WORLD BANK TECHNICAL PAPER NO. 351 Work in proages7 WPTP'351 tor public disCus io!i ;p \ \ 'q/ Povertv and Inconme Distribution inl Latin America I/fi eSlori l' thi/c /9Q( {f Gf0rI ei P / '2 1 ,1 i : ' i /))j/// Jf / f t(/ RECENT WORLD BANK TECHNICAL PAPERS No. 269 Scheierling, Overcoming Agricultural Polluttion of Water: The Challenge of Integrating Agricultutral and Environmental Policies in the European Union No 270 Baneriee, Rehabilitation of Degraded Forests in Asia No 271 Ahmed, Technological Development and Pollution Abatement. A Stuidy of How Enterprises Are Finding Alternatives to Chlorofluorocarbons No. 272 Greaney and Kellaghan, Equity Issues in Public Examinations in Developing Countries No 273 Grimshaw and Helfer, editors, Vetiver Grassfor Soil and Water Conservation, Land Rehabilitation, and Embankment Stabilization: A Collection of Papers and Newsletters Compiled by the Vetiver Network No 274 Govindaraj, Murray, and Chellaraj, Health Expenditutres in Latin America No. 275 Heggie, Management and Financing of Roads An Agendafor Reform No. 276 Johnson, Quiality Review Schemes for Auditors Their Potential for Suab-Saharan Africa No. 277 Convery, Applying Environmental Economics in Africa No. 278 Wijetilleke and Karunaratne, Air Quiality Management. Considerationsfor Developing Countries No 279 Anderson and Ahmed, The Casefor Solar Energy Investments No. 280 Rowat, Malik, and Dakolias, lJudicial Reform in Latin America and the Caribbean: Proceedings of a World Bank Conference No 281 Shen and Contreras-Hermosilla, Environmental and Economic Issues in Forestry Selected Case Studies in Asia No 282 Kim and Benton, Cost-Benefit Analysis of the Onchocerciasis Control Program (OCP) No. 283 Jacobsen, Scobie, and Duncan, Statuttory Intervention in Agricudltural Marketing. A New Zealand Perspective No 284 Valdes and Schaeffer in collaboration with Roldos and Chiara, Suirveillance of Agricultural Price and Trade Policies A Hand book for Urtguay No. 285 Brehm and Castro, The Marketfor Water Rights in Chile: Major Issues No 286 Tavoulareas and Charpentier, Clean Coal Technologies for Developing Countries No. 287 Gillham, Bell, Arin, Matthews, Rumeur, and Heam, Cotton Production Prospectsfor the Next Decade No. 288 Biggs, Shaw, and Srivastiva, Technological Capabilities and Learning in African Enterprises No. 289 Dinar, Seidl, Olem, Jorden, Duda, and Johnson, Restoring and Protecting the World's Lakes and Reservoirs No. 290 Weijenberg, Dagg, Kampen Kalunda, Mailu, Ketema, Navarro, and Abdi Noor, Strengthening National Agricultual Research Systems in Eastern and Central Africa A Frameworkfor Action No. 291 Valdes and Schaeffer in collaboration with Errazuriz and Francisco, Surveillance of Agricultural Price and Trade Policies: A Handbookfor Chile No. 292 Gorriz, Subramanian, and Simas, Irrigation Management Transfer in Mexico Process and Progress No. 293 Preker and Feachem, Market Mechanisms and the Health Sector in Central and Eastern Europe No. 294 Valdes and Schaeffer in collaboration with Sturzenegger and Bebczuk, Surveillance of Agricultural Price and Trade Policies: A Handbookfor Argentina No. 295 Pohl, Jedrzejczak, and Anderson, Creating Capital Markets in Central and Eastern Eutrope No 296 Stassen, Small-Scale Biomass Gasifiersfor Heat and Power A Global Review No. 297 Bulatao, Key Indicatorsfor Family Planning Projects No. 298 Odaga and Heneveld, Girls and Schools in Suib-Saharan Africa. From Analysis to Action No 299 Tamale, Jones, and Pswarayl-Riddihough, Technologies Related to Participatory Forestry in Tropical and Sutbtropical Countries No 300 Oram and de Haan, Technologiesfor Rainfed Agriculture in Mediterranean Climates A Review of World Bank Experiences No 301 Mohan, editor, Bibliography of Publications: Technical Department, Africa Region, Jiuly 1987 to April 1995 No. 302 Baldry, Calamari, and Yam6ogo, Environmental Impact Assessment of Settlement and Development in the Upper Leraba Basin No. 303 Heneveld and Craig, Schools Count: World Bank Project Designs and the Qutality of Primary Edutcation in Sub-Saharan Africa (List continues on the inside back cover) WORLD BANK TECHNICAL PAPER NO. 351 Poverty and Income Distribution in Latin America The Story of the 1980s George Psacharopoulos Samuel Morley Ariel Fiszbein Haeduck Lee Bill Wood The World Bank Washington, D.C. Copyright © 1997 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing April 1997 Technical Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. ISBN 0-8213-3831-5 ISSN: 0253-7494 George Psacharopoulos is senior adviser m the World Bank's Human Development Department. Ariel Fiszbein is senior economist in the Human Resources and Poverty Division, Economic Development Institute. Haeduck Lee is an economist in the Chief Economist's Office of the World Bank's Latin America and the Caribbean Regional Office. Samuel Morley and Bill Wood were consultants to the World Bank at the time this study was written. Library of Congress Cataloging-in-Publication Data Poverty and income distribution in Latin America : the story of the 1980s / George Psacharopoulos . . . [et al.]. p. cm.- (World Bank technical paper, 0253-7494 ; no. 351) Includes bibliographical references. ISBN 0-8213-3831-5 1. Poverty-Latin America-Statistics. 2. Income distribution -Latin America-Statistics. 3. Poverty-Caribbean Area-Statistics. 4. Income distribution-Caribbean Area-Statistics. I. Psacharopoulos, George. II. Series. HC130.P6P68 1996 96-36910 339.2'2'098-dc2l CIP Contents Foreword ..............vii Acknowledgments ............. viii Executive Summary ............. ix Chapter 1: Introduction .1 Background .2 Previous Work .3 Data Sources .5 Survey Coverage .7 Definition of Income Variables .9 Comparability of Poverty and Inequality Results Across Count ries .10 Adjustment of Income Data to National Accounts .10 Unit of Analysis .11 Social Indicators .12 Chapter 2: The Distribution of Per Capita Household Income .15 Changes in Income Distribution, 1980 - 1989 .20 Inequality Trends During the 1980s: A Broader Picture .25 Chapter 3: A Decomposition of Workers' Income Inequality .33 Theil Index of Inequality .34 Changes in the Distribution of Workers' Income During the 1980s .35 A Static Decomposition of Income Inequality .40 Bottom 20 Percent Probability Analysis .47 Chapter 4: Absolute Poverty .51 Adjustment of Income Data for Undereporting .53 Poverty Reference .54 Poverty Measures .59 Headcount Index .61 Poverty Gap .66 FGT Measure .68 Distribution of Poverty in the LAC Region: A Broader Picture .70 Latest Data: 1989 .71 Earlier Data: 1980 .73 Regional Trends Over the 1980s .74 Individual Country Experiences .77 Further Evidence on Poverty over the Decade .79 iv Poverty and Income Distribution in Latin America: The Story of the 1980s Chapter 5: Social Indicators ................................. 83 Health ................................. 84 Infant Mortality ................................. 84 Child Global Malnutrition ................................. 94 Access to Maternal Health Care at Birth ................................. 97 Immunizations ................................. 98 Education ................................. 100 Illiteracy ................................. 100 Years of Schooling ................................. 104 Grade Repetition and Dropout Rates ................................. 105 Demographic and Employment Indicators ................................. 107 Indigenous Population ................................. 107 Female-headed Households ................................. 109 Family Size and the Number of Children ................................. 110 Dependency Ratio ................................. 110 Home Ownership ....11............................. I Crowding ....................................... 111 Labor Participation ................................. 111 Unemployment ................................. 111 Public Sector Employment ................................. 112 Labor Income Disparities ................................. 112 Informal Sector Employment ................................. 112 Conclusions ................................. 113 Chapter 6: Conclusion ................................. 115 Future Directions ................................. 117 ANNEX 1: Household Survey Data Description .121 ANNEX 2: Monthly Household-Level Income Variables in Household Surveys .125 ANNEX 3: Per Capita Household Income Distribution .137 ANNEX 4: Distribution of Workers' Income .187 ANNEX 5: Mean Sample Characteristics of Workers .193 ANNEX 6: Characteristics of Workers' Income by Decile .197 ANNEX 7: Explaining The Probability of a Worker Belonging to the Bottom 20% of the Income Distribution .205 ANNEX 8: Simulated Probability of A Worker Belonging to the Bottom 20% of the Income Distribution .209 ANNEX 9: Correcting for Income Undefeporting .213 ANNEX 10: CEPAL Poverty Lines ..217 ANNEX 11: Country-Specific Poverty Line in US$ ..219 ANNEX 12: Uniform Poverty Lines ......................... 221 Contents v ANNEX 13: Population in Poverty, 1980 and 1989 ........................................................... 223 ANNEX 14: Social Indicators by Income Quintile .......................................................... 231 ANNEX 15: Social Indicators, by Country .......................................................... 245 ANNEX 16: Bibliography on Poverty and Income Distribution in Latin America ............... ................ 273 Tables Table 2.1: Gini Coefficient and Bottom 20 Percent Income Share .......................................................... 18 Table 2.2: Gini Coefficients in Selected Latin American Countries, 1979-1990 ..................................... 26 Table 2.3: Income Inequality and the Economic Cycle ........................................................... 27 Table 3.1: Measures of Inequality in Workers' Income .......................................................... 36 Table 3.2: Changes in Selected Characteristics of the Working Population ............................................. 42 Table 3.3: Contribution of Individual Variables in Explaining Inequality .................................. ............. 43 Table 3.4: Joint Contribution of Variables in Explaining Inequality ........................................................ 44 Table 3.5: Probability of Belonging to the Bottom 20 Percent of Income Distribution ........................... 49 Table 4.1: Percent of Individuals in Poverty and Extreme Poverty .......................................................... 62 Table 4.2: Statistical Significance of Changes in Poverty .......................................................... 65 Table 4.3: Aggregate Poverty Gap .......................................................... 67 Table 4.4: FGT P2 Index .................................................... 69 Table 4.5: Changes in Rural and Urban Poverty, 1980-1989 .................................................... 71 Table 4.6: Actual and Predicted Poverty Levels in 1989 .................................................... 76 Table 4.7: Poverty Headcount Indices During the 1980s .................................................... 80 Table 5.1: Intra-Country Variability of Health Indicators by Urban vs. Rural ............................... .......... 87 Table 5.2: Intra-Country Variability of Health Indicators by Mother's Educational Level .................................................... 89 Table 5.3: Intra-Country Variability of Infant Mortality .................................................... 91 Table 5.4: Illiteracy by Income Quintile (% of 15+ Age Group) .................................................... 103 Figures Figure 2.1: Changes in Per Capita Income Inequality .................................................... 21 Figure 2.2: Trends in Per Capita Income Inequality .................................. 23 Figure 3.1: Changes in Workers' Income Inequality During the 1980s ............................................... 37 Figure 3.2: Trends in Workers' Per Capita Income Inequality ......................................... 39 Figure 3.3: Marginal Contribution of Selected Variables in Explaining Inequality ................ ................. 46 Figure 3.4: Probability of Belonging to the Bottom 20 Percent of Income Distribution .......................... 50 Figure 4.1: Poverty Headcount Trends ........................................................... 63 Figure 5.1: Infant Mortality ........................................................... 85 Figure 5.2: Infant Mortality by Mother's Education ........................................................... 93 Figure 5.3: Illiteracy Prevalence (Percentage of Illiterate Population) ................................................,.101 Figure 5.4: Educational Attainment by Income Quintile ......................................,,,.,,,,.104 Figure 5.5: Grade Repetition in Bolivia ............................. 106 vi Poverty and Income Distribution in Latin America: The Story of the 1980s Figure 5.6: Indigenous Population of Latin America ................................................... 108 Figure 5.7: Population of Indigenous Background by Income Quintile ................................................. 109 Box 1.1: Availability of Income Distribution Data ................................................... 4 Box 1.2: Household Survey Data ................................................... 6 Box 1.3: Assessing Income vs. Consumption Data ................................................... 8 Box 2.1: The Lorenz Curve and the Gini Coefficient .................................................... 17 Box 2.2: Inequality in Latin America vs. the Rest of the World .................................................... 19 Box 2.3: Changes in Per Capita Income During the 1980s ................... ................................ 24 Box 2.4: Real Minimum Wage and Income Inequality ................................................... 29 Box 4.1: Methodology Behind the CEPAL Country-Specific Poverty Lines ................................. ......... 56 Box 5.1: Malnutrition in Guatemala .................................................... 95 Box 5.2: Maternal Health, Infant Health, and Mortality ................................................... 99 Foreword This report presents an update of poverty and income distribution statistics in Latin America and the Caribbean, and examines the trends in these statistics during the 1980s. In addition, a series of non-monetary social indicators are documented in order to present a more complete profile of living conditions in the region. Latin America has historically exhibited a high degree of income inequality relative to other regions of the world, and the results of this study indicate that this continues to be true. However, changes in income inequality are mixed for the time period examined. During the past decade, the Gini coefficient worsened in Argentina (Buenos Aires), Bolivia (urban), Brail, Guatemala, Honduras, Mexico, Panama, Peru (Lima) and Venezuela; the Gini coefficient improved in Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban). A decomposition of workers' income inequality shows that variation in educational attainment has the highest contribution to income inequality out of four variables examined (age, educational attainment, employment category and sector of employment). On average, differences in individual educational levels account for approximately 25 percent of total income inequality in the labor market in Latin America. A further decomposition of the workers' income distribution indicates that low educational attainment is the factor most associated with the probability of belonging to the bottom 20 percent of the labor income distribution. Subject to the qualifications discussed in the report, the poverty headcount index for Latin America and the Caribbean rose from 27 percent to 31 percent for the region as a whole between 1980 and 1989. Poverty increased in Argentina (Buenos Aires), Bolivia (urban), Brazil, Guatemala, Honduras (urban), Mexico, Panamna, Peru (Lima) and Venezuela, and decreased in Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban). Estimates indicate that both income inequality and poverty measures rose during recession and fell during recovery. However this finding should not be interpreted as demonstrating that the poor would be better off had adjustment not been undertaken. Economic conditions after 1980 necessitated substantial reforms in many countries, and to compare poverty levels at the beginning and the end of the decade without taking these conditions into consideration would be inappropriate. In 1989, over 45 percent of the poor lived in Brazil, although that country had only one-third of the region's population. Mexico and Peru had 10 and 9 percent of the poor, respectively, while an additional 19 percent lived in a group of relatively small countries which includes Bolivia, El Salvador, Guatemala, Haiti, Honduras and Nicaragua. Although social indicators generally improved during the 1980s, intra-country breakdowns show high levels of variability across several criteria, including mothers'/women's educational level, urban/rural setting, ethnicity and income group. These criteria should be incorporated into targeted operations and specific policy recommendations in order to raise the efficiency and equity of poverty reduction efforts in the region. Sri-Ram Aiyer Director, Technical Department Latin America and the Caribbean Region Acknowledgments This report is the result of a Regional Study on Poverty and Income Distribution in Latin America and the Caribbean under the general direction of George Psacharopoulos, with the collaboration of Samuel Morley, Ariel Fiszbein, Haeduck Lee, Alex Panagides, Bill Wood, Francisca Castro, Yng Ng and Hongyu Yang. Diane Steele contributed to the editing of the final draft, and Johanna Coenen diligently manipulated the final draft to its present form. The study has benefited from comments on an earlier draft by the following individuals: Sri-ram Aiyer, Oscar Altimir, Alejandra Cox Edwards, Norman Hicks, Nora Lustig, Philip Musgrove, Martin Ravallion, Pedro Sainz, Marcelo Selowsky, Ruben Suarez and Jacques van der Gaag. Executive Summary This report presents the findings of a regional study on poverty and income distribution in Latin America and the Caribbean. The study was undertaken because of the significance of these issues and the paucity of statistical information on recent trends in the region. Moreover, the impact that the economic recession of the 1980s had on the poor needed to be assessed, particularly in light of reduced spending per capita on public social programs. The analysis points to three principal conclusions: * On average, poverty increased and income distribution worsened in Latin America and the Caribbean during the 1980s. Forty-six percent of the increase in poverty took place in the cities of Brazil alone, though part of this reflects the migration of poor rural inhabitants to urban areas. * There is strong evidence that both income inequality and poverty mirrored the economic cycle, rising during recession and falling during recovery. Economies that grew (e.g. Colombia, Costa Rica) performed better with respect to poverty and income distribution than those that stagnated. In particular, countries that failed to stabilize effectively (e.g. Brazil, Peru) experienced substantial increases in poverty. * Educational attainment has the greatest correlation with both income inequality and the probability of being poor. From a policy standpoint, there is a clear association between the provision of education, lessening of income inequality, and poverty reduction. This study is descriptive in nature, in an effort to expand the existing state of statistical knowledge on these important issues. As such, its findings should not be interpreted as demonstrating that the poor would have been better off had adjustment programs not been implemented. Economic conditions after 1980 necessitated substantial reforms in many countries, and to compare poverty levels at the beginning and the end of the decade without taking these conditions into consideration would be inappropriate. Furthermore, prospects for long-run growth, and hence for poverty reduction, are significantly brighter for countries that implemented reforms designed to promote a stable and sustainable macroeconomic framework. x Poverty and Income Distribution in Latin America: The Story of the 1980s Trends in Income Distribution and Poverty Latin America has histonically exhibited a high degree of income inequality relative to other regions of the world. The results of this study indicate that this continues to be true. Income inequality in the LAC region as measured by the Gini coefficient remains well above comparable estimates for other regions. This disparity is further reflected by the fact that, for the countries examined, the bottom twenty percent received only four percent of total income. At the end of the 1980s, total income inequality was worst in Brazil, Guatemala and Honduras. Furthermore, less than three percent of total national income accrued to the poorest twenty percent of the population in these countries. At the other end of the spectrum, Paraguay (Asunci6n) and Uruguay (urban) demonstrated the least degree of inequality. Subject to the qualifications discussed in the report, the poverty headcount index for the LAC region increased from twenty-seven percent to thirty-one percent between 1980 and 1989. For each of the three principal poverty measures presented, Costa Rica and Uruguay (urban) consistently performed best, while Guatemala and Honduras consistently performed worst. This analysis was based on a poverty reference cut-off of $60 per person per month in 1985 purchasing power parity (PPP) U.S. dollars. By the end of the 1980s, the prevalence of poverty was higher in the cities than in the countryside - a reversal from ten years earlier. At the close of the decade, sixty-six million urban inhabitants and sixty-five million rural dweliers were living in absolute poverty in the region. However the probability of being poor remained more than double in rural areas than in urban ones. Most of the poverty in the region is located in a distinct set of countries. In 1989, over forty-five percent of the poor lived in Brazil alone, although that country was home to only one-third of the region's population. Mexico and Peru had ten and nine percent of the poor, respectively, while an additional nineteen percent lived in a group of relatively small countries consisting of Bolivia, El Salvador, Guatemala, Haiti, Honduras and Nicaragua. Income Distribution, Poverty and Growth Trends in income inequality appear to have been significantly influenced by changes in the average level of per capita income. Of the seven countries for which the data span the entire decade, the three that showed reductions in income inequality (Colombia - urban, Costa Rica, Uruguay - urban) had an average growth of three percent in per capita income for the entire decade, while the four that experienced an increase in income inequality (Argentina - Buenos Aires, Brazil, Panama, Venezuela) had an average decline of twelve percent in per capita income for the same period. Likewise, povert decreased in those countries where per capita income increased, and vice versa. Executive Summary Xi Not surprisingly, the income share of the bottom twenty percent of the distribution rose in all countries where inequality declined, and dropped in all countries where inequality increased during the decade. Conversely, changes in the share of the top twenty percent of the distribution demonstrated the exact opposite trends - rising when the Gini coefficient worsened and falling when it improved. Since most countries that experienced a rise in inequality also showed a drop in real per capita income during the decade, these findings indicate that the wealthy were better able to protect themselves from the impact of the recession than the poor. A decomposition of the regional increase in poverty during the 1980s shows that this rise occurred in a concentrated group of countries. A staggering forty-six percent of the total increase in poverty for all of Latin America took place in the cities of Brazil, though part of this undoubtedly reflects the migration of poor rural inhabitants to urban centers. Mexico and Peru also accounted for substantial shares in the total increase in poverty. The results of this study, when combined with outside sources, indicate that income inequality and poverty rise during recession and fall during recovety. An examination of twenty-six economic cycles shows that, in all but three cases, inequality increased during recession and decreased during recovery. A more extensive analysis shows that poverty increased in fifty-five of fifty-eight recession periods, and fell or stayed constant in twenty-five of thirty-two recovery cases. This finding, however, cannot be used as a valid argument to delay necessary adjustments in the macroeconomic structure of a country. Over the course of the decade, countries that failed to effectively implement needed economic reforms experienced the sharpest rise in poverty (e.g. Brazil, Peru). Education and the Income Distribution of Workers An analysis of income inequality for individuals in the labor market shows that, on average, differences in individual educational levels account for approximately twenty-five percent of total inequality. The role of education as an explanatory variable for workers' income inequality is twice as high as that of any other variable examined. Moreover, in nineteen of the twenty country cases, education demonstrates the highest marginal contribution to total income inequality among the labor force. Additional analysis shows that low educational attainment is the factor most associated with the probability of belonging to the bottom twenty percent of the income distribution; this finding was equally applicable at the beginning and the end of the decade for each country. The preeminence of education differentials as a source of income inequality for individuals in the labor market has important policy implications. Given that labor is the principal asset of the poor, improvement in the provision and quality of education represents a key mechanism for reducing overall inequality and lowering the number of individuals living in absolute poverty. xii Poverty and Income Distribution in Latin America: The Story of the 1980s Social Indicators Despite the rise in poverty and a reduction in spending per capita on public social programs, social indicators generally improved during the 1980s. However it is not clear whether the poor have benefitted from these improvements. Breakdowns of these indicators show high levels of variability within countries as well as across countries. Low educational attainment of mothers/women, rural residence, indigenous background and low income level are correlated strongly with poor performance on such indicators as infant mortality and malnutrition. Data limitations severely hamper the assessment of links between income, health, education, region of residence and ethnicity in a comprehensive manner. However, the available evidence suggests that enhancement of the health and education status of the most disadvantaged groups in society will, in turn, contribute to a lessening of income inequality and a reduction in poverty. 