provide a source of cross-subsidy. The ability to meter consumption is key to implementation. The second mechanism, targeted subsidies, provides cash transfers, vouchers, or discounts on electricity, water, and other utilities to households who are judged to be poor. Although metering is not essential to targeted subsidies, they are more costly to administer. The programs in existence to compensate households for the rising utility tariffs have low coverage of the poor. The fifth option, discussed in greater detail in the following section, is to switch fuels. Households heating with electricity may be able to switch to gas in some cases. Alternatively, households may opt to switch to dirty fuels for heating. Many Households Rely on Dirty Fuels for Heating Households continue to rely on dirty fuels for heating, especially in sec- ondary cities and rural areas. The major reasons that households switched from clean to dirty fuels include loss of access to district heat- ing, irregular supply of electricity, high cost of electricity, and lack of access to other clean fuel sources, such as gas. The lack of reliable energy sources and the increased costs of the existing choices (such as electric- ity) pushed many households into lower-quality choices of energy, such as solid fuels. Household reliance on dirty fuels increased sharply in the early transition years. More recently, the pattern has been mixed. Of the countries shown in table 4.3, dirty fuel use increased on a national basis only in Bulgaria and Romania (SEE), while decreasing significantly in Kazakhstan (a middle income CIS country) and Tajikistan (a low income CIS country), as household access to other clean fuels, most notably nat- ural gas and district heating, increased. TABLE 4.3 In Most Countries, Households in Secondary Cities Were More Likely to Heat with Dirty Fuels in 2003 than in 1998 (in percentages) Capital Other Urban Rural Total 1998 2003 1998 2003 1998 2003 1998 2003 EU-8 Hungarya 4 5 16 17 37 37 22 22 SEE Bulgariab 10 5 38 51 91 91 51 57 Romania 23 18 19 24 91 89 51 53 CIS middle income Kazakhstanc 2 2 11 7 36 25 22 15 CIS low income Armenia 42 39 69 78 93 87 72 70 Moldova -- 0 -- 30 -- 93 -- 64 Tajikistand 23 12 62 44 96 93 86 76 Sources: World Bank staff estimates. See appendix table 9 for country-level data and years. . Note: a. 2002 is used instead of 2003; b. 1995 is used instead of 1998; c. 2001 is used instead of 2003; d. 1999 is used instead of 1998; -- = not available. 182 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Households in secondary cities were more likely to use dirty fuels in 2003 than they were in 1998. For the six countries for which time- series data are available, the share of households using dirty fuels to heat increased in four countries and decreased in only two: Kaza- khstan and Tajikistan. Particularly sharp increases in household reliance on dirty fuels in secondary cities from 1998 to 2003 are seen in Armenia and Bulgaria (9 percent and 13 percent, respectively). Regardless of location, the poor rely more on solid fuels than the other groups do. As shown in figure 4.20, the ratio of richest house- holds (q5) to poorest households (q1) is greater than 1 for all countries for both years. In Armenia, Bulgaria, and Romania, poor households were noticeably more likely to rely on dirty fuels in 2003 than in 1998. In Hungary and Kazakhstan, inequality in access to clean heat changed little from 1998 to 2003. Only in Tajikistan has inequality decreased, in part because of reviving the district heating network in Dushanbe. An unexplored dimension of the decaying quality of infrastructure services concerns effects on health and educational outcomes. Anecdo- tal evidence shows that the deteriorating water systems result in levels of waterborne diseases (for example, hepatitis A) that are significantly higher than in the EU (OECD 2003). In addition, the negative effects of in-house pollution due to the use of solid (and especially dirty) fuels on health status have been well documented in other regions (WHO 2002), and it is expected that similar effects may be emerging in the Region, although studies have not fully explored this issue. FIGURE 4.20 Poor Households Are Less Likely than Rich Ones to Use Clean Fuels 10 9 quintile 8 7 poorest 6 to 5 4 3 richest of 2 1 Ratio 0 Hungary Bulgaria Romania Kazakhstan Armenia Moldova Tajikistan Turkey EU-8 SEE Middle Low income CIS Bench- income mark CIS 1998 2003 Sources: World Bank staff estimates using data from ECA Household Surveys Archive. See appendix table 9 for country- level data and years. Note: Share of households using clean fuels in the upper quintile divided by that share in the lowest quintile. Affordable Access to Quality Services 183 Conclusions The Region has achieved major reductions in income poverty after 1999, and some of these improvements have been reflected in other dimensions of poverty. But improvements in income alone have not sufficed to reduce the deprivation of affordable access to quality ser- vices. Reducing poverty in the nonincome dimensions, it appears, is more of a long-term agenda in which the transformation of the pub- lic sector will be critical for several reasons. First, although the Region's countries put increasing resources into social sectors as their economies improved after 1999, spending in many poorer countries is still limited. Spending levels need to be maintained at adequate levels for sustained improvements in service delivery. Second, all formerly socialist countries face a critical challenge in adjusting their service delivery systems to the new environment. Reduced fertility, aging population, and emerging risks such as HIV/AIDS pose new obstacles to the effectiveness and quality of ser- vices. The inherited structure of service delivery in health care and education spreads limited resources too thinly, reducing the quality of service, particularly for the poor. Improving access to, and affordabil- ity of, quality services for the poor hinges on realizing crucial effi- ciency gains within the existing systems. Third, countries are beginning to develop effective governance mechanisms that enhance consumers' voice and improve quality of services. So, for example, decentralization experiences are varied, and the results still limited, but certain experiences in education underscore the positive linkages between effective decentralization, local autonomy, and quality of social services. These experiences are worthy of dissemination and replication. Fourth, the vulnerability of human development outcomes may arise from shocks to health or income or from broader phenomena such as the deterioration of water and gas networks. The lack of insur- ance or coping instruments at the household level represents a major risk. Developing adequate risk management strategies is critical to helping all people, but particularly the poor, cope with risks. A num- ber of promising policy experiments suggest that even countries with limited resources can improve the ability of households, especially poor households, to manage risk. Working along all four dimensions-- fiscal commitment, efficiency gains, improved governance, and better risk management for the poor--will be critical if the public sector is to rise to the challenge of providing affordable access to quality services. 184 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ANNEX TABLE 1 Public Expenditures on Health in the Region % of GDP 1994­99 2000 2001 2002 Armenia 2.8 3.2 3.2 -- Azerbaijan 1.0 0.6 -- -- Belarus 4.6 4.6 4.8 -- Bosnia & Herzegovina 3.2 3.1 2.8 -- Bulgaria 4.7 4.0 3.9 -- Croatia 9.0 7.8 7.3 -- Czech Rep. 6.6 6.5 6.7 6.8 Estonia 5.2 4.5 4.3 4.2 Georgia 1.1 0.4 -- -- Hungary 6.0 5.0 5.1 5.5 Kazakhstan 2.1 2.1 1.9 -- Kyrgyz Rep. -- 2.0 1.9 -- Latvia 3.9 3.5 3.4 3.6 Lithuania 4.5 4.4 4.2 4.1 Macedonia, FYR 4.6 4.2 -- -- Moldova 4.2 2.7 2.5 3.2 Poland 4.1 4.0 4.3 4.4 Romania 3.5 4.1 4.2 -- Russian Fed. 3.9 3.7 3.7 -- Serbia & Montenegro 7.2 5.9 6.5 -- Slovak Rep. 5.2 4.9 5.0 5.1 Slovenia 6.9 6.9 7.1 -- Tajikistan 1.2 0.9 1.0 -- Turkey 3.0 4.2 -- -- Turkmenistan 3.5 3.0 3.0 -- Ukraine 3.5 2.9 2.9 -- Uzbekistan 2.9 2.8 2.7 -- Sources: Public expenditure database and IMF fiscal database. Note: -- = not available. Affordable Access to Quality Services 185 ANNEX TABLE 2 Health Care Utilization Hospitalization admissions per 100,000 1994­99 2000 2001 2002 Armenia 6.9 5.1 4.9 6.2 Azerbaijan 6.2 4.8 4.9 4.9 Belarus 26.5 29.3 30.0 29.3 Bosnia & Herzegovina 8.1 8.0 7.8 6.9 Bulgaria 16.7 15.4 15.3 16.5 Croatia 14.4 15.7 15.8 15.7 Czech Rep. 19.9 20.0 20.3 21.1 Estonia 19.2 20.4 19.7 19.1 Georgia 5.8 4.6 4.5 4.6 Hungary 22.3 23.6 23.9 24.6 Kazakhstan 16.1 14.9 15.5 16.3 Kyrgyz Rep. 16.8 15.9 14.5 12.7 Latvia 21.8 22.1 20.7 19.9 Lithuania 22.8 24.7 24.0 23.6 Macedonia, FYR 9.9 9.7 9.0 -- Moldova 19.0 13.7 12.5 13.8 Poland 13.7 15.5 16.4 -- Romania 20.9 22.4 24.4 25.0 Russian Fed. 21.0 22.0 22.5 22.8 Serbia & Montenegro 11.3 -- -- -- Slovak Rep. 19.5 19.9 19.7 19.0 Slovenia 16.0 16.8 16.6 16.4 Tajikistan 11.8 9.1 9.0 9.2 Turkey 6.9 7.8 7.8 8.0 Turkmenistan 14.2 -- -- -- Ukraine 20.4 19.4 19.8 20.0 Uzbekistan 16.0 13.3 13.8 14.0 Source: Official (administrative) health data for country agencies. Note: -- = not available. 186 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ANNEX TABLE 3 Doctor-Patient Contact per Person per Year 1994­99 2000 2001 2002 Armenia 3.5 2.1 1.8 -- Azerbaijan 6.5 5.0 4.9 4.5 Belarus 11.1 11.7 11.6 11.4 Bosnia & Herzegovina 2.7 -- -- 2.6 Bulgaria 5.6 -- -- -- Croatia 5.9 7.0 -- -- Czech Rep. 14.8 14.8 14.8 14.8 Estonia 6.2 6.7 6.5 6.4 Georgia 2.1 1.4 1.5 1.6 Hungary 10.7 11.1 11.3 11.9 Kazakhstan 6.2 5.5 5.7 6.2 Kyrgyz Rep. 4.6 4.1 4.0 4.5 Latvia 4.7 4.8 4.8 4.6 Lithuania 7.3 6.3 6.5 6.4 Macedonia, FYR 2.9 3.2 3.0 -- Moldova 8.1 6.5 6.2 6.7 Poland 5.3 5.4 5.5 -- Romania 7.5 5.1 5.4 5.7 Russian Fed. 9.2 9.4 9.5 9.6 Serbia & Montenegro 5.3 -- -- -- Slovak Rep. 14.7 16.3 14.6 14.5 Slovenia 7.1 6.8 6.7 6.4 Tajikistan 3.9 3.7 4.7 4.8 Turkey 1.9 2.4 2.6 -- Turkmenistan 5.9 7.0 6.8 -- Ukraine 9.5 10.0 10.1 10.3 Uzbekistan 7.0 8.4 8.3 8.5 Source: Official (administrative) health data. Note: -- = not available. Endnotes 1. The measurement of infant and maternal mortality rates in the Region's countries, particularly in those of the Former Soviet Union, is still affected by the different definition of a "live birth" in the previous sys- tem, the number of deaths at home, the lack of access to registration, and negative incentives to register deaths (particularly in rural areas). These explain the existing discrepancies between official statistics and data col- lected from surveys such as the Demographic and Health Survey (DHS). Notwithstanding these limitations, data suggest that, in some EU-8 coun- tries, maternal mortality rates are comparable to EU-15 levels (about 5­6 maternal deaths per 100,000 live births) (Bos and others 2002; Bonilla- Chacin, Murrugarra, and Temourov 2002; World Bank 2003m). 2. The idea is that while in Central Asia the total number is 12,000, in Ukraine 12,000 is the number of new cases (incidence vs. prevalence). Affordable Access to Quality Services 187 3. See Briceńo, Estache, and Shafik (2004) for a review of quality of infra- structure services in developing countries. 4. The data used are the TIMSS for 1995 and 1999. The microdata at the student, teacher, and school levels for 2003 were not available when this report was written. This section draws from Murrugarra and Sethi (2005). Results for eighth and fourth grades are similar; this section reports eighth grade because of its broader time coverage. 5. TIMSS defines an internationally comparable minimum score based on the minimum level of skills required in eighth grade. 6. See Fiszbein (2001). 7. See Zeikate (2004) for a detailed discussion of the decentralization processes in health and education in the Region's countries. 8. The decomposition divides the variation of mathematics scores into four blocks (individual, teacher, school, and household background), using mean or conditional quantile regression. Once the variation of scores has been estimated at different points of the distribution, each component is the combination of the characteristics and returns for those students in each part of the score distribution. This mimics the use of quantile regres- sion in the decomposition of factors underlying the wage distribution, as applied in Dolado and Llorens (2004). 9. These are the age-standardized rates (ASRs) for populations between 15 and 44. 10. These estimates may differ from official statistics because people may seek attention from private or even informal providers who may not be included in official records. To the extent that most health care is pro- vided at facilities that are publicly organized, this is not a major issue. 11. See Wagstaff and van Doorslaer (2001) for a methodological discussion about estimating the impoverishing effects of health expenditures. 12. Informal payments in Armenia are those that are not considered as the "stipulated" (official) cost in the household survey (World Bank 2002a). 13. Figures for 1989 are from the 1989 Soviet census. Figures for 2002 are based on staff calculations from the household survey data. CHAPTER 5 Prospects for Poverty Reduction This chapter examines the prospects for poverty reduction in the Region. Obviously, how overall growth rates are sustained--and accel- erated--will determine the extent of poverty reduction. Given coun- try differences in the elasticity of poverty reduction to growth, the poverty impact from growth will be variable across countries. This chapter asks three questions. First, given the progress that countries in the Region have made, what level of growth rate would generate sus- tained poverty reduction in the future? Second, given public policies already being implemented, what else needs to be done to promote growth with poverty reduction? Third, how does the future agenda for poverty reduction differ across countries or groups of countries? Alternative Scenarios for Growth, Poverty Reduction, and Inequality Medium-term outlook. One should begin with some simple projections of poverty rates in the medium term (up to 2007). For growth rates, the country-specific projections for personal consumption from the World Bank's Global Economic Prospects 2005 are used. Using the survey data for the latest available year, one then multiplies everyone's con- sumption by the same growth rate and compares it with the poverty 189 190 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union line. Assuming that households in all parts of the income distribution benefit uniformly, poverty is projected to fall in the Region from 12 percent in 2003 to 8 percent in 2007, at an annual rate of 8.5 percent a year, or by about 21 million (figure 5.1). The key variable underpinning these projections is household pri- vate consumption, which is projected to grow by about 6 percent a year between 2003 and 2007. The elasticity of the poverty headcount to growth in average per capita consumption is therefore relatively low (a little more than 1.3), reflecting the increasing concentration of the poor in the low-income countries in the Region, where the effi- ciency of growth in reducing poverty is not high and medium-term growth prospects are below the average of the Region. In the low income CIS group, the average poverty rate is projected to fall from its 2003 level of 47 percent to around 36 percent by 2007, at an annual rate of about 6.5 percent a year. On the other hand, in the middle income CIS group, poverty is projected to fall from 8 per- cent of the population in 2003 to 2.8 percent by 2007, implying an annual rate of poverty reduction of more than 20 percent. Poverty is projected to reach almost zero in the EU-8, but will still affect around 5 percent of the population in SEE. Nevertheless, by 2007, according to these projections, some 40 million people will remain poor. Thus poverty will not disappear, but together with economic vulnerability will affect 30 percent of the population. Faster growth could lead to faster reduction of poverty rates. Sustained economic growth is hence a crucial component of any poverty alleviation strategy. The evidence presented in chapter 2, however, points out that a worsening of the income distribution would undermine a positive impact of growth and significantly reduce the efficiency of growth in reducing poverty. Although it is difficult to discern general trends in inequality over time for a given country, chapter 2 presented evidence that reductions of inequality observed in CIS countries were driven by a unique combination of factors and are unlikely to continue at the same pace. Nevertheless, even the scenario of unchanged inequality offers improvement in poverty rates in the medium term (as shown by figure 5.1). But are the projected reductions in poverty rates sufficient? For this, one needs to see what the longer-term objectives and vision for poverty reduction are. The discussion turns to this subject next. Long-term scenarios contain a vision for poverty reduction in the Region. To assess whether a given rate of poverty reduction is sufficient, one needs to take a broader and a longer-term perspective. The Millennium Prospects for Poverty Reduction 191 FIGURE 5.1 Population of the Region by Poverty Status, 1990­2002, and Outlook for 2007 Projections 100 90 212.1 80 258.2 70 329.2 60 50 population of 40 160.7 % 30 153.3 108.8 20 10 102.0 61.2 40.0 0 1998­9 2002­3 By 2007* Nonpoor: above $4.30 a day Vulnerable: above $2.15 and below $4.30 a day Poor: below $2.15 a day Sources: World Bank staff estimates using data from ECA Household Surveys Archive; outlook for 2007 ­ staff estimates. Note: The projection is based on country-specific growth rates and uses the most recent survey data. Development Goals (MDGs) present such an internationally agreed target for poverty reduction (box 5.3). The first one (MDG1) is to reduce absolute poverty by half by 2015, compared with its level in 1990. Obtaining strictly comparable figures on poverty is difficult, but using external sources, one can arrive at a set of rough estimates of poverty and vulnerability at the country level circa 1990.1 Choosing 2015 as the time horizon for these projections gives a longer-term per- spective and a clearly defined goal. Although this general vision is shared by most nations, countries interpret various targets in a way that makes them more relevant for their level of development and aspi- rations. In particular, for countries in the Region, a poverty line of $2 a day (or, more precisely, $2.15 a day) would appear to be more appro- priate as a standard of material deprivation than the $1 a day ($1.075) currently embodied in MDG1. Using poverty rates from 1990, one can establish a target for each country in the Region with regard to absolute poverty (at $2.15 per day) by 2015 (table 5.1). Halving poverty rates in the poorest CIS countries would mean achieving by 2015 a poverty rate (weighted by 192 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 5.1 EU Accession and Poverty Reduction Objectives Countries in the Region that recently joined the EU (the Czech Republic, Estonia, Latvia, Lithua- nia, Poland, Hungary, Slovenia, and the Slovak Republic), that are about to join the EU (Bulgaria and Romania), or that are in the Stabilization and Association Process with the EU (the Balkan countries) have a clear framework for aligning their policy objectives with the common social agenda of EU. Its key features are listed below. At the European Council of Lisbon in 2000, the European Union member states and the Euro- pean Commission outlined steps to make the EU the world's most competitive economy by 2015 and adopted a new approach to promoting social cohesion across the EU. Specifically, it laid out six key objectives: 1. Promote employment and employability through active labor market measures to help those who have the most difficulty in entering the labor market and a mutually reinforcing system of social protection, lifelong learning, and labor market policies 2. Ensure adequate social protection systems, including minimum income schemes, for all to have a sufficient income for a life with dignity and effective work incentives for those who can work 3. Increase the access of the most vulnerable and those most at risk of social exclusion to de- cent housing conditions, to quality health and long-term care services, and to lifelong learn- ing opportunities, including to cultural activities population) of about 14 percent, which is substantially below the cur- rent poverty incidence of around 52 percent in 2002­3. For the mid- dle income CIS and SEE countries, the target poverty rates would be close to 1.6 percent and 0.2 percent, respectively. (In statistical terms, these rates are not distinguishable from zero.) Thus the goal for these countries can be stated simply as the elimination of absolute poverty. For the EU-8 countries, which had already achieved by 2002­3 an absolute poverty incidence of 2.3 percent (again in statistical terms not very different from zero), a more meaningful target is reducing the size of the vulnerable population--about 25 percent--and pre- venting poverty from reemerging (see box 5.1). Using these differentiated, MDG-related targets, what would it take to achieve these goals for poverty reduction? Table 5.1 reports the average annual growth rates that would be required before 2015 to achieve these differentiated targets, assuming no worsening of the distribution (that is, similar growth rates for the poor and nonpoor alike). This table also reports actuals and projected growth rates for Prospects for Poverty Reduction 193 4. Prevent early exit from schools and formal education and training and facilitate the transition from school to work, in particular of young people leaving school with low qualifications 5. Eliminate poverty and social exclusion among children as a key step to combat the intergen- erational inheritance of poverty, with a particular focus on early intervention and early educa- tion initiatives that identify and support children and poor families 6. Reduce the levels of poverty and social exclusion and increase labor market participation of immigrants and ethnic minorities to the same levels as the majority population The EU also identified a number of monitorable indicators. Every two years, each member state must submit a National Action Plan (NAP) to the European Commission, laying out how it in- tends to fulfill progress on 18 agreed-on "social inclusion indicators" that focus on social out- comes. Primary indicators include poverty rate, inequality indexes, regional employment rates, long-term unemployment rate, prevalence of jobless households, number of early school leavers not in further education/training, life expectancy at birth, and self-defined health status by in- come level. From these sets of objectives and indicators, it is very clear that poverty reduction remains in the center of policy making in the new member states and that both the absolute poverty reduction goals discussed in this report and the implied inequality targets are in line with the broader set of goals accompanying European integration. If anything, EU accession results in a tighter set of requirements than these country-level simulations and scenarios. Sources: Council of the European Union 2004; and Atkinson, Marlier, and Nolan 2004. the period 2002­7, showing that despite recent progress, many coun- tries in the Region are facing a real challenge of poverty reduction. Especially in the poorest CIS countries, growth rates required to achieve needed progress in poverty reduction are significantly above the projected range. Even for the EU-8, more realistic (and challeng- ing) targets would imply the need to accelerate growth rates. Only for the middle income CIS group and SEE are the projected growth rates close to what is needed to eliminate absolute poverty by 2015. But if the political vision and ambition in these countries go well beyond exceeding the welfare levels observed in the past before the breakup of the Soviet block, then accelerated and shared growth would be needed for all groups of countries. A forward-looking agenda could, for example, aim at reducing the risk of reemergence of poverty by reducing the incidence of economic vulnerability to zero (table 5.2) or, with European income convergence in mind, at achieving levels of poverty no higher than those observed in the poorest EU member states today (table 5.3). Both of these visions would suggest a signifi- 194 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 5.1 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, MDG-Related Targets Actual/ Target Current Growth rate projected poverty poverty required to growth rates rate, rate, reach targets, 2002­07, Target , Definition population population Countries GDP weighted GDP weighted MDG-related of Poverty weighted, weighted Low income CIS 5.6% 3.9% Reduce by 1/2 country's Absolute poverty 13.8% 52.3% poverty rate in 1990 ($2.15) Middle income CIS 4.3% 6.8% Eliminate Absolute poverty 1.6% 8.0% absolute poverty ($2.15) SEE 3.4% 5.4% Eliminate Absolute poverty 0.2% 11.4% absolute poverty ($2.15) EU-8 5.9% 4.3% Eliminate Vulnerability 0.0% 24.6% economic vulnerability ($4.30) Source: Staff estimates. TABLE 5.2 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, Country-Specific Targets Focused on Economic Vulnerability Growth rate Actual/projected required to growth rates Current state in reach targets, 2002­07, reaching target Countries GDP weighted GDP weighted Specific target related (poverty rate) Low income CIS 12.0% 3.9% Eliminate absolute poverty ($2.15) 52.3% Middle income CIS 9.7% 6.8% Eliminate economic vulnerability ($4.30) 39.6% SEE 10.8% 5.4% Eliminate economic vulnerability ($4.30) 55.3% EU-8 6.6% 4.3% Reduce poverty to half the incidence 36.6% in the poorest EU-15 member today (10.5%) Source: Staff estimates. TABLE 5.3 Annual Growth Rates of Private Consumption Needed to Achieve Poverty Reduction by 2015, Country-Specific Targets with European Vision Growth rate Actual/projected required to growth rates Specific target related Current state in reach targets, 2002­07, to country level of reaching target Countries GDP weighted GDP weighted development (poverty rate) Low income CIS 12.0% 3.9% Eliminate absolute poverty ($2.15) 52.3% Middle income CIS 8.0% 6.8% Reduce poverty to the incidence in the poorest EU-15 country today (21%) 62.4% SEE 9.9% 5.4% Reduce poverty to the incidence in the poorest EU-15 member today (21%) 65.8% EU-8 6.6% 4.3% Reduce poverty to half the incidence in the poorest EU-15 member today (10.5%) 36.6% Source: Staff estimates. Prospects for Poverty Reduction 195 cant acceleration of growth rates. In fact, they call for a tripling of growth rates in the low income CIS countries and almost a doubling of growth rates in the SEE. They also suggest a significant growth accel- eration in both the EU-8 and the middle income CIS group. Given the assumption of unchanged income distribution, this suggests that the gains from this growth need to be equitably shared in the population. Patterns of Growth: Implications for Growth and Inequality The recent evidence suggests that poverty reduction has been driven largely by a consumption boom. Growth rates for private consump- tion exceeded GDP growth rates in the Region for the initial eco- nomic recovery (figure 5.2). Now, for various reasons (such as needed correction of external account imbalances in SEE or debt burdens in some poor CIS countries), consumption growth rates are likely to slow down and even lag GDP growth rates. What would this development imply for patterns of growth? Inter- national experience suggests that in fast-growing economies, higher GDP growth is driven by increased investment and exports (see figure 5.2 for growth rate projections for the Region). This suggests that to achieve the objective of reducing poverty, the countries of the Region need to step up even more aggressively the potential engines of eco- nomic growth and encourage new investments and net exports. FIGURE 5.2 Trends from Global Projections (Percentage Changes) Growth Rates in the Region 8 7 6 5 4 3 change % 2 1 0 2001 2002 2003 2004 2005 2006 ­1 ­2 ­3 GDP at market prices Private consumption Source: World Bank 2005b. 196 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Inequality dynamics. Although the simulations presented in tables above assumed distributionally neutral growth, what would happen if the distribution actually worsens? Growth incidence curves pre- sented in chapter 2 show that the growth process in the Region trig- gered complex distributional changes. In some countries, they were associated with increase of inequality; in other countries, with inequality reduction. But there were usually systematic differences between the growth rates of consumption achieved by the poor and the nonpoor. Chapter 2 argues that the observed periods of strongly pro-poor growth were driven by the exceptional circumstances and may not be continued in the future. Box 5.2 presents evidence of how, for instance, through the depletion of social capital, the poor may be unlikely to take full advantage of new growth opportunities. What does it imply for projected growth required at the level of the national economy to generate a desired change in poverty? The simplest way to think about the growth rates presented in tables 5.1, 5.2, and 5.3 is to take them as the growth rates in real con- sumption that those groups of the population who are currently poor need to achieve. Thus, if growth rates of the poor systematically devi- ate from the consumption of the rest of the population, it would have implications for both the required overall growth rate and the level of inequality. For Poland and Romania, for example, the growth rate BOX 5.2 Depleted Social Capital of the Poor Limits Opportunities Social capital in the form of networks of information exchange and reciprocity helps people to reach out to new opportunities and to weather difficult times. Although inherently difficult to measure and follow over time, a large body of qualitative evidence from the Region suggests that the social capital of the poor has withered because of not only ruptured workplace networks and unemployment but also the inability of the poor to maintain communications and reciproci- ty.This is due to a variety of reasons. In some cases, geographical isolation, coupled with limited improvements in the standard of liv- ing, has meant that the social capital of the poor has come to be largely confined to people in similar circumstances. For example, in rural Bulgaria, poverty, unemployment, and lack of main- tenance of decayed roads have weakened social cohesion. Trust and social activities are limited to close family members. Community organizations hardly exist, and villages lack leaders able to take initiatives for community development.The loosening of social bonds has also been aggra- vated by the depopulation and aging of depressed rural communities as young families and youth migrate to cities (Blackstone and Agency of Socioeconomic Analysis 2004). Prospects for Poverty Reduction 197 among the poor (vulnerable) in 1999­2002 averaged less than half of the average growth rate. One can therefore simulate what such a sce- nario with unequally shared benefits from growth would imply. For Poland, for example, that would imply achieving an 11 percent overall growth rate, instead of 6.5 percent, to meet the basic target reported in table 5.1 (eliminate economic vulnerability). Other national targets in the context of EU would require even higher growth rates--much higher in fact than the economy was able to sus- tain over the past five years. For inequality, that would imply wors- ening Gini coefficients to a level of 0.46--not entirely implausible, but a level that would put Poland on par with the most unequal Latin American societies. In Romania, assuming the growth rates of the poor to be one-half of the national average, that would imply that the needed growth rate to achieve a complete elimination of absolute poverty by 2015 would be 8 percent a year. The Gini coefficient would worsen to 0.35--a level that is quite conceivable, given the observed inequality dynamics. In Russia, reversal of pro-poor growth would imply the need to sustain a 9-percent-per-year growth rate of mean private consumption until 2015 and see inequality worsen again to 0.37, the highest level observed during Russia's economic transition. Thus, for middle income CIS countries, SEE, and the EU-8, wors- ening of inequality so significant as to reverse completely gains from In others, large-scale migration has meant that the poor have yet to build up the social capital necessary to seek better living conditions. Albania, for example, has seen large-scale migration of rural populations from the impoverished northeast portion of the country. Despite the physi- cal accessibility of improved educational services, however, these populations have far less access to good-quality schools than do wealthier local families, who use their social networks to enroll their children in the better schools (Dudwick and Shahriari 2000). Low social capital offers a partial explanation of why certain households, groups, or communi- ties are less likely than others to benefit from socioeconomic opportunities. Depletion of social capital in the closed networks accessible to the poor goes hand in hand with long-term poverty. As these factors work together to reinforce each other in a vicious cycle, they often work to make poverty resistant to improved job opportunities or to improved physical access to educa- tion and medical care. Sources: Blackstone and Agency of Socioeconomic Analysis 2004; Dudwick and others 2004; and Dudwick and Shahriari 2000. 