1 Introduction This report presents the findings of a regional study of poverty and income distribution in Latin America and the Caribbean. The study has been undertaken because of (i) the significance of the issue, and (ii) the paucity of statistical information on recent trends in poverty and income distribution in the LAC region. Moreover, there is a great need to assess the impact which the economic recession of the 1980s has had on the poor, particularly in light of reduced spending per capita on public social programs in many countries. The analysis put forth in this study is based on micro data obtained from household surveys in a total of eighteen countries in the Latin America and Caribbean region. For thirteen of these countries, the availability of comparable data for two points in time has made it possible to examine how income distribution and the extent of poverty have changed during the decade. For each data set, the principal basic statistics pertaining to income distribution and poverty are presented. These include the decile income distribution, the Gini coefficient, the Lorenz curve, the headcount index, the poverty gap and the FGT measure. These statistics are based on per capita household income from all sources, as determined by the household surveys. As a complement to the per capita household income distributions, an analysis and decomposition of income inequality based on workers' income is presented. The former examines the distribution of income across the entire population, which is a reflection of individual welfare levels. The latter analyzes the prevailing income structure in the labor market. The decomposition of workers' income inequality highlights those characteristics which are associated with success in the labor market, as proxied by individual income. Finally, a set of social and demographic indicators are presented for each country. When possible, these indicators are disaggregated by various criteria, including income quintile and urban/rural place of residence; health indicators are also disaggregated by level of mother's education. These disaggregations highlight the within country variability of social and demographic conditions, and demonstrate associations of these indicators with region of residence, income group, and education level. I A detailed description of the surveys is given in Annex 1. A brief description is given in Box 1.2. 2 Poverty and Income Distribution in Latin America: The Story of the 1980s Backgod Relative income inequality is an issue closely related to poverty and, according to some, responsible for much of the observed social malaise in the region.2 In the case of Latin America, interest in poverty and income distribution has grown during the last part of the eighties with respect to the question, "Has the economic decline in most countries hurt the poor?" Overall, the 1980s was an exceedingly difficult decade for Latin America. The combination of severe balance of payments pressures, terms of trade shocks and real interest rate fluctuations required a reorientation of economic policies for virtually all countries in the region. For most countries, the crisis caused by the external shocks was intensified by economically unsound domestic policies, particularly with respect to fiscal deficits and inflation. As a result of this interaction between imprudent domestic policies and external shocks, average per capita income in the Latin America and Caribbean region fell for six of the ten years in the decade, with an unprecedented drop of 11 percent for the period as a whole. It should be noted, however, that some portion of the drop in per capita income can also be attributed to the excess borrowing that occurred in the 1970s. The effects which the extensive recession and tremendous economic fluctuations during the 1980s have had on income inequality and poverty levels in the Latin America and Caribbean region have been the subject of much debate but limited empirical research. Latin America has historically had very high levels of income inequality relative to other areas in the world.3 However most literature on income distribution in the region is outdated; little is known about the evolution of income inequality during the 1980s. There is also a dearth of information regarding trends in absolute poverty during the entire decade. Few regional studies are available which employ a consistent methodology for all 2 For a further analysis of the interaction between high income inequality and adverse social conditions in Latin America, see Jeffrey D. Sachs, Social Conflict and Populist Policies in Latin America. NBER Working Paper no. 2897. Cambridge, MA: National Bureau of Economic Research, March 1989. For an argument that extreme inequality has led to stagnant economic performance in Peru, see Adolfo Figueroa, Social Policy and Economic Adjustment in Peru. Paper presented at Brookings Institution and Inter-American Dialogue Conference, "Poverty and Inequality in Latin America," 1992. 3 International comparisons in this respect can be found in Gary S. Fields, Income Distribution and Economic Growth. In Ranis, G. and T. P. Schultz, eds., The State of Development Economics: Progress and Perspectives. New York: Basil Blackwell, 1988. Also J. Paukert LeCaillon et al., Income Distribution and Economic Development: An Analytical Survey. Geneva: International Labour Office, 1984. Wouter van Ginneken and Jong-Goo Park, eds., Generating Internationally Comparable Income Distribution Estimates. Geneva: International Labour Office, 1984. Introduction 3 country cases, and none of these assess levels and changes in poverty during the final years of the decade. Finally, there are very few analyses which disaggregate social and demographic indicators by specific characteristics; the tendency is for studies to incorporate social and demographic indicators which are aggregated for the entire country. Yet aggregate social indicators show little about which groups and/or regions suffer the worst conditions in a given country; nor do they highlight those characteristics which are associated with inadequate social indicator performance. Furthermore, there are few available breakdowns of social and demographic indicators by income level. Fortunately, the combination of income data with social and demographic indicator data from the household surveys permits a useful exploration of the association between income level and certain population characteristics. With these issues in mind, this report strives to present a comprehensive picture of changes in income inequality and poverty in Latin America and the Caribbean during the 1980s. In addition, an extensive profile is presented of individual and household characteristics; when possible, individual components of this profile are disaggregated by income quintile and other relevant population characteristics. Previous Work There is a vast literature on the subject of income distribution in the Latin America and Caribbean region.4 But in many cases the most current poverty and income distribution statistics refer to the 1970s and early 1980s. (See Box 1.1) In addition, serious problems of quality emerge as a strong hindrance to meaningful analysis, while varying and vague definitions of crucial details make effective inter-temporal and inter-country comparison all but impossible in many instances. Some of the problems regarding the quality of income distribution statistics are inherent in the current statistical practices in Latin America and the Caribbean. Income concepts may vary considerably between surveys; sample coverage may be weak or nonexistent in certain areas, particularly rural regions. These problems are often compounded due to insufficient articulation of methodology by researchers when presenting their results. By omitting crucial details, researchers do not permit their readers to place the analysis within a relevant context. In particular, the definition of income, the methodology of poverty line construction, and the actual poverty line used are often not reported, despite the fact that slight incongruities in method and/or data definitions can cause significant differences in results. 4 Annex 16 is an updated bibliography of literature pertaining to income distribution and poverty in the Latin America and Caribbean region. 4 Poverty and Income Distribution in Latin America: The Story of the 1980s Box 1.1: Availability of Income Distribution Data The lack of quality data has been a major hindrance in the assessment of poverty conditions in Latin America. In some countries, there are Da income data available at all, while in others the latest data are from the 1960s or 1970s. Given these shortcomings, researchers continuously run into difficulties in assessing the living conditions of the poor and developing effective policy recommendations for improving their welfare. During the past few years, substantial progress has been made in upgrading the data collection efforts in countries throughout the region. In fact, this report could not have been produced in such a comprehensive manner scarcely a decade ago. Yet given the level of development in the region, the statistical capabilities of most countries are surprisingly weak. Many statistical institutes are underfunded, and are plagued by a scarcity of employees with strong technical backgrounds. Geographical coverage, though expanding, leaves much to be desired in some countries. Survey design is often haphazard, with lack of consistency and limited breadth serving as continual barriers to fruitful understanding of the relationship between factors associated with poverty, and their dynamic functioning over time. And in the absence of stronger empirical assessments, poverty alleviation efforts fall short of their potential. The World Bank has been supporting country efforts to improve survey data collection through funding and technical advice; in particular, substantial resources have been channeled into developing and implementing the Living Standards Measurement Study (LSMS) surveys. These surveys are designed to collect a comprehensive profile of income, consumption, employment, education, health, fertility, nutrition, housing and migration for individual households.' The breadth of this data can then be examined, and relationships between these different facets of welfare can be assessed. In this manner, the LSMS surveys allow researchers to explore beyond a simple profile of country conditions, and to analyze the determinants of specific conditions such as poverty, malnutrition, high fertility and low educational attainment. Particularly with recent advances in personal computer technology, it is now possible for governments to collect and analyze data rapidly, and to incorporate these findings into their decision-making process with a minimum of delay. In some instances, the result has been a substantial improvement in both the efficiency and effectiveness of government social programs. To date, LSMS surveys have been conducted in Bolivia, Jamaica, Peru and Venezuela, and are in preparation in Colombia, Ecuador, Guyana, Honduras and Nicaragua. aSee Paul Glewwe, Improving Data on Povertv in the Third World: The World Bank's Living Standards Measurement Study. PRE Working Paper No. 416. World Bank, Washington, D. C, 1990; Paul Glewwe and Jacques van der Gaag, Confronting Povertv in Developing Countries: Definitions. Information and Policies. LSMS Working Paper No. 45. World Bank, Washington, D. C., 1988; and Margaret Grosh, The Household Survev as a Tool for Policy Change. LSMS Working Paper No. 80. Washington, D. C., 1991. Introducion 5 Some of the difficulties and data shortcomings which have plagued past researchers have also been a hindrance in this study. However, a uniform methodology has been applied in each of the country analyses, and this methodology is delineated below. In this respect, the approach used in this study minimizes the problems of non-comparability across time and countries. Data Sources This report is based on an analysis of thirty-one household surveys from a total of eighteen countries. Thirteen of these countries are examined at two points in time, and an assessment of changes between the two survey dates is made. The bulk of the data has been obtained through the Statistical Division of the Economic Commission for Latin America and the Caribbean (CEPAL) in Santiago, Chile. The remaining country data sets are from the data library which is maintained in the Latin America Technical Department at the World Bank. Many countries in the Latin America and Caribbean region routinely send the raw data from their national and/or urban surveys to the Statistical Division of CEPAL. These are not surveys conducted by CEPAL. Rather, they are conducted by national statistical agencies, who then forward their raw data to CEPAL to be included in regional analyses. I CEPAL has kindly supplied the Bank with fully documented original data sets for twelve countries. In the case of ten of these countries, there is comparable income data corresponding to two time periods: the latest survey (1989) and one other survey from an earlier date (circa 1980). For the two remaining countries, the income data from CEPAL is for one time period only.5 These data sets from CEPAL have been supplemented with household surveys for six additional countries from non-CEPAL sources. The analyses of Jamaica and Peru are based on World Bank Living Standards Measurement Study (LSMS) surveys, while the remaining four country data sets come from national statistical offices. For two of these six countries, the data sets have comparable income variables for two points in time. For the remaining four countries, s The ten countries for which CEPAL has supplied data at two points in time are Argentina, Bolivia, Brazil, Colombia, Costa Rica, Guatemala, Honduras, Panama, Uruguay and Venezuela. Data have been supplied for only one point in time for Chile and Mexico. However, data for an additional point in time for Mexico was obtained from a non-CEPAL source and is maintained by the Latin America Technical Department at the World Bank. 6 Poveny and Income Distribution in Latin America: The Story of the 1980s BOx 1.2: Household Survey Data Data sets on which this study is based: Country Year Geographic Coverage No. of Households Argentina 1980; 1989 Metropolitan Area 3,400; 16,759 Bolivia 1986; 1989 4 Cities; 17 Cities 12,226; 37,864 Brazil 1979; 1989 National 88,975; 70,777 Chile 1989 National 32,456 Colombia 1980; 1989 7 Cities; 8 Cities 7,473; 17,949 Costa Rica 1981; 1989 National 6,604; 7,637 Dominican Rep. 1989 National 799 Ecuador 1987 Urban 5,558 El Salvador 1990 Urban 23,773 Guatemala 1987; 1989 National 9,660; 10,934 Honduras 1986; 1989 Urban; National 8,650; 8,648 Jamaica 1989 National 2,725 Mexico 1984; 1989 National 4,963; 11,535 Panama 1979; 1989 National 8,593; 8,817 Paraguay 1983; 1990 Metropolitan Area 5,138; 4,791 Peru 1986; 1990 National; Lima 4,981; 1,385 Uruguay 1981; 1989 Urban 9,506; 21,473 Venezuela 1981; 1989 National 45,421; 61,385 Note: For a more detailed description of the data sets used in this study, see Annex 1. Introduction 7 the data are for only one point in time.6 A detailed description of the country, survey period, survey name, executing agency, survey coverage, sample size and income concept is provided in Annex 1. A brief overview of each survey is also included in Box 1.2. The principal issues concerning the country data in the context of this report are survey coverage, the definition of income, adjustment of income data to national accounts and the unit of analysis. Survey Coverage An important issue to note is the difference in geographical coverage between various surveys. The majority of the surveys are based on a national sample. However there are some important exceptions. The surveys for Argentina and Paraguay cover only the metropolitan areas of Gran Buenos Aires and Gran Asunci6n, respectively, while the surveys for Bolivia, Colombia, Ecuador, El Salvador, Honduras (1986) and Uruguay cover only the urban centers. Of the thirteen countries which have data for two time periods, there are three instances where survey coverage is slightly different between the two dates. In the cases of Bolivia and Colombia, the number of urban centers included in the survey was expanded. The 1986 Bolivia survey covers four major urban centers, while the 1989 survey is based on seventeen urban centers with populations of 10,000 or more. The 1989 Colombia survey includes one additional urban center beyond those in the 1980 survey. The coverage differential is more substantial for Honduras. The 1986 Honduras household survey covers sixteen principal cities, while the Honduras 1989 household survey has national coverage. Therefore, the results obtained for Honduras are not strictly comparable due to differential survey coverage. However, for the income distribution analysis, results for the urban sector are given in the footnotes to Table 2.1, while in the poverty analysis, the urban and rural regions of Honduras (1989) are assessed separately. Only urban poverty is compared between the two dates. Peru is a special case. The Peru 1990 LSMS survey is representative of Lima only, while the 1985-86 LSMS survey is based on a national sample. In order to attain comparable results, only data corresponding to Lima have been used from the 1985-86 LSMS survey. Therefore, both sets of results for Peru are based on Lima only. 6 This report uses data for two points in time for Paraguay and Peru, and data for one point in time for the Dominican Republic, Ecuador, El Salvador and Jamaica. 8 Poverty and Income Distribution in Latin America: The Story of the 1980s Box 1.3: Assessing Income vs. Consumption Data Much has been written about the relative merits of consumption expenditure versus income as a basis for measuring welfare. It is generally acknowledged that consumption expenditure is the preferred choice for determining welfare and formulating policy recommendations. This is true for several reasons. First, welfare is commonly defined as the utility derived from the consumption of goods and services. Though expenditure is only a proxy for actual consumption, it is closer to reflecting the living standards by which individuals currently live than income. Second, and tied into the first argument, is that households tend to smooth consumption through savings and dissavings of income as a way to maintain a steady standard of living. Particularly in societies where income fluctuates during the year, consumption expenditures usually will vary far less than income. Third, consumption expenditure data tends to be more accurate than income data. Most income-expenditure surveys show consumption to be greater than income for the poorer groups of society in a manner which can not be explained simply through dissaving. In contrast, the income of wealthier groups tends to be substantially greater than consumption expenditure, since this strata of society is more likely to save part of its income; this is true even after accounting for the substantial underreporting which is typical of high income groups. Therefore both income inequality and poverty statistics tend to be less severe if consumption data is used instead of income data. Given the general theoretical and practical acceptance of consumption expenditure as superior to income in assessing welfare, why is the welfare concept in this report based predominantly on income? The answer lies in the availability of data. As discussed in Box 1.1, data collection efforts in Latin America are notoriously inadequate for effective policy development, despite improvements in the past few years. Surveys which have been conducted tend to ask comprehensive questions about income, while the consumption section is either non- existent or insufficient for developing a complete household consumption profile. Income- expenditure surveys are conducted only every five to ten years or more, with the last one having taken place during the 1970s in many countries. This reliance on income surveys limits the approach which researchers can take when assessing welfare. In contrast, most Asian countries conduct consumption surveys, but ask very little about income. Although many LAC countries are undertaking LSMS surveys which include both income and consumption components, there is still a great need to step up efforts at collecting information on consumption patterns in order to develop more complete profiles of the poor, and to improve the formulation of social programs designed to help them. a For discussion of the relative merits of income and consumption data in welfare analysis, see Oscar Altimir and J. Sourrouille, Measuring Levels of Living in Latin America. LSMS Working Paper No. 3, World Bank, 1980; Angus Deaton, The Measurement of Welfare. LSMS Working Paper No. 7, World Bank, 1980; Paul Glewwe and J. van der Gaag, Confronting Povertv in Develoing Countries. LSMS Working Paper No. 48, World Bank, 1988; and Martin Ravallion, Poverty Com arisons LSMS Working Paper No. 88, World Bank, 1992. Introduction 9 Definition of Income Variables There is an inherent difficulty in making cross-country regional analyses of income distribution and poverty due to differences in the definition of income embodied in the underlying data.7 Annex 2 presents the descriptions of income as defined in the household surveys and data sets. It is evident that there is a degree of variation in the definition of income across surveys. Some surveys ask only about pnmary wage income, while others include questions about secondary wage income, various types of non-wage income, in-lknd income and even imputed rent. Although it would desirable to have greater uniformity, researchers must contend with the unavoidable discrepancies which prevail from country to country. There is no avoiding that these incongruities limit the comparability of data, and they must be taken into consideration when making inter- temporal and inter-country poverty and income inequality comparisons. To the degree possible, a uniform definition of income over time and across countries has been used. The principal definition in the analysis here is total household income. Generally this refers to monetary income only. In most cases, the income definition captures a fairly identical concept in each country since the variations tend to pertain to a relatively small share of total income. Four countries include the monetary value of in-kind and/or self-production of goods in their survey definitions of total household income: Chile, Colombia, Guatemala and Mexico. These exceptions are particularly important, because poor households tend to have a relatively high share of "income" from these sources. Therefore the per capita income of these groups, in terms of a constant US dollar amount, may be higher for those surveys which include in- kind and self-production items in the definition of total household income. This in turn can result in lower poverty levels and less income inequality than if in-kind income had not been included in the analysis. Consumption expenditure is used instead of income for two countries: Jamaica and Peru. The survey data sets in these two countries are plagued with an excessive number of outliers in the category of reported income. Distributions constructed using the income data from these 7 Problems with respect to comparability of income distribution and poverty statistics have been a continuing source of debate and frustration for researchers. See Oscar Altimir, Income Distribution Statistics in Latin America and their Reliability. Review of Income and Wealth 33, no. 2(1987): 111- 155. Julie DaVanzo, Income Inequality and the Definition of Income: The Case of Malaysia. Santa Monica, CA: RAND Corporation, 1980. Dominique van der Walle, Poverty and Inequality in Latin America and the Caribbean during the 70s and 80s. Views from LATHR No. 22. Latin America and Caribbean Technical Department, World Bank, Washington, D.C., 1991. 10 Poverty and Income Distribution in Latin America: The Story of the 1980s surveys showed an unrealistic concentration of income in the top income decile.8 Because the income distributions based on income variables proved to be unreliable, household consumption expenditures was used as a proxy for income for these two country cases. Comparability of Poverty and Inequality Results Across Countries It cannot be claimed that strict comparability of results has been achieved across countries. In particular, discrepancies in geographical coverage and income concept reduce the robustness of cross-country comparisons. However the results of this study do represent a significant step forward towards a standardized compilation of regional trends regarding these important issues. In the future it is hoped that countries will standardize certain portions of their survey design in order to attain increasing comparability from country to country. The expanded implementation of LSMS surveys mentioned in Box 1.1 and Box 1.3 is an important shift in this direction. Because problems of comparability do exist between countries, particularly due to differences in geographical survey coverage, all results which are based on less than national coverage are labelled as such throughout the text and charts in this report. Furthermore, the reader is encourage to become familiar with the individual income definitions presented in Annex 2 in order to reach a better understanding of the nuances behind the poverty and income distribution statistics for each country. Adjustment of Income Data to National Accounts There is, almost universally, an inescapable degree of under-reporting of income in any household survey. Substantial debate exists over the issue of adjusting survey data for this under-reporting of income by respondents. Two general lines of thought are (i) to leave the data as is and explain the problem to the reader, or (ii) to adjust the data using intricate modelling techniques for matching survey income responses to national account figures according to income type and socioeconomic characteristics of the household.9 a The following income shares were calculated for the top decile using household income per capita: 70.0 percent for the 1985 Peru data, 99.4 percent for the 1990 Peru data, and 64.6 percent for the 1989 Jamaican data. For further description and analysis of the Peruvian LSMS data sets, see Paul Glewwe, The Distribution of Welfare in Peru 1985-86. LSMS Working Paper No. 42. World Bank, Washington, D.C., 1988; And Paul Glewwe and Gillette Hall, Poverty and Inequality during Unorthodox Adjustment: the Case of Peru 1985-1990. LSMS Working Paper No. 86. World Bank, Washington, D.C., 1992. 9 Altimir has developed a detailed methodology of making national account adjustments through the use of weighted coefficients corresponding to sub-groups determined by income type and, in the case on monetary property rents, by socioeconomic characteristics of the household. Altimir, 1987, p. cit. Introduaion 11 This study uses a combination of these techniques. For the distribution and inequality analyses in Chapter 2 and Chapter 3, the data have no been adjusted for the under- reporting of income. The bias which under-reporting causes in income distnbution statistics is difficult to identify and correct accurately. Moreover, the data adjustment process itself may introduce a new bias. Because under-reporting is such a controversial issue, the inequality analyses presented in this study are based on unadjusted data. This allows for a more transparent picture of the methodology followed and the results obtained. However, it was necessary to adjust the income data for calculating the poverty statistics. The under-reporting of income has a limited effect on income distribution statistics, but a strong effect on poverty statistics. This is because an income distribution represents a self- contained sample which is broken down into ten income deciles based on equal population shares. In contrast, poverty levels are determined by the dissection of the income distribution by an exogenous standard which is equivalent for all countries. If one sample has income under-reporting of 20 percent as compared to national accounts, while another may reach under-reporting of 35 percent, the number of people under the poverty line would be vastly different for the two cases, even if the samples are based on the exact same population. The process used for adjusting the data in the poverty analysis is explained in detail in Chapter 4 and Annex 9. Unit of Analysis There are two principal units of analysis that tend to be used for producing estimates of poverty and income distribution statistics: (i) household per capita income for a single household member, and (ii) household per capita income for all household members. The first method ranks households according to household per capita income level by constructing distributions based on one observation per entire household. The second method ranks individuals according to household per capita income levels, and is based on one observation per each person in the household. Both measures are calculated in the same manner by dividing the total household income by the number of individuals in the household. The difference between the measures is the number of observations per household used in constructing the income distributions. The distributions based on all household members give a better reflection of the degree of inequality across the entire population, since family size is usually not constant across income levels and tends to be inversely correlated with income. Chapter 2 presents income distribution statistics which have been determined using the second method; that is, these statistics are based on total household income per capita for all individuals in a household. The income distribution is used as a proxy for consumption distribution, since welfare is generally considered a matter of the consumption of goods and services. In looking at income distribution at the individual level, it is assumed that resources are evenly divided among household members. However, while this assumption is useful for analytic purposes, it should not be interpreted to mean that an equal distribution of consumption is necessarily desirable. In fact, consumption needs are not usually equal throughout the family; a 12 Poverty and Income Distribution in Latin America: The Story of the 1980s sick individual may need a disproporionate share of resources, while a child might need less. The approach used here shows the potential access to income by all individuals; whether income is actually allocated evenly cannot be discerned from the data.'° Chapter 3 presents a narrower construction of the income distribution based on individual workers' income only. The analyses include only those individuals with positive income during the survey reference period and income is assigned only to the person who earned it. This approach examines the prevailing income structure in the labor market of each country. This focus does not direcdy assess individual welfare levels, because it does not take into account any transfers of income between individuals either within a household or between households. The real value of these distributions, in addition to showing overall income inequality in the labor market, is found in the subsequent decomposition which highlights those charactezistics which are most associated with low income among individuals in the labor market. Since labor represents the principal asset of the poor, the decomposition analysis underscores the relevant factors for social programs to examine as potential venues for improving the income potential of the impoverished. Social Indicators The final section of this study presents an analysis of social indicators for the Latin America and Caribbean region. While income is the most common measure of economic well- being, it hints only indirectly at many "quality of life" indicators such as infant mortality, access to health care, nutritional status and educational attainment. Therefore several health, education, demographic and labor market indicators are presented in order to round out the profile of the living conditions profiled in the earlier chapters. 10 For a discussion of intra-household allocation of income/consumption, see Lawrence Haddad and Ravi Kanbur, Is There an Intra-Household Kuznets Curve? World Bank PRE Working Paper No. WPS 466. World Bank, Washington, D.C., 1990. Substantial theoretical and empirical effort has been channeled into the development of equivalency scales which correct for intra-household differentials in income/consumption needs, though there is little consensus on the most appropriate methodology to follow. Ravallion discusses this in detail and highlights the benefits and drawbacks to employing corrective measures for intra-household income/consumption allocation. See Martin Ravallion, Poverty Comparisons. LSMS Working Paper No. 88, World Bank, 1992. See also Marjorie McElroy, The Empirical Content of Nash-Bargained Household Behavior. Journal of Human Resources. 25 (1990): 559-583; Amartya Sen, Resources. Values and Development. Oxford: Basil Blackwell, 1984; and Duncan Thomas, Intrahousehold Resource Allocation: An Inferential Approach. Journal of Human Resources. (1990). This report does not employ equivalence scales in the analysis of income distribution or poverty, since to do so would require the use of arguable assumptions which could potentially introduce more bias than they correct. Introduction 13 This non-income profile of the population is based on two principal sources: Westinghouse Demographic and Health Surveys (DHS) and the household surveys listed in Annex 1. In lieu of historical trends at the aggregate level,"1 this report presents a disaggregated profile of the principal indicators from the latest time period avalable in order to highlight both their absolute level and their within-country variability according to selected characteristics. The health indicators focus on matemal and child health conditions, and are broken down by urban/rual region of residence and mother's educational level. Particular emphasis is placed on the infant mortlity level, since it is reflective of a wide range of conditions, including sanitation, nutrition, access to health care, and mother's educational level. The education, demographic and labor indicators are disaggregated by urban/rural region of residence and income quintile. Not all countries in the region are covered in this analysis due to data shortfalls. 11 Trends of the principal social indicators for countries in the Latin America and Caribbean region are presented in Annex 4 of Human Resources in Latin America and the Caribbean: Priorities and Action. Washington, D.C.: World Bank, 1993. I I 2 The Distribution of Per Capita Household Income As already mentioned, Latin American countries exhibited some of the most unequal distributions of income in the world in the 1970s. 1 This chapter examines the evolution of income inequality during the 1980s, particularly with respect to the severe economic recession which has affected most countries in the region. First, findings based on the most recent available household surveys are presented.2 This is followed by a broader discussion of income inequality in the region which draws on the work of others in order to give a more complete picture of how the distribution of income has evolved. Three principal indicators of income distribution are presented for each household survey: the decile income distribution, the Lorenz curve and the Gini coefficient. (See Box 2.1) Each of these is based on household per capita income for all individuals surveyed. Table 2.1 reports the Gini coefficient and percentage share of income accruing to the bottom 20 percent of the population in eighteen LAC countries for various years since 1979. Again it is emphasized that the results for seven of these countries are for metropolitan or urban regions only. Twelve of these countries have data at two points, which enables us to compare changes in the income distributions over time.3 Annex 3 presents the full decile income distributions as well as the Lorenz curve and Gini coefficient for each individual country. Looking at the latest available data, income inequality as measured by both the Gini coefficient and income share of the bottom 20 percent remains high in Latin America relative to similar statistics for other parts of the world. (See Box 2.2.) Brazil, Chile, Guatemala, Honduras and Panama all have Gini coefficients which exceed 0.55. The bottom 20 percent of the income I See Box 2.2. 2 See Chapter 1 for a discussion of the surveys employed in this study, and an overview of the definition of income, geographical coverage and unit of analysis for determining income distribution statistics. See also Annex 1 for a detailed description of each survey. 3 There is not strict over time comparability for Honduras. The 1986 data for Honduras cover urban areas, while the 1989 data are national in coverage. However a separate Gini coefficient and bottom 20 percent income share have been calculated for urban individuals only for the Honduras (1989) data. These calculations are given in a footnote to Table 2.1. See Chapter 1 for a detailed discussion of the data definitions. 16 Povery and Income Distribution in Latin America: The Story of the 1980s distribution receives less than 3 percent of total income in Brazil, Guatemala, Honduras and Panama. At the other end of the spectrum, Paraguay (Asunci6n) and Uruguay (urban) demonstrate the least degree of inequality, with Gini coefficients of 0.398 and 0.424, respectively, and 5.9 percent and 5.4 percent of income accruing to the bottom 20 percent of the income distribution. The Distribution of Per Capita Household Income 17 Box 2.1: The Lorenz Curve and the Gini Coefficient The Lorenz curve is a cumulative distribution of income across the population, while the Gini coefficient is a measure of income inequality which is derived from the Lorenz curve. The following example from the Argentina 1989 (Buenos Aires) data set highlights how each is determined. Argentina: Income Distribution Argentina: Income Distribution Lorenz Curve Mvay, 198 May, 1989 Millions Household per capita income 5 100 4 3 50 0 0 1400 2800 4200 5600 7000 8400 9800 11200 12600 14000 0 50 100 Populsabon share Household per Capita Income The two distributions above show the exact same thing, but in a different format. The histogram on the left shows the frequency distribution of ten income levels. Each income level spans an equivalent increase in income from the previous income grouping, and is scaled according to an absolute number of individuals. In the Argentine case above, each of the ten levels spans an income range of 1400 australes per month, and is scaled according to millions of individuals. The Lorenz curve on the right simply shows that same distribution, but in a cumulative format based on percentage shares of population and total national income. Thus as the population share grows larger, so does the total income share of that population. At the extremes, zero percent of the population earns zero percent of the national income, while 100 percent of the population earns 100 percent of the income. The Gini coefficient is an inequality index which is defined as the ratio of the area enclosed between the 45 degree line and the Lorenz curve, to the area of the entire triangle enclosed by the 45 degree line. When a large percentage of total national income is concentrated among a relatively small percentage of individuals, the overall Lorenz curve tends to be lower, with a sharp rise in income share at the top of the income distribution. Therefore, in general, the Gini coefficient will increase when the distribution of income becomes more skewed. 18 Poverty and Income Disiribution in Latin America: The Story of the 1980s Table 2.1: Gini Coefficient and Bottom 20 Percent Share of Income at the Individual Level % Share of Income of Bottom 20% of Gini Coefficient Population Circa 1980 Circa 1989 Circa 1980 Circa 1989 Country Year of Survey (or earliest) (or latest) (or earliest) (or latest) Argentina (Buenos Aires) 1980 1989 0.408 0.476 5.3 4.2 Bolivia (Urban) 1986 1989 0.516 0.525 3.9 3.5 Brazil 1979 1989 0.594 0.633 2.6 2.1 Chile .. 1989 .. 0.573 .. 3.7 Colombia (Urban) 1980 1989 0.585 0.532 2.5 3.4 Costa Rica 1981 1989 0.475 0.460 3.3 4.0 Dom. Republic .. 1989 .. 0.503 .. 4.2 Ecuador (Urban) .. 1987 .. 0.445 .. 5.4 El Salvador (Urban) .. 1990 .. 0.448 .. 4.5 Guatemala 1986-7 1989 0.579 0.587 2.7 2.2 Honduras8 1986 1989 0.549 0.591 3.2 2.8 Jamaicab .. 1989 .. 0.435 .. 5.1 Mexico 1984 1989 0.506 0.519 4.1 3.9 Panama 1979 1989 0.488 0.565 3.9 2.0 Paraguay (Asunci6n) 1983 1990 0.451 0.398 4.9 5.9 Peru (Lima)b 1985-6 1990 0.428 0.438 6.2 5.7 Uruguay (Urban) 1981 1989 0.436 0.424 4.9 5.4 Venezuela 1981 1989 0.428 0.441 5.0 4.8 Source: Based on country-specific household surveys described in Annex 1. Notes: Individual income has been calculated by dividing total household income by the nwnber of individuals in the household. a Results are not strictly comparable due to differences in geographical coverage between the 1986 and 1989 surveys. The Gini coefficient based on urban households onlyfor Honduras 1989 is 0.556 whik the bottom 20 percent icome share is 3.5 percent. b Based on consumption data. not availabk Ihe Distribution of Per Capita Household Income 19 Box 2.2: Inequality in Latin America vs. the Rest of the World Latin America has historically had very high levels of income inequality relative to other areas of the world. A joint study by the World Bank and the International Labour Office presents an analysis of income distribution in twenty-three countries, including both industrialized and developing nations.' These distributions are for years during the 1970s and, in some cases, the 1960s. World Bank/ILO Studya This Study Bottom 20% Bottom 20% Gini Share Gini Share Latin American and 0.52 3.1% 0.50 4.0% Caribbean Countries Non-LAC Countries 0.39 6.5% n.a. n.a. Taking out the three Latin America and Caribbean cases in the World Bank/ILO study, the remaining twenty countries have a mean Gini coefficient of 0.39 and a mean bottom 20 percent income share of 6.5 percent of total income. In comparison, the LAC countries in our study have a mean 1989 Gini coefficient of 0.50 and a mean bottom 20 percent income share of 4.0 percent of total income. Although the World Bank/ILO study is based on data from the 1960s and the 1970s, the disparity which they find between LAC countries and non-LAC countries is still quite relevant, and could possibly have increased over time. Wouter van Ginneken and Jong-Goo Park, eds., Generating Internationally ComDarable Income Distribution Estimates. Geneva: LO, 1984. 20 Poverty and Income Distribution in Latin Amerca. The Story of the 1980s. Changes in Income Distribution, 1980 - 1989 Figure 2.1 shows the changes in the Gini index for each of the twelve countries for which data are available at two points in time. The changes in income inequality show mixed results for the time periods examined. The Gini coefficient worsened in Argentina (Buenos Aires), Bolivia (urban), Brazil, Guatemala, Honduras, Mexico, Panama, Peru (Lima) and Venezuela. By this same measure, income distribution improved in Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban). As might be expected, the bottom 20 percent income share went up in all countries whose Gini coefficient declined, and it went down in all countries whose Gini coefficient increased. Of particular interest are the seven country cases for which there are observations at the beginning and the end of the decade. Argentina (Buenos Aires), Brazil and Panama each experienced a sharp rise in inequality, while Venezuela showed a modest increase in inequality. Colombia (urban), Costa Rica and Uruguay (urban) had substantial reductions in inequality. The latter three countries which showed reductions in income inequality had an average growth of 3 percent in per capita income for the entire decade, while the four countries which experienced an increase in income inequality had an average decline of 12 percent in per capita income for the same period. In five other country cases, data are available for the middle and the end of the decade. Bolivia (urban), Guatemala, Honduras, Mexico and Peru (Lima) had rising income inequality during the second half of the decade. In contrast, Paraguay (Asunci6n) showed a substantial drop in income inequality; by 1990, Paraguay demonstrated the least income inequality of the eighteen countries analyzed. Thus, a uniform rise in inequality did not occur across all countries in the region during the 1980s. However, further analysis does show that trends in inequality appear to have been significantly influenced by trends in per capita income. Examining each country separately permits a better assessment of individual country performance. Unfortunately, the surveys on which this study is based do not necessarily correspond to the high and low points of the economic trends of their respective countries. This makes it difficult to examine the relationship between inequality and country economic performance. Nonetheless, there does seem to be some definite links between these two conditions. Argentina (Buenos Aires), Panama and Venezuela all experienced a rise in inequality and negative per capita income growth between 1980 and 1989. All three of these countries were in severe recession in 1989. Argentina had yet to come to terms with its crippling fiscal deficits, while Panama was reeling from the effects of U.S.-imposed sanctions which had begun in 1987. Venezuela was experiencing severe contraction, as a new government sought to bring fiscal policy in line with revenues. On the other hand, Colombia (urban) and Costa Rica both had a lessening of inequality coupled with positive per capita income growth during the decade. Colombia was experiencing slow but positive growth in 1989, while Costa Rica was growing a bit more rapidly. The Distribution of Per Capita Household Income 21 Figure 2.1: Changes in Per Capita Income Inequality 0.7 0.65 WORSE IN 1989 BRA 0.6 GUA HON WE PAN E 0.55 29 BOL * * COL c) MEX* : 0.5 ARG * 0.45 - VENR/ S~~~~~'R PERU *URU BETTER IN 1989 0.4 P PRA 0.35 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 Gini Circa 1980 22 Poverty and Income Distribution in Latin America: The Story of the 1980s In contrast, Uruguay (urban) experienced a fall in income inequality despite a general reduction in per capita income over the decade, while the exact opposite trends occurred in Brazil. Yet these simple observations ignore the specific context of each country. In 1989, Uruguay was undergoing a strong and extended recovery phase which followed a severe recession that had occurred earlier in the decade. At the time of the 1989 survey, Uruguay was beginning to enjoy the benefits of the painful adjustments it had made between 1980 and 1985. On the other hand, Brazil was in the middle of a downtum in 1989 after having experienced very strong growth during 1986. In fact, the economic expansion in the middle of the decade had been so strong that, despite a subsequent severe recession, per capita income in 1989 was still above its 1979 level. Thus it seems that income inequality, as measured by the Gini, is responsive to fluctuations in economic growth. For a given level of per capita income, the degree of inequality appears sensitive to the direction of economic performance. The country cases described above all indicate that recession is associated with rising inequality. Examining the extremes of the income distribution, the bottom 20 percent received a smaller percentage of national income, while the top 20 percent expanded its share, for every case where the country was in recession at the time of the 1989 survey. The opposite is true for all countries which were experiencing recovery. In the cases of Colombia (urban), Costa Rica and Uruguay (urban), the poorest 20 percent of the population actually improved their relative income share during the course of the decade. Figure 2.2 shows a plotting of the Gini coefficient against mean income in constant 1987 dollars for all country cases in this study. As can be readily seen, there is a distinct inverse trend between the Gini coefficient and per capita income in real dollar terms. Argentina (Buenos Aires), Colombia (urban), Costa Rica, Mexico, Panama, Paraguay (Asunci6n), Peru (Lima) and Venezuela all demonstrate a clear negative relationship between changes in real per capita income and income inequality. Guatemala shows no distinct pattern, while Brazil and Uruguay (urban) contradict the dominant trend rather strongly. The Distribution of Per Capita Household Income 23 Figure 2.2: Trends in Per Capita Income Inequality Gini 0.64 * BRA-89 GUAT-89 1 0.59 HON-890 t *COL-80 0 BRA-79 . GUAT-86 \ * CHI-89 41k * PAN-89 HON-860 0.54 0 COL-89 * BOL-89 * MEX-89 tO BOL-86 * DOM R-89 PAN-79* 0.49 MEX94e ; COSTA R-81 0 ARG-89 PARAG-83 *. COSTA R-89 EL S-90 * 0.44 - PERU 900 *ECUA-87 VENURU-81 -81 0.44 PEU-9O.~ JAM-89VE8 PERU-85 URU-8 PARAG-90 ARG-80 0.39 l 400 1000 1600 2200 2800 3400 Per capita income in 1987 constant US$ -24 Poveny and Income Distribution in Latin America: The Story of the 1980s Box 2.3: Changes in Per Capita Income During the 1980s Regional averages mask important differences in individual country economic performance during the decade. Argentina, Bolivia, Haiti, Nicaragua, Peru and Venezuela registered cumulative drops of more than 20 percent in per capita income between 1980 and 1990. In contrast, Chile and Colombia achieved cumulative growth of 8 and 9 percent, respectively, in per capita income over this same period. These cumulative per capita income growth rates correspond to disappointingly low average annual growth rates for the decade. Bolivia, Guatemala, Haiti, Nicaragua, Panama, Peru and Venezuela recorded average annual declines over 2.0 percent. Again, Chile and Colombia realized the best performances, with average annual increases of 1.1 percent each. Disappointingly, Brazil and Costa Rica were the only other LAC countries to achieve positive average annual growth during the period, at 0.6 percent each.a Importantly, most countries in the LAC region experienced highly unstable performances from year to year during the decade. The variability around trends is as notable as the trends themselves. With respect to income per capita, the following single- year drops occurred: Chile by 18 percent in 1981-82; Costa Rica by 12 percent in the same period; Mexico by 7 percent in 1985-86; Uruguay by 12 percent and 14 percent in 1981-82 and 1982-83, respectively; and Venezuela by 12 percent in 1988-89.b Such gyrating conditions necessitated a shortening of planning horizons, a need for continual readjustment to prevailing circumstances and a stark disarticulation of continuity in the level of services provided. This was true for both the private and the public sectors. But trends for per capita income for several countries were markedly different for the final years of the decade than for the decade on the whole. Chile, Costa Rica, Mexico and Paraguay achieved strong economic performance for 1988-90 relative to earlier in the decade. In particular, Chile realized an average annual per capita income growth of 5.6 percent, while Cost Rica, Mexico and Paraguay registered 3.6 percent, 2.3 percent and 2.9 percent growth, respectively, for 1988-90. In contrast, Peru experienced an average annual drop of 8.4 percent, while Argentina, Brazil and Venezuela fell an average of 4.9 percent, 2.7 percent, and 4.5 percent, respectively, for this same period.' However these last three countries were at the dept of economic contraction due to adjustment programs. Argentina and Venezuela have subsequently resumed vigorous growth, and are currently performing quite well. Notes: a This assessment refers only to countries with population over 2 million in 1990. b Panama experienced a drop of 20 percent in 1987-88 due primarily to the disruption caused by the U.S. trade embargo placed on that country. c Annual average per capita income growth rates for these countries for 1980-90 were: -1.8 percent in Argentina; 0.6 percent in Brazil; 1.1 percent in Chile; 0.6 percent in Costa Rica; -0.9 percent in Mexico; -1.3 percent in Paraguay; -2.0 percent in Peru; and -2. 0 percent in Venezuela. The Distribution of Per Capita Household Income 25 Inequality Trends During the 1980s: A Broader Picture Table 2.2 collects all of the available comparable evidence on trends in the Gini coefficients for fourteen Latin American countries, both from the surveys used in this study and from the work of others. Each source represents a set of consistent and comparable estimates of the Gini coefficient for a particular country; however, different sources are often nQt strictly comparable, even for the same country. In most cases, the Gini coefficients are based on total household income per capita, although in some instances they have been calculated using total household income. Each source is footnoted according to the basis of its Gini calculation. The purpose of this table is to present a general picture of trends in inequality throughout the decade for each country. This permits a closer examination of the link between the economic cycle of a country and its level of inequality. The data in Table 2.2 strongly confirm the relationship between the economic cycle of a country and its level of inequality. In the vast majority of cases, economic recession was accompanied by rising inequality while recovery was accompanied by falling inequality. Table 2.3 is a rough measure of this correlation. The table is based on twenty-six country economic cycles. Each cycle corresponds to a period of recovery or recession, with several periods included for those countries for which the data are more complete. Recession is defined as a period of falling per capita income, while recovery is a period of rising per capita income. The few observations which span the entire decade are classified according to the state of the economy in 1989. This makes no difference for Argentina, Colombia, Costa Rica, Panama and Venezuela, but it does cause Uruguay to be classified as in recovery even though average per capita income growth was negative for the decade. By these criteria, twenty-three of the twenty-six observations contained in Table 2.3 fall in either the northeast or the southwest quadrant, suggesting that economic growth reduces inequality. The evidence in these two tables bears on the relationship between income level and the distribution of income. This relationship has been the subject of much debate in the literature. Kuznets analyzed historical data from developed countries and found an inverted U- shaped curve when graphing income against inequality levels."5 The hypothesis behind this inverted U-curve is that countries in the initial stages of development experience increasing income inequality as the labor force moves from a relatively flat wage structure in agriculture to a more stratified wage structure in the higher paying urban sector. Physical and human capital are relatively scarce, and receive high returns for the few individuals who possess them. Over time, however, the number of poor remaining in agriculture drops as the urban labor market absorbs an increasing percentage of the population. The possession of both physical and human capital become diffused among the population, and the returns to each diminish. This Is Simon Kuznets, Economic Growth and Income Inequality. American Economic Review. 