198 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union growth seems rather implausible, but given past trends it cannot be fully ruled out. But where reversal of pro-poor growth may have disastrous impli- cations is in the poorer CIS economies. Because the poor constitute a very sizable share of the population in these countries, if they fail to fully take part in the growth process, that would mean dramatic changes in the distribution and failure to achieve the needed poverty reduction. In Moldova, for example, that would imply stepping up the growth rate to 9 percent annually to just halve the absolute poverty by 2015 (compared with 1990), but the implied change in inequality would drive the Gini coefficient up to 0.57! For Armenia, the Kyrgyz Republic, or Tajikistan, results in inequality increase are similar. Among those countries, only Tajikistan has actually recently experienced rapid growth accompanied by an increase in inequality. The scenario of worsening inequality due to inequitable growth is an entirely plausible proposition for the middle income CIS group, the EU-8, and SEE and less plausible, but extremely worrying, for the low income CIS countries. Because the growth rates it implies are very high and probably not achievable, going along this path would most likely mean that the poverty reduction targets will not be achieved by 2015. Thus, a serious change in policies is required not only to ensure rapid and accelerated growth but also to prevent inequality from rising. Achieving Progress in Nonincome Dimensions The scenarios of poverty reduction discussed above deal only with material poverty. Ending deprivation in other noneconomic dimen- sions is also a challenging task. This challenge is perhaps best under- stood in the prospects of meeting nonincome MDGs (box 5.3). Low income CIS countries face a particularly difficult challenge, but in the middle income CIS countries, SEE, and even the EU-8, the nonin- come MDGs are still relevant because a number of countries are vul- nerable on the health-related goals and because disaggregations of national-level data by region, ethnic group, or gender identify pock- ets where the targets are less likely to be achieved. The child and maternal mortality MDGs represent a challenge in many CIS countries, including the middle income CIS group. Progress with the under-five mortality rate may be limited because utilization rates at secondary hospitals are often lower than international aver- ages and there are concerns about quality and out-of-pocket pay- ments acting as barriers to care. Regarding the maternal mortality rate, poverty, distance, and poorly performing hospital networks Prospects for Poverty Reduction 199 BOX 5.3 Nonincome Dimensions of Poverty and Achieving the MDGs in the Region The Millennium Development Goals (MDGs) grew out of the agreements and resolutions of world conferences organized by the United Nations (UN) over the past decade. Brought togeth- er as a set of "International Development Goals" in 1996, they have since been refined and are now widely accepted as the framework for measuring development progress. At the Millennium Summit in September 2000, the 189 states of the United Nations reaffirmed their commitment to working toward a world of peace and security for all--a world in which sustaining develop- ment and eliminating poverty would have the highest priority. Signed by 147 heads of state, the Millennium Declaration was passed unanimously by the members of the UN General Assembly. The first seven goals are directed at reducing poverty in all its forms: hunger; a lack of income, education, and health care; gender inequality; and environmental degradation. Although each goal is important, collectively they form a comprehensive and mutually reinforcing approach to alleviating poverty. (The MDGs are more relevant to some countries than others in the Region.) · The EU-8 is the least challenged of the Region's subregions by the nonincome dimen- sions of poverty. Most countries have met, or are likely to meet, the nonincome MDGs, ex- cept for the Baltics, where on current trends it does not appear that the spread of HIV/AIDS will be effectively combated. · The nonincome dimensions of poverty are likely to challenge some countries in SEE. It is not clear that countries such as Bulgaria and Romania will be able to combat the spread of HIV/AIDS. Romania may also struggle to meet the water access MDG because only 16 per- cent of the sizable rural population is assessed to have access to an improved water source. · The nonincome dimensions of poverty, particularly related to health, are likely to chal- lenge the middle income CIS countries. None of these countries is assessed as likely to be able to combat the spread of HIV/AIDS. The targets for reductions in child mortality and ma- ternal mortality may also not be achieved in some countries. It should be pointed out, how- ever, that because of the age and epidemiological profile of these countries, proportionately higher gains in life expectancy would accrue from reducing adult mortality through the con- trol of noncommunicable diseases than from achieving targets related to the MDGs. · The low income CIS countries are most severely challenged on nonincome dimensions of poverty. Most MDGs are unlikely to be met in these countries; indeed, only the MDG re- garding attaining gender equity in schooling is on track. (Box continues on the following page.) 200 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union BOX 5.3 (continued) Prospects of Regional Countries Achieving the Global MDGs MDG1 MDG2 MDG3 Increase Increase Reduce school equality poverty enrollment in school EU 8 Czech Rep. Estonia Hungary Latvia Lithuania Poland Slovak Rep. Slovenia SEE Albania Bosnia & Herzegovina Bulgaria Croatia Macedonia, FYR Romania Serbia & Montenegro Turkey Middle income CIS Belarus Kazakhstan Russian Fed. Ukraine Lower Income CIS Armenia Azerbaijan Georgia Kyrgyz Rep. Moldova Tajikistan Uzbekistan Key Likely MDG target likely to be achieved Maybe Too hard to tell whether MDG target will be met or not Unlikely MDG target unlikely to be achieved. No data Inadequate data to predict whether or not MDG target will be met Source: World Bank 2005c. Prospects for Poverty Reduction 201 MDG4 MDG5 MDG6 MDG7 Reduce Reduce Reverse Increase child maternal HIV/AIDS & water mortality mortality TB incidence access 202 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union inhibit progress toward this goal and, in some cases (such as Georgia), maternal deaths are increasing rather than decreasing. The HIV/AIDS MDG represents a challenge in all subregions. Although absolute numbers may be lower than in other parts of the world, the Region is seeing one of the most rapidly growing epidemics in the world. There is a pressing need to improve the effectiveness of disease control through epidemiological and behavioral surveillance systems that can identify the status and trends of HIV infection and its determinants. Denial, stigma, and the institutional challenges of pro- viding services to marginalized and vulnerable subpopulations, such as injecting drug users and persons with disabilities, need to be addressed to combat the spread of HIV/AIDS in the Region. Access to water and sanitation is also an issue in most parts of the Region outside the EU-8. Low investment in, and poor maintenance of, water infrastructure significantly reduced access to sustainable, improved, and safe drinking water since 1990, as well as access to improved or adequate sanitation. Drinking water frequently does not meet biological and chemical standards, supply may be irregular, and there have been outbreaks of waterborne diseases in many countries. As chapter 4 shows, arresting trends toward worsening quality of public services, declining affordability, and exclusion of the poor will require changing the current institutional setup of providing health, education, and infrastructure services. Undertaking such reforms often meets harsh resistance from coalitions of those who benefit from the status quo. Engineering policy change in these areas may prove to be easier in the environment of a rapidly growing economy that generates enough resources to compensate losers than in a more constrained environment. Thus, accelerated shared growth may be an imperative not only for the direct benefits it provides in poverty reduction but also for the indirect impact in opening windows of opportunity to reform the provision of social services. Conclusion Countries of the Region need to accelerate shared growth for contin- ued poverty reduction. Targets for reducing poverty and vulnerability need to be differentiated by country groups, given their current income levels. The low income CIS group needs to generate growth of more than 5 percent a year (at the minimum) to be able to reach even modest targets. Setting more ambitious targets of poverty elim- ination would require equitable and sustainable growth at rates in excess of 10 percent. But even in the middle-income countries, reduc- Prospects for Poverty Reduction 203 ing poverty and economic vulnerability requires growth rates in the range of 6­8 percent, above what is currently forecast for these economies. And all this requires unchanged income distribution; with any worsening of the distribution, even higher growth rates would be needed. The Role for Public Policy It is clear from the above that accelerated and shared growth, along with reform of public service delivery and better targeting of social programs, will be key to making progress on both income and nonin- come dimensions of poverty. It is also important to be able to monitor progress in poverty reduction. Within these four areas, what are the priority actions for public policy? Accelerating Shared Growth It is difficult--based on the experience of the past five years--to overemphasize the importance of raising and sustaining high rates of growth for poverty reduction. As the simulations in overview table 1 suggest, future growth is essential for poverty reduction. The EU-8 is already well placed to take advantage of the new economic opportu- nities and market integration provided by EU accession. Enhanced competition and the mobility of both products and factors of produc- tion that EU accession provides will likely become a dynamic source of growth in the future. This is also true, but perhaps to a more lim- ited extent, for countries with the prospect of accession. But for low and middle income CIS countries that do not yet have such an exter- nal driver for change, domestic catalysts remain crucial. Good eco- nomic governance and responsible leadership must take advantage of the relatively good economic times to put into place policies and insti- tutions that would enhance growth. Understanding the policies and institutions that lead to strong and sustained rates of growth is therefore a first step in reducing poverty. Although this report has less to say on factors that drive growth--which is not the focus of this study--the pursuit of sound economic policies are a necessary precondition. These include sound monetary and fiscal policies (reflected in, for example, moderate-size government and low inflation), a climate conducive for investment, a relatively well-devel- oped financial system, and trade openness. Countries of the Region are, with few exceptions, relatively well integrated into world markets, although more can be done (World Bank Forthcoming-e). 204 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union However, beyond these broad issues around promoting growth, the diagnosis in this report points to a number of areas where more could be done either to increase the assets of the poor or to create greater returns from their assets. These relate to (a) further reforms of the enterprise sector, (b) reforms to agriculture, and (c) promoting opportunities for those in lagging towns and regions. The report con- siders each in turn. Further reforms of the enterprise sector. The failure to reform the enter- prise sector not only in the CIS but also in SEE and parts of the EU-8 is an important underlying cause of poverty in the Region. Differ- ences in income (consumption) between poor and nonpoor are due to several factors such as household dependency, labor force partici- pation, earnings, and other factors (access to social transfers). How- ever, the most important factor by far is earnings, with earnings per employed substantially lower in poor than nonpoor households. The education profile of workers in poor households helps explain part of the earnings gap. The analysis of earnings functions suggests that a worker with primary education faces a wage disadvantage of 20­40 percent across countries of the Region, compared with a worker with secondary education (Yemtsov, Mete, and Cnobloch 2005). Gender is another source of wage differentials in the Region. But even when controlling for the main observable characteristics that distinguish poor and nonpoor households, one is still left with a large "unex- plained" gap in earnings, which is a key determinant of poverty. This unexplained gap is greater than that suggested by estimates from other regions (see, for example, Vandycke 2001; also Munich et al. 2005). One reason for this large unexplained gap is that two observa- tionally equivalent workers in the Region may have very different pay rates, depending on which type of firm they work in. Encouraging the growth of new, more productive firms and strengthening the financial discipline for existing enterprises continues to be important. The typical economy of former socialist countries con- tinues to face significant productivity differences across old, restruc- tured, and new firms within the same sector (World Bank 2002h). New firms are typically the most productive, reflecting not just the more effi- cient use of resources but also the relative dynamics of different kinds of firm and the very different policy environment in which they func- tion. For example, old firms are subject to relatively soft budget con- straints; restructured firms face harder terms, but still carry the legacy of former methods of organization and management; while new firms--which face hard budget constraints--are able and forced to adopt best practices. Many of the poor are trapped in the old, unre- Prospects for Poverty Reduction 205 structured, low-productivity firms. The role of productivity differentials at the micro level, which can in turn be related to a number of institu- tional factors, is unique to the transition process and is a very important factor underlying the phenomenon of the working poor. Reform of the enterprise sector and of the business climate as a whole to create a level playing field across all firms and--in particular--to encourage the entry and growth of new firms is thus an important factor for equalizing the returns to labor and reducing poverty. Boosting agriculture growth and productivity. Many of the poor in the Region are in rural areas, where poverty is proving more resistant to growth than in urban areas. Agriculture is the main activity in rural areas; thus, stimulating agricultural growth is crucial for poverty reduction. Because many of the Region's countries have completed their land reforms, poverty reduction from this type of reform can be expected to be limited in the future. It is worth pointing out, however, that where land reforms have been implemented, espe- cially where initial conditions favor labor-intensive cultivation (for example, low income CIS), land distribution resulted in significant one-time productivity and income gains to rural households. These gains are no different from what was experienced outside the Region (for example, in China in the late 1970s and in Vietnam in the mid-1980s). Indeed, the sharp reduction in rural poverty in the Kyrgyz Republic over the period of this study may be partly attrib- uted to this factor. The more capital-intensive nature of agriculture in other parts of the Region means that where incomes have grown, factors other than land distribution have been important. In many EU-8 countries (for example, the Czech Republic, Estonia, Hungary, and the Slovak Republic), large-scale privatized farms have experienced rapid pro- ductivity growth in large part through labor shedding. But nonfarm growth and a generous social safety net have absorbed excess labor and supported income growth and poverty reduction. In contrast, much of the reduction in rural poverty in the middle income CIS countries (for example, Kazakhstan and Russia) in recent years is due primarily to greater overall liquidity and growth in the economy, which have pushed up agricultural wages and income. Productivity gains have been relatively weak because of limited land reforms and remaining weaknesses in agriculture markets. SEE countries have also failed to see strong productivity growth. This is due to incomplete farm restructuring on account of deficiencies in land markets stem- ming from the restitution process and imperfections in input and credit markets. 206 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Identifying and addressing the key market imperfections in input and output markets are essential for enabling self-employed farmers to lift themselves out of poverty. This is true for all countries. In par- ticular, the integration of rural areas into national credit markets is critical for further investments and productivity growth in agricul- ture. More broadly, improving the investment climate in rural areas is very important. Increasing evidence shows that investments in food processing, agribusiness, trade, and retail companies play a cru- cial role in helping small farmers overcome input and output market imperfections, in helping them upgrade the quality of their products, and in accessing markets (World Bank 2005a). It is also important to promote land reforms, where they are lagging (for instance, in mid- dle income CIS countries), and improve land markets (for instance, in several SEE countries) to facilitate land restructuring. Integration of the rural poor into national labor markets, either through rural off-farm employment or by improving access to urban labor markets and adequate social safety nets, will be crucial for sustained income growth and poverty reduction, particularly in the middle income CIS and SEE countries. Emphasis on rural service delivery and infra- structure is also critical, especially in the low income CIS, not only for its instrumental role in raising rural incomes but also as an aspect of poverty that warrants attention in its own right. Promoting opportunities in lagging regions. Countries in the Region face substantial differences in poverty rates between urban and rural areas and between capital cities and smaller towns. This is not just the case with large countries such as Russia or Ukraine but also with small countries such as Hungary or the Kyrgyz Republic. Although some differentiation is only to be expected, countries need to consider whether, and to what extent, spatial and regional inequalities risk perpetuating intergenerational poverty and inequality traps and act as a drag on economic growth. To the extent that there are implica- tions for equity or efficiency or both, countries need to ask what more can be done to address regional disparities. Most countries seek to address regional inequalities through the maintenance of a stable macroeconomic environment, the creation of a level playing field for businesses, social safety nets, and fiscal trans- fers for targeted programs in lagging regions. But more can be done. First, countries need to enhance labor mobility. When people move to economic nodes that promise a higher expected income, it helps to reduce spatial income disparities. This could be through internal or external migration. Unfortunately, data on internal migration are scarce, but anecdotal stories suggest that the lack of access to housing Prospects for Poverty Reduction 207 in destination areas--itself a result of inadequate housing policy development--credit constraints, and difficulties in accessing benefits and other entitlements and social services in the new location pose a serious constraint to people seeking opportunities to improve their livelihoods. That people are willing to move is evident, for instance, through the high migration from low income CIS countries such as Armenia and Moldova to neighboring EU countries and to Russia. Job-search assistance is also known to be highly cost-effective in matching workers to jobs. Adoption of appropriate policies to encour- age movement, supported by the development of urban housing mar- kets and policies, credit markets, and entitlement reform can provide a strong stimulus to inter- and intraregional mobility and help improve income levels in relatively poorer areas while also boosting competition, productivity, and growth in destination areas. Second, in countries with decentralized fiscal systems, fiscal trans- fers from the center to lagging regions have to be part of the overall package of measures to address regional disparities and promote equalization. In the EU-8, regional disparities are the focus of regional policies promoted by the EU. These provide large transfers to impov- erished regions for rural infrastructure, among other things. Although the evidence is mixed on the results of such policies in reducing regional inequalities (Boldrin and Canova 2001; Funck and Pizzati 2003), the scale of such investments may also not be replicable in other parts of the Region for fiscal and other reasons. Different trans- fer mechanisms could be used such as transfer formulas that provide weights for both equity and efficiency considerations. There may also be a role for introducing competitive allocations to support innova- tive schemes to improve local economic and social conditions, espe- cially those with strong local ownership. Transfers may also be used to improve performance standards in service delivery, especially in the middle income CIS countries and SEE, where budget systems and administrative capacity is stronger. In addition to paying for much- needed improvements, such transfers can also generate demonstra- tion effects for successful policies. Third, investments in human capital have to be an integral part of a strategy to boost economic opportunities for those in the lagging regions. In particular, existing spatial inequalities in access to public ser- vices and quality of services provided need to be addressed as a priority. Strengthening Public Service Delivery Ensuring access and quality of education and health care is critical to promoting opportunity for the poor. However, to improve access and 208 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union quality, it is essential to improve accountability in these sectors. The Region's countries inherited a good network of education and health services, and the erosion in access (or utilization) that had been observed during the 1990s has been partially reversed. However, much remains to be done. Although low levels of spending are an issue, more so in education than in health care, only a few countries spend less than is warranted, given levels of income. Thus, going for- ward, most countries will need to operate within the available resource envelope. Reforms will therefore have to focus on improv- ing the quality and efficiency of public spending. Enhancing education quality and equity. In education, the low income CIS group needs to stem the decline in primary enrollments and qual- ity of education, in particular by ending the situation in which staff are underpaid and complementary expenditures (on textbooks, heat- ing, and repairs) are underfinanced, while at the same time employ- ment and, in some cases, facilities remain well above standards common in much richer countries. In addition, some countries may need to ensure greater equity in education spending across subna- tional regions (for example, the Kyrgyz Republic). Ensuring access to primary education is much less an issue outside the low income CIS countries. Here the main issue is secondary education, where quality and relevance to market demand are often in question. Governance reforms that both strengthen government accountability for out- comes as well as increase participation and voice will be essential to improving outcomes. Lessons from the experience of the EU-8 in raising quality certainly point in this direction. In particular, decen- tralization of services to allow for a greater role for both school admin- istrators and parents has an important role to play in stemming declines in quality. Strengthening access to, and quality of, health care. In health care, low- income countries suffer from having to provide for a range of services when budget resources are limited, but even the available allocations are not spent wisely. This is reflected in the large share of household contributions in total health spending. Improving utilization among the poor is closely linked to financing and quality issues. Tough deci- sions are required on the size of the basic package and a major reallo- cation of expenditure--and greater accountability for its use--to improve access to, and quality of, care (World Bank 2005c). To be sure, public budgets are limited, but the failure to fund a basic pack- age of services for all citizens is much more a result of the failure of citizens to exercise their voice and the ability of politicians to remain Prospects for Poverty Reduction 209 largely unresponsive to poorly expressed demands than resource lim- itations per se. The available allocations are not spent wisely because of the failure of compacts between politicians and providers to create and sustain an environment of incentives in which providers are induced to serve the poor and needy. To improve matters, accounta- bility relationships between politicians and citizens need to become more effective (through such means as more organized voice power of citizens, citizens' report cards, and informed voting), and the accountability relationships between politicians and providers need to be strengthened (through such means as clarifying responsibility, aligning incentives between policy maker as principal and provider as agent, and better enforcement of contracts between organizational and front-line providers). Countries such as Armenia have shown that even with limited resources and high poverty rates, improve- ments in key dimensions such as affordability can be made, albeit at a moderate scale. At the other end of the spectrum, the EU-8 is struggling to main- tain the easy access to a wide range of health services in a context of rising costs (driven up both by aging populations and costlier medical technology). Many countries, like the Slovak Republic, are respond- ing to this situation by reducing the scope of services covered by social health insurance and introducing formal patient copayments in a bid to reduce demand; others, like Estonia, have been fairly successful in finding efficiency gains through reducing hospital beds. Clearly, fur- ther efficiency-enhancing mechanisms will need to be found to con- trol expenditures. More direct attention on quality of services is also warranted, given the growing affluence and expectations of the pop- ulation. Many countries may need to look to the private sector for financial contributions and managerial expertise to take some pres- sure off public provision and improve service delivery. Managing reform of utilities. A looming issue in all countries in the Region is the likely social impact of the increase in the price of utili- ties that will need to be undertaken to attract much-needed invest- ment in the sector. The Region inherited a large network of infrastructure services, including power, water, gas, district heating, and other municipal services, much of which is eroding from lack of maintenance. Service quality is extremely poor in the low income CIS countries, and even within richer countries there are large dis- parities between service quality for the poor and the nonpoor. The infrastructure needs of the poor are unlikely to be met without reform of the utilities sector to bring it to a financially self-sustaining basis, which would encourage much-needed upkeep and maintenance of 210 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union viable infrastructure and improvements in service quality. The EU-8 may well be able to look to the private sector for financial contribu- tions and managerial expertise to improve service delivery. However, in the CIS and SEE, the private sector may well not invest because of regulatory and other uncertainties. Whatever the model, the public sector has a key role to play, either as a regulator or as a provider of services, in improving financial performance and service quality.2 Improving financial performance will involve raising tariffs, which--except for power, where there has been some movement toward cost recovery, and possibly water in the EU-8--are well below cost recovery levels. Further movement toward full cost recovery in power is expected to have a limited impact on poverty, except in the poorest countries. However, across-the-board increases in the full range of utilities is expected to have a more serious impact on poverty (World Bank Forthcoming-b). The social impact of tariff increases will need to be factored into the sequencing and pace of reforms in the event of across-the-board reforms in a range of utili- ties. Where the social safety net is adequate, it can be expected to mitigate the impact on the poor. However, where social safety nets are relatively thin, as for example in the low income CIS group, other options are worthy of consideration. For utilities that can be metered, life-line tariffs can serve as a useful temporary cushion; however, where life-lines are not practical (as, for example, in sec- tors where consumption cannot be measured), reforms would need to be calibrated to affordability. Enhancing Social Protection Strengthening the social safety net. Social safety nets are typically judged using three criteria: coverage, targeting, and adequacy. As shown pre- viously, the