45, (March 1955): 1-28. Also, Simon Kuznets, Modern Economic Growth: Rate. Structure and Speed. New Haven, CT: Yale University Press, 1966. 26 Poverty and Income Distribution in Latin America: lhe Story of the 1980s Table 2.2: Gini Coefficients in Selected Latin American Countries, 19791990 Country/Source 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 ArenXt *enox Alm) this Study# 0.41 04M8 Finzb3ei (1991) 040 0.43 0.40 0.40 A4 0.43 0,45 Bolivia This Study' 0.52 0.52 Brazil nis Study' 0,59 0.63 BonelWSedlace 0.57 0.56 0.59 0.0 0.58 Ahmeida Rein et al 0.59 0.S9 0.60 0.59 0.60 0,61 Hoffinanb 0.59 0.60 0,62 Chile Mujica (Santiago) 0.52 0.53 0.52 0.54 0.54 0.56 0.53 0.54 0.53 Altimir-92-natl 0.47 0.46 Colonbia This Study' o.ss 0.53 Altimir (Bogota) 0.48 0.47 other urban 0.47 0,45 Costa Rica This Study' 0.48 0.46 Gindling/Berfy 0.40 0.40 0.42 0.38 0.38 0.37 0.36 0.42 0.42 Sauma/Trejos 0.42 0.42 Altimir 0.49 0.50 Guatnala l$is Stud9y 0.58 G.S5 WB country memio 0.48 0.53 Honduras This Study' 0.55 0.590 (urban) Miti0co Thb Study' 0.51 OM5 .5sti 0.44 Lustig 0,48 Atmir (1984) 0,48 Panama This Study' 0.49 0.56 Paraguay (Asunei6n) Tis Study' 0.45 O04 Peru This Study' 0.43 0.44 (expenditure) 0.44 GRADE-Lima 0.34 0.39 0.40 0.39 0.41 Uttuguay This Study' 0,4 0.42 Venezuela ThiS Study' 0.43 0.44 IESA 0.40 0.44 0.46 0.44 CEPAL-in Altimirb 0.39 0.42 Notes: Source: Table 2.1 hBased on Total Household Income. All otherfigures arefor per caDita household income. c In order to maintain comparability, the Gini coefficientfor the urban area only is presented here. The Distribution of Per Capita Household Income 27 Table 2.3: Income Inequality and the Economic Cycle Change in Income Distribution Recession Recovery Argentina (1980-82) Brazil (1983-86) Argentina (1982-85) Chile (1983-87) Chile (1987-90) More Equal Colombia (1980-89) Costa Rica (1983-86) Costa Rica (1981-89) Venezuela (1989-91) Uruguay (1981-89) Argentina (1985-88) Guatemala (1986-89) Argentina (1980-89) Bolivia (1986-89) Brazil (1979-83) Brazil (1986-89) Brazil (1979-89) Less Equal Chile (1980-83) Costa Rica (1980-82) Guatemala (1981-86) Mexico (1977-84) Mexico (1984-89) Peru (1981-84) Peru (1984-89) Panama (1979-89) Venezuela (1981-89) Source: Table 2.2 28 Poverty and Income Distribution in Latin America: The Story of the 1980s equalization of returns results in the downward slope in the inverted U-curve, reflecting a lessening in inequality. There has been a substantial amount of theoretical and empirical work focused on determining the validity of the "Kuznets curve".16 One interpretation of the findings in Table 2.2 and Table 2.3 is that certain Latin American countries experienced a backtracking along the Kuznets curve cycle during the 1980s - with inequality rising as per capita income fell and nations suffered a period of "reversed development". However the Kuznets curve reflects a long-run relationship which is based on the movement of the population between agricultural and non-agricultural sectors as society grows over time. Latin America during the 1980s did not demonstrate a reversal of this process. Rather, structural imbalances in the economy were exacerbated by shifts in the external sector; the combined effect was a cyclical downturn for many countries. But there was not a migration of workers from the urban sector back into agriculture. Therefore it is unlikely that the findings of this report have much bearing on the existence or shape of the Kuznets curve. A more likely interpretation of the relationship between income and inequality levels for Latin America during the 1980s is that the intense recessions put strong downward pressure on wages and employment levels, particularly for those at the bottom of the income pyramid and those living in urban areas. With limited unemployment insurance coverage, many workers were forced to accept either severe real wage reductions, unemployment or work in the informal sector. 17 Thus rising inequality was due more to increased intra-sectoral stratification than to movement between sectors, as would be expected in the Kuznets process. The hardest hit groups in the 1980s appear to have been the new entrants into the labor market and the poor. The former suffered from the handicap of little or no experience, 16 For early attempts to establish the Kuznets curve using cross-section data from developing countries, see Irma Adelman and Cynthia T. Morris, Economic Growth and Social Equity in Developing Countries. Stanford, CA: Stanford University Press, 1973; Felix Paukert, Income Distribution at Different Levels of Development: A Survey of Evidence. International Labor Review. 108, nos. 2-3 (1973): 97-125; and Montek S. Ahluwalia, Inequality, Poverty and Development. Journal of Development Economics. 3 (1976). These results have been questioned in recent work by Sudhir Anand and Ravi Kanbur, The Kuznets Process and the Inequality-Development Relationship. Journal of Development Economics. 40, no. 1 (1993); Gary Fields, Changes in Poverty and Inequality in Developing Countries. World Bank Research Observer. 4, (1989): 167-186; and Martin Ravallion, Growth, Inequality and Poverty: New Evidence on Old Questions. World Bank. March 1993. Unpublished mimeo. 17 For a discussion of the interaction of labor markets, wage determinants and economic adjustment, see Susan Horton, Ravi Kanbur and Dipak Mazumdar, Labor Markets in an Era of Adjustment: An Overview. World Bank PRE Working Paper No. WPS 694. World Bank, Washington, D.C., 1991. The Distribution of Per Capita Household Income 29 Box 2.4: Real Minimum Wage and Income Inequality An exercise was conducted whereby the direction of change of the real minimum wage was recorded for each of the country cases in Table 2.3. For the fifteen cases where recession was associated with a rise in inequality (southwest quadrant of Table 2.3), real wages fell in twelve instances and were essentially constant in the other three.' For all cases in which the real wage increased, income inequality declined (upper two quadrants of Table 2.3). Furthermore, with the notable exception of Argentina 1980-82 and 1982-85, all cases which showed an improvement in the real minimum wage were in the recovery stage of the economic cycle (northeast quadrant of Table 2.3). b Finally, three cases showed a trend of falling real minimum wages during recovery (Brazil, Chile and Uruguay in northeast quadrant of Table 2.3). However all three of these examples demonstrated a rise in the real average wage relative to the real minimum wage.' All of this suggests that real minimum wages may have an equalizing effect on the income distribution. However causality between inequality and changes in the real minimum wage is not clear. In most cases, it is probably true that the real minimum wage is also responsive to economic fluctuations, falling during recession and rising in recovery. Thus it is probably not an additional variable, but rather is endogenous to the economic cycle. But there are examples where the economic trend and the real minimum wage do not move together (Argentina 1980-82, and 1982-85, Brazil 1983-86 and Chile 1983-87). In the first two of these cases, the improvement in the real minimum wage may help explain why the distribution in Argentina improved despite the severe recession it suffered in those years. The experience of Brazil between 1979 and 1983 is also instructive in this regard. During those years, per capita income fell a total of 12.3 percent. Yet the real minimum wage was held virtually constant during this period. This may have tempered the rise in inequality, which was relatively moderate compared to the experience of other countries suffering comparable recessions. Notes: a Real minimum wage indices are taken from Cox Edwards (1991). b No real minimum wage was listedfor Venezuela for 1991. ' In Chile, unemploymentfellfrom 22 percent to 12 percent between 1983 and 1987, in Brazil employment rose by 17 percent and in Uruguay unemployment was cut almost in half between 1983 and 1989. Furthermore, although the minimum wage fell during these periods in each country, the average real wage rose in Brazil and Uruguay and was constant in Chile. Real wage data is presented in Alejandra Cox Edwards, Wage Trends in Latin America. Views from LATHR No. 18. World Bank, Latin American and the Caribbean Technical Department, Washington, D.C., 1991. Unemployment rates are presented in CEPAL, Estudio Econ6mico de America Latina y el Caribe 1990. Also, M. Louise Fox and Samuel A. Morley, Adjustment in Brazil: Who Paid the Bill? PRE Working Paper. World Bank, Washington, D. C., 1990. 30 Povey and Income Distribution in Latin Amenica: The Story of the 1980s while the latter tended to be the least-well educated group in the work force.18 New entrants into the labor force accounted for the bulk of the rising unemployment levels during the decade. Furthermore, for those that found work, the evidence suggests that there was a rise in the age-wage differential in most of the countries for which such statistics have been calculated.19 (See Box 2.4 for an analysis of the relationship between real minimum wage and income inequality.) A final issue to highlight in the analysis of inequality in Latin America during the 1980s is the effect of inflation on reported levels of inequality.20 Two issues in particular should be noted. First, surveys which were conducted during periods of high inflation collect income information which may or may not include the latest round of wage adjustments. Since all surveys except Peru 1990 have nominal income data for the entire survey period, changes in the price level during the survey period could bias the inequality calculations.2' There is little that can be done about this. However the problem is most likely minimal, given an assumption that the bias of income level due to inflation was distributed randomly throughout the entire survey sample. The second effect of inflation is on interest income. The 1989 surveys for both Argentina and Brazil both include interest income as part of total household income. But when inflation reaches levels as high as 33 percent per month, as it did in Argentina during the survey, the income due to nominal interest receipts will be quite large. However, a significant portion of this interest income is simply to correct for devaluation of the asset, and does not represent a true increase in welfare. The true value of income due to interest should be the el interest rate times the value of the asset. Since data on the value of the assets held by each household were not included in the surveys, the only way to correct for this overstatement of 18 Chapter 3 of this study examines the characteristics most strongly associated with poverty. This analysis reinforces the well-known correlation between educational level and earnings potential. The poor are often caught in a vicious cycle of little or no education, which results in low earnings, which then perpetuates a host of conditions affecting children in terms of educational opportunities and achievements. 19 For further evidence on age-wage differentials during the 1980s, see Samuel Morley, Policy, Structure and the Reduction of Poverty in Colombia: 1980-1989. Inter-American Development Bank, 1992. Unpublished mimeo. See also Samuel Morley and Carola Alvarez, Recession and Growth of Poverty in Argentina (1991); Poverty and Adjustment in Costa Rica (1992); and Poverty and Adjustment in Venezuela (1992). Inter-American Development Bank. Unpublished mimeos. 20 In particular, Argentina and Brazil had very high rates of inflation in 1989. Venezuela also had a high rate of inflation in 1989, but the Venezuelan household survey measured earned income only, and did not include interest income. 21 Because inflation in Peru during the survey period of the 1990 LSMS survey was so extreme, all figures for this survey have been inflated or deflated to reflect the real value of the Intis as of June 1 of that year. lhe Distribution of Per Capita Household Income 31 true income is to either eliminate or write down the nominal interest receipts included in the total income for each household. In an analysis not presented here, the changes in inequality between the two survey periods were calculated based on earned income only for Argentina and Brazil. The calculations based on earned income only showed the same results as those for total income - namely that inequality increased during the 1980s. In fact, the increase in inequality based on earned income was greater for both countries than the increase based on total household income. 3 A Decomposition of Workers' Income Inequality This chapter examines changes in workers' income inequality during the decade of the 1980s, a period of economic crisis and adjustment for most Latin American countries. The purpose is to identify the principal factors associated with inequality in individual workers' income. Furthermore, a comprehensive profile is developed which highlights the most prominent characteristics of those individuals in the work force who comprise the bottom 20 percent of the income distribution. While the analysis in Chapter 2 was based on the income distribution for the entire population, the analysis presented here is based on the distribution of workers' income only. The data sources used in this chapter are a subset of those used in Chapter 2, and are described in detail in Annex 1.22 The selection of countries has been determined solely on the basis of data availability. The level of coverage varies across countries; in some countries it is national while in others it is limited to urban areas. With the exception of Honduras, coverage between the two survey periods in any single country is identical. In the case of Honduras, the 1986 survey covers urban areas only while the 1989 survey is national in scope. Therefore, the results for 1986 and 1989 in Honduras are not strictly comparable. Ten countries are included in the analysis in this chapter (Argentina - Buenos Aires, Bolivia - urban, Brazil, Colombia - urban, Costa Rica, Guatemala, Honduras, Panama, Uruguay - urban, and Venezuela), and survey observations are compared between two points in time. In most cases, one time point corresponds to the late 1970s/early 1980s while the other corresponds to the late 1980s. In the case of Bolivia (urban), Guatemala and Honduras, the earlier observation is for 1986, while the later one is for 1989. By including two observations for each country, it is possible to assess changes in workers' income inequality during the decade. The worker income distributions are based on all individuals 15 years of age or older who were in the labor force and had positive income at the time of the survey. Given that the analysis uses variables which identify personal rather than household characteristics, it was necessary to relate them to personal rather than household income. Therefore income has been 22 The household surveys which are included in this chapter's analysis are highlighted with an asterisk (*) in Annex 1. 34 Powrty and Income Distribution in Latin America: lhe Story of the 1980s assigned to individual earners. In seven of the ten country cases, per capita income is from all sources, including work, rents, and transfers. For Bolivia (urban), Panama and Venezuela however, the income variable corresponds exclusively to income from work; thus, rents and transfers are not included for these countries. The Panama distributions represent a special case in that they include only employees; no persons who qualify as self-employed or employers are included in the analysis. Theil Index of Inequality This analysis uses the Theil index of inequality as an altenive to the Gini index for measuring income disparities in the distributions, because the former allows a cleaner decomposition than the latter. By using an additively decomposable measure, inequality in the overall income distribution can be broken down into components which reflect the influence of various characteristics of the population.23 The four subgmups included in this chapter are age, education level, employment status, and economic sector of employment. By choosing these sub- groupings, individual characteristics can be assessed in terms of their contribution to total income inequality. The Theil index of inequality M can be defined as follows: N T-Ey.'N- (3.1) where y; is the share of the ith individual in total income, and N is the sample size. This index fluctuates between 0 and In N, so that its value increases with the size of the population. Thus, the index can be standardized (T*) in the following manner: r = T N(3.2) In N where the value of the index is expressed as a percentage of its own maximum value. 23 Theil's entropy coefficient (T) is the only zero-homogeneous measure which satisfies the criteria of "income-weighted decomposability". The less widely used Theil coefficient L is the only measure which satisfies the criteria of "population-weighted decomposability". See Francois Bourguignon, Decomposable Income Inequality Measures, Econometrica 47, no. 4 (1979); and A.F. Shorrocks, The Class of Additively Decomposable InequalityMeasures, Econometrica 48, no. 3 (1980). A Decomposition of Workers ' Income Inequality 35 Changes in the Distribution of Workers' Income During the 1980s The distribution of workers' income by decile for each country case is presented in Annex 4; also included are the Gini coefficient and the Theil index for the overall distributions. Table 3.1 summarizes the main results, and Figure 3.1 provides a graphical overview. Because the Gini is the most widely used measure of income inequality, it is presented here along with the Theil index. But as mentioned earlier, the Theil index will be used in this chapter for the decomposition of inequality. With all the caveats resulting from the differences in coverage between countries, the results in Table 3.1 give a comparative idea of levels of income inequality among the ten countries included in the present sample. The Gini coefficients presented in Table 3.1 differ from those presented in Chapter 2 because they are based on a different distribution of the population. In Chapter 2, income is not assigned solely to the individual who earned it. Rather, the total income for all household members is aggregated and then divided equally to each household member; each household member is assigned the same per capita income. This method gives insight into the welfare of each individual in terms of potential consumption, but does not reflect individual earning levels. The Gini coefficients and the Theil indices presented in Table 3.1 are based on workers' income only. The underlying distribution for this analysis includes only those individuals with positive income during the survey reference period,24 and income is assigned only to the person who earned it. Therefore, the distribution of workers' incomes highlights disparities in the income of individual workers. This is not the same as inequality in per capita household income. When the Theil index is considered, income inequality is found to have increased in six of the ten countries: Argentina (Buenos Aires), Bolivia (urban), Brazil, Honduras, Panama and Venezuela. The increase was the highest in Argentina (Buenos Aires), Brazil and Panama. In the case of Venezuela, there is a discrepancy in the direction of change for the two measures of inequality. In fact, when the Venezuelan distributions by decile are considered, it appears that the top six deciles all improved their share of income at the expense of the seventh to ninth deciles. Finally, the relatively small increase in income inequality found in the case of Honduras cannot be adequately interpreted, given that the survey coverage differs between the two years. The remaining four countries in the sample show a significant reduction in income inequality. The reduction is particularly impressive in the case of Colombia (urban), where the equivalent of six and one half percentage points were transferred from the top two deciles to the rest of the population. Though less dramatic, the reductions in inequality that took place in Costa Rica and Uruguay (urban) are still very significant. 24 See Annex 2 for the reference period for each survey. Unless otherwise specified in the survey questionnaire, the reference period is assumed to be the month prior to the actual survey date. 36 Poverty and Income Distribution in Latin America: The Story of the 1980s Table 3.1: Measures of Inequality in Workers' Income Country Gini Index Theil Theil Index (T) Index (T* %) Theil Change Earlya Late Earlya Latea Earlya Latea (% points) Argentina (Bs. As.) 0.389 0.461 0.264 0.431 3.25 5.02 1.77 Bolivia (Urban) 0.479 0.515 0.443 0.559 5.41 6.02 0.61 Brazil 0.574 0.625 0.691 0.843 5.86 7.27 1.41 Colombia (Urban) 0.578 0.515 0.673 0.526 7.20 5.16 -2.04 CostaRica 0.451 0.410 0.339 0.317 3.76 3.49 -0.27 Guatemala 0.532 0.528 0.612 0.584 6.55 6.19 -0.36 Hondurasb 0.528 0.533 0.551 0.563 5.91 5.96 0.05 Panama 0.376 0.446 0.266 0.361 2.98 4.09 1.11 Uruguay (Urban) 0.452 0.420 0.394 0.329 4.22 3.51 -0.71 Venezuela 0.512 0.498 0.436 0.451 3.95 4.08 0.13 Source: Annex 4 Notes: a See Annex lfor exact "early" and "late" survey dates. b Early and late values are not strictly comparable due to differential survey coverage. A Decomposition of Workers' Income Inequality 37 Flgure 3.1: Changes in Workers' Income Inequality During the 1980s 7.5 -B BRA.o WORSE IN 1989 6.5 BOLE HO/N *GUAT 5.5 - / COL n 2 ARG COL 4.5 -/ VEN/ BETTERIN1989 PAN 3.5 - *URU C.R. 2.5 I I I I 2.5 3.5 4.5 5.5 6.5 7.5 Theil Circa 1980 38 Poverty and Income Distribution in Latin America: The Story of the 1980s Among the countries with the most unequal distnbution of workers' income, Colombia (urban) registered a significant improvement between 1980 and 1989. Having ranked first in mequality dunng the early penod by the measures employed here, Colombia (urban) ranked fifth during the late period. Brazil, which ranked fourth by the Theil index in the early period, exhibited the greatest degree of inequality during the late period. Bolivia (urban), Guatemala and Honduras also continued to demonstrate very high degrees of inequality. There have also been some changes regarding the ranking of countries with the least inequality. In the late 1970sIearly 1980s, Argentina (Buenos Aires) and Panama had the lowest levels of income inequality among the ten countries. This was no longer true by the late 1980s, and by 1989, Argentina (Buenos Aires) presented a moderately high level of income inequality. Costa Rica and Uruguay (urban), which already presented relatively low levels of income inequality in the early 1980s, exhibit the least inequality by the end of the decade. Figure 3.2 shows a plotting of the changes in income inequality (measured in terms of the standardized Theil index) against per capita income for the ten countries. With two exceptions, the evidence indicates the existence of a negative correlation between changes in income and inequality. The two exceptions are Brazil and Uruguay (urban). Brazil shows inequality to have increased at the same time as a rise in income per capita, while Uruguay experienced a reduction in inequality during a period of falling average standards of living. In sum, the different distributive performance of the ten countries suggests a certain relationship between macroeconomic conditions, particularly with respect to per capita income and workers' income inequality. This mirrors the relationship found in Chapter 2 between per capita income of the entire population and income inequality. However, the diverse patterns found within the present samples of countries clearly nlles out a simple relationship between the changes in average income and those in inequality. A Decomposition of Workers' Income Inequality 39 Figure 3.2: Trends in Workers' Per Capita Income Inequality Theil (%) 8 * COL-80 0 BRA-89 GUAT-86 / GUAT-89@ ION8 6 BOL-890 * BRA-79 BOL-86- 5 *COL-89 * ARG-89 4PAN-89 VEN-8 K URU-81 400 COSTA R-81 \ VEN-81 COSTAR-89 URU-89 * ARG-80 3 0 PAN-79 2 I l l l l 200 600 1000 1400 1800 2200 2600 3000 3400 Per capita income in 1987 constant US$ Source Table 3.1: World Tables. 40 Poverty and Income Distribution in Latin America: The Story of the 1980s A Static Decomposition of Income Inequaity This section assesses the relative influence of several variables in explaining the level of income inequality at the time of each country survey. This is done by decomposing the Theil index according to the subgroups which reflect different levels of the vanables selected for analysis. If the population is partitioned into J groups according to the values of a certain variable (i.e., j = 1,2,3,4), the Theil index can be decomposed to show inequality between groups and within groups. It is then conventional to treat the between component as the inequality "explained" by the variable, and the within component as the amount of inequality which is "unexplained" by that variable. The gross contribution of variable j (G) is defined as the inequality between the discrete J groups. Similarly, the joint contribution of variables j and k (G)j, is defined as the inequality between the J and K groups corresponding to those two variables. Finally, the marginal contribution of variable k given variable j (Muq) is the difference between their joint contribution and the gross contribution of variable j: Mth = G,k - Gj (3.3) When more than two variables are included in the decomposition, marginal contributions of several orders can be defined.25 In this analysis, only the marginal contribution of last order will be presented. This can be interpreted as the increase in total "explained" inequality obtained by adding a new variable to the decomposition while including all previous variables as well. Altimir and Pifiera performed a similar analysis for nine Latin American countries using information from around 1970.26 Three of the countries in their paper (Chile, Mexico and Peru) are not included in the present analysis, and their study did not include Bolivia, Guatemala, Honduras and Uruguay. Although their results are not strictly comparable with those given here due to the different variable classifications in the two studies, it is nonetheless interesting to determine the extent to which consistent results are found throughout a time span of two decades. Therefore, relevant comparisons between the results presented here and the earlier results of Altimir and Piniera are made when appropriate. 