system of public transfers in the Region has an important role to play in reducing poverty. Evidence on trends, where available, suggests that these programs cover the poor increasingly well. In EU- 8 countries such as Hungary, nearly 100 percent of the poor receive public transfers of one sort or the other. Many SEE countries also come close to this level of coverage. Coverage is lower in the CIS, but even in the poorest CIS countries, coverage rates are high, exceeding 50 percent of all poor. These levels compare favorably with coverage rates elsewhere. Where data on targeting performance of poverty- focused programs are available, they suggest that targeting perform- ance is improving, albeit at slower rates than one would have hoped. Adequacy varies from the most generous schemes in the EU-8 to the least generous in the low income CIS countries. Prospects for Poverty Reduction 211 Given their importance for poverty reduction and the broad improvements over time, it should be clear that countries need to maintain ongoing social insurance and social assistance reforms, which are largely designed to improve sustainability, and to enhance coverage and targeting of the poor within the available resource enve- lope. In the low income CIS group, the main constraint will continue to be the fiscal means to cover the population adequately. In the mid- dle income CIS group and SEE, although there is more fiscal space for social protection, there is also greater resistance to reforms, as sug- gested (for example) by the difficulties with the monetization of priv- ileges in Russia. Although the objective of the reforms is not in question, the difficulties in implementation serve as a useful reminder of the importance of sequencing with other social and economic reforms, the need to protect the most vulnerable groups, and an appropriate communications strategy to explain the benefits of reforms. Where systems are more generous, as for example in parts of SEE and the EU-8, a balance will need to be struck between the need for social protection and labor market incentives. Strengthening targeted interventions for marginalized groups and minorities. Marginalized groups--typically with tenuous links to the labor mar- ket--are often the hardest to reach and require targeted interven- tions. This may be in the form of assistance in cash or in-kind (such as education, health, or housing), depending on the nature of the group. For the long-term unemployed or nonparticipants, active labor market programs can be particularly relevant. But evidence from suc- cessful training programs suggests that these should be targeted, offered on a selective basis, with clear links to potential employers, and in collaboration with the private sector. It is important to bear in mind that there is limited evidence of successful retraining programs from the low income CIS group. Another group that may require tar- geted interventions is minorities. In this case, interventions may need to be integrated across many fronts. For example, in the Roma minor- ity of the EU-8 and SEE, governments are taking a holistic approach to ending persistent exclusion by setting goals in four areas--educa- tion, employment, health, and housing. Other minorities may require a different approach. The elderly, especially those who are very aged or living alone, may also require special interventions. Many coun- tries make supplementary cash benefits available to them. Other options include provision of assisted living services. For most margin- alized groups, however, additional assistance, whether in cash or kind, need not be provided by the public sector alone. Civil society organizations, community-based groups, and other organizations 212 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union could also be encouraged to come into the sector under the overall direction of the government. Ensuring adequate minimum wages. Minimum wages are an important policy instrument for enhancing the income security of the poor. These can help provide a floor to income, but need to be kept at a rea- sonable level. In this context, the large real increases in the CIS in recent years have brought minimum wages closer to subsistence lev- els. However, future increases in minimum wages need to be consid- ered carefully, so as not to become so high that they have negative effects on growth and employment, with adverse impacts on poverty. Where variations in regional income and labor market profiles are large, governments may need to consider setting region-specific min- imum wages, which may help improve the employability of certain groups of workers (such as younger workers and those in lagging regions). This is an issue particularly in the EU-8 and SEE, where minimum wages represent a relatively high proportion of the average wage and the adverse impact of common minimum wages is particu- larly noticeable. For example, the relatively high minimum wage is found to constrain employment opportunities for the low-skilled in countries such as Lithuania and Poland. Monitoring Progress on Poverty Despite great progress with improving the quality and accessibility of poverty data, there remain significant outstanding challenges. Coun- tries need good survey data to monitor changes in poverty and to eval- uate the impact of specific policy actions on the poor. This report documents huge progress achieved in collecting up-to-date high-qual- ity data across the Region. The previous report on poverty (World Bank 2000a) relied on a single survey for many countries and could produce an estimate of poverty over time for only three countries. With full data sets closed to users outside statistical offices, the report also had to rely on partial data. Since then, many countries have started implementing regular surveys that periodically collect representative data on income and nonincome dimensions of living standards. In addition, data are provided openly to researchers for the purposes of study and policy evaluation. These improvements are not only confined to EU-8 coun- tries (for example, Hungary) but also cover SEE (for example, Roma- nia), middle income CIS countries (for example, Kazakhstan), and low income CIS countries (for example, Georgia and Moldova). Despite this progress, many outstanding challenges remain. First, improvements to data quality and availability are very recent, and for Prospects for Poverty Reduction 213 many countries in the Region, reliable data on poverty changes can be obtained for only a few recent years. The efforts in collecting data need to be maintained. Second, survey coverage and response rates have fallen over time in all countries; and there is a need to strengthen the technical capacity of statistical offices to curb this trend and deal with it appropriately. Third, wide gaps exist in data collection on the nonincome dimensions of poverty: there are practically no attempts to gauge trends in the quality of health care and infrastruc- ture services, and even indicators of access are not consistently col- lected. Given the changing nature of poverty with the increasing role of nonincome components, this gap is the most worrying. Fourth, not all countries have opened their data sets to researchers, undermining the effective use of public funds spent on data collection and moni- toring. These areas--keeping up with periodic surveys to provide comparable data, collecting information on nonincome dimensions, and opening up access to survey data--are priorities for action to ensure adequate information support for poverty reduction efforts. Conclusions The countries in the Region have made significant progress in reduc- ing poverty in the past five years. More than 40 million people moved out of poverty during 1998­2003. Much of this poverty reduction derives from the growth rebound in the CIS countries. But poverty and vulnerability still remain a significant problem: more than 60 million are poor, and more than 150 million are vulnerable. Most of the poor are the working poor. Many others face depriva- tions in access and quality of public services. Regional inequalities both between and within countries are large. The highest levels of absolute poverty are found in poor countries of Central Asia and the South Caucasus, but most of the Region's poor and vulnerable are in middle-income countries. Notwithstanding the tremendous heterogeneity among countries in the Region, reducing poverty and vulnerability requires an acceleration of shared growth, strengthening of public service delivery, better target- ing of social protection, and regular monitoring of progress in poverty reduction across the Region. While much is being done, a number of areas deserve further attention. In promoting accelerated shared growth, there is a need to (a) further reform the enterprise sector to encourage the release of resources from the old, less-productive sectors to the new, more-productive sectors; (b) boost agricultural growth and productivity, especially by addressing remaining imperfections in input and output 214 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union markets, and integrating rural areas into the rest of the economy with regard to capital markets, credit markets, and services; and (c) promote greater opportunity in lagging regions. Public service delivery needs to be improved through increasing the accountability of government and the voice and participation of citizens. This is essential to improving access and quality of social services, which are not only important in their own right but also of instrumental value in helping the poor move out of poverty. There is also the continuing need to further strengthen the social safety net to meet the challenges of restructuring economies. Finally, monitoring progress on poverty reduction on a regular basis needs good-quality data sets that are publicly available. The analysis in this report is based on an extensive multiyear effort to put together a comparable database on the income and nonincome dimensions of poverty in the Region. Data by country for all available years on the different dimensions of poverty are presented in the appendix to the study. It is hoped that this database, these tables, and the accompanying analyses, combined with the attempt to bench- mark the former socialist countries in the Region with other interna- tional comparators (Colombia, Turkey, and Vietnam), will prove to be a useful resource both within and outside the Bank. At the same time, the data-gathering efforts point to a number of shortcomings in the data, not only in basic consumption and income measures but also in access to, and quality of, public services. It is hoped by bringing them to the public eye, this publication will begin a process whereby these shortcomings can be addressed. Endnotes 1. Data issues are discussed in detail in Atkinson and Micklewright (1992). 2. Improving the financial performance of utilities would, in many coun- tries, also have the added value of reducing macroeconomic risks from the quasi-fiscal liabilities of the utilities. Appendix A. Data and Methodology Data: Regional Household Survey Archive To arrive at the internationally comparable assessment of poverty, this report uses primary unit record data from recent household sur- veys to construct a comparable indicator of living standards across all countries in the Region. The comparable indicator of living standards created for this study is described in detail below. Based on the indicator, the study team calculated poverty and inequality data and a set of indicators pre- sented in this appendix (B. Key Poverty Indicators, tables 1­10) and used throughout the book. All data are taken from household-level surveys implemented in the Region's countries during 1997­2003. Survey names, years, and basic characteristics are shown in chart 1). The data of the surveys included in the analyses are nationally representative. To ensure rep- resentativeness, most countries, with a few exceptions (Bosnia and Herzegovina, Kosovo), rely on the most recent census of the popula- tion as a sampling frame and use random multistage sampling tech- niques. Some of the surveys, such as the Russian Longitudinal Monitoring Survey (RLMS) and Bosnia and Herzegovina's Living 215 216 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Standards Measurement Study (LSMS) are implemented as panels designed to follow the same households over time. Countries differ substantially in sample sizes used (as revealed by chart 1). Those that intend to collect regionally representative infor- mation have larger samples than countries that aim at representative- ness by location types only (urban-rural). None of the surveys used relies on simple one-stage random samples, so everywhere sampling weights have to be used to arrive at the estimates based on surveys. Survey data were available for 23 countries, although the data do not span the entire period of 1998­2003 in all countries. Those that do have reasonably comparable data over time spanning the entire period are grouped into cluster A on chart 1 and used as the main source for the dynamic analysis in this book. Overall, there are 10 countries for which data are available and comparable (see discussion of comparability in this appendix, "Comparable Consumption Aggre- gate") over the entire period under review, 1998­2003. Several countries are represented by a single survey (year) or by surveys that are not sufficiently comparable to assess trends in poverty over time: Albania, Azerbaijan, and Serbia and Montenegro. To widen the coverage of the study, the team also relied on data that cover a shorter period, but provide useful information with regard to poverty profile or coverage of nonincome dimensions of poverty (cluster B, chart 1). The countries not covered are Croatia, the Czech Republic, the Slovak Republic, Slovenia, Turkmenistan, and the UN Mission in Kosovo (UNMIK) (cluster C, chart 1). Some of these are not covered based on the assessment of their household survey data as inaccurate: in addition to design flaws in the Czech Republic and Slovak Republic, the response rates to Household Budget Surveys deteriorated to the extent that their representativeness is questioned (EC 2004). Other countries (Croatia, Slovenia, Turkmenistan) make access to data extremely cumbersome and are not willing to provide access to the entire data sets--a precondition to construct a compara- ble welfare aggregate. Finally, for UNMIK, many auxiliary data nec- essary to carry out international comparisons (such as reliable price indexes, or PPP exchange rates) are not available, which preclude the use of data for comparison purposes. Finally, three countries were used as benchmarks (Colombia, Turkey, and Vietnam), and the surveys used are also listed in cluster D on chart 1. The choice of these countries was based on several factors. First, it was important to have a broad geographic coverage. Second, benchmark countries had to have comparable levels of living stan- dards to the Region's countries and some systemic features of the economies that resemble transition settings. Third, survey data had to Appendix: Data and Methodology 217 be easily available and include variables necessary to create consump- tion aggregates (see next section). The team of the World Bank World Development Report 2005 helped to identify such data and kindly pro- vided access to them. Turkey, being the only nontransition country in the Region, was an automatic choice; in fact, its geopolitical and eco- nomic situation is very similar to CEE countries. Vietnam, a country in East Asia transiting to a market economy from a planned system, was identified as a good comparator to the poorest countries in the Region. Colombia, with its large informal sector and dependence on natural resources, was a good match for the middle income CIS group. The previous regional report on poverty, Making Transition Work for Everyone (World Bank 2000a), relied on a single survey data set for many countries and could produce an estimate of poverty over time for only three countries (Hungary, Poland, and Russia); it also had to use partial data, with full data sets closed to any users outside statisti- cal offices. In some countries, there was no option but to rely on non- representative data (Albania, Azerbaijan), and data for some countries were missing (Bosnia and Herzegovina, Serbia and Mon- tenegro). Since the end of the 1990s, many countries in the Region have moved to ongoing surveys that periodically (normally every month or quarter) collect representative data on income and nonincome dimensions of living standards. Because such data collection is often too costly, some countries rely on one-time comparable surveys con- ducted every three to five years. In both cases, data generated can be used to monitor poverty over time. In addition, countries in the Region started to use collected survey data better by providing open access to researchers trying to understand poverty and its causes and to assess public policies. Not only did the middle-income countries in the EU-8 (such as Hungary) achieve the greatest improvements in collecting data and providing open access but also some countries in SEE (such as Bosnia and Herzegovina and Romania), middle income CIS countries (Kazakhstan), and even low income CIS countries (Georgia and Moldova). However, many countries continue to lag behind the leaders in both adequacy and openness of data, as reflected by the survey description provided on chart 1. Some countries are still struggling to start continuous data collection on living standards, or they change the survey design too often to make any comparisons over time (Alba- nia, Azerbaijan, Serbia and Montenegro, Turkmenistan, and UNMIK). Some countries recently introduced new standards in data collection (following recommendations of the Statistical Office of the European Communities [Eurostat]), and the new surveys are not comparable 218 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union with the previously collected data (Bulgaria, Croatia, Estonia, and FYR Macedonia). There are countries in which household surveys of living standards suffer from extremely low response rates, to the point of making them not fully representative (the Slovak Republic). Unfortu- nately, most countries in the Region still do not provide open access to household survey data. They either impose many restrictions on data users and require significant access fees or simply treat household sur- vey data as closed-access information that is not available outside sta- tistical offices, even in the anonymized form. Normally, survey data sponsored by international agencies are open-access information. Such data sets (in the chart, labeled "Open") are simply downloadable through the Internet free of charge or with a minimum fee (after provision of sufficient information on the user). Of course, as with any data set in the public domain, all personal information that would allow identification of individual respondents is removed from such data sets. The Household Budget Survey (HBS) can also be open access and easily accessible to any individual or organization, either by a request that is reviewed by the staff of the statistical office and payment of service fees (in the chart, labeled "Limited"), or it can be restricted to a narrow set of users, sometimes exclusively to statistical offices. Some of the data sets included in the study are not freely available; the World Bank used these data with the permission of the countries' authorities solely for the comparative poverty data analysis presented in this report (in the chart, labeled "Restricted"). None of the data sets used in this report will be disseminated to any user, given the confidentiality and rights restrictions on most of them. Countries differ in types of household data available. In most coun- tries, there are official household surveys that collect information on expenditures and incomes. Such detailed surveys of household budg- ets, with limited information on other aspects of well-being, are classi- fied as Expenditure and Income (EI) Surveys (Household Budget Surveys, in a narrow sense). Nevertheless, such one-topic surveys also normally collect detailed labor market information alongside informa- tion on the characteristics of household members and household con- ditions. Other nonincome indicators are not normally included in such surveys. If they are included, and information on the health status of household members, their social participation, and access to education is also collected, these are classified as Integrated (IN) Surveys. Multi- topic design is a standard inherent feature of a different type of survey, designed and implemented as part of the Living Standards Measure- ment Study (LSMS) of the World Bank (see www.worldbank.org/lsms for details). Several countries in the Region undertook LSMS surveys Appendix: Data and Methodology 219 and rely on them for poverty monitoring (Albania, Bosnia and Herze- govina, Tajikistan), but the mainstream approach is clearly the use of the HBS integrated with additional modules to capture nonmonetary aspects of living standards. The LSMS survey and the HBS rely on two different approaches and types of instrument to collect detailed expenditure and income information from households: the LSMS survey is a recall question- naire asking a household to remember all expenditures over a certain period (usually the past quarter); the HBS is a diary of purchases and incomes that a household has to fill in daily over a certain period (usually two weeks). Sometimes these two types of instrument are combined within the same survey. It is clear that the nature of the data collected by each of the instruments will differ. Unfortunately, few studies are available that compare the relative strengths and weaknesses of both instruments to assess possible biases in each type of data (all of these studies are from outside the Region), so they are treated as comparable. The recall data are normally collected by an integrated type of sur- vey supported by the World Bank: the LSMS program, with some degree of standardization and common features of data collection. LSMS surveys are particularly prevalent in the developing countries. The HBS programs across transition countries have benefited from large flows of international technical assistance (Eurostat and the World Bank being particularly active), also with a useful unification and standardization. Notwithstanding the progress achieved in data quality and accessi- bility, chart 1 documents significant gaps. First, improvements to data quality and availability are very recent, and for many countries in the Region, reliable data on poverty changes could be obtained for only a few recent years. The efforts in collecting data should be maintained to provide policy makers with data on the evolution of poverty and inequality in the future. The survey coverage and response rates have universally fallen in all countries, and there is a need to strengthen the technical capacity of statistical offices to curb this trend and deal with it appropriately. Second, this report documents wide gaps in data collection on non- income dimensions of poverty: there are practically no attempts to gauge trends in the quality of health, education, and infrastructure services, and even indicators of access are not consistently collected. Many countries rely on one-topic surveys, and integrated survey design is not yet mainstreamed. Moreover, there is a worrying ten- dency in some countries to move away from multitopic design to a narrowly focused Expenditure and Income Survey. Given the chang- 220 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ing nature of poverty with the increasing role of nonincome compo- nents, this gap is the most worrying with regard to collecting relevant and useful data. Finally, not all countries have disseminated their data and opened data sets to researchers, undermining the effective use of public funds spent on data collection and monitoring. These areas--keeping up with period surveys providing comparable data, collecting informa- tion on nonincome dimensions, and opening up access to survey data--are priorities for action to ensure adequate information sup- port for poverty reduction efforts. Comparable Consumption Aggregate To examine poverty and inequality, one needs a measure of material well-being. Ideally, this measure should correspond as closely as pos- sible to the way a person experiences his or her standard of living. It is natural to think that a person's standard of living, or material well- being, is a function of all goods and services consumed by that person. But how can one compare different individuals consuming differ- ent quantities of various goods? Economic theory allows one to rank levels of well-being using the cost (monetary value) of the consump- tion bundle consumed in a given period. In theory, any welfare meas- ure should include all of the factors (including health, leisure, social capital, and other desiderata) that contribute to welfare. In practice, however, because of measurement and valuation difficulties, the focus in microdata analysis is on only material well-being, using infor- mation on consumption of goods and services by a household. Even such "simple" measures are, in practice, quite complicated to capture well, and there is debate as to whether income or consumption is the preferable measure (see Deaton and Zaidi 2002). Income is often considered to be the preferred measure because it is an indicator of the "potential" to enhance welfare (including nonmaterial aspects such as leisure). Income data are used by Eurostat to compile EU statistics on poverty and social inclusion, using the integrated sur- vey model that is intended to provide comparable data within the EU: the European Community Household Panel (ECHP) and the European Community Statistics on Income and Living Conditions (EU-SILC). However, measuring income suffers from several defects, both in theory and in practice. First, income can be highly volatile, whereas consumption can be, and is, more readily smoothed by individuals. This smoothing makes consumption a better indicator of welfare than income, because it more accurately represents the welfare level of an Appendix: Data and Methodology 221 CHART 1 Data Sources Country Survey Name Year Access Policy Type Sample A. Countries with Extended Time Series of Comparable Household Data Armenia Integrated Living Conditions Survey 2003 Limited IN 4,600 hh Integrated Living Conditions Survey 2001 Limited IN 4,037 hh LSMS-Integrated Survey 1998/99 Limited IN 3,600 hh Belarus Household Income and Expenditure Survey 1998­2002 Restricted EI 4,882 hh Georgia Integrated Survey of Georgian Households 2000­2003 Open IN 2,800 hh Survey of Georgian Households 1996­1999 Open EI 2,800 hh Hungary Household Budget Survey 1998­2002 Open EI 10,200 hh Lithuania Household Budget Survey 1998­3 Limited EI 7,111 hh Moldova Household Budget Survey 2003 Open IN 4,592 hh Household Budget Survey 1998­2002 Open IN 6,159 hh Poland Household Budget Survey 1998­2002 Limited EI 31,708 hh Romania Family Budget Survey 1998­2003 Open IN 32,000 hh Russian Fed. Household Budget Survey 1997­2002 Restricted EI 49,000 hh RLMS (9-11 rounds) 2002, 01, 1998 Open IN 3-4,000 hh Tajikistan Living Standards Survey (LSMS) 2003 Open IN 4,160 hh Living Standards Survey (LSMS) 1999 Open IN 2,000 hh B. Countries with Limited Time Series Albania Living Standards Survey (LSMS) 2002 Open IN 3,600 hh Azerbaijan Household Budget Survey 2002­3 Restricted EI 8,157 hh Bosnia & Herzegovina Living in BiH Panel 2004 Open IN 3,000 hh BiH Living Standards Survey 2001 Open IN 5,402 hh Bulgaria Household Budget Survey 2003 Limited IN 3,000 hh Integrated Household Survey 2001 Limited IN 2,633 hh Living Standards Survey (LSMS) 1995,1997 Open IN 2,322 hh Estonia Household Budget Survey 2000­3 Limited EI 4,600 hh Kazakhstan Household Budget Survey 2001­2003 Limited EI 12,000 hh Kyrgyz Rep. Household Budget Survey 2001­3 Limited IN 2,857 hh Household Budget Survey 2000 Limited IN 1,894 hh Latvia Household Budget Survey 2002­3 Limited EI 3,600 hh Macedonia, FYR Household Budget Survey 2002­3 Restricted EI 4,100 hh Household Budget Survey 1996­2000 Restricted EI 1,025 hh Serbia & Montenegro Serbia Living Standards Survey 2002­3 Limited IN 6,400 hh Ukraine Household Budget Survey 2002­3 Restricted EI 9,646 hh Uzbekistan Household Budget Survey 2001­3 Restricted EI 9,600 hh C. Countries with Outdated or Limited Availability (Not Used in this Report) Croatia Household Budget Survey On going Restricted EI 3,123 hh Czech Rep. Household Budget Survey On going Restricted EI 3,650 hh Slovak Rep. Household Budget Survey On going No Access EI 1,640­4,700 hh Slovenia Household Budget Survey On going No access EI 2,577 hh Turkmenistan Living Standards Measurement Study 1998 Open IN 2,099hh UNM Kosovo Household Budget Survey Kosovo 2003 Restricted EI 2,800 hh Living Standards Survey (LSMS) 2000 Open IN 2,880 hh D. Benchmark Countries Colombia National Survey of Living Standards 2003 Open IN 23,000 hh Turkey Household Income and Consumption Expenditure Survey 2002 Restricted EI 9,555 hh Vietnam Vietnam Living Standards Survey 1997/98 Open IN 6,000 hh Note: Types of survey: EI- Expenditure and Income survey; IN- integrated surveys, Access Policy: Open ­ data are downloadable for free or limited fee from the statistical office with minimum restraints; Limited ­ policy provides access to data or sets of data to researchers or organizations meeting certain criteria; Re- stricted- data were made available only to the World Bank/selected agencies on exceptional basis; No access ­ no access to data from outside statistical office. 