25 More generally, when z variables are included it is possible to calculate the marginal contributions from order 1 to order z-l. 26 Oscar Altimir and Sebastian Pifiera, Analisis de Descomposiciones de las Desigualdades de Ingreso en America Latina. In 0. Mufloz, La Distribuci6n del Ingreso en America Latina. Buenos Aires: El Cid Editores, 1979. A Decomposition of Workers ' Income Inequality 41 Four variables are used in this analysis: age, education, employment status and economic sector of employment. All variables have been categorized in a discrete fashion, with the criteria being the same for all countries. Although there is an obvious element of arbitrariness in the classification criteria, the important consideration is that the same criteria are used in all cases. This ensures that differences in the explanatory power of each variable across countries and time are not due to the selection method.27 The age and education varables have each been categorized in four groups. These are: Age = 15-25 years, 26-40 years, 41-55, and 56 or more; Education = 0-5 years of schooling, 6-8 years, 9-12 years, and 13 or more. The sector of employment variable has been categorized in eight groups: agriculture, mining, manufacturing, transportation-communication-utilities, construction, commerce, financial services, and services. In some countries, given the survey coverage, the participation of agriculture and mining is nil. The employment status variable has been split into either three or four groups, depending on the country. For Argentina (Buenos Aires), Bolivia (urban), Brazil and Costa Rica, employment status has been categorized in three groups (employees, self-employed, and employers), while in Colombia (urban), Guatemala, Honduras, Uruguay (urban) and Venezuela it has been categorized in four groups (public sector employees, private sector employees, self-employed, and employers). The Panama survey includes only employees; therefore the variable employment status was excluded from the analysis. The frequency distributions corresponding to the four variables used in the decomposition analysis are shown in Annex 5. Table 3.2 gives a summary of the characteristics that have changed most for each country. 27 The variables in our study are categorized following different criteria to that used by Altimir and Piliera, 1979, op. cit. This certainly makes comparisons between this study and their study more problematic. 42 Poverty and Income Distribution in Latin America: The Story of the 1980s Table 3.2: Changes in Selected Characteristics of the Working Population (percent) Country Characteristic Early Value Late Value Argentina Post-secondary education 15 20 Bolivia Self-employed 36 42 Brazil Less-than-primary education 72 58 Colombia Post-secondary education 12 17 Costa Rica Self-employed 16 21 Guatemala Less-than-primary education 70 67 Panarna Post-secondary education 15 24 Uruguay Self-employed 17 20 Venezuela Less-than-primary education 34 23 Source: Annex 5 Tables 3.3 and 3.4 show the main results of the decomposition analysis. Table 3.3 presents the gross and marginal contributions of individual variables to total workers' income inequality. The marginal contributions are calculated from those models which showed the highest joint contributions for each country data set. These joint models are displayed in Table 3.4. With two exceptions, these are three-variable models based on the three most relevant characteristics of the work force. In the cases of Brazil and Venezuela, the large sample size provided enough degrees of freedom to use all four variables in the decomposition exercise. On average, the joint models explained 44.7 percent of total inequality. With the exception of Costa Rica, education and employment status were included in the preferred model in every case. In all but one instance (Venezuela), the "between" component of the decompositions was larger during the earlier period. For the early observations, the average joint contribution was 47 percent, while it was 42.4 percent for the later observations. In other words, the models for the first period in each country tended to explain a higher percentage of total workers' income inequality than the models for the second period. This may reflect an increase in importance of unobserved factors in explaining income inequality throughout the decade. However, a dynamic decomposition analysis would be required to further to assess the evolution of changes in income inequality over time.28 28 L. Ramos, The Distribution of Earnings in Brazil: 1976-1985. Ph.D. Dissertation, University of California, Berkeley, 1990. See also Ariel Fiszbein, Essays on Labor Markets and Income Inequality in Less Developed Countries. Ph.D. Dissertation, University of California, Berkeley, 1991. A Decomposition of Workers' Incone Inequality 43 Table 3.3: Contribution of Individual Variables in Explaining Inequality (Decomposition of Theil Index) Gross Contribution (percent) Marginal Contribution (percent) Employ- Employ- ment Educa- Economic ment Educa- Economic Country Year Age Status tion Sector Age Status tion Sector Argentina 1980 8.6 7.7 18.2 3.6 12.7 5.7 23.7 1989 6.9 6.2 20.2 4.7 11.3 4.2 23.3 Bolivia 1986 4.3 19.7 6.0 10.0 .. 22.5 10.7 13.0 1989 5.0 5.8 8.6 4.3 .. 8.5 10.0 7.0 Brazil 1979 8.3 13.1 28.8 7.7 9.1 8.9 25.6 6.7 1989 7.1 14.1 26.5 5.9 6.9 10.7 23.7 5.9 Colombia 1980 10.1 16.4 35.1 5.4 7.0 8.6 31.1 1989 10.6 15.2 30.2 3.9 6.6 7.1 27.8 Costa Rica 1981 12.4 1.6 26.4 12.8 13.6 .. 20.3 4.3 1989 7.7 3.0 23.6 11.1 8.0 .. 18.3 5.7 Guatemala 1986 2.8 18.9 30.3 8.2 .. 13.2 20.8 7.8 1989 2.6 20.1 29.3 11.4 .. 10.7 15.4 4.5 Honduras 1986 11.2 12.8 37.4 4.9 9.6 3.4 28.7 1989 5.1 20.7 37.3 13.1 6.3 6.6 26.0 Panama 1979 7.5 .. 35.8 10.2 11.0 .. 34.4 7.2 1989 13.3 .. 28.8 10.0 14.9 .. 26.9 6.6 Uruguay 1981 8.1 13.1 13.4 5.0 9.9 10.4 17.2 1989 8.1 11.8 10.1 5.8 10.1 10.8 14.2 Venezuela 1981 6.3 11.9 26.3 6.7 7.2 8.8 21.5 4.2 1989 8.3 19.6 23.1 4.9 6.0 15.7 20.9 4.4 Note: .. not available 44 Poverty and Income Distribution in Latin America: The Story of the 198Qs Table 3.4: Joint Contribution of Variables in Explaining Inequality (ecompo6ition of Theil Index) Country Year Explanatory Variables Contribution (Percent) Argentina 1980 Age + Employment Status + Education 38.1 1989 Age + Employment Status + Education 35.8 Bolivia 1986 Employment Status + Education + Economic Sector 42.1 1989 Employment Status + Education + Economic Sector 22.0 Brazil 1979 Age + Employment Status + Education + Economic Sector 55.5 1989 Age + Employment Status + Education + Economic Sector 50.6 Colombia 1980 Age + Employment Status + Education 53.4 1989 Age + Employment Status + Education 48.5 Costa Rica 1981 Age + Education + Economic Sector 44.9 1989 Age + Education + Economic Sector 36.6 Guatemala 1986 Employment Status + Education + Economic Sector 47.5 1989 Employment Status + Education + Economic Sector 43.4 Honduras 1986 Age + Employment Status + Education 53.5 1989 Age + Employment Status + Education 51.2 Panama 1979 Age + Education + Economic Sector 52.2 1989 Age + Education + Economic Sector 50.4 Uruguay 1981 Age + Employment Status + Education 36.0 1989 Age + Employment Status + Education 32.6 Venezuela 1981 Age + Employment Status + Education + Economic Sector 46.9 1989 Age + Employment Status + Education + Economic Sector 53.0 A Decomposition of Workers' Income Inequality 45 The most striking finding of the decompositions is the overwhelming preeminence of education, as demonstrated in Figure 3.3. In eighteen of the twenty cases (excepting Bolivia and Uruguay 1986), education has an higher gross contribution to inequality than any of the other variables. On average the gross contribution of education is approximately 25 percent; in other words, one fourth of total inequality can be explained as inequality between individuals grouped in just four groups according to their schooling level. In nineteen out of the twenty cases (excepting Bolivia 1986), education also has the highest marginal contribution to total inequality. The finding that education has the highest gross contribution to overall income inequality is consistent with the results of Altimir and Piflera. The only other variable in their paper which had a similar contribution to inequality was occupation, which was not included in the analysis presented here. Unfortunately, Altimir and Piflera did not report marginal contributions in their analysis. Employment status is the second most important variable in the present decomposition analysis. Its average gross contribution to inequality is approximately 13 percent. This is half of the corresponding contribution of education, and is much larger than the finding of Altimir and Pifnera. The contribution of the employment status variable is unusually low in the case of Costa Rica. This is probably due to relatively less stratification of the labor force in Costa Rica as compared with the other countries analyzed. The share of employees in the labor force is very high in Costa Rica, and the average employee income is extremely close to the overall mean income in the sample.29 On average, the contribution of the age variable is somewhat smaller than that of employment status. And in the large majority of cases, the economic sector component had the lowest contribution to income inequality among the four variables assessed on this study. Again, these findings appear to be consistent with those of Altimir and Pinfera. The preeminence of education as a source of income inequality when compared with age, employment status and economic sector has important implications. Relative to the other variables, education is a more permanent characteristic: individuals move from one age group to another throughout their lives, and are able to change sector of employment if sufficient mobility exists in their country. However, in many developing country settings (particularly with respect to the poor), it is not common for people to return to school in order to enhance their earnings capabilities. In general, once an individual reaches adulthood, little further schooling is attained. 29 The average employee income is fully equal to the overall mean per capita income for Costa Rica (1981), and is equal to 97 percent of mean income for Costa Rica (1989). 46 Povety and Income Distribution in Latin America. The Story of the 1980s. Figure 3.3: Marginal Contribution of Selected Variables in Explaining Inequality 25 , 2 22. 20 t. 1 5 10 5 0 Education Employment Category Age Economic Sector Source: Based on Table 3.3 On the other hand, of the four variables considered here, education is the most susceptible to public policies. However, as pointed out earlier, the static decomposition analysis presented here is not the most appropriate method to analyze the dynamic effects of changes in the distribution of education on income inequality. This would require a dynamic decomposition model which estimates the expected change in the returns to education associated with an expansion in the average level and dispersion of schooling. A Decomposition of Workers ' Income Inequality 47 Bottom 20 Percent Probability Analysis The above decomposition of the Theil index assesses the contribution of four principal variables in explaining workers' income inequality for all workers. This section of the chapter takes a more micro/continuous approach, and examines a greater number of variables. The focus is specifically on those individuals who make up the bottom 20 percent of the workers' income distribution. The factors examined have been determined by the available data from the household surveys, and are incorporated into a multivariate model which is tailored to each country case. The various factors which are included in each model are standardized in order to demonstrate the simultaneous contribution of each variable to the probability that an individual belongs to the bottom 20 percent of the workers' income distribution. By definition, this group would include the majority of the poorest of the poor.0 Because the factors which are examined are limited dependent variables, a logit model has been fitted. The model expresses the probability (P) of an individual belonging to the bottom 20 percent of the workers' income distribution, as a function of various personal characteristics (X) such as age, gender, years of schooling and sector of employment: 1 (3.4) The reported coefficients are partial derivatives which indicate the change in the probability of belonging to the bottom 20 percent relative to a unit change in one of the independent variables: d P = pip (I - P) (3.5) where P refers to the dependent variable-probability of the event, B is the logit coefficient and X is the string of independent variables used in the regression. 30 It can not be assumed that those in the bottom 20 percent of the workers' income distribution are the same as those in the bottom 20 percent of the per capita household income distribution for each country case as presented in Chapter 2. Some individuals in the bottom 20 percent of the workers' income distribution may have quite high per capita household incomes. This would be true if they received a substantial intra-household transfer of income. For this reason, the analysis of inequality presented in this chapter serves as a complement to the one put forth in Chapter 2, and not as a substitute for it. However, most individuals in the bottom 20 percent of the workers' income distribution would also rank among the poorest in the total population as well. 48 Poverty and Income Distribution in Latin America: The Story of the 1980s Annex 7 presents the results of the individual country multivariate models. The logit coefficient and the marginal effect are given for each variable included in a country model. In this manner, it is possible to assess the impact which a change in a particular variable would have on the probability of an individual belonging to the bottom 20 percent if aU other variables were kept constant. For example, in the case of Guatemala 1989, every extra year of schooling decreases the probability of a worker belonging to the bottom 20 percent by 3.3 percentage points. Annex 8 presents simulations based on the logit models in Annex 7. These simulations assign differing values to a target characteristic while maintaining all other variables at their mean. Comparisons can then be made both across characteristics and across time. There is a remarkable degree of stability between the early and late probability values for almost all combinations of personal characteistics in each country examined. During the 1980s, no characteristic groupings experienced a substantial decrease in the probability that an individual with those traits would rank in the bottom 20 percent of the workers' income distribution. For the most part, individuals with no education in 1980 were just as likely to belong to the bottom 20 percent of the income distribution as individuals with no education in 1989. The same could be said about the other variables examined as well. Among the various sample characteristics considered, education again has the highest differentiation with respect to the probability of belonging to the bottom of the income distribution. Table 3.5 shows that, for 1989, this ranges on average from 56 percent for those with no education to 4 percent for those with some university education. Table 3.5 also reflects that women have a 34 percent probability of belonging to the bottom 20 percent of the income distribution, as compared to a 15 percent probability for men. Furthermore, indigenous people in Bolivia (urban) and Guatemala have well above a 20 percent probability of belonging to the bottom quintile; the same is true for blacks and mulattoes in Brazil. The simulations also show that public sector workers have consistently lower probabilities of low incomes, as do individuals living in urban rather than rural regions. A Decomposition of Workers' Income Inequality 49 Table 3.5: Probability of Belonging to the Bottom 20 Percent of Income Distribution by Education and Gender, 1989 Educational Level Sex Region Country None Prim. Second. University Male Female Urban Rural Argentina 69 36 13 6 13 37 Bolivia 42 27 14 9 13 33 Brazil 54 19 5 2 14 37 19 35 Colombia 67 32 9 4 16 27 Costa Rica 55 25 8 4 16 34 14 28 Guatemala 36 14 5 2 16 35 Honduras 43 15 4 1 16 34 Panama 83 45 12 4 13 34 Uruguay 65 31 10 4 13 34 Venezuela 50 25 10 5 15 38 LAC Regiona 56 27 9 4 15 34 Source: Based on Annex 8. Secondary Education corresponds to 12 years of schooling. Probabilities are rounded to the nearest full percentage point. Note: a Refers to above countries only. .. not available 50 Povewr and Income Distribution i, Latin America: The Story of the 1980s Figure 3.4: Probability of Belonging to the Bottom 20 Percent of Income Distribution by Education Level and Gender, 1989 70 60 56 50 40 3 30 27 50 0 | He~~~~~~~Bseline 10 15 0 No education Primary Secondary Univ. Male Female Note: Ten country averages Source: Based on Table 5 The findings discussed in this chapter clearly indicate that education is the variable with the strongest impact on income inequality. On average, one-fourth of total income inequality can be attributed to inequalities in the level of schooling. Furthermore, the probability of belonging to the bottom 20 percent of the income distribution diminishes monotonically with schooling in all countries. An equalization in the distribution of education and the subsequent reduction in income differentialreturns to education associated with higher average levels of schooling should contribute significantly to reductions in income disparities and poverty across the region.31 31 For evidence in this respect in a number of Latin American countries, see George Psacharopoulos, Time Trends of the Returns to Educaton: Cross-National Evidence. Economics of Education Review 8, no. 3(1989): 225-231. Also see L. Gomez-Castellanos and George Psacharopoulos, Earnings and 4 Absolute Poverty Closely related to the issue of income distribution is that of absolute poverty. While the formner describes the relative distribution of income accnuing to specific income groups, the latter is a measure of those individuals in the population whose welfare is less than some absolute standard. This contrasts with the relative approach taken in Chapter 3, whereby those characteristics most associated with belonging to the bottom twenty percent of the income distribution were analyzed. However there will always exist twenty percent of the population at the bottom of the income distribution, though the individuals in this group may change over time. In the present chapter, the magnitude of absolute poverty is determined by the point at which the value of the absolute poverty standard intersects the income distribution. Importantly, this absolute poverty standard is exogenous to the income distribution, and may intersect it at any point. The poverty analysis presented here is based on the application of several poverty measures to each data set in the study. A regional absolute poverty standard is developed which represents a uniform welfare level across all countries. Based on this poverty standard, several poverty indices are calculated for each country case. For those countries in the region for which income data are not available, a regression model has been developed in order to assess poverty headcount levels. This permits a regional assessment of poverty levels in 1980 and 1989 based on an aggregation of headcount indices for individual countries. Since the poverty results based on regression analysis are decidedly less robust than those based on hard data, individual country estimates determined by this method have been marked with an asterisk in Annex 13 and are not presented in the main text. However, poverty estimates based on both household surveys and regression analysis are included in the aggregate poverty figures for the region. Given that the surveys from the late 1980s cover approximately eighty percent of the population in Latin America, while the surveys for the early 1980s cover approximately fidy percent of the population, some caution is needed in interpreting the aggregate poverty figures. However, the analysis presented here represents the most comprehensive empirical coverage of Latin America to date for the purpose of assessing poverty. The above results are then placed into a broader framework which incorporates outside estimates of poverty for selected countries in order to develop a more complete picture of the evolution of poverty during the 1980s. Individual country experiences are discussed, and examples of both above and below average performance are highlighted. Furthermore, the context of individual country 52 Poverty and Income Distribution in Latin America: The Story of the 1980s poverty performance is assessed in an effort to determine some of the conditions associated with successfill poverty reduction. Several points emerge from the analysis presented in this chapter. First, subject to the qualifications presented below, there has been an increase in poverty in the Latin America and Canbbean region between 1980 and 1989. According to the poverty standards used in this study, the poverty headcount index for these years rose from 26.5 percent to 31.0 percent for the region as a whole. Second, the rise was not uniform, either across countries or time. At least three countries - Colombia (urban), Costa Rica and Uruguay (urban) - succeeded in reducing their levels of poverty over the decade. Furthermore, in the countries for which the appropriate observations are available, the evidence suggests that poverty foliowed the economic cycle. As with income inequality, there was a distinct tendency for poverty to rise quite sharply during recession and to fall, though less sharply, during recovery. Third, evidence is emerging which shows that the reforms and renewed growth after 1989 are reducing poverty levels in the region. Because many countries experienced recession during 1989, some of the poverty estimates for that year do not reflect the benefits accruing from structural transformations in the countries which implemented reforms. Subsequent growth has had a strongly positive effect in reducing the headcount index in certain countries. Recent poverty estimates for Chile and Venezuela show that economic growth in both of these countries has led to a reduction in their headcount ratios.32 And while empirical poverty measurements are not available for recently recovering countries such as Argentina and Mexico, it appears that poverty is falling in these countries. Poverty levels are not improving in those countries which have been unable to reorient themselves and resume growth. Unfortunately this group represents 70 percent of those living in poverty in the LAC region. Two countries alone - Brazil and Peru - contain approximately 55 percent of all poor in Latin America. An additional 15 percent of the poor live in a group of smaller countries which include Bolivia, Guatemala, Haiti and Honduras. None of these countries is currently growing very rapidly, either because of policy failures and/or external conditions. Until economic conditions change for these countries, there is little likelihood for even the modest reductions in poverty that are presently occurring elsewhere on the continent. 32 CEPAL, La Pobreza en Chile. Santiago, Chile: CEPAL, 1991. Unpublished mimeo. See also Gustavo Marquez, Poverty and Social Policies in Venezuela. Paper presented at Brookings Institution and Inter-American Dialogue Conference, "Poverty and Inequality in Latin America," 1992. Absolute Powvty 53 Adjustment of Income Data for Underreporting Virtually all household income surveys are plagued by some degree of underreporting of income. This may be due to oversight on the part of the respondents, or it may be intentional in order to hide tax evasion. Regardless of the cause, total national income from an expanded sample often falls significantly short of total national income as measured by national accounts.33 Unfortunately it is difficult to assess and correct income underreporting in household surveys. Moreover, the adjustment of income data may introduce new biases. Since inequality measures reflect the relative distribution of an income within a closed sample, the impact of underreporting on these measures is fairly limited. Although the wealthy and the very poor tend to underestirnate their income to a greater degree than the population as a whole, the overall bias due to underreporting tends to be small. For these reasons, the inequality analyses in Chapter 2 and Chapter 3 are based on unadjusted income data from the household surveys. This allows for a more transparent description of the methodology folowed and the results obtained. Absolute poverty statistics, however, reflect the intersection of the income distribution vith an exogenous standard such as a poverty line. Because the value of this standard is determined independently of the income level of a country, any underreporting of income can have a strong effect on final poverty estimates. Underreporting will tend to lower the incomes across the entire distribution, though not necessarily in a uniform manner. In contrast, the poverty line which intersects the distribution remains constant in value. Thus underreporting of income causes the poverty line to intersect the distribution at a much higher point than if there were no underreporting. The result is a poverty estimate which is highly biased in an upward direction. Therefore it has been necessary to adjust the income data of the poverty analysis in order to correct for underreporting. The methodology followed in order to achieve this relies on modelling techniques which match survey income responses to national account figures according to both income type and socioeconomic characteristics of the household.34 To perform this correction directly requires matching a matnx of disaggregated national account figures with a matrix of disaggregated survey income figures for the entire sample from each country. Each subcomponent of the survey matrix is then expanded to be equal to its corresponding subcomponent in the national accounts matrix. Unfortunately, this process requires extensive access to disaggregated national 33 National accounts are usually subject to a system of cross-checking in an effort to determine the most accurate figures possible. While these figures may contain flaws, they ordinarily represent the most accurate data available for each country. For reasons stated above, survey data tend to be less reliable for estimating total national income, though they allow for micro-analysis of income data in a way which national accounts do not. See Altimir, 1987, op. cit. 34 Altimir has developed a detailed methodology of making national account adjustments through the use of weighted coefficients corresponding to sub-groups determined by income type and socioeconomic characteristics of the household. Oscar Altimir, 1987, op. cit.. 54 Poverty and Income Disinbution in Latin America: The Story of the 1980s account figures. In many cases, such disaggregations are not available at all, and must be estimated through a general equilibrium model. To overcome a lack of access to such data, this study utilizes a simpler approach by adopting an income expansion factor for each urban/rural subregion within a country. These expansion coefficients have been calculated using weighted aggregations of the individual income type expansion factors employed by CEPAL.35 Therefore, for each urban/rural region of the country data sets in this study, the income data have been adjusted by a single expansion factor which is applied across all income types. The degree of underreporting, and the expansion coefficients employed for each country and region, are listed in Annex 9; a more detailed description of the methodology foilowed is also included in Annex 9. Poverty Reference The poverty estimates contained in this chapter are based on the same set of thirty-one household surveys used in Chapter 2. These surveys cover eighteen countries for various years during the 1980s, and are described in detail in Annex 1. In thirteen country cases, access to survey data corresponding to both an early year and a late year during the decade permits an assessment of changes which occurred between the two years.36 The barriers to making comparable estimates of poverty across the region are fornidable and are discussed in detail in Chapter 1.37 First, wide variability exists in both the kind of income reported and the degree of income underreporting in each survey. Some countries count only labor income, while less than half of the surveys include in-kind income or the value of owner occupied housing. These differences in income definition in the data sets must be taken into consideration when evaluating the poverty analysis in this chapter. Annex 2 describes the income definition for each survey used in this study, while Annex 9 lists the degree of possible income underreporting in each survey as compared to national account figures. Second, the definition of poverty is the subject of continual debate. The intuitive understanding that poverty is a measure of deprivation belies the complexity of determining formal criteria of a definition. Most poverty definitions rely solely on income for ranking welfare, although it 35 Comisi6n Economica para America Latina y el Caribe (CEPAL), Magnitud de la Pobreza en Am6rica Latina en los Affos Ochenta. LC/G.1653-P. Santiago de Chile, 1991. 36 As in Chapter 2, the Mexico (1989) statistics are based on an unweighted sample which may not accurately reflect the actual population composition. Therefore, extreme caution should be exercised when assessing the figures for Mexico (1989). In particular, comparability with Mexico (1984) data can not be assured. 37 See also Altimir, 1987, op. cit., DaVanzo, 1980, op. cit., and van de Walle, 1991, op. cit. Absolute Povery 55 is possible to create weighted indices which also incorporate non-income attributes such as education, health, nutrition and housing. However, when a poverty definition includes an increasing number of criteria, incomplete and non-comparable data can weaken poverty compansons between countries and regions. In order to minimize problems of comparability, this report defines poverty in terms of per capita household income.3" Although using the single dimension of income as a welfare criterion fails to take into account the importance of benefits received through non-income sources, it is the single-most identifiable factor for assessing welfare levels across the Latin America and Caribbean region through available household surveys. Chapter 5 presents an inter- and intra-country comparison of social and demographic indicators in order to give a more balanced perspective of various non- income "quality of life" factors in each country. Having selected income as the criterion for determining welfare, a cut-off point must be chosen to deternine the bounds of who will be classified as "poor" and who will not. The ideal approach for making poverty assessments is to formulate a constant basket of goods which satisfies a set of minimum basic needs vith respect to nutrition, housing, clothing, education and health. The price value of this basket is then the poverty line. The poor are defined as those individuals whose income, or consumption, is less than the value of the poverty line. (See Box 4.1.) The simplicity of this reasoning neglects several persistent problems when applied in practice. Age, sex and work environment affect individual nutritional requirements; local customs influence dietary choice; while regional supply and demand patterns determine specific food prices. Equally ambiguous factors affect basic non-food requirements. In addition, the specific components of a basket of minimum basic needs have been the subject of much debate. Since all of these factors vary from region to region, there is no definitive poverty line which adequately reflects a set of minimum basic needs for all locations. 38 There is no adjustment for family composition, equivalency scales or disparities in the intra- household allocation of income/consumption. While these issues are relevant for understanding individual welfare levels, adjusting welfare criteria in order to account for intra-household disparities in consumption level and/or need for various household members can introduce unintended biases into the final poverty assessment. Therefore this study assumes equal distribution of household income to all household members. For a discussion of intra-household allocation of income/consumption, see Lawrence Haddad and Ravi Kanbur, Is There an Intra-Household Kuznets Curve? PRE Working Paper No. WPS 466. World Bank, Washington, D.C., 1990 56 Poverty and Income Distribution in Latin America: The Story of the 1980s Box 4.1: Methodology Behind the CEPAL Country-Specific Poverty Lines Countra-snecific poverty lines were developed by CEPAL in a series of background papers for Maynitud de la Pobreza en los Afios Ochenta.' The basic methodology used in determining these poverty lines was to price a nutritionally balanced food basket for each country which met the nutritional recommendations of a joint FAO/WHO expert committee in 1981.b The choice of items to be included in this food basket incorporated prevailing local tastes, as determined by the most recent income-consumption survey for each country. In some countries, these income-expenditure surveys covered only the metropolitan region, while in other countries they covered non-metropolitan and rural regions as well. When data regarding urban and rural variations in intra-country consumption and pricing patterns were not available, the cost per thousand calories of the minimum nutritional basket in urban areas was estimated at 95 percent of that for the corresponding metropolitan area, and the cost per thousand calories in rural areas was estimated at 75 percent of that for the corresponding metropolitan area. In addition, the calorie contents of country and regional baskets were adjusted by CEPAL to account for both inter- and intra-country differences in nutritional requirements. Finally, the poverty lines for each country were assessed as two times the cost of a basic food basket for metropolitan/urban areas, and 1.75 times the cost of a basic food basket in rural areas. These poverty lines were then indexed to the second half of 1988 in the final CEPAL report. The poverty line for Chile was presented in a separate report, but followed the same methodology.0 While the CEPAL poverty lines reflect a fairly uniform nutritional level (adjusted for local needs), they do not reflect a uniform level of purchasing gower. By incorporating differences in local tastes, the CEPAL poverty lines allow inter-country variations in welfare. For example, a small amount of meat may be viewed as a luxury for the poor in one country, but as a necessity in another country. This correlation between mean income and poverty lines which are based on local consumaption patterns is somewhat logical given that the CEPAL poverty lines are not intended to reflect the least-cost option of satisfying nutritional standards, but rather to incorporate the tastes of a sub-group of the national population. In defining this sub-group, CEPAL hoped to capture the consumption patterns of people who are at or slightly above the limit of maintaining nutritional adequacy in each country. Choosing such a subgroup was argued to give a realistic picture of how the poor would structure a diet meeting the FAO/WHO standards. While the exact bounds of this reference group varied from country to country, CEPAL generally chose the consumption pattems of those in the 15-40 percent range of the national income distribution as representative of local food tastes. a CEPAL, Canasta Bdsica de Alimentos v Determinacidn de las Lineas de Indigencia v de Pobreza: Argentina. Brasil. Colombia. Costa Rica. Guatemala. Mexico. Panama. Peru. UruRuav. Venezuela. Unpublished mimeos. Santiago de Chile, 1988/89. b FAO/WHO/UNU (United Nations University); Ener2 and Protein Requirements. WHO Technical Report Series No. 552, Report No. 52. Report of a Joint FA0/WHO Ad Hoc Expert Committee, FAO Nutritional Meetings. Geneva, 1985. CEPAL, Una Estimacidn de la Maznitud de la Pobreza en Chile. 1987. (LC/L.599) Santiago de Chile: CEPAL, 1990. Absolute Poverty 57 In the past, an excessive amount of time and energy has been channelled into developing Iscientifically" derived poverty lines which purport to embody the minimum income necessary to meet the basic human needs for nutrition, housing, clothing, education and health. But ultimately, any poverty cutoff will reflect some degree of arbitrariness due to the subjectivity of how poverty is defined. More importantly, the poverty comparisons presented in this report require that the cut-off point which distinguishes the poor from the non-poor must represent a uniform welfare level in all countries. In other words, the monetary value chosen as the poverty "reference" - or poverty line - should have equal purchasing power across countries. While country-specific poverty lines are more appropriate for single country analyses, a regional analysis must balance the conditions of both the poorest and the richest states when determining the poverty line. What may be a suitable poverty line for a relatively wealthy country such as Argentina is inappropriate for assessing poverty in Guatemala. A uniform regional poverty line needs to focus on the truly poor, and leave considerations of relative poverty aside. In determining such a poverty line, the approach here has deliberately avoided trying to reformulate a functional standard of basic human needs. Rather, the focus has been to determine a single value which embodies a welfare level that can be uniformly applied to all countries in order to assess poverty levels both between and within countries. Therefore this study follows an approach similar to that in the World Development Report. 1990. However the WDR was oriented towards a global analysis, while the present study is concerned only with Latin America and the Caribbean.39 Given the higher level of per capita income in the LAC region relative to the rest of the developing world, it is not surprising that the poverty line developed here is somewhat higher than the one in the WDR The following describes the methodology by which $60 per month in 1985 purchasing power parity (PPP) dollars has been determined as the poverty reference for this study: 1) Original country-specific, nutrition-based poverty lines were obtained from CEPAL for eleven Latin American countries. (Annex 10) 39 For a detailed explanation of how the World Development Report. 1990 povert line was determined, see Martin Ravallion et. al., Quantifying the Magnitude and Severity of Poverty in the Developing World in the Mid-1980s. PRE Working Paper No. WPS 587. World Bank, Washington, D.C., 1991. 58 Povly and Income Distribution in Latin America: The Story of the 1980s 2) Each of these poverty lines was then inflated or deflated to 1985, and converted to U.S. dollars based on purchasing power parity exchange rates.40 A comparison of the CEPAL country-specific poverty lines in 1985 PPP dollars demonstrates that the purchasing power of these poverty lines varied widely across countries, from $67 per person per month in Peru to $146 per person per month in Colombia. This clearly shows that a poverty analysis based on these poverty lines would not be comparable across countries. (Annex I1) 3) A regression of poverty line against mean income was run for the CEPAL poverty lines from ten countries;41 this regression verified a distinct correlation between the PPP value of the poverty line and mean per capita income in each country. Country-specific poverty lines were then estimated using this regression for the seven countries in this study for which nutrition-based CEPAL poverty lines were not available. These estimates spanned a range of $57 per person per month in Bolivia to $71 per person per month in Jamaica. From these results, a uniform $60 per person per month in 1985 PPP dollars was chosen as the national poverty line for the entire Latin America and Caribbean region. (Annex 11) 4) In order to assess levels of extreme poverty, an additional extreme poverty line was chosen at $30 per person per month in 1985 PPP dollars. 5) Finally, these two poverty lines were inflated and deflated to correspond to the time frame of each household survey used in this report.42 (Annex 12) In the following sections of this chapter, a consistent methodology has been applied which uses these uniform poverty lines to assess the magnitude and characteristics of the population 0 The poverty lines were inflated or deflated using the consumer price index listed in the International Financial Statistics, published by the International Monetary Fund. The PPP conversion factors for the exchange rate for each country have been taken from Robert Summers and Alan Heston, A New Set of International Comparisons of Real Product and Prices: Estimates for 130 Countries, 1950-1985. Review of Income and Wealth 34, no. 1(March 1988): 1-26. 41 These ten countries are listed in Annex 11. An eleventh country, Colombia, was omitted from the regression because it proved to have an inexplicably high PPP poverty line relative to income per capita. Though Colombia's position did not contradict the existence of a correlation between the value of the poverty line and mean per capita income, it did impact an otherwise strong relationship between the two. The regression is shown in the notes to Annex 11. 42 The poverty lines were inflated or deflated to the month prior to the actual survey period, because survey participants tend to answer questions about income by referring to the previous month. For example, the poverty lines for the October 1980 survey in Argentina were calculated to correspond to September 1980. Particularly in high inflation scenarios, the choice of reference month has a strong inpact upon poverty measures. Absolute Poverty 59 whose per capita income is below $60 and $30 per month in 1985 PPP dollars. These poverty lines are not intended as a definitive cut-off for determining poverty, and different poverty standards may be recognzed by individual countries. However, for the purposes of cross-country analysis of poverty levels reflecting a uniform welfare level, this report categorizes as poor those indviduals whose per capita income is below $60 and $30 per month in 1985 PPP dollars. These groups are referred to as the "poor" and the "extreme poor", respectively. Poverty Measures There is a vast literature on the measurement of poverty, and individual measures highlight different aspects of poverty conditions.43 The simplest and most commonly used index is the headcount ratio, defined as the fraction of the population whose income is less than the poverty line. However the headcount ratio is insensitive to transfers of income within the poverty population. Furthermore, it will show an improvement if income transfers raise some individuals above the poverty, even if the transferred income was from individuals who are very poor. Therefore it is desirable to supplement the headcount ratio with measures that are sensitive to the depth and intensity of poverty, and that will register an increase whenever there is a transfer from someone who is poor to someone who is less poor. A class of indices have been developed by Foster, Greer and Thorbecke which satisfy these requirements. Known as the FGT class of measures, they are defined as44: Pa= = Y) (4.1) where n = number of individuals in the population q = number of individuals with an income less than the poverty line z = poverty line yi = income of individual i a = degree of poverty aversion 43 See in particular Anthony Atkinson, On the Measurement of Poverty. Econometrica. 55 (1987): 749-764; James Foster, On Economic Poverty: A Survey of Aggregate Measures. Advances in Econometrics. 3 (1984): 215-251; James Foster, J. Greer, and E. Thorbecke, A Class of Decomposable Poverty Measures. Econometrica. 52 (1984): 761-765; Ravi Kanbur, Measurement and Alleviation of Poverty. IMF Staff Papers. 36 (1987): 60-85; Martin Ravallion, Poverty Comparisons: A Guide to Concepts and Methods. LSMS Working Paper No. 88, World Bank, Washington, D.C.; and Amartya Sen, Poverty: An Ordinal Approach to Measurement. Econometrica. 46 (1976): 437-446. 44 For a detailed discussion of the properties and usefulness of the FGT class of measures, see Martin Ravallion, Poverty Comparisons: A Guide to Concepts and Methods. LSMS Working Paper No. 88. World Bank, Washington, D.C., 1992. 60 Poverty and Income Distribution in Latin America: The Story of the 1980s In words, the FGT index is the summation of the percentage gap between the income of each member of the poverty population and the poverty line, raised to a power which reflects the degree of poverty aversion chosen by the researcher. If a is set equal to zero, implying no interest in the intensity of poverty, the index becomes the headcount ratio. That is, Po = q (4.2) n If a is set equal to one, the index is the poverty gap. n , z Equation 4.3 can be rewritten in a more interpretable fashion as: P [zy (4.4) n z whereby the poverty gap is clearly seen as the average shortfall of a poor individual's income from the poverty line, multiplied by the headcount ratio. PI is an improvement on Po since it is sensitive to changes in the depth of poverty; it is dependent not only on how many individuals are below the poverty line, but also how far they are below the poverty line. However PI is still insensitive to transfers within the poverty population that do not alter the average income of the poor. If a is set equal to 2 or higher, then the Pa measure will be distributionally sensitive whereby each individual is weighted by their degree of poverty. In order to give a picture of the level, depth and intensity of poverty, the above three poverty indices will be presented for each country. Table 4.1 presents the poverty headcount indices for each country using the uniform poverty line of $60 per month in 1985 purchasing power parity U.S. dollars established above, while Table 4.3 and Table 4.4 present the poverty gap and FGT P2 index, respectively. Each table also presents poverty measures based on an extreme poverty line of $30 per month. Absolute Poet 61 Although determined by the regression analysis described earlier, the choice of $60 as the poverty cut-off is essentially an arbitrary one. More important than the exact value of welfare chosen is that welfare be kept constant across time and countries if inter-country analysis is to have any meaning. By choosing a uniform poverty line for all countries, effective cross-country comparisons can be made which are based on a single welfare level. While the results might seem low for some of the wealthier countries, this approach highlights those countries whose absolute deprivation is more severe. Headcount Index Examining the results in Table 4.1, the poverty headcount was above 50 percent of the population at the end of the decade for Guatemala, and the urban regions of Bolivia and Honduras. Costa Rica and Uruguay (urban) had the lowest poverty rates at 3.4 percent and 5.3 percent of the population, respectively.45 The headcount indices of extreme poverty give a similar picture. Again Bolivia (urban), Guatemala and Honduras (urban) show headcounts of above 20 percent of the population living in extreme poverty. Argentina (Buenos Aires), Chile, Costa Rica, Jamaica, Paraguay (Asunci6n) and Uruguay (urban) had the lowest incidence of extreme poverty; at the end of the 1980s, less than two percent of the population in each of these countries had an income below $30 per month in 1985 PPP dollars. Examining changes in poverty over time, Argentina (Buenos Aires), Bolivia (urban), Brazil, Guatemala, Honduras (urban), Mexico, Panama, Peru (Lima) and Venezuela show an increase in the percentage of the population below the poverty line. Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban) show a decrease in poverty levels. Furthermore, these changes in poverty levels are found to be highly significant, and as is discussed in greater detail below. The patterns of change are identical when examining extreme poverty as well. Figure 4.1 gives a scatter-plotting of the headcount index for each country based on a uniform poverty line of $60 per month against per capita income in real terms. As with the Gini coefficient, a distinct negative correlation is discernible between the poverty headcount index and real per capita income. This trend is present both between countries, and within a country between two points in time. Argentina (Buenos Aires), Bolivia (urban), Colombia (urban), Costa Rica, Mexico, Panama, Paraguay (Asunci6n), Peru (Lima) and Venezuela all show this tendency. 45 The figures for Argentina, Bolivia, Colombia, Ecuador, El Salvador, Honduras, Paraguay, Peru and Uruguay are for urban areas only. 62 Poverty and Income Distribution in Latin Anerica: The Story of the 1980s Table 4.1: Percent of Individuals in Poverty and Extreme Poverty Poverty Extreme Poverty Headcount Index8 Headcount Indexb (% below $60 (% below $30 poverty line) poverty line) Country Survey Year (%l980)c (z1989)d (#1980)C (;l989)d Argentina (Bs. As.) 1980 1989 3.0 6.4 0.2 1.6 Bolivia (Urban) 1986 1989 51.1 54.0 22.5 23.2 Brazil 1979 1989 34.1 40.9 12.2 18.7 Chile .. 1989 .. 10.0 .. 1.5 Colombia (Urban) 1980 1989 13.0 8.0 6.0 2.9 CostaRica 1981 1989 13.4 3.4 5.4 1.1 Dominican Republic .. 1989 .. 24.1 .. 4.9 Ecuador (Urban) .. 1987 .. 24.2 .. 4.0 El Salvador (Urban) .. 1990 .. 41.5 .. 14.9 Guatemala 1986-7 1989 66.4 70.4 36.6 42.1 Honduras (Urban) 1986 1989 48.7 54.4 21.6 22.7 Jamaica .. 1989 .. 12.1 .. 1.1 Mexico 1984 1989 16.6 17.7 2.5 4.5 Panama 1979 1989 27.9 31.8 8.4 13.2 Paraguay (Asunci6n) 1983 1990 13.1 7.6 3.2 0.6 Peru (Lima) 1985-6 1990 31.1 40.5 f 3.3 10.1 f Uruguay (Urban) 1981 1989 6.2 5.3 1.1 0.7 Venezuela 1981 1989 4.0 12.9 0.7 3.1 Notes: All changes in poverty over timefor an individual country are significant to the I percent level or better. not availabk a 'Poverty' is defined as having an income of $60 per person per month or less. b Extreme poverty" is defined as having an income of $30 per person per month or kss. e or earliest d or latest 'Based on consumption data. f Estimate based on extrapolation from 1985-86 Peru povertyfigure, adjustedfor changes in poverty due to a fall in per capita income according to national accounts. This adjustment assumes an elasticity of poveny with respect to per capita income of -1.60, as determnined by regression in Annex 13. Absolute Povety 63 Figure 4.1: Poverty Headcount Trends Poverty Headcount (%) 70 GUAT-89 65 GUAT-87 * BRA-89 60 - BOL-89 HON-89 BOL-86 HON-86 45 - 40 - EL S-90 * PERU 90 ; BRA-89 35 - PAN-89 * BRA-79 30 - PERU-86 * 25 * * PAN-79 20- D.R.-89 ECU-87 MEX-89 15 P 8 \COL-8 * VEN-89 PAR-83 0 ~~CR-81 l* 10 _ CHI-89 ARG-89 U8 VEN-81 5PAR-90 ' COL-89 * 4-0O URU-81 VEl * CR-89* 0 I IIARGQ80 500 1000 1500 2000 2500 3000 3500 Per capita income in 1987 constant US$ 64 Poverty and Income Distribution in Latin America:: 7Te Story of the 1980s Assessing the significance of poverty differences over time: The poverty estimates presented in Table 4.1 are based on a binomial distribution, whereby individuals are assigned the value of one if they are poor and zero if they are not poor. The poverty rate (P) is then the expected probability of being poor. The standard error of the expected value (P) is given by: SE (P) - X(-P) (4.5) n which is inversely related to the sample size, n. Since the sample sizes for most of the household surveys used in this analysis are very large, it follows that the point estimates of poverty rates are highly significant. Significance tests were performed to determine the robustness of changes in poverty over time. In all thirteen cases for which data were analyzed for two points during the decade, the changes in poverty were found to be highly significant.46 The test statistic is given by: .= (4.6) SE (pi~ where P1 and Pj are the poverty estimates at points i and j in time, and SE (Pi - Pj) is the standard error of the poverty difference. This standard error is asymptotically distributed as standard normal, and is given by: SE (Pi j = - j P ' = Ji (417) n, ni where ni and nj are the respective sample sizes. 46 The statistical tests applied in this analysis are developed in Nanak Kakwani, Testing for Significance of Poverty Differences: With an Application to CMte d'Ivoire. LSMS Working Paper No. 62. World Bank, Washington, D.C., 1990. Absolute Poverty 65 The large sample sizes for most countries has resulted in small standard errors for (Pi - Pj), and very high Z values. As can be seen in Table 4.2, the Z-statistic exceeds the critical absolute value of 2.58 for significance at the one percent level or better in each of the thirteen country cases. Table 4.2: Statistical Significance of Changes in Poverty Headcount Rates Test Statistic Country (;t1980)- (# 989)' (Z) Argentina (Bs. As.) 3.0 6.4 13.157 Bolivia (Urban) 51.1 54.0 5.538 Brazil 34.1 40.9 58.261 Colombia (Urban) 13.0 8.0 - 24.280 Costa Rica 13.4 3.4 - 37.529 Guatemala 66.4 70.4 12.983 Honduras (Urban) 48.7 54.4 13.129 Mexico 16.6 17.7 3.894 Panama 27.9 31.8 8.922 Paraguay (Asunci6n) 13.1 7.6 - 9.088 Peru (Lima) 31.1 40.5 11.416 Uruguay (Urban) 6.2 5.3 - 4.904 Venezuela 4.0 12.9 105.083 Notes: See Annex Ifor the exact year of each data set. aPoverty is defined as having an income of $60 per person per month or less. or earliest or latest Therefore the changes observed in poverty levels in this report are highly significant. This is true even in those country cases where the change over time was small. 