222 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union individual at any given time. In transition economies, people are paid very irregularly, with several months of wage arrears being common. In this context, relatively steady consumption-based welfare meas- ures give a more accurate picture than often erratic income-based measures. Second, regardless of the measure, it is essential that it be compre- hensive, that no aspect of income or consumption be omitted. Other- wise, erroneous conclusions may be drawn about the numbers and characteristics of the poor. If, for example, the value of home-pro- duced food were omitted from an income aggregate (total income measure), then rural populations would look much poorer than they actually are. Or if a consumption aggregate is constructed using only monetary expenditures, those who receive in-kind benefits from employment would look poorer than they actually are. Measurement problems are more severe in transition countries for income than for consumption. Income underreporting is common for many reasons, including sometimes because survey respondents are not willing to fully disclose illegal or semilegal income sources. Expe- rience in many countries showed that households were not willing to provide information on unregistered businesses and informal sector activities. Repeatedly, practical experience suggests that the quality of consumption-based data obtained from households is better than the quality of income-based data. At the top end of the income distribu- tion, households tend to underreport their income, reflecting a lack of faith in the confidentiality of the survey, concerns about the tax authority, complexity of earnings that would lengthen an interview, and the like. At the other end of the income distribution, the problem is less one of willingness to provide accurate data and more one of inability to do so. Households engaged in informal activities or with household businesses of a subsistence nature often cannot separate out what is "household" income and what is "business income," thus undermining the reliability of the data collected. This specificity of countries in the Region with regard to quality of income data collected through the regular surveys is recognized by Eurostat and by the countries themselves, which continue to use con- sumption expenditure to monitor poverty. It also raises an important question about whether the EU-SILC will provide credible and com- parable data on the context of poverty in transition economies for the new EU member states. In summary, given the difficulties of defining a total welfare meas- ure, the problems noted above with income-based measurement, and the practices of countries in the Region to measure poverty, this report relies on measuring welfare here with consumption. Appendix: Data and Methodology 223 Consumption is being used as a measure of well-being in most poverty assessments undertaken in the Region over the past five years (14 countries covered) and is accepted as the main base to monitor poverty officially in a number of countries (for example, in Armenia, Bulgaria, the Kyrgyz Republic, and FYR Macedonia). Making Transition Work for Everyone, the previous regional report on poverty and inequality (World Bank 2000a), relied on any welfare indicator (for example, income where consumption was not avail- able) supplied with the data to carry out the analysis. This study fol- lows a different approach: it re-creates consumption aggregates from unit record data, using the same set of rules and definitions (see appendix, A. Data and Methodology, chart 2). Why was it deemed necessary? There are significant differences in the details of how aggregates of consumption expenditures are constructed in different countries. The list of items included in consumption expenditures differs across countries (for example, inclusion or exclusion of purchases of durables). The procedure for imputing the value of goods and services consumed in-kind (housing, flow of services from durables, or own food production) differs a great deal. The treatment of outliers is also strikingly different. Finally, different versions of price indexes to cor- rect for regional and intertemporal prices are applied. All of these differences in procedures to construct consumption aggregates imply that some part of observed differences in outcomes could be attributed to differences in procedure and that only to some extent do they reflect real differences. Of course, data comparability depends not only on consistency of processing the data but also on underlying data quality, which may differ. Clearly this aspect was beyond the control of the team, but to the extent that it was possible to set up the data in the most comparable way, it is reflecting a stan- dard approach to international comparative studies. For example, the Luxembourg Income Study (LIS), which relies on income survey data to carry out social welfare comparisons between OECD countries, relies on a set of strictly and uniformly defined rules to construct an income aggregate. In relying on uniformly defined consumption of goods and services by a household as the measure of living standards, there were a num- ber of conceptual and practical issues that needed to be addressed. First, unlike food, consumer durables and housing are consumed over a long time. It is customary, therefore, to include the imputed value of the consumption flow associated with the possession of con- sumer durables (including housing), but exclude the expenditure on 224 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union CHART 2 Structure of Consumption Aggregate Constructed, Percentage Included components, structure Food, beverages, Utilities Transport and tobacco Clothing (w/o rent) Furnishingsa communicationa Education COICOP Divisions I, II III IV V VII,VIII IX Albania 2002 61.7 4.5 10.8 4.3 7.3 1.6 Armenia 2003 72.3 5.5 6.2 2.3 4.1 5.1 Azerbaijan 2003 64.3 7.6 6.0 3.1 6.8 1.0 Belarus 2002 68.1 7.6 6.0 2.8 6.1 1.3 Bulgaria 2002 58.7 3.5 18.3 1.1 7.3 4.7 Georgia 2003 67.7 5.5 5.7 2.7 10.7 1.9 Estonia 2003 42.2 7.4 16.0 3.1 12.0 1.8 Hungary 2002 38.7 7.0 16.2 6.7 15.1 1.6 Kazakhstan 2003 60.5 8.5 10.2 3.5 6.9 2.0 Kyrgyz Rep. 2003 65.0 11.9 8.2 0.5 6.1 1.6 Latvia 2003 41.0 8.5 13.5 2.8 16.0 1.4 Lithuania 2003 44.5 10.3 14.9 2.8 15.6 1.1 Macedonia, FYR 2003 54.2 8.9 12.0 3.6 11.6 0.3 Moldova 2003 66.4 9.0 12.4 2.5 3.5 0.8 Poland 2002 39.8 6.4 19.7 3.3 13.5 1.5 Romania 2003 57.8 6.4 15.6 2.2 10.1 1.0 Russian Fed. 2002 55.8 13.7 8.0 2.3 6.9 1.1 Serbia 2002 60.8 5.9 13.0 3.7 8.1 1.2 Tajikistan 2003 71.2 5.4 6.4 3.3 4.8 4.1 Ukraine 2003 72.2 6.0 9.9 1.0 4.5 1.3 Uzbekistan 2003 72.3 6.8 1.6 2.6 9.6 0.3 Colombia 2002 41.1 6.9 8.5 4.6 14.7 6.3 Turkey 2002 38.8 7.3 14.2 3.8 12.9 1.7 Vietnam 1998 56.0 6.3 3.4 5.7 3.8 5.5 Note: a = excluding durables; b = at 2000 PPP, top and bottom coded; c = ratio to total consumption aggregate; -- = not available. the purchase of these goods. However, for the Region, data availabil- ity limits the application of this approach to all countries. The authors did not, therefore, include estimates of flow of services of durables, nor have they added in durable purchases or rents. Catastrophic health expenditures were excluded from the estimate of current con- sumption on similar grounds. Second, when consumption is used as a measure of well-being, higher consumption should indicate a higher level of well-being. For most consumption items, this correspondence is reasonable; how- ever, for some categories such as health expenditures, this correspon- dence is questionable. As a result, health expenditures were not included as a part of consumption (Deaton and Zaidi 2002). Third, given the significance of spatial differences, the authors adjusted for spatial price differences, employing survey-data-based Paasche price indexes using the same set of information in all coun- Appendix: Data and Methodology 225 Included components, structure Excluded components, ratios Hotels & Recreation All Total consumption, restaurants and othera components $ per capitab Healthc Rentc Durablesc XI X, XII VI 2.9 6.8 100 1,388.00 7.7 0.6 -- 1.2 3.2 100 913.72 11.5 -- 0.8 5.2 6.0 100 1,429.84 3.5 0.5 7.5 1.5 6.6 100 2,704.13 2.3 -- 1.7 0.0 6.4 100 2,248.00 3.1 0.7 0.6 1.7 4.0 100 972.89 5.7 0.3 1.3 3.5 14.0 100 2,752.61 4.1 1.2 7.3 3.1 11.6 100 2,890.00 4.4 0.9 8.1 0.7 7.7 100 1,476.00 2.4 0.2 3.6 1.0 5.6 100 708.00 1.6 0.5 2.6 5.7 11.1 100 3,401.00 3.8 0.9 8.1 2.4 8.4 100 2,762.00 4.7 0.0 10.0 3.3 6.1 100 3,171.00 3.2 0.1 4.1 0.8 4.6 100 1,045.81 4.8 1.0 4.4 2.0 13.8 100 2,611.00 5.5 4.4 6.1 1.5 5.4 100 1,624.00 3.0 0.3 1.6 2.5 9.6 100 2,179.00 2.4 0.8 11.0 1.2 6.1 100 1,992.98 8.0 1.0 1.6 0.6 4.3 100 670.00 5.8 0.2 -- 1.6 3.5 100 2,496.30 2.8 0.4 2.5 3.3 3.5 100 1,042.00 3.0 -- 0.6 4.6 13.3 100 4,398.00 2.9 5.9 4.0 3.0 18.3 100 1,816.00 2.8 3.8 9.1 5.9 13.4 100 1,078.00 18.0 0.2 18.8 tries (see Deaton and Zaidi [2002] for a detailed discussion; also see Price Deflators section below). Fourth, households in the Region cope with poverty by relying on an array of nonmarket strategies, including producing their own food and engaging in reciprocal exchange with other households and insti- tutions. A consistent approach was used in assigning a monetary value to these components of consumption. Median local prices were relied on to impute the value of in-kind food consumption from own agri- cultural activities, and households' own estimations of the value of gifts and transfers in-kind for food and nonfood items were used. Fifth, the same procedure, which conforms to methods used in other international household survey data depositories (such as the Luxembourg Income Study), was used to clean the data of outliers across all data sets. The data are "bottom-coded" at 1 per cent of per capita mean real consumption and "top-coded" at 10 times the median 226 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union of household consumption, following a similar approach to income survey data proposed by Gottschalk and Smeeding (1997, 661). This procedure limits the effect of extreme values at either end of the dis- tribution. The final data set excludes all records with zero consump- tion. This decision is consistent with Atkinson and Micklewright (1995) and with the method used and recommended by the LIS Key Figures reported on the LIS Web site (http://www.lisproject.org/). Because the authors have followed a consistent approach across all data sets, they are reasonably confident that differences across coun- tries in the final consumption measure are due to differences in the primary data and are not due to the method of aggregation. The basic descriptive statistics for all countries covered (latest available year) are presented in chart 2. The constructed consumption aggregate produces a ranking that fits closely with the macroeconomic data (see appendix, B. Key Poverty Indi- cators, table 1 for GDP per capita data). Richer countries tend to have lower food shares. The excluded rent, being a small component, makes little effect on the ranking of households (note that utilities included in the consumption are large, reflecting the climatic and infrastructure fea- tures of the Region). Finally, food share behaves in a standard fashion across deciles of the distribution, falling with higher welfare. (Full details on consumption aggregates' components are provided on the dedicated Web site: http://www.worldbank.org/eca/.) All these features show the validity of the approach chosen to con- struct a measure of living standards. Some countries stand out some- what, and the comparability may not be fully taken for granted. For example, Ukraine has a food share comparable to the poorest coun- tries in the Region, yet its consumption per capita measured at PPP puts the country solidly in the middle-income range. Vietnam, on the other hand, has a rather low food share, which would imply a higher living standard than suggested by other data. However, these peculi- arities, while important, are not in themselves undermining the com- parability of consumption aggregates. There are also some persistent differences between countries in the consumption aggregate as measured by different types of survey; for example, education and health expenditures seem to have much higher shares and ratios in the LSMS surveys than in the HBS. Given that the information on these types of expenditure is collected in LSMS surveys in a contextual section (that is, in a module on the use of edu- cation services) and with much longer recall periods than in the HBS, it is not a surprising outcome. The ongoing research project conducted by the LSMS group intends to answer the question about the effect of dif- ferent designs of data collection instruments on the welfare indicators. Appendix: Data and Methodology 227 The constructed estimate of consumption has several shortcomings that reflect some persistent data problems in the Region. Over time, there has been a considerable deterioration in response rates in many countries. Countries deal with this problem in different ways, which may have (as yet unknown) implications for survey-based poverty and inequality measures. There are other issues (reliability of diaries versus recall estimates and so forth) that are behind the research and are less evident, but affect the quality of data. This is the first time to the authors' knowledge that comparable con- sumption aggregates have been constructed for countries in the Region. Poverty Lines and Purchasing Power Parity This report uses an absolute concept of poverty, which is consistent with a large body of literature from both outside and within the World Bank, in which poverty is seen as the inability to meet basic material needs. Although the notion of basic needs differs across countries, it can be rea- sonably well defined as the current cost of the subsistence consumption basket. In practically all countries in the Region, one finds groups of the population unable to meet such basic needs. This group, and the group who are "nearby" in income terms, are the focus of this book. The alternative measure of deprivation--relative poverty--has also been used in the literature. It is also a mainstay of the official poverty and social exclusion statistics used by the European Commission to monitor the situation in the EU member states (see Atkinson, Mar- lier, and Nolan 2004). However, the difficulties that it creates for monitoring differences across countries and changes over time within countries, combined with the still relatively fragile economic situa- tion in many countries of the Region, make the authors favor the absolute poverty approach. The noncomparability of poverty lines based on the relative concept is also admitted by the EC and Eurostat (Dennis and Guio 2003). An absolute poverty line, as the name implies, attempts to estab- lish the value of consumption that a person needs to stay out of poverty, regardless of time and place. Clearly there are difficulties with doing this. The first widely accepted global poverty estimates, produced by the World Bank's World Development Report 1990, chose a poverty line measured in 1985 purchasing power parity (PPP). Chen and Ravallion (2001) have since updated these numbers, using an expanded database of household surveys based on 1993 PPP exchange rates for consumption. This report uses the most recent PPP numbers from 2000, as reported in OECD (2003). The report uses data from most countries in the Region 228 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union and provides PPP exchange rates between national currencies and the euro. The report also provides the PPP conversion factor from the euro to U.S. dollars of 2000. Thus, these data can be used to convert national currencies to U.S. dollars of 2000, based on their PPP for that year. To make it comparable and relevant for global poverty monitoring, the 2000 U.S. dollars are then converted to the 1993 ones, using the U.S. consumer price inflation index. The final set of factors then represents the amount of national currency needed to buy a bundle of consumer goods that one dollar in 1993 would have bought. To get current value for the survey years, national inflation rates (CPI) are used to inflate (deflate) these 2000 exchange rates. Some countries excluded from OECD (2003) (such as Bosnia and Herzegovina) were estimated based on the PPP from neighboring countries and exchange rates. PPP data for some came from EBRD (Sanfey and others 2004), and data for Albania have been estimated based on the 1996 PPP set. More recent data on PPP are more relevant for the transition economies of the Region because they reflect contemporary (in many cases, liberalized) prices, as opposed to the administered prices of the past rounds of international price comparison surveys. Using 2000 PPPs also provides more plausible estimates of absolute poverty (see chapter 1, annex 1 for a detailed discussion). For example, it is highly implausi- ble that poverty in Uzbekistan is negligible (which is the impression that one gets using the 1993 and 1996 PPPs). Errors can also go the other way (that is, overstate poverty), as appears to be the case when the 1993 PPP is used for Georgia. In addition to issues with relative prices, Geor- gia experienced one of the worst hyperinflations in its history in 1993, which would have made measurement of prices problematic. On one hand, the total poverty headcount for the Region does not change much whether one uses 1993 PPP or 2000 PPP; however, individual country assessments are affected. On the other hand, 1996 PPP, with few exceptions--Bulgaria, Estonia, and Lithuania--pro- duces a lower poverty count than 2000 PPP. The use of 2000 PPP has additional drawbacks. The OECD (2003) reported both consumption and GDP PPP euro exchange rates, but con- version from euros to dollars were provided only for GDP numbers. The choice was made to use GDP 2000 PPP figures (which are also more eas- ily available for countries outside the Region) as a baseline. The full set of the PPP exchange rate estimates used for this study is provided on the dedicated Web site: http://www.worldbank.org/eca/. There is some arbitrariness inherent in setting the level of poverty lines. The absolute poverty line attempts to gauge a standard in a way that is comparable across time and space (Ravallion 1994). The World Appendix: Data and Methodology 229 Bank often uses $1 a day for cross-country comparisons, which (in 1985 PPPs) was chosen around 1990 because it was the most typical poverty line among the low-income countries (later updated to $1.075 a day, using 1993 PPPs). None of the Region's countries was considered when coming up with this estimate. This $1-a-day poverty line has since come to be regarded as providing the absolute minimum standard of living. Much has changed since 1990, particularly in the Region. Com- paring national poverty lines for a group comprising countries of the Region, outside the Region, and in the EU, one sees that, as else- where in the world, there is a close correlation in the Region between the average standard of living and the national minimum needs defi- nition (see box 1.1 for a detailed discussion). However, no country in the Region has a poverty line close to $1 a day. On the contrary, the lowest poverty lines cluster around the $2 mark. (See appendix B, table 2.) The study therefore uses $2 a day (or, more accurately, $2.15, which is exactly double $1.075) as an absolute poverty line for the purposes of this report. The study also uses a higher poverty line ($4.30 a day) as a proximate vulnerability threshold to identify households that are not suffering absolute material deprivation, but are vulnerable to poverty. Although it seems somewhat arbitrary, it does bear some rela- tion to empirically observed vulnerability to poverty. Analysis of panel data from the Region suggests that households with per capita con- sumption at least twice the poverty line face less than a 50 percent chance of becoming poor in the foreseeable future (World Bank 2002). To provide data comparable to other regions and test for robustness of findings based on a single estimated PPP, the study also calculated a full set of poverty indexes (reported in appendix B, Key Poverty Indicators, table 2). For the 1993 and 1996 PPP revisions, they are provided for any interested reader on http://www.worldbank.org/ eca/. There one can also find estimated poverty rates for the Region and benchmark countries with $1-a-day international poverty lines not reported in this book. Price Deflators In the cases in which data were collected over a long period of time, it was also necessary to adjust for changes in prices over time. Quar- terly CPI (IMF) indexes were used to compute real values. This measure ignores the differential impact of price increases on the poor and nonpoor. No price indexes for low-income groups are routinely available in the Region that would allow this study to address this issue. 230 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Regional price differences can cause the same bundle of goods to be more expensive in one region than in another. However, the dif- ference in expenditure caused by these regional price differences does not reflect differences in material well-being. Hence, these regional price differences need to be corrected. As discussed in Deaton and Zaidi (2002), the Paasche price index offers a most reliable way to measure spatial price differences. Note that this index involves not only the prices faced by a household in relation to the reference prices but also its expenditure pattern, some- thing that is not true of a Laspeyres index. The distinction is an impor- tant one: to convert total expenditure into the welfare index, the price index must be tailored to the household's own demand pattern, a demand pattern that varies with the household's income, demo- graphic composition, location, and other characteristics. The refer- ence prices are the median of the prices observed from all households. Based on these deflators estimated for each household, an indicator of regional price levels is calculated, which is then aggregated (using the median) to the regional level in each country. The resulting indexes usually range within 0.9­1.1 of the national price level. Because nonfood prices are usually not avail- able for the Region's countries from statistical offices and unit val- ues for nonfood are normally not collected by household surveys, the spatial price deflator is based entirely on differences in food prices. This clearly is an issue but given data limitations, nothing can be done to deal with it. Equivalence Scales Consumption data from household surveys are collected at the level of the household rather than of the individual; however, to deter- mine the welfare levels of people, total household consumption must be divided among household members. Consumption cannot, how- ever, be explicitly assigned to individual household members using the data. Instead, an adjustment based on some allocation rule must be imposed to attribute their share of a household's resources to indi- viduals within the household. One such allocation rule is simply to divide total household con- sumption by the number of household members. This yields per capita consumption. This report relies on per capita measure of con- sumption, and, as discussed in Deaton and Zaidi (2002), it is a rea- sonable choice. This is the most commonly applied method, and it implies that all family members receive an equal share of household resources. For some findings that are sensitive to the equivalence-of- Appendix: Data and Methodology 231 scale assumptions (especially demographic profiles), the robustness checks have been carried out, and the results reported in the study. Poverty Indexes The sections that follow describe a set of poverty indicators to assess the extent of the deprivation of individuals in the income dimension. Measuring Poverty (appendix B. Key Poverty Indicators, table 2) The simplest and most commonly used measure of poverty is the headcount index, which is given by the fraction of individuals with equivalent consumption below the poverty line (Foster, Greer, and Thorbecke 1984).1 This measure, however, does not show whether the poor are only slightly below the poverty line or whether their consumption falls substantially short of the poverty line. The head- count measure also does not reveal whether all the poor are about equally poor or whether some are very poor and others just below the poverty line. To examine these three dimensions of poverty--headcount, short- fall, and inequality among the poor--an FGT class of poverty meas- ures is used. This class is described by the following equation: a 1 n Č Ę z ^ P(a) = - c i  , 0~Ż ° n i =1 ÍÎ ÍmaxÁË z where a is parameter (explained below), z is the poverty line, ci is consumption of individual i, and n is the total number of individuals. If one sets a equal to 0, P(0), or the poverty headcount index, is obtained. P(0) simply measures the fraction of individuals below the poverty line. If one sets a equal to 1, P(1), or the poverty deficit, is obtained. The poverty deficit is a poverty measure that takes into account how far the poor, on average, are below the poverty line. One can show the following equation: P(1) = P(0) * ( Average Deficit) in which the average deficit is the amount, measured as a percentage of the poverty line, by which the consumption of the poor on average fall short of the poverty line. Finally, if one sets a equal to 2, P(2), sometimes also called the severity of poverty or FGT(2), is obtained. This poverty measure also 232 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union takes into account whether some of the poor are deeper into poverty than others. In this presentation of the poverty results in the appendix, B. Key Poverty Indicators, table 2, the authors rely on the headcount index, P(1), and P(2) indexes. Measuring Inequality (appendix, B. Key Poverty Indicators, table 2) Because inequality has many aspects, there are many ways to meas- ure it. This report relies mainly on two types of inequality measure: quintile shares and Gini coefficients. Quintile shares are straightforward indicators of inequality that are easy to interpret because they depict the share of total consumption that goes to each of the 20 percent groups of equal size ranked by per capita consumption. The report uses the share of the bottom 20 per- cent in the total consumption. Quintiles are used to construct tables on nonincome dimensions to highlight socioeconomic differences by income groups. Quintile shares do not reflect what happens in other parts of the distribution. To address this shortcoming, the Gini coefficient is also used. It is given by the following equation: 2 n Ę G =  ri - n + 1^ c i m n2 i =1ÁË 2 ~Ż , in which there are n individuals indexed by i, their consumption per capita is given by ci, and mean consumption is denoted by m and in which ri is household's i rank in the consumption ranking (that is, for the household with lowest consumption, ri equals 1, while for the household with the highest consumption, ri equals n). The Gini coef- ficient is bounded between 0 and 1, with 0 indicating absolute equal- ity and 1 indicating absolute inequality. The Gini coefficient is especially sensitive to changes in inequality in the middle of the con- sumption distribution. Poverty Profiles Characteristics Identifying the key characteristics of the poor is an important first step in understanding causes of poverty. Because poverty in the Region is a multifaceted phenomenon, the multiple poverty profiles show simple correlations between household and individual charac- teristics and poverty and the contribution of each group to total poverty (structure of poverty). Appendix: Data and Methodology 233 Regional Characteristics (appendix, B. Key Poverty Indicators, table 3) All surveys used in the analysis are representative by urban-rural location or by the size of the population center, because these nor- mally form the strata for survey samples. In several countries, surveys are also conducted to produce representative data by regions. Capital cities are often distinguished as a "region." In countries where sam- ples are representative by region (note that definition of a statistical region differs from that of an administrative one), poverty statistics could also be computed by these regions. Given the importance of spatial dimensions of poverty as produced in this report, poverty indexes are reported for both poverty line ($2.15 at 2000 PPP) and vulnerability to poverty line ($4.30 at 2000 PPP) on panels A and B. To give a sense of the variation of poverty across regions and the con- centration of poverty in the poorest regions, maximum and mini- mum poverty rates and contribution to national poverty are presented for countries where data allow this breakdown. Names of the poorest and richest regions are given for reference. Demographic Dimension of Poverty (appendix, B. Key Poverty Indicators, table 4) The table reports both the poverty rates and contributions to the national poverty by age groups, gender, and number of children in a family. Age groups are defined based on the common age brackets and use the age of respondents at the time of the survey. Education Dimension of Poverty (appendix, B. Key Poverty Indicators, table 5) To make sure that those in the compulsory education process are not influencing the distribution (because computation of com- pleted level may be problematic for them), the table uses only indi- viduals above 15 years old. It classifies the detailed educational categories that differ across countries into a set of standard classifi- cations (a simplified version of the International Standard Classifi- cation of Education [ISCED] 1997) proposed by Sullivan and Smeeding (1997). It distinguishes five levels of education: persons without education or with incomplete primary education, those with completed primary only, those with general secondary, those with specialized (vocational or technical) secondary, and those with university education. The Labor Market Profile of Poverty (appendix, B. Key Poverty Indicators, table 6) To construct labor market indicators, the labor section of the house- 234 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union hold surveys are used. Sometimes they are detailed enough to build employment and unemployment categories, following the ILO defini- tions. In other cases, they rely entirely on respondents' self-identifica- tion. Employment is normally defined in Household Expenditure and Income Surveys in a very inclusive way: all types of gainful activity over the reference period (usually a month, but sometimes even a year) are considered as employment, however short that work period was. The table is built for all individuals above 15 years old. In addi- tion, those above 65 years old are considered outside the working age and reported as "retired." Classification into self-employment or wage (salaried) employment is based on the type of job in the primary occu- pation of a household member. All unpaid family workers reporting work for a family farm or business are classified as "self-employed." Nonincome Poverty Indicators Different aspects of poverty--income and nonincome--interact and reinforce each other in ways that often exacerbate the deprivation that poor people face. Poor health outcomes and low educational achievement not only decrease well-being but also limit people's income-earning potential. Defining poverty as multidimensional, however, also raises the question of how to measure these different dimensions. There are no strict and agreed-on standards that would fit every country. This report uses the survey data and other sources of information to obtain individual or household-level measures of deprivation in the follow- ing dimensions: education, health, housing, and infrastructure. To see whether deprivation in these dimensions is different across groups, all four tables highlight spatial differences (reporting value of indexes by location). They also show the correlation between poverty in the monetary dimension with the nonincome space by reporting nonincome indicators for the top and bottom quintiles. The Profile of Poverty: Health Dimension (appendix, B. Key Poverty Indicators, table 7) Only a few surveys in the Region allow the construction of health indicators. The most common question is about a health condition over the reference period (which differs across countries) serious enough to limit daily activities. This indicator (morbidity rate) is reported by groups described above. The use of services is computed for those with a health condition as a share of respondents that used formal medical services (clinics, hospitals, private doctors) to take care of their illness in a reference period. There are significant differences Appendix: Data and Methodology 235 across countries in the length of the reference period; hence, large dif- ferences in the value may be partly ascribed to these design features. The Profile of Poverty: Education Dimension (appendix, B. Key Poverty Indicators, table 8) Surveys allow construction of basic enrollment rates (for two age groups), which are reported by quintiles and locations. The Profile of Poverty: Infrastructure Dimension (appendix, B. Key Poverty Indicators, table 9) Two indicators are selected: access to water (connection to a water pipe) and use of clean fuels (electricity, liquid fuels, and gas) for heat- ing. The use of dirty fuels (coal and firewood) is shown to have sig- nificant health effects. Overall, there is no consistently collected information on the quality of infrastructure services (with a few exceptions discussed in the report). This is a major gap in the data. The Profile of Poverty: Housing Dimension (appendix, B. Key Poverty Indicators, table 10) Housing dimension is reported as the ownership rights on the dwelling the household is residing in (thus a refection of inequality in asset ownership), and the measure of quality is determined by report- ing the share of population living in overcrowded dwellings (with more than three persons per room or with a total living space per per- son of less than six square meters). Endnote 1. The exposition of the poverty and inequality measures is phrased in equivalent consumption, but the same measures could be applied to equivalent income. 236 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union B. KEY POVERTY INDICATORS TABLE 1 Macroeconomic Environment GDP Inflation Government Real wages GDP per capita, (CPI, annual expenditure, total index, Country Year Population growth 2000 PPP % change) (% of GDP) 1998 = 1.00 Albania 2002 3,150,265 5 4,113 5.2 -- -- Armenia 1998/99 3,162,500 5 2,206 4.65 -- 1.06 Armenia 2001 3,086,704 10 2,669 3.1 -- 1.25 Armenia 2002 3,067,953 13 3,019 1.1 -- 1.38 Armenia 2003 3,055,630 14 3,468 4.6 -- 1.46 Azerbaijan 2002 8,172,000 11 3,096 2.8 -- 1.93 Azerbaijan 2003 8,233,000 11 3,417 2.2 -- 2.29 Belarus 1998 10,069,000 8 4,369 73.0 30.37 1.00 Belarus 1999 10,035,000 3 4,527 294.0 30.86 1.07 Belarus 2000 10,005,000 6 4,802 168.6 28.90 1.20 Belarus 2001 9,970,260 5 5,043 61.1 29.58 1.56 Belarus 2002 9,925,000 5 5,331 42.6 -- 1.68 Bosnia & Herzegovina 2001 4,057,056 4 5,378 3.2 58.8 -- Bosnia & Herzegovina 2004 4,158,000 6 6,267 0.4 50 1.87 Bulgaria 1995 8,400,000 3 6,285 62.1 40.96 -- Bulgaria 2001 7,910,000 4 6,585 7.5 34.38 1.08 Bulgaria 2003 7,823,000 4 7,304 2.3 -- 1.20 Estonia 2000 1,369,500 8 10,253 4.0 29.52 1.14 Estonia 2001 1,364,000 6 11,064 5.8 28.02 1.21 Estonia 2002 1,358,000 7 11,907 3.6 -- 1.30 Estonia 2003 1,353,000 5 12,790 1.3 -- 1.40 Georgia 1997 5,320,000 11 1,725 7.1 17.32 -- Georgia 1998 5,307,000 3 1,755 3.6 15.16 1.00 Georgia 1999 5,289,000 3 1,814 19.2 14.98 1.02 Georgia 2000 5,262,000 2 1,880 4.0 12.26 1.05 Georgia 2001 5,224,000 5 2,040 4.7 10.93 1.32 Georgia 2002 5,177,000 5 2,169 5.6 12.35 1.50 Georgia 2003 5,126,000 11 2,445 4.8 -- 1.58 Hungary 1998 10,114,000 5 11,503 14.3 43.99 1.00 Hungary 1999 10,068,000 4 12,010 10.0 43.42 1.02 Hungary 2000 10,024,000 5 12,705 9.7 41.37 1.06 Hungary 2001 10,187,000 4 13,105 9.2 41.45 1.14 Hungary 2002 10,159,000 3 13,391 5.3 -- 1.28 Kazakhstan 2001 14,909,200 14 5,206 8.4 14.63 1.34 Kazakhstan 2002 14,875,000 10 5,672 5.8 -- 1.49 Kazakhstan 2003 14,878,100 9 6,302 6.4 -- 1.60 Kyrgyz Rep. 2000 4,915,000 5 1,560 18.7 18.00 0.90 Kyrgyz Rep. 2001 4,955,000 5 1,599 6.9 17.73 1.00 Kyrgyz Rep. 2002 5,003,900 0 1,570 2.1 -- 1.14 Kyrgyz Rep. 2003 5,052,000 7 1,654 3.1 -- 1.25 Latvia 2002 2,338,000 6 8,922 1.9 -- 1.13 Latvia 2003 2,321,000 7 9,702 2.9 -- 1.19 Lithuania 1998 3,555,000 7 8,464 5.1 29.38 1.00 Lithuania 1999 3,531,000 ­2 8,384 0.8 30.58 1.05 Lithuania 2000 3,505,000 4 8,716 1.0 27.34 1.03 Appendix: Data and Methodology 237 TABLE 1 (continued) GDP Inflation Government Real wages GDP per capita, (CPI, annual expenditure, total index, Country Year Population growth 2000 PPP % change) (% of GDP) 1998 = 1.00 Lithuania 2001 3,482,000 6 9,313 1.3 26.08 1.02 Lithuania 2002 3,469,000 7 9,955 0.3 -- 1.07 Lithuania 2003 3,454,000 9 11,055 ­1.2 -- 1.17 Macedonia, FYR 2002 2,038,000 1 6,257 1.8 -- 1.03 Macedonia, FYR 2003 2,049,000 3 6,419 1.2 -- 1.07 Moldova 1998 4,299,000 ­7 1,337 7.7 35.86 1.00 Moldova 1999 4,288,000 ­3 1,294 39.3 29.71 0.87 Moldova 2000 4,278,000 2 1,290 31.3 29.58 0.89 Moldova 2001 4,270,000 6 1,332 9.8 22.80 1.08 Moldova 2002 4,255,000 8 1,420 5.2 25.52 1.31 Moldova 2003 4,237,600 6 1,426 11.7 -- 1.51 Poland 1998 38,666,152 5 9,159 11.8 35.19 1.00 Poland 1999 38,658,000 4 9,529 7.3 33.24 1.28 Poland 2000 38,648,000 4 9,935 10.1 32.72 1.30 Poland 2001 38,251,000 1 10,125 5.5 34.65 1.33 Poland 2002 38,232,000 1 10,299 1.9 -- 1.36 Romania 1998 22,503,000 ­5 5,751 59.1 33.33 1.00 Romania 1999 22,457,990 ­1 5,699 45.8 35.06 1.00 Romania 2000 22,443,000 1 5,715 45.7 34.09 1.01 Romania 2001 22,132,000 5 6,098 34.5 30.40 1.12 Romania 2002 21,803,000 4 6,476 22.5 -- 1.15 Romania 2003 21,744,000 5 6,875 15.3 -- 1.25 Russian Fed. 1997 147,304,000 1 6,427 14.8 0.00 1.15 Russian Fed. 1998 146,899,008 ­5 6,244 27.7 26.14 1.00 Russian Fed. 1999 146,308,992 6 6,642 85.7 21.68 0.78 Russian Fed. 2000 145,555,008 10 7,242 20.8 22.89 0.94 Russian Fed. 2001 144,752,000 5 7,559 21.5 24.63 1.13 Russian Fed. 2002 144,070,800 5 7,993 15.8 -- 1.31 Serbia & Montenegro 2002 8,160,000 4 -- 21.2 -- 1.53 Tajikistan 1999 6,160,000 4 740 27.5 12.41 1.06 Tajikistan 2003 6,304,700 10 1,045 16.4 -- 1.67 Ukraine 2002 48,717,300 5 4,719 0.8 -- 1.38 Ukraine 2003 48,355,700 9 5,188 5.2 -- 1.61 Uzbekistan 2000/01 24,808,500 4 1,539 26.1 -- 1.67 Uzbekistan 2002 25,271,000 4 1,604 27.3 -- 1.93 Uzbekistan 2003 25,590,000 4 1,648 10.2 -- 1.98 Colombia 2003 44,584,000 4 6,331 7.13 -- -- Turkey 2002 69,626,000 8 6,145 45.0 -- 1.36 Vietnam 1998 76,520,000 6 1,855 7.27 20.34 1.00 Source: WDI 2005. Note: -- = not available. 238 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 2 Main Poverty and Inequality Indicators Poverty indices, $PPP 2.15/day Poverty indices, $PPP 4.30/day Poverty rate Poverty depth Poverty severity Poverty rate Poverty depth Poverty severity Country Year P0 P1 P2 P0 P1 P2 Albania 2002 24 5 2 71 28 14 Armenia 1998/99 58 19 8 92 49 30 Armenia 2001 59 19 9 91 50 31 Armenia 2002 55 17 7 91 48 28 Armenia 2003 50 14 5 93 46 26 Azerbaijan 2002 5 1 0 74 19 7 Azerbaijan 2003 4 1 0 70 17 6 Belarus 1998 9 2 1 48 15 6 Belarus 1999 7 2 1 42 13 5 Belarus 2000 6 1 0 38 11 4 Belarus 2001 4 1 0 27 7 3 Belarus 2002 2 0 0 21 5 2 Bosnia & Herzegovina 2001 5 1 0 40 10 4 Bosnia & Herzegovina 2004 4 1 0 35 9 4 Bulgaria 1995 3 1 1 20 6 3 Bulgaria 2001 10 3 1 36 13 6 Bulgaria 2003 4 1 0 33 9 4 Estonia 2000 4 1 0 26 7 3 Estonia 2001 4 1 0 28 8 3 Estonia 2002 4 1 0 27 8 3 Estonia 2003 5 1 0 26 8 3 Georgia 1997 45 18 10 80 42 26 Georgia 1998 42 16 9 80 40 25 Georgia 1999 50 20 11 84 45 29 Georgia 2000 53 22 12 86 48 31 Georgia 2001 55 22 12 88 49 32 Georgia 2002 49 19 10 84 45 28 Georgia 2003 52 21 11 85 46 30 Hungary 1998 1 0 0 20 4 1 Hungary 1999 1 0 0 19 4 1 Hungary 2000 1 0 0 18 4 1 Hungary 2001 1 0 0 13 3 1 Hungary 2002 0 0 0 12 2 1 Kazakhstan 2001 31 9 31 73 32 73 Kazakhstan 2002 26 7 26 71 29 71 Kazakhstan 2003 21 5 21 66 25 66 Kyrgyz Rep. 2000 78 32 78 97 61 97 Kyrgyz Rep. 2001 74 29 74 97 59 97 Kyrgyz Rep. 2002 73 28 73 97 59 97 Kyrgyz Rep. 2003 70 24 70 96 56 96 Latvia 2002 3 1 0 18 5 2 Latvia 2003 3 1 0 17 5 2 Lithuania 1998 3 1 0 24 6 3 Lithuania 1999 3 1 0 25 7 3 Lithuania 2000 4 1 0 29 8 3 Lithuania 2001 4 1 0 29 8 3 Lithuania 2002 4 1 0 30 8 3 Lithuania 2003 4 1 0 24 7 3 Appendix: Data and Methodology 239 Inequality indices Mean per capita $PPP 2.15 in National poverty Gini coefficient Share of the annual consumption local currency, line in local currency, Country Year (per capita) lowest 20% in local currency annual annual Albania 2002 0.3194 8 96,518 54,553 58,359 Armenia 1998/99 0.3208 8 141,940 128,484 140,584 Armenia 2001 0.3249 8 143,828 131,800 143,783 Armenia 2002 0.3102 9 151,498 133,250 145,364 Armenia 2003 0.2850 10 162,286 139,379 152,051 Azerbaijan 2002 0.1812 13 2,026,126 1,150,929 1,478,966 Azerbaijan 2003 0.1822 13 2,143,867 1,176,639 1,512,004 Belarus 1998 0.2908 9 38,983 16,699 27,263 Belarus 1999 0.2994 8 167,415 65,796 107,417 Belarus 2000 0.2933 8 475,632 176,741 288,544 Belarus 2001 0.3008 8 902,610 284,730 464,845 Belarus 2002 0.2918 9 1,399,101 406,025 662,869 Bosnia & Herzegovina 2001 0.2634 9 2,687 1,064 2198 Bosnia & Herzegovina 2004 0.2951 9 3,026 1,077 -- Bulgaria 1995 0.3261 8 52,249 14,070 20,287 Bulgaria 2001 0.3368 7 1,496 522 753 Bulgaria 2003 0.2774 9 1,562 565 815 Estonia 2000 0.3386 7 21,384 6,023 14,953 Estonia 2001 0.3323 8 21,709 6,372 15,890 Estonia 2002 0.3350 7 22,800 6,602 16,900 Estonia 2003 0.3301 7 23,457 6,687 17,167 Georgia 1997 0.4041 5 586 415 543 Georgia 1998 0.3855 6 614 430 562 Georgia 1999 0.3936 6 655 512 670 Georgia 2000 0.3970 6 646 533 697 Georgia 2001 0.3825 6 646 558 729 Georgia 2002 0.3901 6 763 589 770 Georgia 2003 0.3906 6 765 617 807 Hungary 1998 0.2498 10 261,938 81,765 219,572 Hungary 1999 0.2589 10 294,765 89,942 241,530 Hungary 2000 0.2540 10 322,988 98,666 264,958 Hungary 2001 0.2510 10 390,509 107,743 289,334 Hungary 2002 0.2496 10 417,447 113,346 304,379 Kazakhstan 2001 0.3458 7 67,472 40,178 37,876 Kazakhstan 2002 0.3297 8 74,844 42,548 40,111 Kazakhstan 2003 0.3183 8 85,163 45,271 42,678 Kyrgyz Rep. 2000 0.2993 9 6,578 8,322 7,548 Kyrgyz Rep. 2001 0.2902 9 7,390 8,897 8,069 Kyrgyz Rep. 2002 0.2924 9 7,663 9,083 8,238 Kyrgyz Rep. 2003 0.2761 10 8,445 9,365 8,494 Latvia 2002 0.3403 7 982 236 416 Latvia 2003 0.3503 7 1,051 243 437 Lithuania 1998 0.3029 8 4,524 1,348 2,822 Lithuania 1999 0.3035 8 4,503 1,359 2,845 Lithuania 2000 0.3057 8 4,298 1,373 2,873 Lithuania 2001 0.3052 8 4,286 1,391 2,911 Lithuania 2002 0.3050 8 4,285 1,395 2,919 Lithuania 2003 0.3251 8 4,850 1,378 2,884 (Table continues on the following page.) 240 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 2 (continued) Main Poverty and Inequality Indicators Poverty indices, $PPP 2.15/day Poverty indices, $PPP 4.30/day Poverty rate Poverty depth Poverty severity Poverty rate Poverty depth Poverty severity Country Year P0 P1 P2 P0 P1 P2 Macedonia, FYR 2002 4 1 0 23 7 3 Macedonia, FYR 2003 4 1 0 24 7 3 Moldova 1998 67 29 16 93 56 38 Moldova 1999 79 37 22 96 64 46 Moldova 2000 77 35 19 96 62 44 Moldova 2001 70 29 15 94 57 39 Moldova 2002 56 20 10 90 48 30 Moldova 2003 43 13 5 85 41 23 Poland 1998 2 0 2 23 5 2 Poland 1999 2 0 2 25 6 2 Poland 2000 2 0 2 26 6 2 Poland 2001 2 0 2 26 7 2 Poland 2002 3 0 3 27 7 3 Romania 1998 14 3 3 63 21 9 Romania 1999 19 4 4 69 25 12 Romania 2000 20 5 5 72 26 13 Romania 2001 16 4 4 64 22 10 Romania 2002 16 4 4 62 22 10 Romania 2003 12 3 3 58 19 9 Russian Fed. 1997 10 3 1 41 14 7 Russian Fed. 1998 13 4 1 46 17 8 Russian Fed. 1999 21 6 3 59 24 12 Russian Fed. 2000 17 5 2 54 20 10 Russian Fed. 2001 11 3 1 47 16 7 Russian Fed. 2002 9 2 1 41 13 6 Serbia & Montenegro 2002 6 1 1 42 12 5 Tajikistan 1999 91 45 26 100 71 53 Tajikistan 2003 74 30 15 96 59 40 Ukraine 2002 3 1 0 31 8 3 Ukraine 2003 1 0 0 22 5 2 Uzbekistan 2000/01 54 19 54 89 48 89 Uzbekistan 2002 42 12 42 86 41 86 Uzbekistan 2003 47 14 47 86 43 86 Colombia 2003 6 2 0 24 8 4 Turkey 2002 20 6 2 58 23 12 Vietnam 1998 41 10 4 85 39 21 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: P0 reported in %. P1 and P2 are multiplied by 100. -- = not available. Appendix: Data and Methodology 241 Inequality indices Mean per capita $PPP 2.15 in National poverty Gini coefficient Share of the annual consumption local currency, line in local currency, Country Year (per capita) lowest 20% in local currency annual annual Macedonia, FYR 2002 0.3678 6 73,520 17,784 59,881 Macedonia, FYR 2003 0.3732 6 72,735 17,997 58,644 Moldova 1998 0.3710 6 1,368 1,454 1,359 Moldova 1999 0.3653 7 1,536 2,026 1,894 Moldova 2000 0.3500 7 2,095 2,660 2,487 Moldova 2001 0.3571 7 2,674 2,920 2,730 Moldova 2002 0.3449 8 3,482 3,072 2,872 Moldova 2003 0.3280 8 4,573 3,432 3,208 Poland 1998 0.2960 9 4,967 1,467 2,592 Poland 1999 0.3024 9 5,236 1,574 2,781 Poland 2000 0.3050 8 5,722 1,733 3,062 Poland 2001 0.3072 8 6,039 1,828 3,230 Poland 2002 0.3197 8 6,199 1,863 3,291 Romania 1998 0.2736 9 5,311,954 2,736,291 3,372,638 Romania 1999 0.2834 9 7,100,299 3,989,513 4,917,307 Romania 2000 0.2820 9 9,979,185 5,812,720 7,164,516 Romania 2001 0.2862 9 14,849,780 7,818,109 9,636,274 Romania 2002 0.2939 8 18,767,966 9,577,183 11,804,435 Romania 2003 0.2878 9 22,797,434 11,013,761 13,575,100 Russian Fed. 1997 0.3527 7 6,417,704 2,276,173 3,059,599 Russian Fed. 1998 0.3694 6 7,760 2,907 3,907 Russian Fed. 1999 0.3566 7 11,412 5,398 7,256 Russian Fed. 2000 0.3488 7 14,961 6,521 8,765 Russian Fed. 2001 0.3392 7 20,256 7,922 10,648 Russian Fed. 2002 0.3381 7 25,467 9,173 12,331 Serbia & Montenegro 2002 0.2920 9 85,313 33,593 44,940 Tajikistan 1999 0.2890 8 147,001 253,064 245,296 Tajikistan 2003 0.3274 8 519 608 589 Ukraine 2002 0.2736 9 2,551 891 1,691 Ukraine 2003 0.2681 10 2,982 937 1,779 Uzbekistan 2000/01 0.3549 7 103,091 87,324 -- Uzbekistan 2002 0.3260 8 178,741 132,397 -- Uzbekistan 2003 0.3545 7 193,781 145,902 -- Colombia 2003 0.4877 4 2,645,884 472,157 -- Turkey 2002 0.3932 6 1,436,459,264 620,648,442 916,489,450 Vietnam 1998 0.3110 9 2,443 1,779 1,794 242 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 3 Panel A (Based on 2.15 $ PPP) Poverty Profile: Spatial Dimension Poverty rate (%), Regional poverty rates (%), $PPP 2.15/day $PPP 2.15/day Country Year Capital Other urban All urban Rural Maximum Minimum Albania 2002 16 19 18 27 32 (Mountain) 16 (Tirana) Armenia 1998/99 58 66 62 52 77 (Shirak) 35 (Tavush) Armenia 2001 55 64 60 59 74 (Tavush) 27 (Siunik) Armenia 2002 48 67 57 51 81 (Shirak) 40 ( Siunik) Armenia 2003 35 57 46 57 70 (Armavir) 35 (Yerevan) Azerbaijan 2002 5 7 6 5 .. .. Azerbaijan 2003 4 8 6 3 .. .. Belarus 1998 8 11 10 7 12 (Gomel) 7 (Minsk) Belarus 1999 5 9 8 6 10 (Vitebsk) 5 (Minsk) Belarus 2000 3 8 7 4 8 (Mogilev) 3 (Minsk) Belarus 2001 2 5 4 3 5 (Gomel) 2 (Minsk) Belarus 2002 1 3 2 2 4 (Vitebsk) 1 (Minsk) Bosnia & Herzegovina 2001 3 5 5 5 -- -- Bosnia & Herzegovina 2004 2 6 4 4 -- -- Bulgaria 1995 3 3 3 4 -- -- Bulgaria 2001 0 8 10 17 -- -- Bulgaria 2003 0 5 6 6 -- -- Estonia 2000 2 5 4 6 -- -- Estonia 2001 3 4 4 6 -- -- Estonia 2002 2 5 4 5 -- -- Estonia 2003 4 5 5 5 -- -- Georgia 1997 34 42 38 52 63 (Guria) 34 (Tbilisi) Georgia 1998 35 41 38 47 59 (Samtskhe-Javakheti) 35 (Tbilisi) Georgia 1999 44 51 48 52 61 (Samtskhe-Javakheti) 39 (Adjara) Georgia 2000 46 53 50 57 80 (Samtskhe-Javakheti) 42 (Samegrelo) Georgia 2001 40 60 51 60 72 (Samtskhe-Javakheti) 40 (Tbilisi) Georgia 2002 32 54 43 56 68 (Kakheti) 32 (Tbilisi) Georgia 2003 32 50 41 62 67 (Kakheti) 32 (Tbilisi) Hungary 1998 1 1 1 2 .. .. Hungary 1999 1 1 1 1 .. .. Hungary 2000 1 1 1 2 .. .. Hungary 2001 0 1 1 1 .. .. Hungary 2002 0 0 0 0 .. .. Kazakhstan 2001 7 25 24 40 54 (Jambyl) 7 (Astana) Kazakhstan 2002 5 19 18 35 44 (Kyzylorda) 5 (Almaty) Kazakhstan 2003 2 14 13 31 39 (South_Kaz) 2 (Astana) Kyrgyz Rep. 2000 55 79 68 84 94 (Naryn) 55 (Bishkek) Kyrgyz Rep. 2001 50 77 65 79 96 (Naryn) 50 (Bishkek) Kyrgyz Rep. 2002 49 72 62 79 94 (Naryn) 49 (Bishkek) Kyrgyz Rep. 2003 42 68 57 77 95 (Naryn) 37 (Chui) Latvia 2002 1 3 2 4 -- -- Latvia 2003 1 3 2 5 -- -- Lithuania 1998 1 3 2 5 -- -- Lithuania 1999 1 3 2 7 -- -- Lithuania 2000 1 2 2 9 -- -- Lithuania 2001 3 3 3 8 -- -- Lithuania 2002 1 3 2 9 -- -- Lithuania 2003 1 2 1 8 -- -- Appendix: Data and Methodology 243 Contribution to poverty (%), Regional contributions poverty (%), $PPP 2.15/day $PPP 2.15/day Country Year Capital Other urban All urban Rural Maximum Minimum Albania 2002 8 25 33 67 49 (Central) 8 (Tirana) Armenia 1998/99 28 34 62 38 28 (Yerevan) 2 (Vayots D) Armenia 2001 26 34 60 40 26 (Yerevan) 2 (Siunik) Armenia 2002 26 35 62 38 26 (Yerevan) 3 (Vayots Dzor) Armenia 2003 20 34 54 46 20 (Yerevan) 3 (Vayots Dzor) Azerbaijan 2002 33 25 57 43 .. .. Azerbaijan 2003 11 45 56 44 .. .. Belarus 1998 14 62 76 24 22 (Gomel) 11 (Grodno) Belarus 1999 11 64 75 25 20 (Gomel) 9 (Grodno) Belarus 2000 9 72 80 20 19 (Mogilev) 9 (Minsk) Belarus 2001 10 66 76 24 20 (Gomel) 10 (Minsk) Belarus 2002 7 67 75 25 22 (Vitebsk) 3 (Grodno) Bosnia & Herzegovina 2001 11 49 60 40 -- -- Bosnia & Herzegovina 2004 7 53 60 40 -- -- Bulgaria 1996 12 47 59 41 -- -- Bulgaria 2001 0 42 42 58 -- -- Bulgaria 2003 0 60 60 40 -- -- Estonia 2000 14 45 59 41 -- -- Estonia 2001 20 39 60 40 -- -- Estonia 2002 16 48 64 36 -- -- Estonia 2003 24 42 66 34 -- -- Georgia 1997 19 26 46 54 25 (Imereti) 4 (Samtskhe-Javakheti) Georgia 1998 22 27 50 50 23 (Imereti) 4 (Guria) Georgia 1999 23 28 52 48 23 (Tbilisi) 4 (Guria) Georgia 2000 22 27 48 52 22 (Tbilisi) 5 (Guria) Georgia 2001 18 29 47 53 18 (Tbilisi) 4 (Guria) Georgia 2002 17 28 45 55 18 (Imereti) 5 (Samtskhe-Javakheti) Georgia 2003 15 23 39 61 17 (Imereti) 3 (Swaneti) Hungary 1998 12 30 43 57 .. .. Hungary 1999 17 37 53 47 .. .. Hungary 2000 11 36 47 53 .. .. Hungary 2001 4 43 46 54 .. .. Hungary 2002 13 52 65 35 .. .. Kazakhstan 2001 1 43 44 56 21 (South_Kaz) 1 (Astana) Kazakhstan 2002 1 40 40 60 21 (South_Kaz) 1 (Astana) Kazakhstan 2003 0 36 36 64 26 (South_Kaz) 0 (Astana) Kyrgyz Rep. 2000 11 19 30 70 27 (Osh) 5 (Talas) Kyrgyz Rep. 2001 10 20 31 69 28 (Osh) 5 (Talas) Kyrgyz Rep. 2002 10 19 29 71 29 (Osh) 5 (Talas) Kyrgyz Rep. 2003 9 19 28 72 31 (Osh) 6 (Talas) Latvia 2002 17 39 56 44 -- -- Latvia 2003 11 34 45 55 -- -- Lithuania 1998 15 27 42 58 -- -- Lithuania 1999 12 27 39 61 -- -- Lithuania 2000 14 17 32 68 -- -- Lithuania 2001 24 17 41 59 -- -- Lithuania 2002 13 19 31 69 -- -- Lithuania 2003 10 18 28 72 -- -- (Table continues on the following page.) 244 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 3 (continued) Panel A (Based on 2.15 $ PPP) Poverty Profile: Spatial Dimension Poverty rate (%), Regional poverty rates (%), $PPP 2.15/day $PPP 2.15/day Country Year Capital Other urban All urban Rural Maximum Minimum Macedonia, FYR 2002 5 3 4 5 -- -- Macedonia, FYR 2003 4 5 5 3 -- -- Moldova 1998 42 70 56 74 -- -- Moldova 1999 57 80 69 85 -- -- Moldova 2000 54 83 69 82 -- -- Moldova 2001 49 78 64 74 -- -- Moldova 2002 33 61 47 61 -- -- Moldova 2003 27 48 37 47 -- -- Poland 1998 1 1 1 2 .. .. Poland 1999 1 1 1 3 .. .. Poland 2000 1 2 2 3 .. .. Poland 2001 1 2 2 3 .. .. Poland 2002 2 2 2 3 .. .. Romania 1998 6 9 8 20 23 (North-East) 6 (Bucharest) Romania 1999 7 13 12 27 31 (North-East) 7 (Bucharest) Romania 2000 10 15 14 28 30 (North-East) 10 (Bucharest) Romania 2001 7 10 9 25 26 (North-East) 7 (Bucharest) Romania 2002 6 9 8 24 25 (North-East) 6 (Bucharest) Romania 2003 4 7 6 20 18 (North-East) 4 (Bucharest) Russian Fed. 1997 1 9 8 14 28 (Tyva Republic) 1 (Moscow) Russian Fed. 1998 3 13 12 15 75 (Ingushetiya Rep.) 0 (Belgorod oblast) Russian Fed. 1999 18 20 20 24 45 (Dagestan Rep.) 3 (Belgorod oblast) Russian Fed. 2000 15 16 15 20 68 (Ingushetiya Rep.) 2 (Belgorod oblast) Russian Fed. 2001 5 10 10 16 40 (Dagestan Rep.) 1 (Yamalo- Nenetskiy Aut. Reg.) Russian Fed. 2002 5 7 7 14 39 (Taimyr Aut. Reg.) 0 (St. Petersburg) Serbia & Montenegro 2002 6 4 4 9 9 (South-East Serbia) 5 (Vojvodina) Tajikistan 1999 73 90 85 92 -- -- Tajikistan 2003 54 73 67 76 -- -- Ukraine 2002 2 3 3 4 .. .. Ukraine 2003 0 1 1 2 .. .. Uzbekistan 2000/01 24 50 44 60 85 (Kashkadarya) 24 (Tashkent city) Uzbekistan 2002 11 39 33 47 66 (Kashkadarya) 10 (Tashkent city) Uzbekistan 2003 4 43 34 55 72 (Syrdarya) 4 (Tashkent city) Colombia 2003 1 3 3 14 13 (Pacific) 1 (Bogota D.C.) Turkey 2002 8 21 18 24 39 (SE Anatolia) 8 (Aegean) Vietnam 1998 3 19 14 48 65 (Rur. N. Mountains 3 (Hanoi and & Midlands) Ho Chi Minh City) Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Latvia, FYR Macedonia, Poland (1998), Romania, Serbia & Montenegro column "Capital" includes some rural dwellings. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains five urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Russian Federation: data for Ingushetiya Republic is not available for 1999. Turkey: "Capital" contains estimates for Ankara and Istanbul and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. .. = Negligible; -- = not available. Appendix: Data and Methodology 245 Contribution to poverty (%), Regional contributions poverty (%), $PPP 2.15/day $PPP 2.15/day Country Year Capital Other urban All urban Rural Maximum Minimum Macedonia, FYR 2002 33 29 62 38 -- -- Macedonia, FYR 2003 30 42 71 29 -- -- Moldova 1998 11 20 31 69 -- -- Moldova 1999 13 20 32 68 -- -- Moldova 2000 13 21 34 66 -- -- Moldova 2001 13 21 33 67 -- -- Moldova 2002 11 21 31 69 -- -- Moldova 2003 11 20 31 69 -- -- Poland 1998 3 49 52 48 .. .. Poland 1999 4 35 38 62 .. .. Poland 2000 3 42 45 55 .. .. Poland 2001 3 44 47 53 .. .. Poland 2002 6 46 52 48 .. .. Romania 1998 5 31 36 64 29 (North-East) 5 (Bucharest) Romania 1999 4 32 35 65 28 (North-East) 4 (Bucharest) Romania 2000 5 34 38 62 25 (North-East) 5 (Bucharest) Romania 2001 4 27 31 69 27 (North-East) 4 (Bucharest) Romania 2002 4 26 30 70 27 (North-East) 4 (Bucharest) Romania 2003 3 23 27 73 25 (North-East) 3 (Bucharest) Russian Fed. 1997 1 61 62 38 4 (Rostov oblast) 0 (Belgorod oblast) Russian Fed. 1998 1 67 68 32 5 (Moscow oblast) 0 (Murmansk oblast) Russian Fed. 1999 5 65 70 30 6 (Moscow oblast) 0 (Kamchatka oblast) Russian Fed. 2000 5 63 68 32 5 (Moscow) 0 (Krasnoyarsk territory) Russian Fed. 2001 3 59 62 38 5 (Moscow oblast) 0 (Evenkiyskiy Aut. Reg.) Russian Fed. 2002 4 54 58 42 6 (Moscow oblast) 0 (Evenkiyskiy Aut. Reg.) Serbia & Montenegro 2002 19 27 46 54 22 (Vojvodina) 12 (East Serbia) Tajikistan 1999 5 15 21 79 -- -- Tajikistan 2003 7 18 24 76 -- -- Ukraine 2002 3 59 61 39 .. .. Ukraine 2003 2 54 56 44 .. .. Uzbekistan 2000/01 4 26 30 70 14 (Kashkadarya) 2 (Syrdarya) Uzbekistan 2002 2 26 28 72 15 (Kashkadarya) 2 (Navoi) Uzbekistan 2003 1 26 27 73 12 (Andizhan) 1 (Tashkent city) Colombia 2003 3 31 34 66 23 (Atlantica) 0 (San Andres & Providencia) Turkey 2002 6 48 54 46 19 (Mediterranian) 5 (Aegean) Vietnam 1998 0 7 8 92 26 (Rur. N. Mountains 0 (Hanoi and Ho Chi Minh & Midlands) City) 246 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 3 Panel B (Based on 4.30 $ PPP) Poverty Profile: Spatial Dimension Poverty rate (%), Regional poverty rates (%), $PPP 4.30/day $PPP 4.30/day Country Year Capital Other urban All urban Rural Maximum Minimum Albania 2002 63 67 66 75 80 (Mountain) 63 (Tirana) Armenia 1998/99 91 95 93 91 88 (Siunik) 84 (Tavush) Armenia 2001 87 94 91 92 98 (Shirak) 53 (Siunik) Armenia 2002 87 94 91 92 99 (Vayots Dzor) 81 (Siunik) Armenia 2003 86 94 90 96 100 (Siunik) 86 (Yerevan) Azerbaijan 2002 72 78 74 74 90 (Nakhchivan) 66 (Sheki- Zagatala) Azerbaijan 2003 66 74 69 70 84 (Nakhchivan) 62 (Baku) Belarus 1998 44 52 50 45 53 (Gomel) 44 (Minsk) Belarus 1999 38 44 42 42 45 (Vitebsk) 38 (Minsk) Belarus 2000 28 42 39 34 44 (Mogilev) 28 (Minsk) Belarus 2001 19 32 29 23 30 (Mogilev) 20 (Minsk) Belarus 2002 14 26 23 18 26 (Brest) 14 (Minsk) Bosnia & Herzegovina 2001 39 43 37 42 -- -- Bosnia & Herzegovina 2004 21 39 38 33 -- -- Bulgaria 1995 17 18 18 24 -- -- Bulgaria 2001 23 30 29 49 -- -- Bulgaria 2003 20 34 31 36 -- -- Estonia 2000 17 27 23 32 -- -- Estonia 2001 23 27 25 33 -- -- Estonia 2002 21 27 24 32 -- -- Estonia 2003 21 27 24 28 -- -- Georgia 1997 76 79 78 83 91 (Guria) 72 (Adjara) Georgia 1998 78 78 78 82 86 (Samtskhe-Javakheti) 78 (Tbilisi) Georgia 1999 84 86 85 84 91 (Guria) 75 (Kvemo Kartli) Georgia 2000 81 89 85 88 95 (Samtskhe-Javakheti) 81 (Tbilisi) Georgia 2001 80 92 86 89 92 (Samtskhe-Javakheti) 80 (Tbilisi) Georgia 2002 73 89 81 87 92 (Guria) 73 (Tbilisi) Georgia 2003 75 87 81 89 94 (Kvemo Kartli) 75 (Tbilisi) Hungary 1998 14 18 16 25 25 (Nograd) 10 (Csongrad) Hungary 1999 14 16 16 26 29 (Baranya) 11 (Csongrad) Hungary 2000 14 17 16 22 27 (Baranya) 12 (Heves) Hungary 2001 7 12 10 18 25 (Nograd) 5 (Csongrad) Hungary 2002 8 11 10 16 20 (Baranya) 5 (Csongrad) Kazakhstan 2001 40 66 65 82 91 (Jambyl) 40 (Astana) Kazakhstan 2002 37 63 62 82 90 (South) 37 (Astana) Kazakhstan 2003 29 57 55 79 90 (South) 29 (Astana) Kyrgyz Rep. 2000 94 98 96 97 100 (Naryn) 92 (Chui) Kyrgyz Rep. 2001 92 97 95 98 100 (Naryn) 92 (Bishkek) Kyrgyz Rep. 2002 93 97 95 98 100 (Naryn) 89 (Chui) Kyrgyz Rep. 2003 89 96 93 98 100 (Naryn) 88 (Chui) Latvia 2002 11 19 14 26 -- -- Latvia 2003 7 19 12 27 -- -- Lithuania 1998 15 20 17 40 -- -- Lithuania 1999 14 21 17 42 -- -- Lithuania 2000 18 27 22 44 -- -- Lithuania 2001 17 28 22 46 -- -- Lithuania 2002 16 30 22 47 -- -- Lithuania 2003 11 21 15 42 -- -- Appendix: Data and Methodology 247 Contribution to poverty (%), Regional contributions poverty (%), $PPP 4.30/day $PPP 4.30/day Country Year Capital Other urban All urban Rural Maximum Minimum Albania 2002 10 28 38 62 47 (Central) 10 (Tirana) Armenia 1998/99 27 31 58 42 27 (Yerevan) 2 (Vayots Dzor) Armenia 2001 27 32 60 41 27 (Yerevan) 2 (Siunik) Armenia 2002 28 30 58 42 28 (Yerevan) 3 (Vayots Dzor) Armenia 2003 27 30 58 42 28 (Yerevan) 3 (Vayots Dzor) Azerbaijan 2002 33 21 54 46 23 (Baku) 5 (Nakhchivan) Azerbaijan 2003 32 21 53 47 22 (Baku) 5 (Nakhchivan) Belarus 1998 15 55 70 30 17 (Gomel) 13 (Grodno) Belarus 1999 14 54 68 32 17 (Gomel) 12 (Mogilev) Belarus 2000 12 60 72 28 16 (Brest) 12 (Minsk) Belarus 2001 12 62 73 27 17 (Brest) 12 (Minsk) Belarus 2002 11 63 74 26 19 (Gomel) 10 (Grodno) Bosnia & Herzegovina 2001 18 42 60 40 -- -- Bosnia & Herzegovina 2004 12 43 55 45 -- -- Bulgaria 1995 12 48 60 40 -- -- Bulgaria 2001 9 44 53 47 -- -- Bulgaria 2003 8 57 66 34 -- -- Estonia 2000 19 41 60 40 -- -- Estonia 2001 24 39 63 37 -- -- Estonia 2002 23 40 63 37 -- -- Estonia 2003 25 39 64 36 -- -- Georgia 1997 24 27 51 49 24 (Tbilisi) 4 (Guria) Georgia 1998 26 28 54 46 26 (Tbilisi) 4 (Guria) Georgia 1999 26 28 54 46 26 (Tbilisi) 4 (Guria) Georgia 2000 23 28 51 49 23 (Tbilisi) 4 (Guria) Georgia 2001 22 28 50 50 22 (Tbilisi) 4 (Guria) Georgia 2002 22 27 49 51 22 (Tbilisi) 4 (Guria) Georgia 2003 22 25 46 54 22 (Tbilisi) 3 (Swaneti) Hungary 1998 13 39 52 48 13 (Budapest) 2 (Csongrad) Hungary 1999 13 39 53 47 13 (Budapest) 2 (Vas) Hungary 2000 14 43 57 43 14 (Budapest) 2 (Zala) Hungary 2001 9 41 50 50 12 (Borsod-Abauj-Zemplen) 2 (Csongrad) Hungary 2002 11 42 54 46 11 (Budapest) 2 (Veszprem) Kazakhstan 2001 2 49 51 49 17 (South) 2 (Astana) Kazakhstan 2002 2 48 49 51 28 (South) 2 (Astana) Kazakhstan 2003 2 46 48 52 20 (South) 2 (Astana) Kyrgyz Rep. 2000 15 19 34 66 25 (Osh) 4 (Talas) Kyrgyz Rep. 2001 15 20 34 66 25 (Osh) 4 (Talas) Kyrgyz Rep. 2002 15 19 34 66 26 (Osh) 4 (Talas) Kyrgyz Rep. 2003 14 20 34 66 26 (Osh) 4 (Talas) Latvia 2002 23 34 57 43 -- -- Latvia 2003 16 35 51 49 -- -- Lithuania 1998 24 23 47 53 -- -- Lithuania 1999 22 24 46 54 -- -- Lithuania 2000 24 27 51 49 -- -- Lithuania 2001 23 27 50 50 -- -- Lithuania 2002 21 27 49 51 -- -- Lithuania 2003 19 24 42 58 -- -- (Table continues on the following page.) 248 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 3 (continued) Panel B (Based on 4.30 $ PPP) Poverty Profile: Spatial Dimension Poverty rate (%), Regional poverty rates (%), $PPP 4.30/day $PPP 4.30/day Country Year Capital Other urban All urban Rural Maximum Minimum Macedonia, FYR 2002 26 20 20 26 -- -- Macedonia, FYR 2003 26 25 24 25 -- -- Moldova 1998 83 93 88 96 -- -- Moldova 1999 88 97 93 98 -- -- Moldova 2000 87 98 93 98 -- -- Moldova 2001 84 97 90 96 -- -- Moldova 2002 75 93 84 93 -- -- Moldova 2003 75 88 81 87 -- -- Poland 1998 9 18 17 32 45 (Przemysl) 9 (Warsaw) Poland 1999 13 20 19 35 34 (Podkarpackie) 20 (Slaske) Poland 2000 15 21 21 34 34 (Podkarpackie) 18 (Lodzkie) Poland 2001 16 22 21 34 34 (Podkarpackie) 20 (Slaske) Poland 2002 20 22 22 36 35 (Podkarpackie) 22 (Lodzkie) Romania 1998 47 56 55 72 73 (North-East) 47 (Bucharest) Romania 1999 55 62 61 79 79 (North-East) 55 (Bucharest) Romania 2000 59 66 65 81 80 (North-East) 59 (Bucharest) Romania 2001 53 54 53 77 74 (North-East) 53 (Bucharest) Romania 2002 46 52 51 75 70 (North-East) 46 (Bucharest) Romania 2003 41 47 45 72 64 (North-East) 41 (Bucharest) Russian Fed. 1997 19 41 39 47 84 (Ingushetiya Rep.) 18 (Belgorod oblast) Russian Fed. 1998 25 48 46 48 91 (Ingushetiya Rep.) 22 (Belgorod oblast) Russian Fed. 1999 51 58 58 63 89 (Chita oblast) 33 (Tumen oblast) Russian Fed. 2000 51 52 52 58 92 (Ingushetiya Rep.) 25 (Tumen oblast) Russian Fed. 2001 37 44 43 56 82 (Dagestan Rep.) 14 (Khanty-Mansiyskiy Aut. Reg.) Russian Fed. 2002 36 37 37 53 79 (Ingushetiya Rep.) 10 (Yamalo-Nenetskiy Aut. Reg.) Serbia & Montenegro 2002 36 37 35 51 56 (South-East Serbia) 36 (Belgrade) Tajikistan 1999 98 99 99 100 -- -- Tajikistan 2003 89 96 93 97 -- -- Ukraine 2002 18 30 29 37 63 (Zakarpatska) 12 (Sevastopol) Ukraine 2003 11 20 20 28 32 (Rivnenska) 10 (Mykolaivska) Uzbekistan 2000/01 70 86 82 93 99 (Kashkadarya) 70 (Tashkent city) Uzbekistan 2002 51 83 75 92 95 (Syrdarya) 51 (Tashkent city) Uzbekistan 2003 39 83 73 93 99 (Syrdarya) 39 (Tashkent city) Colombia 2003 8 17 15 47 43 (Pacific) 3 (San Andres & Providencia) Turkey 2002 34 57 52 67 77 (Eastern Anatolia) 41 (Aegean) Vietnam 1998 45 71 63 91 96 (Rur. N. Central Coast) 45 (Hanoi and Ho Chi Minh City) Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Latvia, FYR Macedonia, Poland (1998), Romania, Serbia and Montenegro. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains five urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Russian Fed.: data for Ingushetiya Republic are not available for 1999. Turkey: "Capital" contains estimates for Ankara and Istanbul, and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. -- = not available. Appendix: Data and Methodology 249 Contribution to poverty (%), Regional contributions poverty (%), $PPP 4.30/day $PPP 4.30/day Country Year Capital Other urban All urban Rural Maximum Minimum Macedonia, FYR 2002 31 31 62 39 -- -- Macedonia, FYR 2003 29 38 67 33 -- -- Moldova 1998 15 19 34 66 -- -- Moldova 1999 16 19 36 64 -- -- Moldova 2000 17 20 36 64 -- -- Moldova 2001 16 19 35 65 -- -- Moldova 2002 15 19 35 65 -- -- Moldova 2003 16 19 35 65 -- -- Poland 1998 2 44 46 54 7 (Katowice) 1 (Chelmno) Poland 1999 4 41 45 55 10 (Wielkopolskie) 3 (Lubuskie) Poland 2000 5 43 47 53 10 (Mazowieckie) 3 (Lubuskie) Poland 2001 5 43 48 52 10 (Mazowieckie) 3 (Opolskie) Poland 2002 6 42 48 52 12 (Mazowieckie) 3 (Lubuskie) Romania 1998 8 41 49 51 20 (North-East) 8 (Bucharest) Romania 1999 8 41 49 51 19 (North-East) 8 (Bucharest) Romania 2000 8 42 50 50 19 (North-East) 8 (Bucharest) Romania 2001 8 38 46 54 20 (North-East) 8 (Bucharest) Romania 2002 7 38 46 54 19 (North-East) 7 (Bucharest) Romania 2003 7 36 43 57 19 (North-East) 7 (Bucharest) Russian Fed. 1997 3 67 69 31 5 (Moscow oblast) 0 (Magadan oblast) Russian Fed. 1998 3 69 72 28 5 (Moscow oblast) 0 (Magadan oblast) Russian Fed. 1999 5 66 71 29 5 (Moscow) 0 (Krasnoyarsk territory) Russian Fed. 2000 6 66 71 29 6 (Moscow) 0 (Krasnoyarsk territory) Russian Fed. 2001 5 63 68 32 5 (Moscow) 0 (Evenkiyskiy Aut. Reg.) Russian Fed. 2002 5 60 66 34 5 (Moscow) 0 (Evenkiyskiy Aut. Reg.) Serbia & Montenegro 2002 19 35 54 46 25 (Vojvodina) 9 (East Serbia) Tajikistan 1999 6 15 22 78 -- -- Tajikistan 2003 9 18 26 74 -- -- Ukraine 2002 3 58 61 39 8 (Donetska) 0 (Sevastopol) Ukraine 2003 3 55 58 42 8 (Donetska) 0 (Sevastopol) Uzbekistan 2000/01 7 28 35 65 12 (Samarkand) 10 (Bukhara) Uzbekistan 2002 5 27 32 68 11 (Samarkand) 3 (Navoi) Uzbekistan 2003 4 28 32 68 12 (Fergana) 3 (Navoi) Colombia 2003 5 42 48 52 27 (Atlantica) 0 (San Andres & Providencia) Turkey 2002 9 46 55 45 19 (Marmara) 9 (Aegean) Vietnam 1998 4 13 17 83 18 (Rur. Mekong 4 (Hanoi and Ho Chi River Delta) Minh City) 250 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 4 Poverty Profile: Demographic Dimension Poverty rate by gender Poverty rate by number Poverty rate by age (%), of HH head (%), of children in HH (%), $PPP 2.15/day $PPP 2.15/day $PPP 2.15/day Children Adults Elderly No One or Country Year (<16 y.o.) (17­65) (>65 y.o.) Male Female children two Three Albania 2002 30 21 19 18 13 5 17 43 Armenia 1998/99 62 56 54 52 54 40 56 67 Armenia 2001 64 58 57 54 55 43 59 70 Armenia 2002 59 53 52 50 51 41 55 63 Armenia 2003 54 49 47 46 42 34 48 65 Azerbaijan 2002 6 5 5 4 3 1 4 8 Azerbaijan 2003 5 4 4 3 2 1 4 6 Belarus 1998 11 7 6 6 8 7 10 26 Belarus 1999 9 6 5 6 6 5 8 22 Belarus 2000 8 4 3 4 5 4 7 28 Belarus 2001 4 3 2 3 3 3 4 15 Belarus 2002 3 2 1 1 2 1 2 10 Bosnia & Herzegovina 2001 6 4 5 4 2 3 4 10 Bosnia & Herzegovina 2004 -- -- -- 4 3 3 5 -- Bulgaria 1995 6 3 3 2 3 2 2 26 Bulgaria 2001 13 7 5 5 5 3 7 36 Bulgaria 2003 8 4 2 3 3 2 5 28 Estonia 2000 6 4 3 3 4 3 5 11 Estonia 2001 6 4 3 3 4 3 5 10 Estonia 2002 6 4 3 3 4 3 6 8 Estonia 2003 6 5 4 4 5 4 5 8 Georgia 1997 49 43 44 41 37 33 44 57 Georgia 1998 49 40 41 38 35 30 43 59 Georgia 1999 56 48 48 43 42 37 48 72 Georgia 2000 60 51 51 48 46 38 53 74 Georgia 2001 62 53 53 50 47 39 55 79 Georgia 2002 55 47 50 44 45 38 50 66 Georgia 2003 57 49 53 46 45 39 51 71 Hungary 1998 3 1 0 1 1 0 1 7 Hungary 1999 2 1 0 1 0 0 1 5 Hungary 2000 3 1 0 1 0 0 1 5 Hungary 2001 2 1 0 1 0 0 1 4 Hungary 2002 1 0 0 -- -- 0 0 42 Kazakhstan 2001 38 29 18 26 18 9 25 54 Kazakhstan 2002 33 24 14 22 14 7 21 49 Kazakhstan 2003 28 19 11 18 10 5 17 46 Kyrgyz Rep. 2000 85 74 71 75 55 40 72 93 Kyrgyz Rep. 2001 82 70 62 72 55 37 69 90 Kyrgyz Rep. 2002 81 68 53 70 50 29 69 91 Kyrgyz Rep. 2003 80 66 51 67 36 20 64 90 Latvia 2002 4 3 2 2 2 2 2 7 Latvia 2003 5 3 2 2 2 1 3 13 Lithuania 1998 4 2 1 1 2 1 3 8 Lithuania 1999 6 3 2 2 3 1 3 13 Lithuania 2000 7 3 2 3 3 1 4 17 Lithuania 2001 7 4 2 3 3 2 4 14 Appendix: Data and Methodology 251 Structure of poverty by Structure of poverty by number Structure of poverty by age (%), gender of HH head (%), of children in HH (%), $PPP 2.15/day $PPP 2.15/day $PPP 2.15/day Children Adults Elderly No One or Country Year (<16 y.o.) (17­65) (>65 y.o.) Male Female children two Three Albania 2002 41 52 7 91 9 9 48 43 Armenia 1998/99 34 59 8 72 28 27 49 24 Armenia 2001 30 61 9 73 27 34 49 17 Armenia 2002 29 62 9 72 28 33 52 15 Armenia 2003 29 63 8 73 27 31 53 16 Azerbaijan 2002 37 57 7 79 21 13 48 39 Azerbaijan 2003 34 58 8 87 13 12 55 33 Belarus 1998 31 56 13 39 61 72 27 1 Belarus 1999 30 55 15 40 60 72 26 1 Belarus 2000 30 56 14 39 61 73 26 1 Belarus 2001 26 59 15 46 54 76 24 1 Belarus 2002 27 57 16 41 59 76 23 1 Bosnia & Herzegovina 2001 27 60 13 84 16 50 35 15 Bosnia & Herzegovina 2004 -- -- -- 77 23 90 10 -- Bulgaria 1995 33 56 12 73 27 48 31 21 Bulgaria 2001 31 58 11 77 23 40 41 18 Bulgaria 2003 30 63 7 72 28 43 46 11 Estonia 2000 28 61 11 41 59 53 38 9 Estonia 2001 28 62 10 43 57 52 41 7 Estonia 2002 27 62 11 47 53 52 42 6 Estonia 2003 22 63 15 44 56 63 34 4 Georgia 1997 25 62 13 67 33 45 44 10 Georgia 1998 25 61 14 68 32 47 43 10 Georgia 1999 24 63 13 66 34 50 41 9 Georgia 2000 26 62 12 65 35 42 48 10 Georgia 2001 24 62 13 67 33 42 48 10 Georgia 2002 22 63 15 66 34 47 45 8 Georgia 2003 24 62 14 69 31 45 45 10 Hungary 1998 44 52 4 76 24 25 41 34 Hungary 1999 44 50 6 81 19 39 31 30 Hungary 2000 46 53 1 88 12 20 48 32 Hungary 2001 48 50 1 95 5 4 59 37 Hungary 2002 66 56 2 -- -- 19 24 57 Kazakhstan 2001 37 58 4 63 37 18 52 30 Kazakhstan 2002 37 59 5 61 39 17 54 29 Kazakhstan 2003 37 59 4 63 37 14 54 31 Kyrgyz Rep. 2000 42 53 5 73 27 16 48 36 Kyrgyz Rep. 2001 42 53 4 70 30 14 49 36 Kyrgyz Rep. 2002 41 55 4 72 28 13 54 33 Kyrgyz Rep. 2003 41 55 4 79 21 11 54 36 Latvia 2002 27 63 10 38 62 55 35 10 Latvia 2003 31 61 8 35 65 41 46 13 Lithuania 1998 36 58 6 41 59 30 56 15 Lithuania 1999 41 52 7 47 53 25 54 21 Lithuania 2000 39 54 7 46 54 32 47 20 Lithuania 2001 36 59 5 48 52 37 47 16 (Table continues on the following page.) 252 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 4 (continued) Poverty Profile: Demographic Dimension Poverty rate by gender Poverty rate by number Poverty rate by age (%), of HH head (%), of children in HH (%), $PPP 2.15/day $PPP 2.15/day $PPP 2.15/day Children Adults Elderly No One or Country Year (<16 y.o.) (17­65) (>65 y.o.) Male Female children two Three Lithuania 2002 7 4 3 3 3 2 4 17 Lithuania 2003 6 3 2 2 3 2 3 14 Macedonia, FYR 2002 6 4 2 3 1 1 4 9 Macedonia, FYR 2003 6 4 2 3 2 1 4 9 Moldova 1998 74 65 64 63 58 52 68 83 Moldova 1999 85 76 79 77 68 66 80 94 Moldova 2000 84 74 77 74 65 64 79 92 Moldova 2001 78 67 71 67 58 57 73 87 Moldova 2002 66 52 54 51 45 39 60 80 Moldova 2003 53 40 38 38 32 27 47 68 Poland 1998 3 1 0 1 1 0 1 6 Poland 1999 4 1 1 1 1 0 1 7 Poland 2000 4 2 1 1 1 0 2 7 Poland 2001 5 2 1 2 1 0 2 9 Poland 2002 5 2 1 2 2 0 2 9 Romania 1998 22 12 7 9 8 4 12 43 Romania 1999 29 17 11 13 12 7 18 55 Romania 2000 32 18 10 14 12 7 20 61 Romania 2001 27 14 9 11 10 6 14 58 Romania 2002 26 14 8 11 9 5 14 54 Romania 2003 21 11 7 8 7 4 11 47 Russian Fed. 1997 13 9 8 10 7 4 10 25 Russian Fed. 1998 18 12 11 12 9 6 14 32 Russian Fed. 1999 26 19 20 19 16 12 22 43 Russian Fed. 2000 21 15 17 15 12 9 17 40 Russian Fed. 2001 -- -- -- -- -- -- -- -- Russian Fed. 2002 13 8 8 9 6 4 9 27 Serbia & Montenegro 2002 7 5 8 6 5 5 5 18 Tajikistan 1999 92 90 89 88 87 61 84 92 Tajikistan 2003 76 72 72 69 61 39 61 78 Ukraine 2002 6 3 1 2 2 1 4 19 Ukraine 2003 2 1 1 1 1 0 2 5 Uzbekistan 2000/01 58 51 48 50 39 20 42 63 Uzbekistan 2002 45 40 35 38 27 15 33 49 Uzbekistan 2003 50 45 40 43 29 15 38 56 Colombia 2003 8 4 5 4 4 2 3 11 Turkey 2002 29 16 14 14 16 5 12 41 Vietnam 1998 49 36 37 38 30 20 32 59 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: -- = not available. Appendix: Data and Methodology 253 Structure of poverty by Structure of poverty by number Structure of poverty by age (%), gender of HH head (%), of children in HH (%), $PPP 2.15/day $PPP 2.15/day $PPP 2.15/day Children Adults Elderly No One or Country Year (<16 y.o.) (17­65) (>65 y.o.) Male Female children two Three Lithuania 2002 35 57 8 49 51 39 42 18 Lithuania 2003 35 56 9 41 59 45 38 16 Macedonia, FYR 2002 34 61 6 92 8 16 55 28 Macedonia, FYR 2003 34 60 6 87 13 18 55 27 Moldova 1998 31 60 9 66 34 44 46 10 Moldova 1999 29 60 11 67 33 48 44 8 Moldova 2000 28 60 12 66 34 52 40 7 Moldova 2001 28 59 13 65 35 53 41 6 Moldova 2002 29 58 13 62 38 50 44 7 Moldova 2003 28 60 12 63 37 47 45 7 Poland 1998 51 46 3 69 31 14 37 49 Poland 1999 51 46 4 63 37 10 40 50 Poland 2000 46 51 3 61 39 15 44 40 Poland 2001 46 52 2 65 35 17 42 42 Poland 2002 43 55 2 58 42 17 46 37 Romania 1998 38 57 5 78 22 29 46 25 Romania 1999 35 59 7 77 23 34 46 20 Romania 2000 35 59 6 79 21 32 49 20 Romania 2001 35 59 7 77 23 33 46 21 Romania 2002 35 58 7 76 24 33 47 20 Romania 2003 35 58 7 76 24 34 45 21 Russian Fed. 1997 34 60 6 38 62 28 60 12 Russian Fed. 1998 32 61 7 35 65 32 58 10 Russian Fed. 1999 29 63 8 32 68 38 56 6 Russian Fed. 2000 29 63 9 32 68 38 55 8 Russian Fed. 2001 -- -- -- -- -- -- -- -- Russian Fed. 2002 30 63 8 32 68 34 58 8 Serbia & Montenegro 2002 14 55 31 64 36 81 16 3 Tajikistan 1999 48 48 4 83 17 5 26 69 Tajikistan 2003 45 51 4 82 18 8 30 62 Ukraine 2002 38 57 5 48 52 20 62 18 Ukraine 2003 34 57 8 42 58 27 63 10 Uzbekistan 2000/01 45 52 4 81 19 7 37 56 Uzbekistan 2002 42 54 4 82 18 8 40 52 Uzbekistan 2003 42 55 4 83 17 7 44 49 Colombia 2003 49 46 5 71 29 13 36 51 Turkey 2002 47 49 4 89 11 12 38 50 Vietnam 1998 45 49 6 78 22 13 43 44 254 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 5 Poverty Profile: Education Dimension Poverty rate by education (%), $PPP 2.15/day (adults, >15 y.o.) None/ unfinished Primary/ Secondary Secondary Country Year primary basic general special Tertiary Albania 2002 12 27 14 11 3 Armenia 1998/99 56 58 57 55 41 Armenia 2001 69 62 61 56 44 Armenia 2002 48 57 58 51 40 Armenia 2003 63 58 52 47 35 Azerbaijan 2002 8 6 4 5 4 Azerbaijan 2003 8 5 4 4 3 Belarus 1998 7 11 10 8 4 Belarus 1999 6 9 8 7 3 Belarus 2000 4 6 6 5 2 Belarus 2001 4 5 4 3 1 Belarus 2002 2 3 2 2 0 Bosnia & Herzegovina 2001 5 6 4 2 5 Bosnia & Herzegovina 2004 -- -- -- -- -- Bulgaria 1995 7 4 1 0 1 Bulgaria 2001 16 9 2 3 1 Bulgaria 2003 14 5 2 1 0 Estonia 2000 8 6 4 4 1 Estonia 2001 4 6 4 4 1 Estonia 2002 6 6 4 4 1 Estonia 2003 8 7 3 5 1 Georgia 1997 -- -- -- -- -- Georgia 1998 44 48 44 38 30 Georgia 1999 54 55 51 51 38 Georgia 2000 59 59 55 53 41 Georgia 2001 62 62 57 55 40 Georgia 2002 61 57 51 49 33 Georgia 2003 66 62 56 48 33 Hungary 1998 2 1 0 1 0 Hungary 1999 2 1 0 0 0 Hungary 2000 2 1 0 0 0 Hungary 2001 1 1 0 0 0 Hungary 2002 1 0 0 0 0 Kazakhstan 2001 25 32 36 24 13 Kazakhstan 2002 23 27 31 19 9 Kazakhstan 2003 19 22 26 14 7 Kyrgyz Rep. 2000 80 75 80 69 56 Kyrgyz Rep. 2001 67 71 75 67 49 Kyrgyz Rep. 2002 66 72 74 62 48 Kyrgyz Rep. 2003 92 65 74 57 41 Latvia 2002 14 4 3 1 0 Latvia 2003 7 5 3 2 0 Lithuania 1998 4 4 3 . . . 1 Lithuania 1999 4 4 4 . . . 1 Lithuania 2000 5 4 4 . . . 1 Lithuania 2001 7 5 5 . . . 1 Lithuania 2002 7 6 4 . . . 1 Lithuania 2003 8 5 2 4 0 Appendix: Data and Methodology 255 Structure of Poverty by Education (%), $PPP 2.15/day (adults, >15 y.o.) None/ unfinished Primary/ Secondary Secondary Country Year primary basic general special Tertiary Albania 2002 1 77 13 8 1 Armenia 1998/99 2 23 44 19 12 Armenia 2001 3 15 44 26 13 Armenia 2002 2 14 48 25 12 Armenia 2003 2 14 49 25 11 Azerbaijan 2002 3 20 47 21 10 Azerbaijan 2003 2 21 47 21 10 Belarus 1998 16 13 27 37 7 Belarus 1999 8 18 32 35 6 Belarus 2000 8 16 32 38 6 Belarus 2001 12 17 28 37 6 Belarus 2002 8 24 29 35 3 Bosnia & Herzegovina 2001 32 7 46 5 11 Bosnia & Herzegovina 2004 -- -- -- -- -- Bulgaria 1995 61 24 7 4 4 Bulgaria 2001 51 28 5 14 2 Bulgaria 2003 41 39 6 12 2 Estonia 2000 2 35 17 41 4 Estonia 2001 1 40 19 36 4 Estonia 2002 1 40 18 38 3 Estonia 2003 1 38 14 43 4 Georgia 1997 -- -- -- -- -- Georgia 1998 5 11 43 30 11 Georgia 1999 5 11 42 22 20 Georgia 2000 6 11 42 22 20 Georgia 2001 5 12 43 21 18 Georgia 2002 6 11 45 20 18 Georgia 2003 6 14 45 18 16 Hungary 1998 25 43 3 28 1 Hungary 1999 30 46 1 23 0 Hungary 2000 26 49 1 24 0 Hungary 2001 24 56 2 18 1 Hungary 2002 35 33 5 28 0 Kazakhstan 2001 7 15 43 29 7 Kazakhstan 2002 8 17 43 27 5 Kazakhstan 2003 7 16 45 26 5 Kyrgyz Rep. 2000 2 13 54 22 9 Kyrgyz Rep. 2001 1 14 52 24 9 Kyrgyz Rep. 2002 2 15 56 18 10 Kyrgyz Rep. 2003 1 16 56 18 9 Latvia 2002 7 52 22 18 0 Latvia 2003 2 55 22 19 2 Lithuania 1998 49 34 11 . . . 11 Lithuania 1999 44 37 14 . . . 14 Lithuania 2000 39 42 14 . . . 14 Lithuania 2001 40 42 14 . . . 14 Lithuania 2002 47 37 11 . . . 11 Lithuania 2003 6 49 27 16 3 (Table continues on the following page.) 256 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 5 (continued) Poverty Profile: Education Dimension Poverty rate by education (%), $PPP 2.15/day (adults, >15 y.o.) None/ unfinished Primary/ Secondary Secondary Country Year primary basic general special Tertiary Macedonia, FYR 2002 0 5 2 0 1 Macedonia, FYR 2003 2 4 3 2 1 Moldova 1998 70 75 68 58 40 Moldova 1999 83 84 79 73 53 Moldova 2000 81 81 77 70 54 Moldova 2001 72 76 70 63 41 Moldova 2002 59 61 55 46 28 Moldova 2003 46 47 42 33 18 Poland 1998 1 2 0 1 0 Poland 1999 2 2 0 1 0 Poland 2000 2 3 1 2 0 Poland 2001 1 3 0 2 0 Poland 2002 2 3 1 2 0 Romania 1998 17 15 8 9 1 Romania 1999 23 22 12 14 1 Romania 2000 24 23 13 15 1 Romania 2001 23 20 8 11 1 Romania 2002 22 19 7 11 1 Romania 2003 18 16 5 8 1 Russian Fed. 1997 . . . 7 8 6 3 Russian Fed. 1998 . . . 14 17 13 8 Russian Fed. 1999 . . . 17 19 15 10 Russian Fed. 2000 . . . 14 15 12 7 Russian Fed. 2001 -- -- -- -- -- Russian Fed. 2002 . . . 7 8 6 3 Serbia & Montenegro 2002 11 7 1 4 1 Tajikistan 1999 91 91 91 88 79 Tajikistan 2003 75 76 75 63 50 Ukraine 2002 2 3 3 . . . 9 Ukraine 2003 2 2 1 . . . 0 Uzbekistan 2000/01 57 51 58 43 30 Uzbekistan 2002 46 40 45 33 20 Uzbekistan 2003 48 44 52 37 24 Colombia 2003 13 7 2 0 0 Turkey 2002 29 17 7 6 1 Vietnam 1998 43 37 31 . . . 50 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Kyrgyz Rep.: individuals reporting incomplete higher education are included in "secondary general" category. Lithuania 1998­2002; Ukraine, Vietnam: separation of secondary education into secondary general and secondary special was not possible. Russian Federation 1997­2000: "Primary/basic" includes individuals with none/unfinished primary education; 2002: only 5 people reported to have "none/unfinished primary" education. -- = not available. . . . = negligible. Appendix: Data and Methodology 257 Structure of Poverty by Education (%), $PPP 2.15/day (adults, >15 y.o.) None/ unfinished Primary/ Secondary Secondary Country Year primary basic general special Tertiary Macedonia, FYR 2002 0 76 22 0 1 Macedonia, FYR 2003 0 63 32 2 2 Moldova 1998 17 24 34 19 6 Moldova 1999 16 24 34 20 7 Moldova 2000 15 26 35 18 7 Moldova 2001 14 26 36 17 6 Moldova 2002 14 28 38 15 5 Moldova 2003 14 30 40 10 5 Poland 1998 3 46 3 48 1 Poland 1999 3 47 2 47 0 Poland 2000 3 47 3 46 1 Poland 2001 1 49 2 47 1 Poland 2002 1 49 3 46 1 Romania 1998 26 32 23 18 1 Romania 1999 25 34 22 19 1 Romania 2000 23 33 23 21 0 Romania 2001 27 36 17 20 1 Romania 2002 27 38 15 20 1 Romania 2003 26 39 14 20 0 Russian Fed. 1997 . . . 7 43 40 11 Russian Fed. 1998 . . . 7 40 40 13 Russian Fed. 1999 . . . 6 38 40 16 Russian Fed. 2000 . . . 6 40 40 14 Russian Fed. 2001 . . . -- -- -- -- Russian Fed. 2002 . . . 5 46 38 12 Serbia & Montenegro 2002 41 28 2 28 1 Tajikistan 1999 13 18 46 17 6 Tajikistan 2003 7 20 58 9 6 Ukraine 2002 2 34 6 -- 6 Ukraine 2003 3 35 5 -- 5 Uzbekistan 2000/01 5 16 57 16 5 Uzbekistan 2002 4 16 56 19 5 Uzbekistan 2003 3 15 59 18 5 Colombia 2003 25 57 18 0 0 Turkey 2002 33 60 5 1 0 Vietnam 1998 50 11 0 . . . 0 258 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 6 Poverty Profile: Labor Market Poverty rate by labor market status (%), $PPP 2.15/day (adults, >15 y.o.) Wage Inactive in Country Year employee Self-employed Unemployed Retired Student working age Albania 2002 13 24 31 16 12 23 Armenia 1998/99 49 45 67 60 50 60 Armenia 2001 47 52 69 59 53 62 Armenia 2002 45 51 63 52 45 57 Armenia 2003 38 53 51 49 43 54 Azerbaijan 2002 4 . . . 3 5 4 7 Azerbaijan 2003 3 . . . 3 5 4 6 Belarus 1998 7 6 17 7 8 13 Belarus 1999 6 6 13 6 7 10 Belarus 2000 5 4 12 4 5 9 Belarus 2001 3 3 7 3 4 7 Belarus 2002 2 2 5 2 2 4 Bosnia & Herzegovina 2001 2 3 8 4 2 6 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 1 0 10 3 0 10 Bulgaria 2001 2 4 16 4 2 17 Bulgaria 2003 2 2 12 3 2 11 Estonia 2000 2 2 12 4 5 7 Estonia 2001 3 2 11 3 4 9 Estonia 2002 2 1 12 3 5 8 Estonia 2003 3 2 10 4 6 8 Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 38 47 56 56 44 54 Georgia 2000 45 51 58 58 46 60 Georgia 2001 44 54 61 59 47 61 Georgia 2002 39 49 49 52 40 53 Georgia 2003 -- -- -- -- -- -- Hungary 1998 -- -- -- -- -- -- Hungary 1999 -- -- -- -- -- -- Hungary 2000 1 0 0 1 0 1 Hungary 2001 0 0 3 0 1 3 Hungary 2002 0 0 1 0 0 1 Kazakhstan 2001 19 35 39 19 29 37 Kazakhstan 2002 16 29 37 15 25 33 Kazakhstan 2003 12 24 31 13 21 30 Kyrgyz Rep. 2000 -- -- -- -- -- -- Kyrgyz Rep. 2001 -- -- -- -- -- -- Kyrgyz Rep. 2002 56 80 75 54 69 75 Kyrgyz Rep. 2003 49 81 69 54 67 76 Latvia 2002 1 3 9 2 2 6 Latvia 2003 1 3 9 2 2 8 Lithuania 1998 2 . . . 6 1 2 4 Lithuania 1999 2 . . . 8 2 2 6 Lithuania 2000 3 . . . 8 2 2 5 Lithuania 2001 3 . . . 11 2 3 7 Lithuania 2002 3 . . . 6 4 4 8 Lithuania 2003 1 6 9 2 3 7 Appendix: Data and Methodology 259 Structure of poverty by labor market status (%), $PPP 2.15/day (adults, >15 y.o.) Wage Inactive in Country Year employee Self-employed Unemployed Retired Student working age Albania 2002 12 41 11 13 3 20 Armenia 1998/99 23 21 17 17 5 17 Armenia 2001 20 14 21 21 6 18 Armenia 2002 21 15 20 21 6 17 Armenia 2003 20 18 15 20 7 20 Azerbaijan 2002 54 . . . 0 17 3 29 Azerbaijan 2003 46 . . . 0 21 6 27 Belarus 1998 50 5 10 21 6 8 Belarus 1999 50 6 8 21 5 10 Belarus 2000 51 5 9 20 5 11 Belarus 2001 46 4 8 22 5 14 Belarus 2002 44 5 11 21 7 13 Bosnia & Herzegovina 2001 9 7 10 24 4 47 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 21 2 30 29 0 18 Bulgaria 2001 10 3 41 23 2 21 Bulgaria 2003 17 10 24 16 3 30 Estonia 2000 28 2 23 20 11 17 Estonia 2001 30 1 22 16 10 20 Estonia 2002 27 1 23 17 14 19 Estonia 2003 35 1 15 19 14 16 Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 20 41 10 15 7 7 Georgia 2000 20 44 8 13 7 8 Georgia 2001 18 46 9 12 7 7 Georgia 2002 17 38 9 12 6 18 Georgia 2003 -- -- -- -- -- -- Hungary 1998 -- -- -- -- -- -- Hungary 1999 -- -- -- -- -- -- Hungary 2002 46 2 0 0 1 51 Hungary 2001 16 0 21 6 9 48 Hungary 2002 17 0 31 19 6 28 Kazakhstan 2001 22 27 20 11 11 10 Kazakhstan 2002 23 28 17 11 12 10 Kazakhstan 2003 25 27 13 11 13 10 Kyrgyz Rep. 2000 -- -- -- -- -- -- Kyrgyz Rep. 2001 -- -- -- -- -- -- Kyrgyz Rep. 2002 28 38 5 9 13 7 Kyrgyz Rep. 2003 24 37 4 10 15 9 Latvia 2002 19 5 34 25 7 10 Latvia 2003 23 7 31 19 5 14 Lithuania 1998 61 . . . 15 8 5 11 Lithuania 1999 53 . . . 20 11 5 11 Lithuania 2000 56 . . . 20 10 6 8 Lithuania 2001 50 . . . 23 9 7 10 Lithuania 2002 47 . . . 18 11 10 13 Lithuania 2003 18 19 17 16 11 19 (Table continues on the following page.) 260 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 6 (continued) Poverty Profile: Labor Market Poverty rate by labor market status (%), $PPP 2.15/day (adults, >15 y.o.) Wage Inactive in Country Year employee Self-employed Unemployed Retired Student working age Macedonia, FYR 2002 2 1 6 2 . . . 6 Macedonia, FYR 2003 2 2 5 2 . . . 6 Moldova 1998 62 67 72 65 62 76 Moldova 1999 74 81 78 79 71 80 Moldova 2000 71 82 85 76 68 76 Moldova 2001 63 74 71 70 65 67 Moldova 2002 47 59 71 54 46 52 Moldova 2003 -- -- -- -- -- -- Poland 1998 1 1 4 1 1 3 Poland 1999 1 1 5 1 1 3 Poland 2000 1 1 7 1 2 4 Poland 2001 1 2 6 1 2 3 Poland 2002 1 2 7 1 2 3 Romania 1998 6 17 23 9 23 6 Romania 1999 7 25 28 13 29 9 Romania 2000 8 25 30 15 34 10 Romania 2001 10 21 23 10 26 7 Romania 2002 9 21 24 10 26 7 Romania 2003 7 16 21 8 20 5 Russian Fed. 1997 8 10 17 8 9 -- Russian Fed. 1998 10 11 20 11 13 -- Russian Fed. 1999 17 17 29 19 21 -- Russian Fed. 2000 14 10 24 15 17 -- Russian Fed. 2001 -- -- -- -- -- -- Russian Fed. 2002 7 7 16 7 8 -- Serbia & Montenegro 2002 3 6 10 5 2 12 Tajikistan 1999 91 88 93 90 85 90 Tajikistan 2003 72 68 79 74 61 74 Ukraine 2002 2 1 5 2 4 6 Ukraine 2003 1 1 2 1 3 0 Uzbekistan 2000/01 -- -- -- -- -- -- Uzbekistan 2002 -- -- -- -- -- -- Uzbekistan 2003 -- -- -- -- -- -- Colombia 2003 3 5 5 0 3 7 Turkey 2002 12 21 13 5 11 18 Vietnam 1998 28 40 30 33 17 28 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Azerbaijan, Lithuania 1998-2002: "Wage employee" contains data for both wage employee and self-employed (disaggregation was not possible). FYR Macedonia 2002­2003: "Inactive in working age" includes students (disaggregation was not possible). Moldova, Poland, and Romania: ILO definition of labor force is used. Russian Federation 1997­2000, 2002: Working pensioners were assumed to be wage employees. Not possible to estimate inactive in working age. -- = not available; . . . = negligible. Appendix: Data and Methodology 261 Structure of poverty by labor market status (%), $PPP 2.15/day (adults, >15 y.o.) Wage Inactive in Country Year employee Self-employed Unemployed Retired Student working age Macedonia, FYR 2002 13 5 25 9 . . . 49 Macedonia, FYR 2003 15 9 19 6 . . . 47 Moldova 1998 45 10 4 28 4 9 Moldova 1999 43 15 3 29 4 6 Moldova 2000 40 18 3 30 4 5 Moldova 2001 37 21 2 31 5 4 Moldova 2002 33 25 3 31 4 3 Moldova 2003 -- -- -- -- -- -- Poland 1998 24 16 17 13 9 20 Poland 1999 21 19 19 14 11 18 Poland 2000 20 10 27 12 11 19 Poland 2001 21 13 27 9 11 19 Poland 2002 20 10 29 9 14 17 Romania 1998 17 39 13 6 15 11 Romania 1999 15 43 13 6 12 11 Romania 2000 15 40 14 6 14 11 Romania 2001 27 31 11 5 16 10 Romania 2002 27 29 11 5 17 11 Romania 2003 28 29 10 6 16 11 Russian Fed. 1997 55 1 21 14 9 -- Russian Fed. 1998 53 1 20 17 9 -- Russian Fed. 1999 54 0 17 19 10 -- Russian Fed. 2000 54 0 17 18 11 -- Russian Fed. 2001 -- -- -- -- -- -- Russian Fed. 2002 53 2 18 16 11 Serbia & Montenegro 2002 17 7 17 31 2 27 Tajikistan 1999 31 16 9 8 6 29 Tajikistan 2003 31 20 3 8 5 34 Ukraine 2002 31 0 25 18 12 13 Ukraine 2003 34 20 8 2 36 0 Uzbekistan 2000/01 -- -- -- -- -- -- Uzbekistan 2002 -- -- -- -- -- -- Uzbekistan 2003 -- -- -- -- -- -- Colombia 2003 17 38 7 0 4 33 Turkey 2002 18 37 1 2 4 39 Vietnam 1998 10 77 1 6 3 4 262 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 7 Poverty Profile: Health Dimension Morbidity rate (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 4 12 16 14 15 14 Armenia 1998/99 19 17 17 21 18 18 Armenia 2001 17 13 12 15 15 14 Armenia 2002 18 11 13 16 14 14 Armenia 2003 14 10 10 14 10 11 Azerbaijan 2002 -- -- -- -- -- -- Azerbaijan 2003 -- -- -- -- -- -- Belarus 1998 -- -- -- -- -- -- Belarus 1999 -- -- -- -- -- -- Belarus 2000 -- -- -- -- -- -- Belarus 2001 -- -- -- -- -- -- Belarus 2002 -- -- -- -- -- -- Bosnia & Herzegovina 2001 14 24 26 26 24 23 Bosnia & Herzegovina 2004 17 20 18 18 17 18 Bulgaria 1995 6 10 13 9 9 10 Bulgaria 2001 10 15 15 14 14 15 Bulgaria 2003 20 14 12 17 12 14 Estonia 2000 -- -- -- -- -- -- Estonia 2001 -- -- -- -- -- -- Estonia 2002 -- -- -- -- -- -- Estonia 2003 -- -- -- -- -- -- Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 -- -- -- -- -- -- Georgia 2000 25 18 17 22 18 19 Georgia 2001 14 14 11 16 10 13 Georgia 2002 10 9 8 11 7 9 Georgia 2003 11 7 9 12 8 9 Hungary 1998 -- -- -- -- -- -- Hungary 1999 -- -- -- -- -- -- Hungary 2000 -- -- -- -- -- -- Hungary 2001 -- -- -- -- -- -- Hungary 2002 -- -- -- -- -- -- Kazakhstan 2001 18 17 11 20 9 14 Kazakhstan 2002 14 13 7 16 6 10 Kazakhstan 2003 9 13 8 15 7 11 Kyrgyz Rep. 2000 -- -- -- -- -- -- Kyrgyz Rep. 2001 -- -- -- -- -- -- Kyrgyz Rep. 2002 -- -- -- -- -- -- Kyrgyz Rep. 2003 -- -- -- -- -- -- Latvia 2002 -- -- -- -- -- -- Latvia 2003 -- -- -- -- -- -- Lithuania 1998 -- -- -- -- -- -- Lithuania 1999 -- -- -- -- -- -- Lithuania 2000 -- -- -- -- -- -- Lithuania 2001 -- -- -- -- -- -- Lithuania 2002 -- -- -- -- -- -- Lithuania 2003 -- -- -- -- -- -- Appendix: Data and Methodology 263 Health care utilization rate (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 28 39 34 40 29 35 Armenia 1998/99 40 38 33 43 31 37 Armenia 2001 34 37 32 37 28 34 Armenia 2002 30 35 33 37 34 32 Armenia 2003 38 31 31 39 23 33 Azerbaijan 2002 -- -- -- -- -- -- Azerbaijan 2003 -- -- -- -- -- -- Belarus 1998 -- -- -- -- -- -- Belarus 1999 -- -- -- -- -- -- Belarus 2000 -- -- -- -- -- -- Belarus 2001 -- -- -- -- -- -- Belarus 2002 -- -- -- -- -- -- Bosnia & Herzegovina 2001 51 33 24 25 27 31 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 56 66 55 71 48 61 Bulgaria 2001 73 66 53 68 51 62 Bulgaria 2003 79 84 85 81 85 83 Estonia 2000 -- -- -- -- -- -- Estonia 2001 -- -- -- -- -- -- Estonia 2002 -- -- -- -- -- -- Estonia 2003 -- -- -- -- -- -- Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 -- -- -- -- -- -- Georgia 2000 69 71 83 76 77 75 Georgia 2001 85 85 85 84 82 85 Georgia 2002 90 89 87 89 88 88 Georgia 2003 92 93 93 95 89 93 Hungary 1998 -- -- -- -- -- -- Hungary 1999 -- -- -- -- -- -- Hungary 2000 -- -- -- -- -- -- Hungary 2001 -- -- -- -- -- -- Hungary 2002 -- -- -- -- -- -- Kazakhstan 2001 50 59 47 59 47 55 Kazakhstan 2002 61 56 54 59 50 56 Kazakhstan 2003 54 55 56 56 55 55 Kyrgyz Rep. 2000 -- -- -- -- -- -- Kyrgyz Rep. 2001 -- -- -- -- -- -- Kyrgyz Rep. 2002 -- -- -- -- -- -- Kyrgyz Rep. 2003 -- -- -- -- -- -- Latvia 2002 -- -- -- -- -- -- Latvia 2003 -- -- -- -- -- -- Lithuania 1998 -- -- -- -- -- -- Lithuania 1999 -- -- -- -- -- -- Lithuania 2000 -- -- -- -- -- -- Lithuania 2001 -- -- -- -- -- -- Lithuania 2002 -- -- -- -- -- -- Lithuania 2003 -- -- -- -- -- -- (Table continues on the following page.) 264 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 7 (continued) Poverty Profile: Health Dimension Morbidity rate (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 -- -- -- -- -- -- Macedonia, FYR 2003 -- -- -- -- -- -- Moldova 1998 -- -- -- -- -- -- Moldova 1999 -- -- -- -- -- -- Moldova 2000 -- -- -- -- -- -- Moldova 2001 -- -- -- -- -- -- Moldova 2002 -- -- -- -- -- -- Moldova 2003 15 10 13 19 8 13 Poland 1998 -- -- -- -- -- -- Poland 1999 -- -- -- -- -- -- Poland 2000 -- -- -- -- -- -- Poland 2001 -- -- -- -- -- -- Poland 2002 -- -- -- -- -- -- Romania 1998 14 13 10 15 8 12 Romania 1999 12 14 11 16 8 12 Romania 2000 11 14 10 15 7 12 Romania 2001 10 13 10 14 8 12 Romania 2002 11 13 10 13 8 12 Romania 2003 12 15 11 15 8 13 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 43 37 32 38 35 36 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 46 38 36 39 34 38 Russian Fed. 2001 42 39 35 42 35 38 Russian Fed. 2002 37 39 33 42 33 37 Serbia & Montenegro 2002 -- -- -- -- -- -- Tajikistan 1999 10 10 8 9 7 8 Tajikistan 2003 6 9 6 8 6 7 Ukraine 2002 87 71 63 73 66 69 Ukraine 2003 -- -- -- -- -- -- Uzbekistan 2000/01 10 7 4 8 4 5 Uzbekistan 2002 7 4 2 6 2 3 Uzbekistan 2003 11 4 2 7 2 4 Colombia 2003 9 13 10 12 9 11 Turkey 2002 -- -- -- -- -- -- Vietnam 1998 38 39 43 37 50 42 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Latvia, FYR Macedonia, Poland (1998), Romania, and Serbia and Montenegro: column "Capital" includes some rural dwellings. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains five urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Russian Federation: data for Ingushetiya Republic is not available for 1999. Turkey: "Capital" contains estimates for Ankara and Istanbul and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. Morbidity rate reference periods are: 1 month for Albania, Armenia, Bosnia & Herzegovina, Bulgaria, Colombia, Georgia, Kazakhstan, Moldova, Romania, Russian Federation, Tajikistan, Uzbekistan, and Vietnam; and 12 months for Ukraine. -- = not available. Appendix: Data and Methodology 265 Health care utilization rate (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 -- -- -- -- -- -- Macedonia, FYR 2003 -- -- -- -- -- -- Moldova 1998 -- -- -- -- -- -- Moldova 1999 -- -- -- -- -- -- Moldova 2000 -- -- -- -- -- -- Moldova 2001 -- -- -- -- -- -- Moldova 2002 -- -- -- -- -- -- Moldova 2003 67 54 45 53 52 51 Poland 1998 -- -- -- -- -- -- Poland 1999 -- -- -- -- -- -- Poland 2000 -- -- -- -- -- -- Poland 2001 -- -- -- -- -- -- Poland 2002 -- -- -- -- -- -- Romania 1998 22 19 21 21 23 20 Romania 1999 57 56 54 55 52 55 Romania 2000 64 60 59 62 53 60 Romania 2001 68 62 59 66 54 61 Romania 2002 77 63 60 66 57 63 Romania 2003 78 67 60 68 60 65 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 42 42 40 45 34 41 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 42 37 37 42 36 37 Russian Fed. 2001 31 27 28 27 29 28 Russian Fed. 2002 33 29 27 31 28 29 Serbia & Montenegro 2002 -- -- -- -- -- -- Tajikistan 1999 48 47 46 50 41 46 Tajikistan 2003 57 52 53 53 50 53 Ukraine 2002 -- -- -- -- -- -- Ukraine 2003 -- -- -- -- -- -- Uzbekistan 2000/01 49 68 72 65 67 67 Uzbekistan 2002 55 69 70 65 69 67 Uzbekistan 2003 53 65 72 60 63 65 Colombia 2003 73 70 65 78 66 69 Turkey 2002 -- -- -- -- -- -- Vietnam 1998 37 37 38 33 41 37 266 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 8 Poverty Profile: Access to Education Dimension Primary enrollment rate, age 7­14 (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 90 92 92 93 90 92 Armenia 1998/99 98 98 98 99 96 98 Armenia 2001 98 98 98 99 96 98 Armenia 2002 99 99 98 99 99 99 Armenia 2003 98 99 98 98 98 98 Azerbaijan 2002 99 98 98 99 98 98 Azerbaijan 2003 98 99 97 99 98 98 Belarus 1998 -- -- -- -- -- -- Belarus 1999 -- -- -- -- -- -- Belarus 2000 -- -- -- -- -- -- Belarus 2001 -- -- -- -- -- -- Belarus 2002 -- -- -- -- -- -- Bosnia & Herzegovina 2001 96 97 97 99 96 97 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 90 92 87 94 77 91 Bulgaria 2001 97 96 89 99 77 93 Bulgaria 2003 99 97 94 100 90 96 Estonia 2000 99 96 96 97 95 97 Estonia 2001 95 97 97 97 96 97 Estonia 2002 98 98 97 97 96 98 Estonia 2003 99 99 99 99 99 99 Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 -- -- -- -- -- -- Georgia 2000 97 98 99 99 97 98 Georgia 2001 98 97 98 99 97 98 Georgia 2002 99 96 97 97 95 97 Georgia 2003 100 99 99 100 99 99 Hungary 1998 94 94 94 97 92 94 Hungary 1999 94 94 95 94 93 94 Hungary 2000 96 94 94 97 93 94 Hungary 2001 100 100 100 100 100 100 Hungary 2002 100 100 100 100 100 100 Kazakhstan 2001 -- -- -- -- -- -- Kazakhstan 2002 91 93 92 95 91 93 Kazakhstan 2003 96 91 92 93 90 92 Kyrgyz Rep. 2000 100 99 99 99 98 99 Kyrgyz Rep. 2001 91 97 97 96 97 96 Kyrgyz Rep. 2002 99 98 96 95 96 97 Kyrgyz Rep. 2003 95 92 91 92 94 92 Latvia 2002 99 99 98 99 98 99 Latvia 2003 99 100 100 98 99 100 Lithuania 1998 -- -- -- -- -- -- Lithuania 1999 -- -- -- -- -- -- Lithuania 2000 -- -- -- -- -- -- Lithuania 2001 -- -- -- -- -- -- Lithuania 2002 98 98 97 99 97 98 Lithuania 2003 98 98 99 96 99 99 Appendix: Data and Methodology 267 Secondary enrollment rate, age 15-17 (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 84 75 39 69 34 53 Armenia 1998/99 80 74 69 77 70 73 Armenia 2001 80 76 72 81 72 75 Armenia 2002 72 78 74 86 69 75 Armenia 2003 75 74 78 84 72 76 Azerbaijan 2002 93 85 84 86 89 87 Azerbaijan 2003 96 88 81 90 85 87 Belarus 1998 -- -- -- -- -- -- Belarus 1999 -- -- -- -- -- -- Belarus 2000 -- -- -- -- -- -- Belarus 2001 -- -- -- -- -- -- Belarus 2002 -- -- -- -- -- -- Bosnia & Herzegovina 2001 97 92 88 96 80 91 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 92 83 50 93 44 77 Bulgaria 2001 95 90 56 94 40 83 Bulgaria 2003 100 90 61 100 56 85 Estonia 2000 98 97 98 98 96 98 Estonia 2001 98 99 97 100 95 98 Estonia 2002 100 99 98 100 97 99 Estonia 2003 92 99 98 100 93 97 Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 -- -- -- -- -- -- Georgia 2000 96 87 83 90 83 87 Georgia 2001 91 90 90 93 87 90 Georgia 2002 98 94 92 98 89 94 Georgia 2003 96 93 86 94 85 90 Hungary 1998 93 95 90 96 87 93 Hungary 1999 95 97 93 100 88 95 Hungary 2000 100 98 93 99 92 96 Hungary 2001 100 100 100 100 100 100 Hungary 2002 100 100 100 100 100 100 Kazakhstan 2001 -- -- -- -- -- -- Kazakhstan 2002 100 99 99 100 98 99 Kazakhstan 2003 100 99 100 100 100 99 Kyrgyz Rep. 2000 92 81 72 85 68 76 Kyrgyz Rep. 2001 97 82 80 88 80 83 Kyrgyz Rep. 2002 98 86 89 91 89 90 Kyrgyz Rep. 2003 97 76 88 85 83 87 Latvia 2002 94 97 98 98 94 96 Latvia 2003 97 96 96 98 91 96 Lithuania 1998 -- -- -- -- -- -- Lithuania 1999 -- -- -- -- -- -- Lithuania 2000 -- -- -- -- -- -- Lithuania 2001 -- -- -- -- -- -- Lithuania 2002 99 98 96 97 96 98 Lithuania 2003 100 100 98 100 98 99 (Table continues on the following page.) 268 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 8 (continued) Poverty Profile: Access to Education Dimension Primary enrollment rate, age 7­14 (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 -- -- -- -- -- -- Macedonia, FYR 2003 -- -- -- -- -- -- Moldova 1998 93 90 95 92 94 94 Moldova 1999 94 96 94 93 94 95 Moldova 2000 96 94 96 93 96 96 Moldova 2001 95 94 95 100 93 95 Moldova 2002 96 96 96 99 97 96 Moldova 2003 99 98 97 99 96 98 Poland 1998 99 98 98 98 98 98 Poland 1999 98 98 98 98 98 98 Poland 2000 97 98 98 98 98 98 Poland 2001 99 97 98 98 98 98 Poland 2002 98 98 98 98 98 98 Romania 1998 96 96 93 98 90 95 Romania 1999 94 94 92 96 89 93 Romania 2000 96 95 92 96 90 94 Romania 2001 93 94 91 94 89 93 Romania 2002 96 95 92 96 90 93 Romania 2003 94 95 92 96 89 93 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 -- -- -- -- -- -- Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 -- -- -- -- -- -- Russian Fed. 2001 -- -- -- -- -- -- Russian Fed. 2002 -- -- -- -- -- -- Serbia & Montenegro 2002 100 98 98 99 97 98 Tajikistan 1999 91 94 96 97 92 95 Tajikistan 2003 88 90 91 91 87 90 Ukraine 2002 -- -- -- -- -- -- Ukraine 2003 -- -- -- -- -- -- Uzbekistan 2000/01 94 87 88 89 89 88 Uzbekistan 2002 97 96 97 97 97 97 Uzbekistan 2003 99 99 99 99 99 99 Colombia 2003 97 96 86 99 86 93 Turkey 2002 97 92 90 96 88 92 Vietnam 1998 95 96 92 95 85 92 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Latvia, FYR Macedonia, Poland (1998), Romania, and Serbia and Montenegro: column "Capital" includes some rural dwellings. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains five urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Russian Federation: data for Ingushetiya Republic is not available for 1999. Turkey: "Capital" contains estimates for Ankara and Istanbul and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. -- = not available. Appendix: Data and Methodology 269 Secondary enrollment rate, age 15-17 (%) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 -- -- -- -- -- -- Macedonia, FYR 2003 -- -- -- -- -- -- Moldova 1998 72 75 62 55 54 67 Moldova 1999 68 78 60 62 54 65 Moldova 2000 82 80 60 80 64 69 Moldova 2001 71 72 73 87 73 72 Moldova 2002 67 85 76 89 74 77 Moldova 2003 94 90 76 97 68 81 Poland 1998 100 98 95 99 94 97 Poland 1999 97 100 96 100 94 97 Poland 2000 98 98 95 99 92 97 Poland 2001 100 99 97 100 95 98 Poland 2002 99 99 98 99 97 99 Romania 1998 90 85 54 89 53 74 Romania 1999 90 91 65 95 61 81 Romania 2000 92 94 68 98 66 83 Romania 2001 95 93 68 98 61 83 Romania 2002 90 95 68 97 62 83 Romania 2003 98 95 70 96 63 84 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 91 88 64 98 80 79 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 90 90 77 85 88 86 Russian Fed. 2001 81 87 77 93 79 84 Russian Fed. 2002 76 86 76 91 66 82 Serbia & Montenegro 2002 100 98 97 99 98 98 Tajikistan 1999 65 61 64 63 58 64 Tajikistan 2003 61 64 67 73 60 66 Ukraine 2002 -- -- -- -- -- -- Ukraine 2003 -- -- -- -- -- -- Uzbekistan 2000/01 76 58 54 64 46 57 Uzbekistan 2002 84 66 71 78 67 70 Uzbekistan 2003 87 78 78 82 75 79 Colombia 2003 81 73 48 81 52 67 Turkey 2002 63 61 49 78 39 57 Vietnam 1998 77 78 56 75 42 60 270 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 9 Poverty Profile: Infrastructure Dimension Household access to water Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 100 98 98 99 96 98 Armenia 1998/99 99 96 77 87 91 89 Armenia 2001 100 95 81 93 90 91 Armenia 2002 99 95 73 88 88 87 Armenia 2003 100 98 78 90 88 90 Azerbaijan 2002 92 76 32 68 62 62 Azerbaijan 2003 96 84 42 75 69 69 Belarus 1998 99 91 46 78 81 78 Belarus 1999 100 88 47 78 78 77 Belarus 2000 99 89 48 79 79 78 Belarus 2001 99 91 52 80 80 80 Belarus 2002 98 89 55 79 79 80 Bosnia & Herzegovina 2001 100 100 100 100 100 100 Bosnia & Herzegovina 2004 99 99 99 100 99 99 Bulgaria 1995 100 100 99 100 99 100 Bulgaria 2001 100 100 99 100 98 99 Bulgaria 2003 99 97 80 98 78 92 Estonia 2000 99 94 71 95 80 89 Estonia 2001 100 95 70 95 81 88 Estonia 2002 99 94 70 96 82 88 Estonia 2003 99 93 74 94 83 89 Georgia 1997 96 81 69 80 72 79 Georgia 1998 92 68 69 75 69 75 Georgia 1999 97 82 79 90 79 84 Georgia 2000 94 78 74 85 75 80 Georgia 2001 98 85 72 89 73 82 Georgia 2002 99 86 70 89 71 81 Georgia 2003 99 85 54 82 56 72 Hungary 1998 100 95 89 99 82 93 Hungary 1999 100 95 90 99 83 94 Hungary 2000 100 96 90 99 85 95 Hungary 2001 100 100 100 100 100 100 Hungary 2002 100 100 100 100 100 100 Kazakhstan 2001 100 96 86 95 89 92 Kazakhstan 2002 100 96 89 97 88 93 Kazakhstan 2003 100 96 87 97 89 92 Kyrgyz Rep. 2000 92 96 77 91 79 83 Kyrgyz Rep. 2001 92 96 73 88 70 81 Kyrgyz Rep. 2002 92 96 72 91 75 80 Kyrgyz Rep. 2003 87 97 77 95 82 83 Latvia 2002 97 85 56 95 64 81 Latvia 2003 98 87 59 95 66 83 Lithuania 1998 96 94 60 93 67 84 Lithuania 1999 97 93 61 93 66 84 Lithuania 2000 96 92 59 94 64 83 Lithuania 2001 98 93 62 95 69 85 Lithuania 2002 97 94 63 95 66 85 Lithuania 2003 97 91 56 95 61 82 Appendix: Data and Methodology 271 Use of clean fuels for heating Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 80 60 24 56 30 41 Armenia 1998/99 58 31 7 31 28 28 Armenia 2001 51 29 16 42 24 30 Armenia 2002 54 21 17 36 24 29 Armenia 2003 61 22 13 44 19 30 Azerbaijan 2002 -- -- -- -- -- -- Azerbaijan 2003 -- -- -- -- -- -- Belarus 1998 -- -- -- -- -- -- Belarus 1999 -- -- -- -- -- -- Belarus 2000 -- -- -- -- -- -- Belarus 2001 -- -- -- -- -- -- Belarus 2002 -- -- -- -- -- -- Bosnia & Herzegovina 2001 48 8 5 22 10 14 Bosnia & Herzegovina 2004 -- -- -- -- -- -- Bulgaria 1995 90 62 9 62 40 49 Bulgaria 2001 84 46 3 57 16 37 Bulgaria 2003 95 49 9 59 26 43 Estonia 2000 -- -- -- -- -- -- Estonia 2001 -- -- -- -- -- -- Estonia 2002 -- -- -- -- -- -- Estonia 2003 -- -- -- -- -- -- Georgia 1997 -- -- -- -- -- -- Georgia 1998 -- -- -- -- -- -- Georgia 1999 -- -- -- -- -- -- Georgia 2000 -- -- -- -- -- -- Georgia 2001 -- -- -- -- -- -- Georgia 2002 -- -- -- -- -- -- Georgia 2003 -- -- -- -- -- -- Hungary 1998 96 84 63 93 54 78 Hungary 1999 95 85 63 92 55 79 Hungary 2000 94 83 63 91 57 78 Hungary 2001 95 84 63 91 55 79 Hungary 2002 95 83 63 89 56 78 Kazakhstan 2001 98 89 64 91 63 78 Kazakhstan 2002 98 93 73 94 70 85 Kazakhstan 2003 98 93 75 94 72 85 Kyrgyz Rep. 2000 -- -- -- -- -- -- Kyrgyz Rep. 2001 -- -- -- -- -- -- Kyrgyz Rep. 2002 -- -- -- -- -- -- Kyrgyz Rep. 2003 -- -- -- -- -- -- Latvia 2002 -- -- -- -- -- -- Latvia 2003 -- -- -- -- -- -- Lithuania 1998 -- -- -- -- -- -- Lithuania 1999 -- -- -- -- -- -- Lithuania 2000 -- -- -- -- -- -- Lithuania 2001 -- -- -- -- -- -- Lithuania 2002 -- -- -- -- -- -- Lithuania 2003 -- -- -- -- -- -- (Table continues on the following page.) 272 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 9 (continued) Poverty Profile: Infrastructure Dimension Household access to water Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 94 99 89 96 90 94 Macedonia, FYR 2003 97 99 85 97 90 94 Moldova 1998 99 54 2 46 15 29 Moldova 1999 98 55 1 50 16 29 Moldova 2000 99 56 2 49 22 30 Moldova 2001 99 57 3 46 21 30 Moldova 2002 99 52 3 48 18 29 Moldova 2003 100 54 3 43 22 30 Poland 1998 98 99 90 98 90 95 Poland 1999 98 99 91 98 92 96 Poland 2000 98 99 91 99 91 96 Poland 2001 98 99 93 99 93 97 Poland 2002 97 99 94 99 93 97 Romania 1998 85 89 18 76 36 57 Romania 1999 85 90 19 78 34 58 Romania 2000 83 89 20 78 36 58 Romania 2001 96 90 21 82 33 60 Romania 2002 85 90 20 82 31 58 Romania 2003 85 90 19 80 31 57 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 99 89 44 84 72 77 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 100 90 42 87 65 77 Russian Fed. 2001 100 90 49 90 68 80 Russian Fed. 2002 100 90 49 89 68 80 Serbia & Montenegro 2002 96 98 79 96 82 90 Tajikistan 1999 95 96 91 93 93 92 Tajikistan 2003 100 93 56 78 57 67 Ukraine 2002 100 83 23 74 56 64 Ukraine 2003 100 86 22 76 54 66 Uzbekistan 2000/01 99 85 55 78 57 68 Uzbekistan 2002 100 89 57 82 64 70 Uzbekistan 2003 100 90 60 85 65 72 Colombia 2003 99 97 52 96 70 86 Turkey 2002 97 99 85 99 84 94 Vietnam 1998 99 78 57 78 52 63 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Azerbaijan, Belarus, Estonia, Hungary, Kazakhstan, Latvia, Lithuania, Moldova, Romania, Serbia & Montenegro, Ukraine, and Turkey: running/piped water. Latvia, FYR Macedonia, Poland (1998), Romania, Serbia and Montenegro: column "Capital" includes some rural dwellings. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains 5 urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Russian Federation: data for Ingushetiya Republic is not available for 1999. Turkey: "Capital" contains estimates for Ankara and Istanbul and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. -- = not available. Appendix: Data and Methodology 273 Use of clean fuels for heating Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 -- -- -- -- -- -- Macedonia, FYR 2003 -- -- -- -- -- -- Moldova 1998 -- -- -- -- -- -- Moldova 1999 -- -- -- -- -- -- Moldova 2000 -- -- -- -- -- -- Moldova 2001 -- -- -- -- -- -- Moldova 2002 -- -- -- -- -- -- Moldova 2003 100 70 7 48 27 36 Poland 1998 -- -- -- -- -- -- Poland 1999 -- -- -- -- -- -- Poland 2000 -- -- -- -- -- -- Poland 2001 -- -- -- -- -- -- Poland 2002 -- -- -- -- -- -- Romania 1998 77 81 9 67 27 49 Romania 1999 78 81 9 69 25 49 Romania 2000 79 82 10 70 27 50 Romania 2001 94 82 13 74 26 53 Romania 2002 82 77 11 72 21 48 Romania 2003 82 76 11 71 20 47 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 -- -- -- -- -- -- Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 -- -- -- -- -- -- Russian Fed. 2001 -- -- -- -- -- -- Russian Fed. 2002 -- -- -- -- -- -- Serbia & Montenegro 2002 -- -- -- -- -- -- Tajikistan 1999 77 38 4 28 10 14 Tajikistan 2003 88 56 7 33 19 24 Ukraine 2002 -- -- -- -- -- -- Ukraine 2003 -- -- -- -- -- -- Uzbekistan 2000/01 -- -- -- -- -- -- Uzbekistan 2002 -- -- -- -- -- -- Uzbekistan 2003 -- -- -- -- -- -- Colombia 2003 -- -- -- -- -- -- Turkey 2002 43 9 1 27 3 11 Vietnam 1998 -- -- -- -- -- -- 274 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 10 Poverty Profile: Housing Dimension Tenancy rights Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 88 90 97 94 94 94 Armenia 1998/99 94 86 93 93 91 92 Armenia 2001 94 90 94 94 92 93 Armenia 2002 89 93 94 92 92 92 Armenia 2003 88 90 95 92 91 91 Azerbaijan 2002 79 89 97 87 90 89 Azerbaijan 2003 80 93 97 89 92 90 Belarus 1998 30 49 75 63 45 54 Belarus 1999 36 50 77 64 47 56 Belarus 2000 58 62 78 75 56 67 Belarus 2001 59 63 75 72 56 66 Belarus 2002 58 70 75 77 62 69 Bosnia & Herzegovina 2001 75 75 84 83 66 79 Bosnia & Herzegovina 2004 86 83 89 87 78 86 Bulgaria 1995 91 93 96 95 89 93 Bulgaria 2001 91 90 92 88 89 91 Bulgaria 2003 77 87 90 86 82 87 Estonia 2000 79 86 88 85 78 84 Estonia 2001 86 88 88 87 84 87 Estonia 2002 86 88 88 91 81 87 Estonia 2003 83 89 91 91 81 88 Georgia 1997 88 77 97 91 89 89 Georgia 1998 91 80 97 92 90 91 Georgia 1999 92 81 96 91 90 91 Georgia 2000 94 92 99 96 96 96 Georgia 2001 94 95 99 97 96 97 Georgia 2002 92 93 98 95 95 95 Georgia 2003 91 89 99 94 94 95 Hungary 1998 77 89 95 93 82 89 Hungary 1999 78 90 95 93 82 89 Hungary 2000 80 89 96 93 81 90 Hungary 2001 83 89 95 92 82 90 Hungary 2002 83 89 95 93 83 90 Kazakhstan 2001 93 93 97 96 94 95 Kazakhstan 2002 90 94 98 95 96 96 Kazakhstan 2003 92 95 98 96 97 96 Kyrgyz Rep. 2000 83 92 99 94 97 95 Kyrgyz Rep. 2001 87 95 98 96 95 96 Kyrgyz Rep. 2002 88 96 98 96 96 96 Kyrgyz Rep. 2003 84 98 97 94 95 95 Latvia 2002 79 66 81 85 62 75 Latvia 2003 82 70 80 85 62 78 Lithuania 1998 88 89 85 90 81 87 Lithuania 1999 87 88 85 89 79 87 Lithuania 2000 89 88 87 91 81 88 Lithuania 2001 88 87 84 88 79 86 Lithuania 2002 85 86 84 85 79 85 Lithuania 2003 85 86 83 85 80 85 Appendix: Data and Methodology 275 Overcrowded housing (more than 3 per room or less than 6 sq m per person) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Albania 2002 10 16 19 4 35 17 Armenia 1998/99 23 17 14 7 31 18 Armenia 2001 19 11 6 7 24 11 Armenia 2002 20 10 8 9 18 12 Armenia 2003 19 15 8 10 18 13 Azerbaijan 2002 7 4 11 4 14 8 Azerbaijan 2003 5 3 6 2 8 5 Belarus 1998 11 11 7 4 19 10 Belarus 1999 14 9 9 4 17 10 Belarus 2000 12 9 6 3 18 9 Belarus 2001 12 9 6 3 15 9 Belarus 2002 11 8 7 3 17 8 Bosnia & Herzegovina 2001 5 13 6 2 17 8 Bosnia & Herzegovina 2004 4 4 2 2 3 3 Bulgaria 1995 5 3 4 1 10 4 Bulgaria 2001 3 4 7 1 21 5 Bulgaria 2003 5 6 8 2 15 6 Estonia 2000 1 1 1 1 3 1 Estonia 2001 1 1 1 0 3 1 Estonia 2002 1 0 2 0 3 1 Estonia 2003 1 1 1 0 2 1 Georgia 1997 27 16 10 12 19 16 Georgia 1998 20 12 5 6 13 11 Georgia 1999 19 9 4 5 13 9 Georgia 2000 16 12 6 7 11 10 Georgia 2001 16 13 4 7 11 9 Georgia 2002 16 11 6 7 13 10 Georgia 2003 15 13 5 6 13 9 Hungary 1998 2 1 1 0 3 1 Hungary 1999 2 1 1 0 4 1 Hungary 2000 1 0 1 0 2 0 Hungary 2001 1 0 0 0 2 0 Hungary 2002 1 1 1 0 3 1 Kazakhstan 2001 8 11 16 4 27 13 Kazakhstan 2002 13 9 14 4 25 11 Kazakhstan 2003 13 8 11 3 19 9 Kyrgyz Rep. 2000 25 22 14 9 28 17 Kyrgyz Rep. 2001 24 20 14 9 27 17 Kyrgyz Rep. 2002 28 23 14 10 33 18 Kyrgyz Rep. 2003 28 20 13 7 30 17 Latvia 2002 2 2 3 0 8 3 Latvia 2003 2 1 3 0 6 2 Lithuania 1998 4 3 6 1 12 4 Lithuania 1999 4 3 5 1 8 4 Lithuania 2000 3 3 3 1 7 3 Lithuania 2001 3 3 4 0 8 3 Lithuania 2002 3 2 4 1 7 3 Lithuania 2003 3 2 3 0 7 3 (Table continues on the following page.) 276 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union TABLE 10 (continued) Poverty Profile: Housing Dimension Tenancy rights Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 100 100 100 100 100 100 Macedonia, FYR 2003 100 100 100 100 100 100 Moldova 1998 47 78 98 80 90 85 Moldova 1999 53 81 99 80 89 87 Moldova 2000 55 82 98 80 89 87 Moldova 2001 57 80 99 83 89 88 Moldova 2002 67 86 100 87 91 91 Moldova 2003 70 91 99 90 93 92 Poland 1998 56 49 89 68 64 65 Poland 1999 56 50 89 69 65 66 Poland 2000 63 55 90 74 65 70 Poland 2001 60 53 89 73 60 68 Poland 2002 59 54 89 74 63 69 Romania 1998 93 93 95 95 90 94 Romania 1999 92 93 96 95 91 94 Romania 2000 91 94 96 95 91 95 Romania 2001 94 94 97 96 92 95 Romania 2002 94 95 97 96 93 96 Romania 2003 93 94 97 96 93 95 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 94 90 92 91 89 91 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 92 90 92 92 90 91 Russian Fed. 2001 90 90 92 89 91 90 Russian Fed. 2002 91 92 93 91 94 92 Serbia & Montenegro 2002 87 85 92 86 88 88 Tajikistan 1999 86 83 94 91 87 92 Tajikistan 2003 82 86 97 93 92 94 Ukraine 2002 74 80 97 87 81 86 Ukraine 2003 74 79 97 87 84 85 Uzbekistan 2000/01 94 91 96 94 94 95 Uzbekistan 2002 97 93 97 96 96 96 Uzbekistan 2003 95 95 98 96 96 97 Colombia 2003 55 55 64 63 60 57 Turkey 2002 62 66 89 73 77 74 Vietnam 1998 77 93 98 93 96 96 Source: World Bank staff estimates using the ECA Household Surveys Archive. Note: Latvia, FYR Macedonia, Poland (1998), Romania, and Serbia and Montenegro: column "Capital" includes some rural dwellings. Lithuania: column "Capital" contains estimates for the five largest cities. Poland (1999­2002): column "Capital" includes urban dwellings of Mazowieckie vojevodship, which contains five urban counties--Warsaw, Radom, Plock, Siedlce, Ostroleka. Turkey: "Capital" contains estimates for Ankara and Istanbul and includes some rural dwellings. Vietnam: column "Capital" contains estimates for Hanoi and Ho Chi Minh City. -- = not available. Appendix: Data and Methodology 277 Overcrowded housing (more than 3 per room or less than 6 sq m per person) Country Year Capital Other urban Rural Top quintile Bottom quintile Average Macedonia, FYR 2002 11 7 12 4 22 10 Macedonia, FYR 2003 9 8 7 2 19 8 Moldova 1998 17 13 7 6 15 10 Moldova 1999 18 12 5 10 11 9 Moldova 2000 16 10 5 7 13 8 Moldova 2001 17 9 5 6 12 8 Moldova 2002 16 7 5 5 12 7 Moldova 2003 18 7 4 5 13 7 Poland 1998 6 5 10 1 18 7 Poland 1999 6 5 10 1 18 7 Poland 2000 6 5 9 1 20 7 Poland 2001 6 5 9 1 18 6 Poland 2002 7 5 8 1 18 6 Romania 1998 7 8 14 2 29 10 Romania 1999 7 8 15 1 32 11 Romania 2000 7 9 14 1 33 11 Romania 2001 6 7 14 1 31 10 Romania 2002 8 9 14 1 32 11 Romania 2003 8 8 14 1 30 10 Russian Fed. 1997 -- -- -- -- -- -- Russian Fed. 1998 9 13 11 8 19 12 Russian Fed. 1999 -- -- -- -- -- -- Russian Fed. 2000 8 12 12 7 21 12 Russian Fed. 2001 16 12 8 8 18 11 Russian Fed. 2002 18 10 10 7 18 11 Serbia & Montenegro 2002 4 3 2 2 6 3 Tajikistan 1999 37 36 42 26 57 40 Tajikistan 2003 36 33 32 17 50 32 Ukraine 2002 3 3 3 0 8 3 Ukraine 2003 5 4 4 1 9 4 Uzbekistan 2000/01 10 14 11 7 20 12 Uzbekistan 2002 9 10 9 7 13 9 Uzbekistan 2003 13 9 7 7 11 8 Colombia 2003 5 6 15 0 22 8 Turkey 2002 1 6 7 0 19 5 Vietnam 1998 14 21 27 20 31 25 Bibliography Adams, R. 2002. "Economic Growth, Inequality and Poverty: Findings from a New Data Set." Policy Research Working Paper 2972, World Bank, Washington, DC. Andren, Daniela, John S. Earle, and Dana Sapatoru. 2004. "The Wage Effects of Schooling under Socialism and in Transition: Evidence from Romania, 1950­2000." Staff Working Paper 04-108, W.E. Upjohn Institute for Employment Research, Kalamazoo. Andrienko, Yuri, and Sergei Guriev. 2004. "Determinants of Interregional Mobil- ity in Russia: Evidence from Panel Data." Economics of Transition 12 (1): 1­27. Aslund, Anders. 2001. "The Myth of Output Collapse after Communism." Working Paper 18, Carnegie Endowment for International Peace. Wash- ington, DC. Atkinson, Anthony B., Eric Marlier, and Brian Nolan. 2004. "Indicators and Targets for Social Inclusion in the European Union." Journal of Common Market Studies 42 (March): 47­75. Atkinson, Anthony B., and John Micklewright. 1992. Economic Transformation in Eastern Europe and the Distribution of Income. Cambridge, U.K.: Cambridge University Press. Atkinson, Tony, Bea Cantillon, Eric Marlier, and Brian Nolan. 2002. Social Indicators: The EU and Social Inclusion. Oxford: Oxford University. Babetskii, Ian, Alexandre Kolev, and Mathilde Maurel. 2003. "Kyrgyz Labour Market in the Late 1990s: The Challenge of Formal Job Creation." Com- parative Economic Studies 45 (4, December): 493­519. 279 280 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Bartelsman, Eric, John Haltiwanger, and Stefano Scarpetta. 2004. "Microeco- nomic Evidence of Creative Destruction in Industrial and Developing Countries." Discussion Paper 1374, Institute for the Study of Labor (IZA), Bonn. Bashmakov, I. 2004. "Analysis of Situation with Collectability of Payments for Housing and Public Utility Services." Photocopy, Center for Energy Efficiency, Moscow. Bazen, Stephen, and Patrick Moyes. 2003. "International Comparisons of Income Distribution." Luxembourg Income Study Working Paper No. 341, Luxembourg. Blackstone Corporation (Canada) and Agency of Socioeconomic Analysis (Bulgaria). 2004. "The Economic and Social Impact of Farm Restructuring in ECA: Bulgaria." An input to Nora Dudwick, Karin Fock, and David Sedick: "A Stocktaking of Land Reform and Farm Restructuring in Bul- garia, Moldova, Azerbaijan, and Kazakhstan." Social Development Unit, Europe and Central Asia Region, World Bank, Washington, DC. Bloem, Adriaan M., Paul Cotterell, and Terry Gigantes. 1998. "National Accounts in Transition Countries: Balancing the Biases?" Review of Income and Wealth 44 (1): 1­24. Boldrin, Michele, and Fabio Canova. 2001. "Inequality and Convergence in Europe's Regions: Reconsidering European Regional Policies." Economic Policy 16 (32, April): 205­53. Bonilla-Chacin, Maria E., Edmundo Murrugarra, and Moukim Temourov. 2002. "Health Care during Transition and Health Systems Reform: Evi- dence from the Poorest CIS Countries." Human Development Depart- ment, Europe and Central Asia Region, World Bank, Washington, DC. Borak, Neven, and Lovrenc Pfajfar. 2002. "Inequalities in Income Distribu- tion in Slovenia." Post-Communist Economies 14 (4): 455­68. Bornhorst, Fabian, and Simon Commander. 2004. "Regional Unemployment and its Persistence in Transition Countries." Discussion Paper 1074, IZA, Bonn. Bos, E., A. Ryskulova, S. Patel, and M. Hiraga. 2002. "Infant and Child Mor- tality Rates in Eastern Europe and Central Asia: What's Wrong with the Data?" Photocopy, World Bank, Washington, DC. Bourguignon, F. 2003. "The Growth Elasticity of Poverty Reduction: Explain- ing Heterogeneity across Countries and Time Periods." In Inequality and Growth: Theory and Policy Implications, ed. T. Eicher and S. Turnovsky. Cam- bridge: The MIT Press. Brainerd, Elizabeth, and David M. Cutler. 2005. "Autopsy on an Empire: Understanding Mortality in Russia and the Former Soviet Union." Discus- sion Paper 1472, IZA, Bonn. Briceńo, C., A. Estache, and N. Shafik. 2004. "Infrastructure Services in Developing Countries: Access, Quality, Costs, and Policy Reform." Policy Research Working Paper 3468, World Bank, Washington, DC. Brown, J. David, and John S. Earle. 2004a. "Economic Reforms and Produc- tivity-Enhancing Reallocation in the Post-Soviet Transition." Discussion Paper 1044, IZA, Bonn. Bibliography 281 ------. 2004b. "Job Reallocation and Productivity Growth in the Ukrainian Transition." Discussion Paper 1349, IZA, Bonn. Campos, Nauro F., and Dean Jolliffe. 2003. "After, Before and During: Returns to Education in Hungary (1986­1998)." Economic Systems 27: 377­90. Caselli, Francesco, and Silvana Tenreyro. 2004. "Is Poland the Next Spain?" Federal Reserve Board of Boston Public Policy Paper No. 04-8 (June). Boston, MA. Cazes, Sandrine, and Alena Nesporova. 2003. Labour Markets in Transition: Balancing Flexibility & Security in Central and Eastern Europe. Geneva: Inter- national Labour Office. Chen, Shaohua, and Martin Ravallion. 2001. "How Did the World's Poorest Fare in the 1990s?" Review of Income and Wealth 47 (3, September): 283­300. Chernetsky, Taras. Forthcoming. "Magnitudes and Distribution of Remit- tances in Eastern Europe: Evidence from Household Surveys Data and Balance of Payment Statistics." World Bank, Washington, DC. Clert, C., and E. Gomart. 2004. "Human Trafficking in South Eastern Europe: Beyond Crime Control, an Agenda for Social Inclusion and Development." Photocopy, Social Development Unit, Europe and Central Asia Region, World Bank, Washington, DC. Coady, David, Margaret Grosh, and John Hoddinott. 2004. The Targeting of Transfers in Developing Countries: Review of Experience and Lessons. Regional and Sectoral Studies Series. World Bank, Washington, DC. Commander, Simon, and Janos Köllő. 2004. "The Changing Demand for Skills: Evidence from the Transition." Discussion Paper 1073, IZA, Bonn. Commander, Simon, and Yulia Rodionov. 2005. "A Model of the Informal Economy with an Application to Ukraine." IZA-EBRD Conference on Labour Market Dynamics, Role of Institutions and Internal Labour Mar- kets, University of Bologna, May 5­8, 2005. Corak, M. 2005. "Principles and Practicalities in Measuring Child Poverty for the Rich Countries." Working Paper 2005-01, UNICEF, Innocenti Research Centre, Florence. Council of the European Union. 2004. Joint Report by the Commission and the Council on Social Inclusion. Brussels, Belgium: Council of the European Union. Davidow, S. 1996. "Observations on Health Care Issues in the Former Soviet Union." Journal of Community Health 21 (1, February): 51­60. Deaton, A. 2004. "Measuring Poverty in a Growing World (or Measuring Growth in a Poor World)." Revised version of the Review of Economics and Statistics Lecture, presented at Harvard University, April 15, 2003. http://www.wws.princeton.edu/%7Erpds/downloads/deaton_measuring poverty_204.pdf. Deaton, A., and S. Zaidi. 2002. "Guidelines for Constructing Consumption Aggregates for Welfare Analysis." LSMS Working Paper 135, World Bank, Washington, DC. Delteil, Violaine, Ariane Pailhé, and Dominique Redor. 2004. "Comparing Individual Wage Determinants in Western and Central Europe: On the 282 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Way to Convergence? The Cases of France and Hungary." Journal of Com- parative Economics 32 (3, September): 482­99. Dennis, Ian, and Anne-Catherine Guio. 2003. "Monetary Poverty in EU Acceding and Candidate Countries." Statistics In Focus. Theme 3: 21/2003. Luxembourg: European Commission, Office for Official Publications of the European Communities. ------. 2004. "Poverty and Social Exclusion in the EU." Statistics In Focus. Population and Social Conditions: 16/2004. Luxembourg: European Com- mission, Office for Official Publications of the European Communities. Dikhanov, Yuri. 1996. "Decomposition of Inequality Based on Incomplete Information." Paper presented at the International Association for Research in Income and Wealth (IARIW) 24th General Conference, Lille- hammer, Norway, August 18­24, 1996. Dolado, Juan José, and Vanesa Llorens. 2004. "Gender Wage Gaps by Educa- tion in Spain: Glass Floors versus Glass Ceilings." Discussion Paper 4203, Centre for Economic Policy Research (CEPR), London. Dudwick, N., K. Fock, D. Sedik, and J. Swinnen. 2004. "The Economic and Social Impacts of Farm Restructuring in Eastern Europe and Central Asia." Social Development Unit, Europe and Central Asia Region, World Bank, Washington, DC. Dudwick, N., and H. Shahriari. 2000. "Education in Albania: Changing Atti- tudes and Expectations." Photocopy, Social Development Unit, Europe and Central Asia Region, World Bank, Tirana. Dutz, Mark, Celine Kauffmann, Serineh Najarian, Peter Sanfey, and Ruslan Yemtsov. 2001. "Labour Market States, Mobility and Entrepreneurship in Transition Economies." Working Paper 66, European Bank for Recon- struction and Development (EBRD), London. Epaulard, Anne. 2003. "Macroeconomic Performance and Poverty Reduc- tion." Working Paper 03/72, International Monetary Fund Institute, Washington, DC. Eriksson, Tor, and Mariola Pytlikova. 2004. "Firm-Level Consequences of Large Minimum-Wage Increases in the Czech and Slovak Republics." Labour: Review of Labour Economics and Industrial Relations 18 (1, March): 75­103. Eurelectric. 2003. "Electricity Tariffs as of 1 January 2003 (Published Tariffs)." Brussels: Eurelectric. Web site: http://public.eurelectric.org/. European Commission (EC). 2002. "Health Statistics: Key Data on Health 2002." Luxembourg: Office for Official Publications of the European Com- munities. ------. 2003. "Employment in Europe 2003: Recent Trends and Prospects." Directorate-General for Employment and Social Affairs, Unit EMPL/A.1, Employment and European Social Fund. Luxembourg: Office for Official Publications of the European Communities. ------. 2004. "Household Budget Survey in the Candidate Countries: Methodological Analysis 2003." Luxembourg: Office for Official Publica- tions of the European Communities. Bibliography 283 Feldmann, Horst. 2004. "How Flexible Are Labour Markets in the EU Acces- sion Countries Poland, Hungary, and the Czech Republic?" Comparative Economic Studies 46: 272­310. Fidrmuc, Jan. 2004. "Migration and Regional Adjustment to Asymmetric Shocks in Transition Economies." Journal of Comparative Economics 32: 230­47. Fiszbein, Ariel, ed. 2001. Decentralizing Education in Transition Societies: Case Studies from Central and Eastern Europe. Washington, DC: World Bank Insti- tute (WBI). Fleisher, Belton, ed. 2005. "Symposium on Returns to Schooling in Transi- tion: The Chinese, European, and Russian Experiences." Journal of Com- parative Economics, 33 (June): 223­370. Fleisher, Belton, Klara Sabirianova Peter, and Xiaojun Wang. 2004. "Returns to Skills and the Speed of Reforms: Evidence from Central and Eastern Europe, China, and Russia." Discussion Paper 1182, IZA, Bonn. June. Förster, M., D. K. Jesuit, and T. M. Smeeding. 2005. "Regional Poverty and Income Inequality in Central and Eastern Europe: Evidence from the Lux- embourg Income Study." In Spatial Inequality and Development, ed. R. Kan- bur and A. J. Venables. UNU-WIDER Studies in Development Economics. Oxford: Oxford University Press. Foster, James, Joel Greer, and Erik Thorbecke. 1984. "A Class of Decompos- able Poverty Measures." Econometrica 52 (3): 761­66. Funck, Bernard, and Lodovico Pizzati, ed. 2003. European Integration, Regional Policy, and Growth. Washington, DC: International Bank for Reconstruction and Development, World Bank. Geweke, J., G. Gowrisankaran, and R. Town. 2003. "Bayesian Inference for Hospital Quality in a Selection Model." Econometrica 71(4, July): 1215­1238. Godinho, Joana, Adrian Renton, Viatcheslav Vinogradov, Thomas Novotny, George Gotsadze, Mary-Jane Rivers, and Mario Bravo. 2005. "Reversing the Tide: Priorities for HIV/AIDS Prevention in Central Asia." Working Paper 54, World Bank, Washington, DC. Gomart, E. 2000. "Social Assessment of Poverty in Croatia, December 1999­March 2000." Photocopy, World Bank, Washington, DC. Gorodnichenko, Yuri, and Klara Sabirianova Peter. 2004. "Returns to School- ing in Russia and Ukraine: A Semiparametric Approach to Cross-Country Comparative Analysis." Discussion Paper 1325, IZA, Bonn. Gottschalk, Peter, and Timothy M. Smeeding. 1997. "Cross-National Com- parisons of Earnings and Income Inequality." Journal of Economic Literature 35 (June): 633­87. Grammenos, Stefanos. 2003. Illness, Disability and Social Inclusion. European Foundation for the Improvement of Living and Working Conditions. Cen- tre for European Social and Economic Policy. Luxembourg: Office for Offi- cial Publications of the European Communities. Gregory, Paul R., and Valery Lazarev. 2004. "Structural Change in Russian Transition." Discussion Paper 896, Yale University Economic Growth Cen- ter, New Haven. 284 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Haltiwanger, John, Harmut Lehmann, and Katherine Terell. 2003. "Sympo- sium on Job Creation and Job Destruction in Transition Economies." Eco- nomics of Transition 11: 205­19. Hamilton, Ellen, Deniz Baharoglu, Bill Denning, Christine Kessides, Alexan- dre Kolev, and Maka Lomaia. 2004. "Dimensions of Urban Poverty in ECA." Photocopy. Europe and Central Asia Region, Energy and Infra- structure Unit. World Bank, Washington, DC. Hinnosaar, Marit, and Tairi Rőőm. 2003. "The Impact of Minimum Wage on the Labor Market in Estonia: An Empirical Analysis." Working Paper 8, Eesti Pank (Bank of Estonia), Tallinn, Estonia. Holtzman, S. B., and T. Nezam. 2004. "Living in Limbo. Conflict-Induced Dis- placement in Europe and Central Asia." Social Development Unit, Europe and Central Asia Region, World Bank, Washington, DC. International Labour Organization (ILO). 2003. Key Indicators of the Labour Market. CD-ROM, Geneva: ILO. ------. 2004. World Employment Report 2004­05: Employment, Productivity, and Poverty Reduction. Geneva: ILO. Ivaschenko, Oleksiy. 2005. "The Patterns and Determinants of Longevity in Russia's Regions: Evidence from Panel Data." Photocopy, World Bank, Washington, DC. Jurajda, Stěpán, and Katarína Mathernová. 2004. "How to Overhaul the Labor Market: Political Economy of Recent Czech and Slovak Reforms." Background Paper prepared for the World Development Report 2005, World Bank, Washington, DC. Kakwani N., and Z. Sajaia. 2004. "Internationally Comparable Poverty Lines for CIS Countries." Photocopy. World Bank, Poverty Reduction and Eco- nomic Management Unit, Europe and Central Asia Region, DFID, United Kingdom. Kapsos, Steven. 2004. "Estimating Growth Requirements for Reducing Work- ing Poverty: Can the World Halve Working Poverty by 2015?" Employ- ment Strategy Paper 2004/14, ILO, Geneva. Kertesi, Gábor, and János Köllő. 2001. "Economic Transformation and the Revaluation of Human Capital ­ Hungary, 1986­1999." Working Paper on the Labour Market 2001/4, Institute of Economics, Hungarian Academy of Sciences, Budapest. ------. 2003. "The Employment Effects of Nearly Doubling the Minimum Wage: The Case of Hungary." Working Paper on the Labour Market 2003/06, Institute of Economics, Budapest University of Economics, Budapest. Kolenikov, Stanislav, and Anthony Shorrocks. 2003. "A Decomposition Analysis of Regional Poverty in Russia." UNU/WIDER Discussion Paper 2003/74, Helsinki. Kolev, Alexandre. 2003. "Joblessness and Precarious Work in Bulgaria: Addressing the Multiple Aspects of Vulnerability in the Labour Market." SP Discussion Paper 0303, World Bank, Washington, DC. Kolev, Alexandre, and Sabine Bernabe. 2003. "Identifying Vulnerable Groups in the Kyrgyz Labour Market: Some Implications for the National Poverty Bibliography 285 Reduction Strategy." CASE Paper 71, Centre for Analysis of Social Exclu- sion (CASE), London School of Economics, London. Lehmann, Hartmut, and Jonathan Wadsworth. 2001. "Wage Arrears and the Distribution of Earnings in Russia." William Davidson Institute Working Paper 421, University of Michigan, Stephen M. Ross Business School, Ann Arbor. Lekhan, Valery, Volodymyr Rudiy, and Ellen Nolte. 2004. Health Care Systems in Transition: Ukraine. Copenhagen: WHO Regional Office for Europe on behalf of the European Observatory on Health Systems and Policies. Macours, Karen, and Johan F. M. Swinnen. 2004. "Rural Poverty in ECA." Background Paper for the present study, World Bank, Washington, DC. Marshall, Martin, Sheila Leatherman, Soeren Mattke, and the Members of the OECD Health Promotion, Prevention, and Primary Care Panel. 2004. "Selecting Indicators for the Quality of Health Promotion, Prevention and Primary Care at the Health Systems Level in OECD Countries." Health Technical Paper 16, OECD, Paris. Mencini, L., and G. Redmond. Child Income Poverty in CIS and SEE. Innocenti Social Monitor 2005. Florence: UNICEF, Innocenti Research Centre. Mete, Cem, and Shirley H. Liu. 2005. "Predictors and Implications of Sickness in a Transition Economy." Photocopy, World Bank, Washington DC. Mincer, J. 1974. Schooling, Experience and Earnings. New York: Columbia Uni- versity Press. Mullis, Ina V. S., Michael O. Martin, Eugenio J. Gonzalez, and Stephen J. Chrostowski. 2004. TIMSS 2003 International Mathematics Report. Interna- tional Association for the Evaluation of Education Achievement. Chestnut Hill, MA: TIMSS and PIRLS International Study Center, Lynch School of Education, Boston College. Munich, Daniel. 2003. Human Capital in Transition: Czech Students and Workers Adapting to the Market. Center for Economics Research and Graduate Edu- cation (CERGE), Charles University, Economics Institute. Prague: Acad- emy of Sciences of the Czech Republic (ASCR). Munich, Daniel, Jan Svejnar, and Katherine Terrell. 2005. "Returns to Human Capital under the Communist Wage Grid and during the Transi- tion to a Market Economy." Review of Economics and Statistics 87 (1, Febru- ary): 100­123. Murrugarra, E., and A. Sethi. 2005. "Quality of Education in Transition Economies: Household, Teachers, and School Factors across the Educa- tional Achievement Distribution." Background Paper for the present study, World Bank, Washington, DC. Murthi, Mamta, and Erwin Tiongson. Forthcoming. "Attitudes to Equality: The Socialist Legacy Revisited." World Bank, Washington, DC. Narayan, Deepa, Raj Patel, Kai Schafft, Anne Rademacher, and Sarah Koch- Schulte. 2000. Voices of the Poor: Can Anyone Hear Us? New York: Oxford University Press. Nesperova, Alena. 2002. "Why Unemployment Remains So High in Central and Eastern Europe." Employment Paper 2002/43, International Labour Organization, Geneva. 286 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Newell, Andrew, and Barry Reilly. 1997. "Rates of Return to Educational Qualifications in the Transitional Economies." Discussion Paper in Eco- nomics 03/97, University of Sussex, Brighton, United Kingdom. Newell, Andrew, and Mieczyslaw W. Socha. 2005. "The Distribution of Wages in Poland, 1992­2002." Discussion Paper 1485, IZA, Bonn. Nolte, Ellen, Martin McKee, and Anna Gilmore. 2004. "Morbidity and Mor- tality in Transition Countries in the European Context." Paper prepared for the European Population Forum 2004: Population Challenges and Pol- icy Responses, January 12­14, 2004, United Nations Economic Commis- sion for Europe (UNECE), Geneva. OECD. 2003. "Summary Results of ECP 2000: A Note by the OECD Secre- tariat." Working Paper 6, Statistical Commission and Economic Commission for Europe, Conference of European Statisticians, Joint Consultation on the European Comparison Programme, Geneva, March 31­April 2, 2003. O'Higgins, Niall. 