66 Poverty and Income Distribution in Latin Amrnica: The Story of the 1980s Poverty Gap Table 4.3 shows the FGT poverty gap measure for the $60 poverty and $30 extreme poverty lines. Not surprisingly, the countries with the highest headcount indices also had the largest poverty gaps; at the end of the decade, Bolivia (urban), Guatemala and Honduras (urban) all had poverty gaps in excess of twenty percent. Costa Rica, Paraguay (Asunci6n) and Uruguay (urban) were at the other end of the spectrum with gaps of less than two percent each. (See Table 4.3.) Bolivia (urban), Guatemala and Honduras (urban) also ranked worst with respect to their extreme poverty gaps. In contrast, Argentina (Buenos Aires), Chile, Costa Rica, Jamaica, Paraguay (Asunci6n) and Uruguay (urban) all had extreme poverty gaps equal or less than 0.5 percent. Examining changes in the poverty gap, Argentina (Buenos Aires), Bolivia (urban), Brazil, Guatemala, Honduras (urban), Mexico, Panama, Peru (Lima) and Venezuela show increases during the decade. Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban) demonstrate decreases. These patterns are identical for the extreme poverty gap. Moreover, the direction of change is the same in each country case for the poverty gap as it is for the headcount index. Absolute Poverty 67 Table 4.3: Aggregate Poverty Gap Poverty Gap' Extreme Poverty Gaph (Aggregate, Local Currency) (Aggregate, Local Currency) Country -- (ml1980)' (%w1989)' (- 1980)' (ftl989)d Argentina (Bs. As.) 0.6 2.1 0.0 0.5 Bolivia (Urban) 22.8 24.4 7.6 9.3 Brazil 13.7 18.8 3.9 7.1 Chile 2.8 .. 0.4 Colombia (Urban) 6.2 3.3 3.1 1.3 Costa Rica 5.9 1.3 2.2 0.4 Dominican Republic .. 8.0 .. 1.3 Ecuador (Urban) .. 6.9 1.1 El Salvador (Urban) .. 16.9 5.6 Guatemala 34.6 39.3 16.0 20.6 Honduras (Urban) 22.3 24.2 8.3 8.3 Jamaica! .. 3.2 .. 0.2 Mexico 4.9 5.8 0.6 1.3 Panama 10.5 14.3 2.8 6.1 Paraguay (Asunci6n) 3.8 1.8 0.9 0.2 Peru (Lima)' 8.6 13.3 0.7 2.6 Uruguay (Urban) 1.9 1.4 0.3 0.2 Venezuela 1.1 4.2 0.2 1.1 Notes: .. not available See Annex I for the exact year of each data set. a TPoverty' is defined as having an income of $60 per person per month or less. b 'Extreme poverty'M is defined as having an income of $30 per person per month or less. e or earliest d or latest e Based on consumption data. 68 Poverty and Income Distribution in Latin America: The Story of the 1980s FGT Measure The FGT P2 index gives greater weight to those further below the poverty line, and is presented for each country in Table 4.4. As with the two previous indices, Bolivia (urban), Guatemala and Honduras (urban) rank lowest, highlighting that the severity of poverty in these three countries is the worst for all countries included in this study. Brazil also ranks poorly with respect to the P2 index. Again, Costa Rica and Uruguay (urban) demonstrate the best performances. Rankings of the FGT P2 index for extreme poverty also indicate the dismal situation in Bolivia (urban), Guatemala and Honduras (urban). However several countries show low outcomes with respect to the P2 index for extreme poverty. In addition to Argentina (Buenos Aires), Chile, Costa Rica, Jamaica, Paraguay (Asunci6n) and Uruguay (urban), countries demonstrating relatively low levels of the severity of extreme poverty are the Dominican Republic, Ecuador (urban) and Venezuela. Finally, changes over time show exactly the same patterns for the P2 index as for the headcount index and the poverty gap. Colombia (urban), Costa Rica, Paraguay (Asunci6n) and Uruguay (urban) improved, while Argentina (Buenos Aires), Bolivia (urban), Brazil, Guatemala, Honduras (urban), Mexico, Panama, Peru (Lima) and Venezuela got worse. The same was true with respect to the P2 index for extreme poverty. Absolute Poverty 69 Table 4.4: FGT P2 Index FGT P2 Index FGT P2 Index Povertya Extreme Povertyb Country (;1980)c (zl98 (;1980)c (#1989) Argentina (Bs. As.) 0.2 1.0 0.0 0.3 Bolivia (Urban) 13.1 14.6 3.6 5.4 Brazil 7.4 11.2 1.8 3.8 Chile .. 1.2 .. 0.2 Colombia (Urban) 4.1 1.9 2.1 0.8 Costa Rica 3.5 0.7 1.2 0.2 Dominican Republic .. 3.5 .. 0.4 Ecuador (Urban) .. 3.0 .. 0.5 El Salvador (Urban) .. 9.5 .. 3.1 Guatemala 22.4 26.9 9.3 13.1 Honduras (Urban) 13.2 14.0 4.4 4.4 Jamaica .. 1.2 .. 0.0 Mexico 2.0 2.8 0.2 0.6 Panama 5.4 8.8 1.3 3.7 Paraguay (Asunci6n) 1.8 0.7 0.4 0.1 Peru (Lima)' 3.3 6.2 0.2 1.0 Uruguay (Urban) 0.8 0.6 0.2 0.1 Venezuela 0.5 2.1 0.1 0.5 Notes: .. not availabk See Annex 1 for the exact year of each data set. a TPoverty" is defined as having an income of $60 per person per month or less. b TExtreme poverty n is defined as having an income of $30 per person per month or less. 'or earliest or latest 'Based on consumption data. 70 Poverty and Income Distribution in Latin America: The Story of the 1980s Distribution of Poverty in the LAC Region: A Broader Picture The household surveys which serve as the foundation of this report cover over three-quarters of the population living in the Latin America and Caribbean region. While this represents the bulk of the population, it leaves out a significant number of countries in the region for which data are not available. To overcome this lack of data, it has been necessary to estimate poverty levels for the remaining LAC countries in order to assess the levels and changes in poverty for the region as a whole during the decade. Because poverty levels tend to differ substantially across urban-rural regions, separate urban and rural estimates are given for each country. Estimated poverty headcount indices for the entire Latin America and Caribbean region for 1980 and 1989 are presented in Table 4.5. The 1980 estimate is based on surveys which covered 52 percent of the total LAC population, while the 1989 estimate is based on surveys which covered over 79 percent of the LAC population. Poverty levels in the remaining areas are estimated using the regression models given in the footnotes to Table A13.1 and Table A13.2 in Annex 13.47 The poverty estimates presented in Table A13.1 and Table A13.2 in Annex 13 were determined by one of three ways. First, those estimates based on hard data have no asterisk highlighting them. Second, because certain country data sets do not correspond to a year near 1980 or 1989, the mid-decade poverty estimate based on hard data is adjusted to reflect changes in poverty between the survey date and either 1980 or 1989 since it would be unreasonable to assume that mid-decade poverty levels are identical to 1980 estimates. This adjustment assumes an elasticity of poverty with respect to per capita income of -1.6 as determined by regression in below Table A13.2 in Annex 13; changes in per capita income during the relevant time period were determined from national account data. These cases are marked by a double asterisk, and include Bolivia (1986), Guatemala (1986-87), Honduras (1986), Mexico (1984), Paraguay (1983) and Peru (1985-86).48 Third, the poverty estimates for all remaining countries were determined through regression analysis as presented in Annex 13. Countries whose poverty levels were determined by this method are highlighted by a single asterisk. 47 Each regression included an urban dummy variable and a Brazil dummy variable. The urban dummy variable allowed for separate urban and rural poverty estimates. The Brazil dummy variable was included because Brazil is a significant outlier in terms of both size of population and poverty level relative to mean income; each of these factors is well above the norm for Latin America and would have unduly influenced the outcome of the regression without a dummy variable to correct for them. 48 The poverty estimates in these six cases in Annex 13 are for 1980, not the year of the survey. However, the poverty estimates for Bolivia (urban), Guatemala, Honduras (urban), Mexico, Paraguay (urban) and Peru (Lima) in Table 4.1 are for the year of the survey. Absolute Poverty 71 The poverty estimates which are determined by regression analysis are substantially less reliable than those based on household surveys. For this reason, individual country regression poverty estimates are presented only in Annex 13, and are highlighted by an asterisk. Only country poverty estimates based on hard data are included in the main text of this report. However, aggregate poverty figures for the entire Latin America and Caribbean region are presented in this chapter, and these poverty headcount indices incorporate both actual and regression estimates for individual countries. Table 4.5: Changes in Rural and Urban Poverty, 1980-1989 Headcount Index Year LAC Total Population Population in Poverty (% points) Region 1980 All 345,400,000 91,400,000 26.5 Urban 227,400,000 38,200,000 16.8 Rural 118,000,000 53,200,000 45.1 1989 All 421,400,000 130,900,000 31.0 Urban 300,100,000 66,000,000 22.0 Rural 121,300,000 64,800,000 53.4 Change- All 76,000,000 39,500,000 +17.0 1980-89 Urban 72,700,000 27,800,000 +31.0 Rural 3,300,000 11,600,000 +18.4 Source: Annex 13 - Tables A13. 1 and A13.2 Latest Data: 1989 Table 4.5 presents estimates of the size and distribution of the total population in Latin America whose income was less than the $60 per month poverty reference in 1980 and 1989.49 Overall, 31.0 percent of the population is estimated to have been living on less than $60 per month in 1989. In absolute terms, 131 million people in Latin America had a per capita income which was less than the poverty reference defined in this study. Table A13.1 in Annex 13 presents the disaggregated country estimates of poverty for 1989. 49 As explained earlier, this $60 poverty reference is in constant 1985 purchasing power parity U.S. dollars. 72 Poverty and Income Distribution in Latin America: The Story of the 1980s However, the prevailing economic circumstances of each country should be taken into consideration when evaluating the data and pattems presented in Table A13.1. In 1989, poverty was probably near its highest level for the entire decade, as many countries experienced poor economic performance. Argentina, Panama, Peru and Venezuela all had sharp declines in per capita income during 1989. Per capita income was stagnant in Brazil and Mexico, and both countries were impacted by high inflation rates. Some of the above countries have subsequently experienced strong economic growth, which has been accompanied by declining poverty levels. Keeping this context in mind, several conclusions can be drawn from Table 4.5. First, the percentage of people living in poverty in 1989 was more than double in rural regions (53.4 percent) than in urban regions (22.0 percent). But because Latin America has become increasingly urban, the absolute number of people in poverty in 1989 was higher in the cities (66.0 million) than in the countryside (64.8 million). Second, poverty has not been homogeneous across either countries or urban/rural regions. In 1989, over forty-five percent of the poor in Latin America lived in just one country - Brazil - even though that country had only one-third of the region's population. This reflects the extreme inequality which has historically characterized the Brazilian income distribution. As was seen in Chapter 2, at the end of the decade Brazil had the greatest degree of income inequality of the eighteen countries analyzed in this study. Thus, even though the 1989 picture is affected by the recession in Brazil during that year, the Brazilian poverty level was extremely high given the average per capita income for the country.50 Both the Brazilian rural and urban headcount ratios were more than double those of Mexico, even though the former has a higher per capita income level than the latter. These deviations are significant. If the Brazilian poverty rates were reduced to those of Mexico, the overall poverty headcount index for Latin America and the Caribbean would be cut by more than one-fourth; this is the equivalent of raising thirty-eight million people out of poverty. Several other countries contained a disproportionately large share of the poverty in the region as compared to their population share. Over nine percent of the poor were concentrated in Peru. An additional nineteen percent of the poor lived in a group of small, relatively impoverished countries which, while not a part of the debt crisis per se, had falling or stagnant per capita income over the decade for a variety of reasons.5" Altogether the above sub-group of countries accounted for over seventy percent of total poverty in Latin America even though they contained only forty-eight percent of the population. Furthermore, none of the countries has resumed a stable growth trajectory since 5o Note that in the headcount regression in Annex 13, a Brazil dummy was included and was found to be positive and highly significant. The coefficient on the dummy was 36.3 implying that other things being equal, Brazil had a headcount ratio 36 percent higher than the average LAC country. 51 Included in this group are Bolivia, El Salvador, Guatemala, Haiti, Honduras and Nicaragua. Absolute Poverty 73 1989, so their situation is probably worse today than it was at the end of the eighties. As a result, the poverty problem is undoubtedly even larger and more concentrated at the present than it was in 1989. Third, the incidence of poverty differs widely across countries. The main cause of this is sharp differences in the per capita income between countries. Because the poverty reference is an exogenous standard, countries with relatively low per capita income levels are likely to have high poverty indices. Looking at the latest year, it is not surprising that Bolivia, El Salvador, Guatemala and Honduras had high poverty headcount ratios; their per capita income levels were below the regional average. However disparities in the distribution of income also account for variations in the poverty measures. Countries with relatively equal distributions have lower poverty headcounts than countries at the same per capita income level but with greater inequality. Brazil, as discussed above, is an excellent example of this. Given its favorable per capita income, the Brazilian poverty headcount is very high for the region. This is due to the extreme inequality which characterizes the country. At the other extreme, Costa Rica and Jamaica have low levels of poverty incidence, despite the fact that neither has a particularly high level of per capita income. Earlier Data: 1980 Examining the earlier data, 26.5 percent of the population is estimated to have been living on less than $60 per month in 1980. By this measure, 91 million individuals in Latin America were living on less than the poverty reference defined in this study. Disaggregated country estimates of poverty are presented in Table A13.2 in Annex 13. Because the survey coverage for 1980 is not as comprehensive as for 1989, the estimates for the earlier period are somewhat more tentative. Household data were available for nine countries for around 1980, representing forty-six percent of the total population in the Latin America and Caribbean region.52 Poverty in the remaining countries was estimated by a regression of headcount against per capita income for the initial nine countries. The regression is presented in Table A13.2 in Annex 13. As in 1989, the percentage of people living in poverty in 1980 is more than double in rural regions (45.1 percent) than in urban regions (16.8 percent). Unlike 1989, however, the 52 Data from Bolivia (1986), Guatemala (1986-87), Honduras (1986) and Paraguay (1983) were not used in constructing the 1980 regression, since they pertained to time periods which were too distant from 1980. The data upon which the regression is based are identified in Annex 9, and the regression itself is presented in Table A13.2 in Annex 13. In Table 4.4 and Table A13.2 (Annex 13), the poverty levels for Bolivia, Guatemala, Honduras and Paraguay are estimated using both actual data and this regression. The figures for the countries in Tables 4.1-4.3, however, are based on hard data only. 74 Povr and Income Distribution in Latin America: The Story of the 1980s majority of the poor in Latin America in 1980 still lived in rural areas. At the beginning of the decade, approximately 38 million urban inhabitants were poor, while almost 53 million rural dwellers were poor. Regional Trends Over the 1980s For most of the last decade, Latin America has struggled with the aftermath of the debt crisis. During this difficult period, average per capita income in the region fell by eleven percent, real wages declined substantially and there was a sharp increase in unemployment and/or underemployment. Given these conditions, it is not surprising that poverty levels for the region increased during the decade. While national studies have documented this for individual countries, their methodologies and poverty lines vary. This study represents a unique assessment of poverty in the LAC region which utilizes a uniform and comparable methodology across the entire continent for the entire decade.'3 As Table 4.5 demonstrates, the total population living below the $60 poverty reference in Latin America and the Caribbean increased by 43 percent, from 91 million in 1980 to 131 million in 1989. This corresponds to a jump in the regional headcount index from 26.5 percent to 31.0 percent of the population. The incidence of poverty rose much faster in the cities than in the countryside. The incidence of urban poverty grew by almost a third, from 16.8 percent to 22.0 percent, while the incidence of rural poverty increased from 45.1 percent to 53.4 percent. In absolute terms, the number of urban poor rose by 28 million while the number of rural poor grew by 12 million. The relatively better performance of the rural sector does not necessarily mean that conditions there fared better. Over 95 percent of the increase in total population during the 1980s occurred in the cities; this indicates that there was substantial internal migration from the countryside to the cities. In the large majority of country cases, the rise of rural poverty was substantially less than that of urban poverty. This may simply reflect the transfer of poverty from the countryside to the city, rather than any real improvement in living conditions in the rural 53 CEPAL, 1991, op. cit. analyzes changes in poverty levels between 1980 and 1986 for ten countries, and gives an overview of income distribution in 1986. Altimir develops a deeper analysis of the 1980-86 period and extends his inquiry to 1989 for several countries with available data. Morley analyzes the 1980-89 period for Colombia, while Morley and Alvarez do the same for Argentina, Costa Rica and Venezuela. See Oscar Altimir, Latin American Poverty in the Last Two Decades. 1992. Unpublished mimneo. See also Samuel Morley, Policy, Structure and the Reduction of Poverty in Colombia: 1980- 1989. Inter-American Development Bank (1992); Samuel Morley and Carola Alvarez, Recession and the Growth of Poverty in Argentina. (1991); Poverty and Adjustment in Costa Rica. (1992); Poverty and Adjustment in Venezuela. Inter-American Development Bank (1992). While each of these studies is useful, they do not permit a comprehensive overview of what happened over the continent as a whole. Absolute Poverty 75 sector. Heavy internal migration may indicate exactly the opposite; that is, rural conditions may have been bad enough to induce people to move to the cities despite rising poverty in the cities. However detailed data are needed on income differentials and movements in relative wages to adequately assess which of these two possibilities more accurately reflects conditions in the rural sector during the 1980s. Given an exogenous standard such as the poverty reference used in this study, there are two factors which can cause poverty to rise over time: drops in per capita income and changes in the distribution of income. For example, if the relative distribution stays constant but the level of per capita income drops, poverty will increase. Similarly, if average per capita income remains constant but inequality is worsened in such a way that more individuals earn less than the poverty reference, then obviously poverty will increase. Much effort has been put into decomposing changes in poverty over time into the effect due to each of these two factors.54 A counterfactual exercise can shed light on the separate effects of falling per capita income and increasing inequality on changes in poverty. By uniformly inflating (deflating) the earlier income distribution by the percentage increase (decrease) in per capita national income between two time periods, it is possible to calculate an hypothetical poverty headcount ratio based on the early distribution transfonned to reflect the later income levels. The degree to which the actual headcount ratio is higher (lower) than this hypothetical headcount ratio reflects the contribution of changes in the distribution of income to any increase (decrease) in poverty. Table 4.6 shows such a decomposition of the changes in poverty for each of the countries in this study which there are two survey periods. Column 1 and Column 3 repeat the actual headcount ratios from Table 4.1. Column 2 is the hypothetical poverty index in 1989 derived by inflating or deflating the 1980 distribution by the observed per capita national income change between the two survey dates. Column 4 is the difference between Column 2 and Column 1, and reflects the change in poverty due solely to the change in income. Column 5 is the difference between Column 3 and Column 2, and is the change in poverty due to shifts in the distribution of income. Examining Column 5, Argentina, Bolivia, Brazil, Guatemala, Honduras and Venezuela all show higher urban income inequality as adversely affecting poverty levels during the time periods assessed, while Colombia, Costa Rica, Mexico, Panama, Paraguay and Uruguay demonstrate positive effects on poverty levels due to lower urban income inequality. With respect to the rural sector, only Costa Rica and Panama show a beneficial impact on poverty due to improvements in the income distribution. Brazil, Guatemala, Mexico and Venezuela all show deteriorations in rural poverty due to a worsening of the income distribution. 54 Martin Ravallion and Guarav Datt, Growth and Redistribution Components of Changes in Povertv Measures: A Decomposition with Applications to Brazil and India in the 1980s. World Bank LSMS Working Paper No. 83. World Bank, Washington, D.C., 1991. 76 Pove and Income Distribution in Latin America: The Story of the 1980s Table 4.6: Actual and Predicted Poverty Levels in 1989 (1) (2) (3) (4) (5) Poverty Change due to: Po Po Po Change in Change in Country/ Actual Hypothetical Actual Income Distribution Region 1980 1989 1989 (2) - (1) (3) - (2) Argentina (Bs. As.) 3.0 5.4 6.4 2.4 1.0 Bolivia (urban) 51.1 53.8 54.0 2.7 0.2 Brazil (urban) 23.9 22.0 33.2 - 1.9 11.2 Brazil (rural) 55.0 52.0 63.1 - 3.0 11.1 Colombia (urban) 13.0 12.1 8.0 - 0.9 - 4.1 Costa Rica (urban) 9.9 9.1 3.5 - 0.8 - 5.6 Costa Rica (rural) 16.7 15.7 3.2 - 1.0 - 12.5 Guatemala (urban) 48.7 48.3 54.8 - 0.4 6.5 Guatemala (rural) 71.8 71.4 79.4 - 0.4 8.0 Honduras (urban) 48.7 47.5 54.4 - 1.2 6.9 Mexico (urban) 12.9 15.8 9.1 2.9 - 6.7 Mexico (rural) 27.0 30.1 31.6 3.1 1.5 Panama (urban) 26.0 31.7 25.9 5.7 - 5.8 Panama (rural) 33.0 39.5 36.8 6.5 - 2.7 Paraguay 13.1 14.6 7.6 1.5 - 7.0 (Asunci6n) 6.2 6.9 5.3 0.7 - 1.6 Uruguay (urban) 2.5 5.9 10.8 3.4 4.9 Venezuela (urban) 9.0 19.5 23.5 10.5 4.0 Venezuela (rural) Absolute Poverty 77 Individual Country Experiences Table A13.3 in Annex 13 presents a decomposition of the overall changes in poverty in Latin America and the Caribbean according to country and urban/rural region. The change in both absolute number of poor and the percentage share of the total change in poverty is given for each country. This portrait indicates that poverty in Latin America is fairly localized. A staggering 46.3 percent of the total increase in poverty for all of Latin America occurred in the cities of Brazil. An additional 14.5 percent of the total increase took place in Peru. The combination of these two countries with the group of small, stagnant countries identified earlier comprised 47 percent of the regional population in 1980, yet they accounted for 84 percent of the rise in poverty during the decade.55 Mexico contributed an additional 11 percent of the total poverty increase. However unlike the other countries in the group, subsequent economic growth in Mexico has undoubtedly induced a reduction of poverty since 1989. The regressions which were used to estimate the level of poverty for countries with missing observations serve another useful function. By comparing the actual level of poverty with that predicted by the regression, the relative poverty conditions of individual countries can be assessed. Countries with actual poverty levels below those predicted by the regression have a relatively more equal distribution of income between the poor and the non-poor than the regional norm. The income distribution is relatively less equal for countries where the actual poverty headcount is above the predicted level. Obviously this comparison can only be conducted for the subset of countries for which actual observations are available. Table A13.4 of Annex 13 displays both the actual and predicted headcount ratios for 1980 and 1989. A ratio greater than one indicates that the actual level of poverty was higher than would be predicted on the basis of its income per capita according to the respective regression. For example, the table shows that in 1980 the observed level of poverty in urban Bolivia was only 74 percent of what would be predicted by the 1980 regression for an urban region at that income level; in 1989, actual poverty for this same region was 94 percent of the level predicted by the 1989 regression. Examining individual country experiences in 1980 shows that there were three countries with poverty levels much higher than expected: Brazil, Panama and Peru (Lima). Brazil was included in this group even though its ratio of actual to expected was close to one because the expected value included the effect of the Brazil dummy. Without this dummy variable, the expected level of urban poverty would have been 10 percent, which is significantly less than the 24 percent which was observed. This highlights the degree to which Brazilian poverty, relative the country's per capita income, was well above the norm for Latin America. Guatemala, Honduras (urban), Mexico and Uruguay (urban) also had higher than expected poverty ratios in 1980, but these differences were not as great as for the first three countries. Argentina (Buenos 55 These countries are Bolivia, El Salvador, Guatemala, Haiti, Honduras and Nicaragua; they accounted for 22.9 percent of the increase in poverty in the region during the 1980s. 78 Powrty and Income Distribution in Latin America: The Story of the 1980s Aires), Colombia (urban), Costa Rica and Venezuela had much lower than expected rates of poverty given their per capita income levels. The list of countries with worse than expected performance grew between 1980 and 1989. In addition to Brazil, Guatemala, Honduras, Panama and Peru (Lima), the countries of Argentina (Buenos Aires), Mexico and Venezuela also experienced poverty levels which were higher than would be predicted based on mean income levels.56 The group of countries with better than expected performance continued to include Colombia (urban) and Costa Rica (particularly the rural sector of Costa Rica.) The urban sector in Uruguay also showed better than predicted poverty rates. The disaggregated poverty picture implied by Table A13.4 supports the evidence from earlier tables. Brazil started and finished the decade with higher than expected levels of poverty, as did Peru. These two countries alone accounted for 61 percent of the change in poverty in the region between 1980 and 1989, as reflected in Table A13.3. Argentina and Venezuela started the decade with low levels of poverty relative to their incomes, but ended the decade in the opposite situation. Both had very rapid increases in poverty which were due to sharp declines in per capita income and a regressive shift in income inequality. The picture is less clear for the smaller countries which accounted for most of the balance of poverty in 1989. Both Guatemala and Honduras started and finished the decade with higher than predicted poverty levels. By contrast Bolivia had lower than expected levels of poverty at both observation points, although poverty did increase slightly during the decade. Colombia and Costa Rica are the benign counter-examples to the countries discussed above. Both had less poverty than expected in 1980 and 1989, and an overall decline in poverty over the decade. 