2003. "Trends in the Youth Labour Market in Developing and Transition Countries." Social Protection Discussion Paper 0321, World Bank, Washington, DC. Paci, Pierella. 2002. Gender in Transition. Washington DC: World Bank. Paci, Pierella, and Barry Reilly. 2004. "Does Economic Liberalization Reduce Gender Inequality in the Labor Market? The Experience of the Transition Economies of Europe and Central Asia." Photocopy, World Bank, Wash- ington, DC. Parkin, D. M., S. R. Whelan, J. Ferlay, and H. Storm. 2005. Cancer Incidence in Five Continents. Vols. I­VII. IARC CancerBase No. 7. IARC: Lyon. Peter, Klara Sabirianova. 2003. "Skill-Biased Transition: The Role of Markets, Institutions, and Technological Change." Discussion Paper 893, IZA, Bonn. Raiser, Martin, Mark Schaffer, and Johannes Schuchhardt. 2004. "Bench- marking Structural Change in Transition." Structural Change and Economic Dynamics 15 (1): 47­81. Rashid, Mansoora, and Jan Rutkoswki. 2001. "Labor Markets in Transition Economies: Recent Developments and Future Challenges." Social Protec- tion Discussion Paper 0111, World Bank, Washington, DC. Ravallion, Martin. 1994. Poverty Comparisons. Fundamentals of Pure and Applied Economics 56. Chur, Switzerland: Harwood Academic Press. ------. 2001a. "Growth, Inequality, and Poverty: Looking Beyond Averages." Policy Research Working Paper 2558, World Bank, Washington, DC. ------. 2001b. "Measuring Aggregate Welfare in Developing Countries: How Well Do National Accounts and Surveys Agree?" Policy Research Working Paper 2665, World Bank, Washington, DC. ------. 2004. "Pro-Poor Growth: A Primer." Policy Research Working Paper 3242, World Bank, Washington, DC. Ravallion, Martin, and Shaohua Chen. 1997. "What Can New Survey Data Tell Us about Recent Changes in Distribution and Poverty?" World Bank Economic Review 11 (2): 357­82. Ravallion, Martin, and Michael Lokshin. 1999. "Subjective Economic Wel- fare." Policy Research Working Paper 2106, World Bank, Washington, DC. Bibliography 287 Revenga, A., D. Ringold, and W. M. Tracy. 2002. "Poverty and Ethnicity: A Cross-Country Study of Roma Poverty in Central Europe." Technical Paper 531, World Bank, Washington, DC. Ringold, Dena, Mitchell A. Orenstein, and Erika Wilkens. 2003. "Roma in an Expanding Europe: Breaking the Poverty Cycle." Conference Edition. IBRD. World Bank: Washington, DC. Rozelle, Scott, and Johan F. M. Swinnen. 2004. "Success and Failure of Reform: Insights from the Transition of Agriculture." Journal of Economic Literature 42 (June): 404­56. Rutkowski, Jan. 2001. "Earnings Inequality in Transition Economies of Cen- tral Europe: Trends and Patterns during the 1990s." Social Protection Dis- cussion Paper 0117, World Bank, Washington, DC. ------. 2003a. "Does Strict Employment Protection Discourage Job Cre- ation? Evidence from Croatia." Policy Research Paper 3104, World Bank, Washington, DC. ------. 2003b. "Firms, Jobs, and Employment in Moldova." Policy Research Paper 3253, World Bank, Washington, DC. Sanfey, Peter, Elisabetta Falcetti, Anita Taci, and Sladjana Tepic. 2004. "Spot- light on South-Eastern Europe: An Overview of Private Sector Activity and Investment." EBRD Publications Desk, London. Sanfey, Peter, and Utku Teksoz. 2005. "Does Transition Make You Happy?" Working Paper 91, EBRD, London. Schneider, Friedrich. 2002. "The Size and Development of the Shadow Economies of 22 Transition and 21 OECD Countries." Discussion Paper 514, IZA, Bonn. Schneider, Friedrich, and Robert Klinglmair. 2004. "Shadow Economies around the World: What Do We Know?" Discussion Paper 1043, IZA, Bonn. Schreyer, Paul, and Dirk Pilat. 2001. "Measuring Productivity." OECD Eco- nomic Studies 33: 127­70. Shore-Sheppard, Lara D. 2005. "Stemming the Tide? The Effect of Expanding Medicaid Eligibility on Health Insurance." Working Paper 11091, National Bureau of Economic Research (NBER), Cambridge, MA. Sierminska, Eva, and Thesia Garner. 2002 "A Comparison of Income, Expen- ditures, and Home Market Value Distributions Using Luxembourg Income Study Data from the 1990s." Luxembourg Income Study Working Paper No. 343, Luxembourg. Signoret, Jose, and Edmundo Murrugarra. 2003. "Vulnerability in Consump- tion, Education, and Health: Evidence from Moldova during the Russian Crisis." Policy Research Working Paper 3010, World Bank, Washington, DC. Sullivan, Dennis H., and Timothy M. Smeeding. 1997. "Educational Attain- ment and Earnings Inequality in Eight Nations." Discussion Paper, Lux- embourg Income Study Working Paper 164 (July), Luxembourg. Szalai, Julia. 2002. "Social Outcasts in 21st Century Hungary." Review of Soci- ology of the Hungarian Sociological Association 8 (2, December): 35­52. 288 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Toralieva, Gulnura, and Dina Maslova. 2004. "Kyrgyzstan: Disabled Workers Written Off." On Institute of War and Peace Web site: http://www.iwpr.net/index.pl?archive/rca/rca_200401_255_2_eng.txt. Treisman, Daniel, and Vladimir Gimpelson. 2001. "Political Business Cycles and Russian Elections, or the Manipulations of `Chudar.'" British Journal of Political Science 31: 225­46. United Nations. 2000. "Copenhagen Declaration on Social Develop- ment." Agreements, World Summit for Social Development, Geneva, June 2000. Document A/CONF 166/9, United Nations, New York. United Nations Millennium Project. 2005. A Home in the City. Task Force on Improving the Lives of Slum Dwellers. London and Sterling, VA: Earth- scan. van de Walle, Dominique. 2002. "The Static and Dynamic Incidence of Viet- nam's Public Safety Net." Policy Research Working Paper 2791, World Bank, Washington, DC. Vandycke, N. 2001. "Access to Education for the Poor in Europe and Central Asia: Preliminary Evidence and Policy Implications." Technical Paper 511, World Bank, Washington, DC. Vecernik, Jiri. 2001. "Earnings Disparities in the Czech Republic: Evidence of the Past Decade and Cross-National Comparison." Prague Economic Papers 10 (3, September): 201­21. Vodopivec, Milan. 2004. "Labor Market Developments in the 1990s." In Slove- nia: From Yugoslavia to the European Union, ed. Mojmir Mrak, Matija Rojec, and Carlos Silva-Jauregui. 292­314. Washington, DC: World Bank. Wagstaff, Adam, and Eddy van Doorslaer. 2001. "Paying for Health Care: Quantifying Fairness, Catastrophe, and Impoverishment, with Applica- tions to Vietnam, 1993­98." Policy Research Working Paper 2715. World Bank, Washington, DC. Weissman, Ellen M. 1996. "Using Performance Monitoring to Improve Com- munity Health: Conceptual Framework and Community Experience." National Academy of Sciences. Washington, DC: National Academy Press. http://books.nap.edu/html/concept. World Bank. 1990. World Development Report 1990. New York: Oxford Univer- sity Press. ------. 2000a. Making Transition Work for Everyone: Poverty and Inequality in Europe and Central Asia. Washington, DC: World Bank. ------. 2000b. Republic of Latvia, The: Poverty Assessment. 2 vols. Washington, DC: World Bank. ------. 2001a. Hungary: Long-Term Poverty, Social Protection, and the Labor Mar- ket. 2 vols. Washington, DC: World Bank. ------. 2001b. Poland Labor Market Study: The Challenge of Job Creation. Wash- ington, DC: World Bank. ------. 2001c. Slovak Republic: Living Standards, Employment, and Labor Market Study. Washington, DC: World Bank. ------. 2001d. World Development Report 2000/2001: Attacking Poverty. Wash- ington, DC: World Bank. Bibliography 289 ------. 2002a. Armenia: Poverty Update. Washington, DC: World Bank. ------. 2002b. Bulgaria: Poverty Assessment. Washington, DC: World Bank. ------. 2002c. Georgia: Poverty Update. Washington, DC: World Bank. ------. 2002d. Labor Market in Postwar Bosnia and Herzegovina: How to Encour- age Businesses to Create Jobs and Increase Worker Mobility. Washington, DC: World Bank. ------. 2002e. Republic of Tajikistan: Poverty Assessment. Washington, DC: World Bank. ------. 2002f. Russian Labor Market: Moving from Crisis to Recovery. Europe and Central Asia Region, Human Development Sector Unit. Washington, DC: World Bank. ------. 2002g. Slovak Republic: Poverty and Welfare of Roma in the Slovak Republic. Bratislava: World Bank, Foundation S.P.A.C.E., Ineko, and Open Society Institute. ------. 2002h. Transition: The First Ten Years: Analysis and Lessons for Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. ------. 2002i. Turkmenistan: A Profile of Living Standards in Turkmenistan. Washington, DC: World Bank. ------. 2003a. Albania: Poverty Assessment. Washington, DC: World Bank. ------. 2003b. Armenia: Poverty Assessment. Washington, DC: World Bank. ------. 2003c. Azerbaijan Republic: Poverty Assessment. 2 vols. Washington, DC: World Bank. ------. 2003d. Bosnia and Herzegovina: Poverty Assessment. 2 vols. Washington, DC: World Bank. ------. 2003e. FYR Macedonia: Country Economic Memorandum Tackling Unem- ployment. Washington, DC: World Bank. ------. 2003f. Guatemala: Poverty in Guatemala. Washington, DC: World Bank. ------. 2003g. Inequality in Latin America and the Caribbean: Breaking with the Past? Washington, DC: World Bank. ------. 2003h. Kosovo Labor Market Study: Policy Challenges of Formal and Infor- mal Employment. Washington, DC: World Bank. ------. 2003i. Kyrgyz Republic: Enhancing Pro-Poor Growth. Washington, DC: World Bank. ------. 2003j. "Review of Parent-Community Participation in Schools in ECA." Background Paper for Social Development in Europe and Central Asia Region: Issues and Directions. ECA, Socially and Sustainable Development Department. Washington, DC: World Bank. ------. 2003k. Romania: Poverty Assessment. 2 vols. Washington, DC: World Bank. ------. 2003l. Serbia and Montenegro: Poverty Assessment. 2 vols. Washington, DC: World Bank. ------. 2003m. Tracking Nonincome Dimensions of Poverty in ECA. Poverty Reduction and Economic Management Unit and Human Development Sector Unit, Europe and Central Asia Region. Washington, DC: World Bank. 290 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union ------. 2003n. Uzbekistan: Living Standards Assessment. Washington, DC: World Bank. ------. 2004a. "Azerbaijan Municipal Services Study." Photocopy, Infra- structure and Energy Services Department (ECSIE), World Bank, Wash- ington, DC. ------. 2004b. Belarus: Poverty Assessment. 2 vols. Decision Draft. Washington, DC: World Bank. ------. 2004c. Bosnia and Herzegovina: Country Economic Memorandum. Wash- ington, DC: World Bank. ------. 2004d. The Governance of Service Delivery in ECA: A Regional Study. ECA Vice Presidency. Washington, DC: World Bank. ------. 2004e. Kazakhstan: Living Standards Assessment in Kazakhstan. 2 vols. Washington, DC: World Bank. ------. 2004f. Latvia: Poverty Assessment. Decision Draft. Washington, DC: World Bank. ------. 2004g. Latvia: The Quest for Jobs and Growth: A World Bank Country Eco- nomic Memorandum. Washington, DC: World Bank. ------. 2004h. Poland: Growth, Employment and Living Standards in Pre-Acces- sion Poland. Washington, DC: World Bank. ------. 2004i. Recession, Recovery, and Poverty in Moldova. Washington, DC: World Bank. ------. 2004j. World Development Report 2005. Washington, DC: World Bank. ------. 2005a. Dynamics of Vertical Coordination in Agri-Food Supply Chains in ECA. Washington, DC: World Bank. ------. 2005b. Global Economic Prospects 2005: Trade, Regionalism, and Develop- ment. Washington, DC: World Bank. ------. 2005c. MDGs in Europe and Central Asia: Performance and Prospects. Washington, DC: World Bank. ------. 2005d. Operationalizing MDGs: How Effective Is the Long Route? A Case Study of Georgia. Washington, DC: World Bank. ------. 2005e. "Poverty and Social Exclusion of the Roma in Serbia and Montenegro." Photocopy, World Bank, Washington, DC. ------. 2005f. "Pro-Poor Growth in the 1990s: Lessons and Insights from 14 Countries." Washington, DC: World Bank. ------. 2005g. Russian Federation: Reducing Poverty through Growth and Social Policy Reform. Washington, DC: World Bank. ------. 2005h. "Serbia: The Social Safety Net and the Poor." Photocopy, World Bank, Washington, DC. ------. 2005i. World Development Indicators. Washington, DC: World Bank. ------. 2005j. World Development Report 2006: Equity and Development. Wash- ington, DC: World Bank. ------. Forthcoming-a. Enhancing Job Opportunities in Transition Countries: Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. Bibliography 291 ------. Forthcoming-b. Foundations for Sustainable Infrastructure in Europe and Central Asia: Tariffs, Safety Nets, Regulation, Commercialization. Washington, DC: World Bank. ------. Forthcoming-c. From Disintegration to Reintegration: Eastern Europe and the Former Soviet Union in International Trade. Washington, DC: World Bank. ------. Forthcoming-d. Doing Business 2006. Washington, DC: World Bank. World Health Organization (WHO). 2002. Reducing Risks, Promoting Healthy Life: World Health Report 2002. Geneva: UN, WHO. Wu, X., J. Lampietti, and A. S. Meyer. 2004. "Coping with the Cold: Space Heating and the Urban Poor in Developing Countries." Energy Economics 26: 345­57. Yemtsov, Ruslan, Cem Mete, and Stefania Rodica Cnobloch. 2005. "Evolu- tion of the Predictors of Earnings during the Transition." Background Paper for the present study, World Bank, Washington, DC. Yoon, Yang-Ro, Barry Reilly, Gorana Krstic, and Sabine Bernabe. 2003. "A Study of Informal Labour Market Activity in the CIS." Washington, DC: World Bank. Zeikate, S. 2004. "Decentralization of Service Delivery and the Access for the Impoverished People." Background Paper for the present study, World Bank, Washington, DC. Index accountability, 208, 209, 214 utilities, 172­173 health care, 37­38 Azerbaijan, utilities, 173 age, 62, 76, 211 agriculture Belarus employment, 114, 115 catastrophic health expendi- growth and productivity, tures, 168­169 35­36, 131­132, 144, health care spending, 167 205­206, 213­214 trends, 53 promotion, 41 birth control, 69 transition to market, role in, Bulgaria 116­118 education, 156 value added per worker, 115 health care utilization rates, Armenia 166 catastrophic health expendi- tures, 169 catastrophic health expendi- electricity, 180 tures, 168­169 growth rate and inequality, cervical cancer, 163, 187n.9 198 children, 8, 9, 11, 56, 57, 62­64, health care spending, 167 193 health care utilization, 165, mortality, 77n.9 166 see also infant mortality 293 294 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union chronic health conditions, 70 credit market, 206 CIS. See Commonwealth of Inde- rural areas, 36 pendent States Czech Republic, education, 155, cities, 3­4, 14 156, 157, 160 capital, 57, 59, 60, 93­94 capital vs secondary, 10, 14 data secondary, dirty fuels and, availability, 212­213, 181­182 217­219, 220 climate, 4, 76 benchmark countries, Colombia, 48 216­217 health care utilization and consistency, 225­226 spending, 166, 167 countries covered and not Commonwealth of Independent covered, 216 States (CIS) quality, 83, 101, 212­213, defined, 43n.6 217­219, 220, 222 growth, 2 regional household survey low income, 43n.5, 48, 51, archive, 215­220 77n.3 sources, 221 middle income, 43n.4, 48, 51 database, income and nonin- communicable diseases, 148 come poverty, 214 consumption, 44n.12, 105n.12, data collection, 219­220 223 household survey, 42­43 changes, survey data and, nonincome dimensions, 213 97­102 decentralization, 183, 187n.7 durables, 223­224 delivery networks, 147 equivalence scales, 230­231 demographics, 4, 76, 233, equivalent consumption, 250­253 235n.1 health care and, 25­26, 162 estimation, 98­100, 102 deprivation, absolute, 5, 6­7, 8 GDP vs, 195 disabled persons, 65 goods and services, 223 displaced persons, internally growth rates and, 193­195 (IDPs), 65 indicator of welfare, 220, 222 see also migration measurement, 222 distribution per capita, 52, 82, 190 changes in, 87­90 poor vs nonpoor, 134­136 effect, 85 private, 193­195 share, 20, 90 well-being and, 224 doctor-patient contact per per- consumption aggregates, 223, son by country, 186 226 drug abuse, 25, 69 structure of, 224­225 consumption poverty, 51­66 earnings, 109, 134­135, 204 water access, and health, 71 unexplained gap, 204 Index 295 earnings, (continued) employment, (continued) see also wages growth, wage growth vs, economic growth. See growth 110­111 economic opportunities. See low-productive, 140, 142 opportunities productivity and, 142­145 economically inactive poor. See structure of, 111­113 nonworking poor employment rate, 44n.16, 126 economies of scale, 63 survey data vs ILO statistics, education, 2, 22­23, 30, 37, 57, 113 59, 72, 135, 148, 149­161, women, 44n.15 187n.4, 193, 200, 204, employment-to-population 207­208, 233, 235, ratios, 3, 17, 18 254­257, 266­269 energy, 171­182 coverage of, 149­154 enterprises, 35 equity, 37, 208 old, restructured, and new, mathematics performance, 139, 204­205 155 enterprise sector reform, 35, 41, performance, 156­157, 204­205, 213 160­161, 187n.5 entrepreneurship, 89, 105n.10 primary, 150­151 epidemics, 69 quality, 23, 37, 72, 154­161, equivalence scales, 230­231 208 ethnicity, 11, 193 returns to, 123, 145n.7 see also minorities secondary, 151­154 EU-8, 2, 3, 48, 51 subnational governments, Europe and Central Asia (ECA), 158 2 teachers, 159 European Union (EU) test scores, 154­156 accession, 34, 192­193 see also school enrollment; first members, 44n.10 teachers new members, 42­43n.2 elasticity, 45n.20, 85, 190 poverty reduction objectives, inequality and, 85­87 192­193 partial, poverty reduction exchange, reciprocal, 225 and, 94­96 exchange rates, 16 poverty reduction, 85­86 Expenditure and Income (EI) elderly, social protection, 211 Surveys, 218 electricity, 26­27, 172 household expenditure, 176 families. See children quality of service, 174 farm, household 117 tariffs, 176­177, 178 financial sector, 38 employment, 13, 57, 58, 109, firms, 35 192 old, restructured, and new, discouragement, 140 139, 204­205 296 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union food share, 226 health care services, (continued) fuel, dirty, 27­28, 181­182 access to, 37­38, 161­171, see also heating; utility services 208­209 full cost recovery, 210 catastrophic care, 168­169 cost, 187n.12, 209 gas, tariffs, 178­179 efficiency, 209 gender, 124, 136, 204 expenditures, 28, 29, 187n.11 gap, 145n.6 mechanisms to protect the inequality, 124 poor, 169­171 secondary education, 153 noncommunicable diseases, 4 Georgia, 51 paying for, 165­168 consumption poverty, water providers, 162 access, and health, 71 public funding, 170­171, 184 health care spending, 167 quality of, 37­38, 162, 164, income, 44n.17 208­209 trends, 54 service usage, 234­235 utilities, 173 spending, 162, 166­168 governance, 183 utilization, 70, 164­171, 185, gross domestic product (GDP), 187n.10 16, 80 tilization rates, 166 growth rates, future, 33 health insurance, 170, 171 private consumption vs, 195 health status, 2­3, 67­71, 125, growth, 3, 32, 41, 51, 79­105, 262­265 104n.4 indicators, 234­235 accelerating, 203­207 inequalities, 125 connecting the poor to, perceptions, 70 107­110 heating, 73 distribution, 83­84, 90­91 dirty fuels, 181­182 effect, 84­85 district, 172­173, 182 elasticities, 81­87, 105n.7 household expenditure, 176 elasticity averages, 82, 83 tariffs, 178­179 inequality and, 14 high poverty, 5 patterns, 108, 195­203 HIV/AIDS, 3, 25, 69­70, 148, shared, 20, 90­91, 213­214 186n.2, 199, 201, 202 targets, 202­203 hospital utilization, 26, 164, 198, growth rates, 17, 80, 81 202 global projections, 195 Household Budget Surveys private consumption and , (HBS), 218, 219, 226 193­195 household consumption, 102 shared, 34­37 measurement, 99, 101 household data, variability, health care services, 23, 25­26, 218­219 30, 37, 207­208 households, 44n.12, 148­149 Index 297 households, (continued) inequality, (continued) heterogeneity, 160 international perspective, 103 private consumption, 190 measuring, 44n.13, 232 regional survey archive, pattern, 89 215­220 perception of, 44n.14 utility services expenditures, regional, 36­37, 206, 207, 175­179 213 household surveys, 48 transfers and, 22 data for this report, 42­43 trends, 13­16 health care utilization, 165 infant mortality, 68­69, 148, housing, 73­74, 192, 235, 186n.1, 198, 201 274­277 goals, 44­45n.19 labor mobility and, 206­207 see also maternal mortality human capital informal employment, 133 characteristics, 134­136 transition economies and, investment in, 207 134­135 Hungary, 5 infrastructure, 26­28, 38, 72­74, health care spending, 167 202, 235, 270­273 profile of the poor, 65 heating, 182 rural, 206, 207 illiteracy, 22, 72 schools, 157­158, 160 income, 18­19, 44n.12, 94­95, institutional factors, 136­140 104n.5, 204 institutionalized populations, distribution, 32, 190 formerly, 65 growth, 15­16, 84­85 isolation, geographical, 196 indicator of welfare, 220, 222 inequality, 44n.12 jobless, 109 measurement, 15­16 jobs, 124­125 reporting, 222 creation, 3, 17 income poverty, 183 opportunities, enhancing, 141 inequality, 3, 18­20, 30, 78n.15, 85, 86­87, 93, 102­104, Kyrgyz Republic, 205 104­105n.6, 125, 190, 206 growth rate and inequality, aversion, 74 198 between-group, 89 trends, 53 changes, 87­90, 105n.9 decomposition, 89 labor markets, 107­145, dynamics, 196­198 158­161, 205, 206, elasticity and, 85 233­234 factors, 103­104 enhancing job opportunities, growth and, 195­203 141 growth rate and, 197 mobility, 36, 139­140, indicators, 238­241 206­207 298 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union labor markets, (continued) Millennium Development Goals profiles of the poor, 108 (MDGs), (continued) reallocation, 17­18, 112­114 prospects by region, 200­201 land reform, 35­36, 205, 206 minimum wage, 40, 144 language, education and, 154 adequate, 212 Latvia, education and schools, minorities, 39, 193, 211­212 156, 160 moderate poverty, 5 life expectancy, 67, 68, 77n.8, 148 Moldova, 51 life satisfaction, 75, 78n.15 education, 59, 151, 156 literacy, 22, 72 growth rate, distribution, and Lithuania, 51 inequality, 19, 84, 198 education, 156 inequality, 90 urban-rural gap, 92, 94 trends, 53 Living Standards Measurement urban-rural gap, 92 Study (LSMS), 218­219, wage increases, productivity 226 and, 143 living standards, 215, 226 water service, 175 low poverty, 5 morale, 31, 74 morbidity, 25, 163 Macedonia, education, 156 mortality, 25, 67, 163, 198 macroeconomic environment, see also infant mortality; 236­237 maternal mortality Making Transition Work for Every- one, data collection, 217, noncompetitive pressures, 223 142­144 marginalized groups, 39, nonemployment, 105n.10 211­212 nonincome dimensions market-based incentives, 36­37 challenges, 33­34 markets, input and output, 206 data collection, 213 maternal mortality, 68­69, indicators, 234­235 77n.10, 186n.1, 198, 201, of poverty, 32, 183, 198­202 202 of well-being (1998­2003), goals, 44­45n.19 22­29 see also infant mortality nonworking poor, 66, 108, 123 mathematics performance, 155, 156, 157, 187n.8 opportunities, 59, 197 migrants, 65, 193 expansion, 110­119 migration, 25, 69, 139­140, 153, the poor and, 119­129 197, 206­207 promoting, 41, 206­207, 214 ethnicity, education and, 154 output, 16­18, 80 Millennium Development Goals (MDGs), 44­45n.19, pay rate variations, 137 191­193, 199 pension spending, 119 Index 299 Poland, 20, 51 poverty line, 7­8, 44nn.9, 11, distribution changes, 91 48, 49­50, 54, 77n.1, education, 59 78n.12, 105n.8, 227­229 growth rate, distribution, and absolute, 6­8 inequality, 19, 84, 197 arbitrariness, 228­229 health care, 163 future, 191 inequality, 90, 197 lowest, 45n.21 trends, 54 poverty profile, 55­66 wage increases, productivity characteristics, 232­234 and, 143 demography, 250­253 policy, 183 education, 254­257, 266­269 recommendations, 140­145 health status, 262­265 reforms, targeting and, housing, 274­277 130­131 infrastructure, 270­273 see also public policy labor market, 158­261 politicians, health care and, spatial dimension, 242­249 208­209 poverty rates, 7, 30­31, 44n.8, poor people, composition, 105n.11, 191­192 11­12, 66 projected, 189­190 see also poverty profile poverty reduction, 47, 51, 76, population 200, 213 distribution, 51 challenges, 33, 40­41 poverty vs, 191 elasticity, 85­86 Portugal, poverty line, 45n.21 factors contributing to poverty, 4 (1998­2003), 16­22 absolute vs relative, 227 growth and, 20, 80­81, clusters, 54­55, 56­57, 76 90­91, 104nn.1­3, coping mechanisms, 225 193­195 country-level trends, 51­55 long-term, 190­195 depth, 14, 51­55, 56 medium-term, 189­190 distribution, 3, 77n.6, 86 monitoring progress, 40­41, elasticity, 92­93, 190 212­213 incidence, 51, 55, 77n.7, 190 private consumption and, indexes, 231­232 193­195 indicators, 236­277 prospects, 29­34, 189­214 levels, 2 regional, 2­4 material vs nonmaterial, 50 trends, 31­32 measuring, 1­2, 231­232 variability, 51, 55 nonincome dimensions, poverty risk, 8­9, 11, 12, 31, 41, 67­75 44n.11, 56 trends, 4­16, 43­44n.7, age and, 62 53­54 regional, 61 variation, 11 poverty wage, 110, 145n.2 300 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union power, full cost recovery, 210 Romania PPP. See purchasing power parity consumption poverty, water price indexes, 230 access, and health, 71 prices, 229­230 education, 59, 156, 157, 160 private consumption, GDP vs, growth, 19, 84, 197 195 health care spending, 167 private resources, 147 inequality, 90, 197 private transfers, 119, 126­127, urban poverty, 94 129 wage increases, productivity poverty reduction and, 21­22 and, 143 production, inefficiency, 4 rural areas, 57, 95­96, 205 productivity, 131­132 growth, 41 distribution, 139 infrastructure, 206, 207 employment and, 142­145 investment, 36 wage increases and, 140, 143 opportunities, 36­37 public health service delivery, 206 activities, 162 populations, 9, 10, 11, indicators, 69­70 59­60 providers, 25­26 rural-to-urban poverty rate, public policy, 34­41 92­96, 97, 118 role for, 203­213 Russia see also policy financial crisis, 16, 30, 47, public services, 4, 28­29, 37­38, 104­105n.6 41­42, 202 growth, 80, 197 delivery, 207­210, 214 heating, 173 public transfers, 20­21, 22, 118 inequality, 90, 104, 197 purchasing power parity (PPP), teachers and schools, 160 4­5, 6, 7, 48, 49­50, 77n.4, trends, 53 227­229 urban-rural gap, 92 errors, 228 wages, 145n.4 Russian Federation reform, structural, 16 consumption poverty, 71 regions and subregions, 48 education, 59 characteristics, 233 growth, 19, 84 disparities, 61­62 health care, 71, 166 household survey archive, productivity, 143 215­220 urban poverty, 94 variation, 61, 138 wage increases, 143 remittances, role of, 120 water access, 71 resources, delivery and, 147 Roma, 64 safety nets, 107­145 education, 154 sanitation services, 73­74, 202 social protections, 211 household expenditure, 176 Index 301 school enrollment, 152­153, Southeastern Europe (SEE), 2, 200 3, 43n.3 , 48, 51 primary, 22­23, 24 spatial dimension, 242­249 secondary, 23, 24 spatial price differences, see also education 224­225 schools spending levels, 183 autonomy, 158 standard of living, 5­6 governance and manage- structural reforms, 80 ment, 157­158, 161 subjective poverty, 74­75 infrastructure, 157­158, 160 subregions. See regions and sub- performance outcomes, regions 155­161 subsistence activities, 105n.10 SEE. See Southeastern Europe survey data, 48, 216 self-employment, 112, 132­133 consumption changes and, services and service delivery, 38, 97­102 183 countries not covered, 77n.2 access to quality services, need for, 212 147­187 utilities, 172 employment, 114 rural, 206 Tajikistan, 5, 51 sewerage growth rate and inequality, household expenditure, 176 198 sanitation services, 73­74, utilities, 173 202 water service, 175 sexually transmitted infections targeting (STIs), 25, 69, 70 interventions, 211­212 sex work, 69 policy reforms and, 130­131 shortages, 147­148 tariffs skills, 136 household coping options, Slovak Republic, 155, 156 179­181 Slovenia, education, 155, 156 utilities, 176­179, 210 social capital, opportunities and, teachers, 160 196­197 aging, 159 social exclusion, 193 see also education social protection, 39­40, transfers, 125, 126­127, 129, 207 127­131, 144­145, 192 changes in value and direc- benefits, 20­21, 130 tion, 118­119 enhancing, 210­212 payments, 21, 131 utility services, 180­181 private, 21­22, 119, 126­127, social safety net, 38, 214 129 strengthening, 39, 210­211 transition shock, 74, 76 social services, 214 tuberculosis (TB), 25, 69, 148, socially excluded, 30 201 302 Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union Turkey, 44n.8, 48, 77n.5 Vietnam, (continued) nonincome dimensions of wage increases, productivity poverty, 32 and, 143 wage growth vs employment vulnerability, 32, 41, 183, 213 growth, 111 growth rate to eliminate, 194 wage increases, productivity variation in change, 52, 54 and, 143 see also poverty risk Turkmenistan, 77n.3 vulnerable groups, 51, 63­66 Ukraine, 80, 104n.2 wages, 110, 112, 139, 145n.2, unemployed poor, 108 204 unemployment, 9, 11, 12, 56, arrears, 19 57, 66, 125 changes, 121 incentives, 140 determination, 138­139 inequality among, 89­90 dispersion, 137­138 long-term, 30, 123­124 distribution, 123 urban areas, 11, 94 employment, 90, 105n.10 opportunities, 36­37 gap, poor vs nonpoor, 138, poverty and growth, 92 139 see also cities growth, 16­18, 110­111, urban-rural gap, 60 120­123 utility services, 171­182 employment growth and, access, quality, and affordabil- 110­111 ity, 172 increases, productivity and, consumption, 180 140, 143 cost, 175­179 inequality, 44n.18, 90, 123 coverage rates, 172­173 minimum, 40, 144, 212 expenditures, 28, 29 poor vs nonpoor, 133 financial performance, 214n.2 Russia, 145n.4 nonpayment patterns, 180 see also earnings quality, 174­179, 209 water, 172 reform, 38, 209­210 access to, 27, 28, 71, 73, 201, social protection system, 202 180­181 household expenditure, 176 tariffs, 210 quality of services, 174­175 Uzbekistan, health care utiliza- tariffs, 176­178, 179 tion rates, 166 women, 136 wages, 124 Vietnam, 48 see also gender; maternal mor- health care utilization rates, tality 166 working, 109 wage growth vs employment working poor, 64­65, 108, 129 growth, 111 global trends, 132 This report is part of a series undertaken by the Europe and Central Asia Region of the World Bank. The series draws on original data, the World Bank's operational experience, and the extensive literature on the Region. Poverty, jobs, trade, migration, and infrastructure will be among the topics covered. "....a most interesting report. I have read it with considerable interest, and have learned a lot. It tells a clear story, and it contains a lot of interesting material." Sir Anthony Atkinson, Warden of Nuffield College, Oxford University, United Kingdom "The key conclusion of the report is that rapid economic growth is fundamentally important for job creation and, consequently, reducing poverty." Ewa Balcerowicz, President of the Board, Center for Social and Economic Research, Warsaw, Poland W hile the countries of Eastern Europe and the Former Soviet Union have made significant progress in reducing poverty during 1998­2003, poverty and vulnerability remain significant problems. More than 60 million are poor and more than 150 million are vulnerable. Most of the poor are the working poor. Many others face deprivations in terms of access and quality of public services. Regional inequalities both between and within countries are large. The highest levels of absolute poverty are found in the poor countries of Central Asia and the South Caucasus, but most of the region's poor and vulnerable are in middle- income countries. Growth, Poverty, and Inequality examines these important issues and recom- mends that public policies focus on: accelerating shared growth and job creation; improving public service delivery; strengthening social protection; and enhancing the monitoring of progress in poverty reduction. This book will be especially useful for policy makers and social scientists working in the Region. ISBN 0-8213-6193-7