56 The inclusion of Mexico in this group is slightly misleading. The actual poverty index for Mexico rose between 1980 and 1989, yet the predicted level of urban poverty fell from 12.1 percent to 3.6 percent. (See Table A13.4) The dramatic shift in the predicted level occurred because the 1989 regression estimates lower levels of poverty for a given level of income than the 1980 regression, particularly in the income range in which Mexico falls. Thus the dramatic shift in actual/predicted performance is due primarily to a strong downward shift in the regression line, as is shown in Figure 4.2. The same explanation applies to the case of Panama. Absolute Poverty 79 Further Evidence on Poverty over the Decade The surveys examined to this point permit an assessment of poverty at the beginning and the end of the decade. But they have limited benefit for evaluating the evolution of poverty during the decade, particularly in relation to the economic cycles of each country. For example, the Brazilian observations are for 1979 and 1989. The former corresponds to a time period before the debt crisis, while the latter assesses conditions after an entire recession- recovery cycle which began in 1981 and peaked in 1986. In order to effectively examine the impact of recession and recovery on poverty in Brazil, observations would be needed for 1983 and 1986 since these match the trough and peak of the economic cycle, respectively. Similar mismatches between observation time-frame and economic performance occur for many of the countries examined earlier in this chapter. In order to develop a broader perspective of the relationship between poverty and general economic conditions in each country, Table 4.7 supplements the poverty estimates from this study with those from outside sources.57 Caution should be exercised when interpreting Table 4.7. Poverty estimates by different researchers can vary dramatically, even when the results are based on the same underlying data set. This occurs due to differential treatment of income underreporting and poverty line construction. In some cases, the elasticity of poverty indices with respect to the poverty line can be quite high. For example, the poverty reference line for this study is $60 per person per month in 1985 PPP dollars. If this line were raised to $72 per month, the 1979 urban poverty headcount in Brazil would rise from 24 percent to 31 percent. 57 The sources for Table 4.5 are as follows: All CEPAL references except for Chile are CEPAL, 1991, op. ci.; Argentina - INDEC, La Pobreza en Argentina, Buenos Aires, 1990; Morley and Alvarez, 1991, cit.; Brazil - M. Louise Fox and Samuel Morley, Who Paid the Bill? Adjustment and Poverty in Brazil, 1980-1985, PRE Working Paper No. WPS 648, World Bank, Washington, D.C.; Ricardo Moran, Income Distribution and Poverty, mimeo; Colombia - Juan Londolio, Income Distribution and StructuralAdjustment: Colombia, 1938-88, Ph.D. Dissertation, Harvard University, 1990; Morley, 1992, Qp. cit.; Costa Rica - T.H. Gindling and A. Berry, lhe Performance of the Labor Market During Recession and Adjustment: Costa Rica in the 1980s. World Development, (forthcoming); Morley and Alvarez, 1992, o .; Chile - PREALC, Pobreza y Empleo: un Analisis del Periodo 1969-87 en el Gran Santiago, #348, 1990; CEPAL, La Pobreza en Chile en 1990, mimeo; Uruguay - Oscar Altimir, Latin American Poverty in the Last 7Wo Decades, mimeo; Venezuela - Gustavo Marquez, Poverty and Social Policies in Venezuela, mimeo, IESA, 1991. 80 Poverty and Income Distribution in Latin America: The Story of the 1980s Table 4.7: Poverty Headcount Indices During the 1980s (percent) Country/Source 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Argentina This Stdy (metro) 3.0 6.4 Metro4lWDEC 10.1 28.0 20.6 25.2 27.9 Marley/Alv. (metre) 6.3 10.9 21.5 CEPAL (metro) 6.0 11 0 CEPAL (totat) 10.0 16.0 Brazil This Study (urban) 23.9 33.2 Fox/Morley 24.8 30.9 25.4 16.1 23.3 26.9 Moran 22.0 28.0 12.0 19.0 Colombia This Study (urban) 13.0 810 CEPAL 42.0 42.0 Londoflo 29.0 25.0 Morley (urman) 34.0 33.0 34,0 Costa Rica This Study (urban) 9.9 3.5 This Study (rural) 10.4 3.2 Gindling/Berry 48.0 62.0 78.0 69.0 58.0 63.0 52.0 45.0 Morley/Alvarez 25.4 26.8 10.2 CEPAL 24.0 27.0 Chile PreakWOmetro 36.Q 40.3 31.2 48.5 45A4 50.9 48.6 CEPAL (total) 38.0 35.0 CEPAL (urban) 37.0 34.0 Guatemala This Study (urban) 48.7 54.8 This Study (rural) 71.8 79.4 CEPAL (urban) 47.0 60.0 CEPAL (rural) 84.0 80.0 Mexico CEPAL (total) (1977) 40.0 37.0 Panama This Study (urban) 26.0 25.9 This Study (rural) 33.0 36.8 CEPAL (urban) 36.0 36.0 CEPAL (rural) 50.0 52.0 Peru This Study (tnetro) 31.1 CEPAL (total) 53.0 60.0 Uruguay This Study (urban) 6.2 5.3 Altimir (urban/house) 9.0 14.0 10.0 Altimir (rural/house) 21.0 23.0 23.0 Venezuela This Study (urban) 2.5 10.8 This Study (rural) 9.0 23.5 CEPAL (urban) 22. 33,0 CEPAL (rural) 43.0 42.0 CEPAL (total) 25.0 32.0 Morley/Alvarz (total) 24.0 29.0 48.2 Maquez (total) _7 28. 31.8 41,3 34,6 Absolute Poveny 81 However, intertemporal poverty estimates which utilize a constant methodology and a constant poverty line over time are comparable. This is the approach taken in this study. Poverty estimates for each country are comparable over time and, with the caveats discussed in Chapter 1, across countries as well. Each line in Table 4.7 provides an internally consistent set of poverty estimates for a given country based on a uniform methodology and poverty line. For the reasons just described, comparisons can not be made between lines, except in a very rough sense. But taking the different series together can give a fairly consistent and robust idea of what happened to poverty in individual countries over various sub-periods of the decade. In those countries for which there are observations at the appropriate time periods, the evidence strongly supports the hypothesis that poverty is countercyclical to macroeconomic conditions. It tends to rise rather sharply during recession and to fall, though usually less sharply, during recovery. Three country cases demonstrate this. It was already mentioned that Brazil experienced a severe recession from 1981 to 1983 followed by recovery from 1983 to 1986. The poverty series labelled Moran in Table 4.7 closely corresponds to this cycle: it rises from 1981 to 1983 during the recession, then falls between 1983 and 1986 as the economy recovered, and finally rises to a new peak in 1989 as inflation and recession reappeared. A similar pattern can be seen in the Gindling-Berry series for Costa Rica.59 In that country, the recessionary bottom and the poverty peak both occurred in 1982. Subsequently the economy recovered, and the poverty indices declined. A particularly noteworthy facet concerning Costa Rica is that by 1987 the absolute poverty level appears to have fallen below its preadjustment 1979 level, despite the fact that per capita income was lower throughout the decade than it was in 1980. This indicates that, in at least this case, the conditions affecting the poor may have significantly improved in the long run after the 1981-82 adjustment process. The picture is similar in Uruguay, where a recession during the early 1980s bottomed out in 1985. The Altimir observations do not match this time frame exactly, but they are close enough to suggest that this recession caused an increase in poverty. Subsequently the economy recovered between 1986 and 1989, with real per capita income growing by a total of 10 percent during that time span. According to Altimir, urban poverty fell 30 percent during the same three year period.60 58 For example, some of the household surveys covered only urban areas, while others were national in coverage. 59 T.H. Gindling and A. Berry, The Performance of the Labor Market During Recession and Structural Adjustment: Costa Rica in the 1980s. World Development, (forthcoming). 60 Altimir, 1992, op cLit. 82 Poverty and Income Distribution in Latin America: The Story of the 1980s In an effort to further investigate the connection between poverty levels and general economic performance, an exercise was conducted whereby all the available information on urban and rural poverty levels for individual Latin American and Caribbean countries were compiled. (Most of this is presented in Table 4.7.) Recession periods were then determined as any time period in which there were at least two years of falling per capita income. All other time periods of two years or longer were classified as recoveries. This exercise resulted in 58 observations of recession periods during the decade, and 32 recovery periods. In 55 out of the 58 recessions cases, poverty increased. Of the 32 recovery periods, poverty fell in 22 cases, was essentially constant in 3 cases, and increased in the remaining seven cases.61 Thus the evidence supports the not very surprising proposition that poverty is sensitive to the level of income.62 61 The seven cases were the metropolitan areas of Colombia 1980-86, Argentina, 1985-88, Chile 1984-86, the urban sector of Colombia 1986-89, and all three observations of Costa Rica by CEPAL between 1981 and 1988. It should noted that these CEPAL estimates are outliers. All estimates by other sources for the same period show poverty to be declining in Costa Rica. 6 In a recent paper, Morley tried to say more than this. He estimated a regression model using cross-sectional observations from the 1980s in which changes in the headcount ratio were the dependent variable, and changes in per capita income (along with several other variables) were used as explanatory variables. The regression was run on recovery and recession observations separately in order to get an estimate of the elasticity of poverty with respect to income. It was found that the change in income was a highly significant determinant of changes in poverty during recession but not during recovery. In the recession regressions, the income elasticity of poverty was found to be slightly larger than -2, indicating that poverty rises by two percent for every one percent decline in income. This is a large effect, and highlights that recession, regardless of its cause, has a heavy impact on the poor. Somewhat unexpectedly, Morley found no such relationship between changes in poverty and changes in income during the recovery phase. He argued, however, that this was due to the nature of the available observations as opposed to a total lack of improvement in poverty during recovery. Most of the recovery observations came from Brazil, Colombia and Costa Rica; of these three, only Brazil had strong income growth during the recovery phase, and this effect was captured by a dummy variable. In the other cases, the subsequent growth rate tended to be low and relatively invariant. As a result, only a small portion of the variation in poverty reduction across countries was attributed to variations in national growth rates by the regressions. See Samuel A. Morley, Structural Adjustment and the Determinants of Poverty in Latin America. Paper presented at Brookings Institution and Inter-American Dialogue Conference, "Poverty and Inequality in Latin America," 1992. 5 Social Indicators The previous chapters of this study examined economic inequality and poverty as defined by income. However, while income is the most common measure of economic well-being, there are other "quality of life" indicators which contribute to assessing the welfare of an individual. Social indicators such as life expectancy, access to health care, nutritional status and educational attainment all serve to complement income measures, clarifying the picture of individual well- being. Though movements in social indicators do tend to be correlated with movements in public social spending levels, this correlation is imperfect. During the 1980s, real per capita spending on health and education fell in absolute terms in many Latin American countries, yet social indicators continued a trend of steady improvement throughout the decade.63 This chapter will examine the available information on non-income measures of welfare, i.e. social indicators. This chapter will not attempt an in depth examination of this subject; rather it will provide a brief overview of the general state of social indicators, supplementing the poverty and income distribution figures in order to form a better picture of living standards within the Latin America and Caribbean region. The analysis will look at regional variations of infant mortality, child global malnutrition, access to maternal health care, immunization and education while the measures of poverty and inequality serve as a backdrop of income measures. The data are based on a variety of sources in addition to the household surveys described in Annex 1. The examination will highlight correlates, patterns and changes in these indicators at the national aggregate and sub-regional level, comparing characteristics across regional, educational, and income quintile grounds." The previously examined trends of inequality and poverty indicators during the last decade will be juxtaposed to trends in the various social indicators in an attempt to illicit observable correlations and patterns. The chapter is divided into three principal sections: health, education and demographic/employment indicators. The health section is based primarily on Westinghouse Demographic and Health Surveys (DHS) supplemented by World Bank and other documentation. The education section is based on World Bank and other sources, as well as the household survey 63 Margaret E. Grosh, Social Spending in Latin America: The Story of the 1980s. World Bank Discussion Papers No. 106. World Bank, Washington, D.C., 1990. 64 For most cases examined, correlates are mother's education, urban/rural division, and ethnicity. 84 Povery and Income Distribution in Latin America: The Story of the 1980s data described in Annex 1. The household surveys fulfill the need for timely micro-level data on the distribution of social indicators in the region - information that is otherwise surprisingly lacking. The surveys were taken in fourteen Latin Amencan countries: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Guatemala, Honduras, Mexico, Panama, Trinidad & Tobago, Uruguay and Venezuela. The data for Argentina, Bolivia, Colombia and Uruguay cover only the urban areas. Therefore, the survey analysis has been disaggregated by urban and rural area in order to maintain comparability as well as to examine interesting differences.65 The last section, examining demographic/employment indicators, is based solely on the household surveys and is analyzed by income quintiles. By examining the relationships of quintiles within urban/rural demographic categories and between quintiles of equivalent urban/rural composition, the role of income on each social indicator can be better understood. Annex 14 presents the quintile distribution of key social indicators, disaggregated by urban and rurl categories. Health Infant Mortality Infant mortality is one of the most sensitive indicators of general socioeconomic well-being. Infant mortality serves as a good indicator of socioeconomic development due to its correlation with other social variables such as sanitation, education, nutrition, matemal and infant health care. In the absence of more specific data, the infant mortality rate (IMR) is a strong reflection of general levels of well-being within a population.66 In addition to the preservation of life, it is commonly understood that with a decrease in infant mortality there is a concomitant decrease in fertility. 67 In an age pressured with explosive populations and the resulting difficult economic consequences, this link has provided additional impetus for the international and national attention given towards infant mortality reduction. In general, mortality as a whole has decreased significantly over the last several decades, but most dramatic has been the drop in infant mortality. The LAC region is no 65 Rural and urban poverty contain well documented income disparity. Rural areas tend to possess lower average incomes than urban areas largely due to crucial labor market differences between the two categories. Direct aggregate comparisons of samples without rural data to those with rural data would be inconsistent. 66 The Infant Mortality Rate (IMR) is expressed as the number of infant deaths (under 1 year of age) per one-thousand live births. 67 Julie DaVanzo and Jean-Pierre Habicht. Quantitative Studies of Mortality Decline in the Developing World. Staff Working Papers No. 683. World Bank, Washington, D.C., 1985. Social Indicators 85 Figure 5.1: Infant Mortality IBRD 28451 ano Ed 40 OCEA OCEA .~~~~~E S or rau Trinidad aDd ~~ No data shownota R Th bonaree coos deon,hn an .n ote snifo; 'Dornot TIATI shown on th.s soap do not swidlr on the part oE *he World ,Bok OCEAN El araweptone of ss .bo Tnidarana Souce DHS 1urey and Worl Bank Soia Indicator111 nn 1_ s. r Under 40 111111O111|1111 111111 II Q I / CEAC R996 ThEisl1 map 11 wasprduedbyth Map1 DesignUnt of The> _nodBak TH1 1 bo/ui coor, /nmmhos/ndonr thr nfm ob or accptnc ofsuhioudnta Source: Ecu doS/ Suvy_n ol akScaniaos Guot~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~EEMBER 199 86 Povenj and Income Distribution in Latin America: The Story of the 1980s exception where in Brazil, for example, infant mortality has decreased 46 percent since the late 1g4oS.68 According to the DHS surveys, infant mortality rates were worst in Bolivia, Brazil and Guatemala (see Figure 5.1 and Table 5.1).69 Colombia and Trinidad & Tobago had some of the best observed rates. Though the drop in aggregate figures is encouraging, the aggregate figures conceal stark intra-country regional variation. In 1987, Guatemala had an overall infant mortality rate of 79.1, yet IMR ranged from 48.0 in the more temperate north-central region to 119.3 in the tropical south-central region (see blowup in Figure 5.1). In 1992, Peru showed high inter-regional variability in infant mortality, with some areas (Inka, Mariatequi, Loreto) having an infant mortality rate more than 3 times that of Lima (103, 101, 98 vs. 30 in Lima, respectively). Though this disparity is not uniform across aU of Latin America, it is not uncommon (see Table 5.3), and the region's poor overaU average IMR of 54 warrants concern.70 Infant mortality rates tend to be strongly correlated with other social indicators. Understanding the interaction between IR and the social indicators influencing its variation is an important step towards developing more efficient targeting strategies. Four correlates of IM were examined: mother's education, domicile location, ethnicity and access to health care. With few exceptions, mother's educational level played the greatest role in IMR variation, though the other factors were important as well. 68 Westinghouse DHS Survey, op cit. 69 In Figure 5.1, Argentina, Chile, Costa Rica, Guyana, Haiti, Honduras, Nicaragua, Panama, Suriname, Uruguay and Venezuela infant mortality rates are based on the 1990 figures from Social Indicators of Development. Diskettes, World Bank, Washington, D.C., 1992. All other country infant mortality rates are based on the Demographic and Health Surveys (DHS). 70 United Nations, World Population Prospects. New York: United Nations, 1990. Social Indicators 87 Table 5.1: Intra-Country Variability of Health Indicators by Urban vs. Rural Immunizationsb Child Access to Infant Under 5 __ Global Maternal Survey Mortality Mortality BCG Measles Polio3 DPT3 Malnu- Care at Country Year Rate Rate tritionc Birthd Bolivia 1989 95.8 141.8 62.3 72.5 57.1 51.2 13.3 37.6 Urban 78.6 114.1 69.1 74.1 64.8 564 10.7 58.2 Rural 112.0 168.4 52.5 70.1 46.1 43.8 15.9 18.3 Brazil 1986 86.0 99.0 74.6 83.6 84.4 79.8 12.7e 80.5 Urban 76.0 88.0 81.6 87.5 89.8 86.4 9.8c 91.8 Rural 107.0 121.0 58.9 74.8 72.2 64.8 15.5° 57.8 Colombia 1990 26.9 34.9 93.3 81.2 822 81.0 i. 76.3 Utba 29.1 36.0 Rural 22.8 33.0 .. .. .. .. Colombia 1986 39.4 51.6 51.0 39.0 43.0 43.0 11.9 urban 38.5 47.8 52.0 41.0 44.0 44.0 10.2 Rural 40.9 57.8 50.0 35.0 41.0 41.0 14.7 Domnimcan Rep. 1986 67.7 , 56.6 18.6 28.8 37.4 12.5 $9.7 Urban 69.2 . 58.3 18.2 31.1 39.8 9.6 95.0 Rural 65.8 ,I. 526 19.4 26.1 31.7 16.6 82,8 Ecuador 1987 65.2 89.5 .. .. .. Urban 52.0 65.3 .. .. .. Rural 77.3 111.7 .. .. .. El Salvador 1988 50.0 66.0 77,3 %1.0 61.5 61.4 ., $2.5 Metpolitaxn 42.0 42.0 85,1 85.6 87.7 59.2 .. 87.9 Urban 40.0 63.0 84.9 81.9 65.6 64.7 .. 0.3 Rural 56.0 74,0 71.6 79.1 61.1 60.8 ., 33.7 El Salvador 1985 67.5 .. 71.1 72.7 62.8 64.4 Metropolitan 47.8 .. .. .. .. Urban 63.4 .. .. .. .. Rural 80.9 .. .. .. Guatbmu 1987 79.1 1204 51,0 68.7 52.1 47.0 33.5 Urba 65.3 98.6 43.6 74.6 61.9 60.2 257. Rual 84.5 129.6 46,6 66.7 48.7 42.5 M6.$ Mexico 1987 56.2 70.9 .. .. .. .. .. 63.3 pop. < 2,500 78.8 103.5 .. .. .. .. .. 67.1 2,500-19,999 61.6 77.7 .. .. .. .. .. 80.0 20,000+ 40.2 47.7 .. .. ., .. .. 80.0 Metropolitan 28.6 31.5 .. .. .. .. .. 80.0 - Continued 88 Poverty and Income Distribution in Latin Ainerica: The Story of the 1980s Table 5.1: Intra-Country Variability of Health Indicators by Urban vs. Rural (continued) Child Access to Infant Under 5 Immunizationsb Global Maternal Survey Mortality Mortality Malnu- Care at Country Year Ratea Rate tritionc Birtid BCG Measles Polio3 DPT3 Paraguay 1990 35.4 45.0 662 56.6 51.7 51.9 164 536 Urban 32.0 43.0 85.5 64.0 62,1 64.0 10.2 - Ral 32 38.2 50.8 50.6 43.4 42.1 21.5 Peru 1991/ 64.0 92.0 87.9 70.5 66.3 66.4 10.8 Urban 1992 47.0 67.0 92.8 75.6 75.1 75.3 6.4 Rural 90.0 131.0 80.6 62.8 53.2 53.1 17.8 Peru 1986 76.0 112.0 58.9 70,9 64,7 65.5 Urban 54.0 74.0 * Rural IL0.0 153. *- .. .. Trin. & Tob. 1987 31.1 34.4 .. 44.0 87.0 87.1 6.9 95.9 Urban 36.3 40.9 .. 51.9 58.6 85.4 5.0 96.2 Rural 27.5 29.9 .. 38.4 88.0 88.3 8.2 95.7 Source: Westinghouse Demographic and Health Surveys, with the exception of El Salvador (1988) which was conducted by the Salvadoran Demographic Association with the assistance of the U.S. Centerfor Disease Control. Notes: a Number of infants, per thousand live births, who die before reaching one year of age. b Percent of children immunized. For Bolivia, Brazil, Colombia (1990), Guatemala, Paraguay, Peru (1991/92),figures are for the 12-23 month age group. For Colombia (1986), Dominican Republic, El Salvador (1988), El Salvador (1985), Peru (1986), Trinidad and Tobago,figures arefor the under 5 years age group. c Refers to 0-59 month age group in most cases, although the age group variesfor afew countries. Child Global Malnutrition is expressed as a percentage of the child population (usually those under 5 years of age, though age groups vary in some cases) who are 2 or more standard deviationsfrom the average weightfor their age. d Percentage of births occurring in hospitals, clinics, or other professional medical establishments. e Brazilian malnutritionfigures refer to the Northeast only. Social Indicators 89 Table 5.2: Intra-Country Variability of Health Indicators by Mother's Educational Level Immunizations' Child Access to Infant Under 5 Global Maternal Survey Mortality Mortality BCG Measles Polio3 DPT3 Malnu- Care at Country Year Rate" Rate tritiono Birthd Bolivia 1989 95.8 141.8 62.3 12.5 57.1 51.2 13.3 37.6 None 123.6 181.3 36.7 66.6 38.4 36.8 231 7.6 Primay 108.1 162.1 56.6 61.1 45.4 40.4 13.1 265 secondary 64.8 100.0 64.4 757 65.8 55.7 11.3 58.5 Higher 46.4 60.S 78.6 88.6 76.0 69.2 6.6 82.9 Brazil 1986 86.0 99.0 74.6 83.6 84.4 79.8 12.7c 80.5 None 122.0 140.0 59.3 75.5 63.8 52.9 20.5 64.5 1-3 years 75.0 85.0 71.3 77.5 75.7 69.2 14.3 67.1 4 years 75.0 85.0 77.8 84.7 85.8 85.8 10.6 84.5 5+ years 38.0 43.0 79.4 89.1 94.6 90.8 5.4 95.8 Colombia 1990 26.9 34.9 03.3 81.2 82.2 81.0 .. 76.3 None 60.5 74A4 82.6 70,1 62.Z 62.2 53.2 Primary 27.2 36.5 94.8 85.0 78.9 78.6 .. 67.4 SIOCoolary 21.9 26.5 93.5 78.1 87.4 $5.0 $8.2 Higher 11.5 19.0 . .. .. .. ., 97.7 Colombia 1986 39.4 51.6 51.0 39.0 43.0 43.0 11.9 None 60.2 81.8 40.0 38.0 35.0 36.0 16.5 Primary 40.5 55.1 52.0 39.0 44.0 44.0 14.0 Secondary 28.4 30.3 53.0 40.0 44.0 44.0 8.0 Higher .. .. 34.0 28.0 28.0 28.0 0.0 Dominican Rep. 1986 61.7 .. 56.6 18.6 28.8 37.4 12.5 89.7 NOn 101.7 .. $3.9 10.1 13,4 13.4 17.8 73 Primary 76.0 .. 57.3 18.0 29.7 34.8 15.1 88.4 Secondary 6.5 . S1.3 16.2 29,6 37.7 9.8 97.6 university 34.4 .. 65.1 26.6 31.2 49.0 2.5 98.6 Ecuador 1987 65.2 89.5 .. .. .. None 105.5 159.7 .. .. .. Primary 68.1 92.0 .. .. .. Secondary 44.4 54.9 .. .. .. University 22.0 26.4 .. .. .. El Salvador 1988 50.0 66.0 77.3 81.0 61.5 61.4 52.5 NOn 68.0 880 ,. ,, ,. 1-3 years 61.0 71.0 ,. ,, L. 4-6 yews 45.0 66,0 ,. .. .. 7+ years 35.0 43.0 .. .. .. El Salvador 1985 67.5 .. 71.1 72.7 62.8 64.4 None 100.0 .. .. .. .. 1-3 years 70.6 .. .. .. .. 4-6 years 47.7 .. .. .. .. 7-9 years 48.6 .. .. .. .. 10+ years 32.1 .. .. .. .. - Continued 90 Poverty and Income Distribution in Latin America: The Story of the 1980s Table 5.2: Intra-Country Variability of Health Indicators by Mother's Educational Level (continued) Immunizations' Child Access to Infant Under 5 Global Maternal Survey Mortality Mortality Malnu- Care at Country Year Rate Rate trition0 Birthe BCG Measles Polio3 DPT3 GuteSma 197 79.1 120.8 51.0 68.7 521 47.0 33.5 wone 8145 136.2 39.8 64.6 45.1 41.7 42.o Pdmy $86.4 119.S 58,9 12.0 53.0 47.0 3010 Secondey 40.5 .. 64.6 81.3 70.8 60.4 20.3 Supetirtr ,. 43.3 60,0 60 62.9 60.0 1t0 7 . Mexico 1987 56.2 70.9 .. .. .. .. .. 63.3 None 82.7 113.7 .. .. .. .. .. 25.0 Inc. Primary 63.9 81.2 .. .. .. .. .. 52.7 Cornp. Primary 45.9 51.0 .. .. .. .. .. 75.9 Secondary + 27.4 28.6 .. .. .. .. .. 92.1 Paraguay 1990 35.4 45.0 66.2 56.6 51.7 51.9 *6,6 53.3 0.2 years 44.8 65.2 43.3 41.7 34.2 32.2 27.6 25. 3,5 yewr 41.7 51A4 50,3 48.7 43.3 42.1 216 34.1 Comp. Primary 33,3 40.6 76.5 60.9 52.9 54.6 13.7 60.1 $ocoauty + 22A 27.4 89.0 70.6 70.7 72.8 6,9 462 Peru 1991/ 64.0 92.0 87.9 70.5 66.3 66.4 10.8 None 1992 . 75.2 65.1 50.0 50.0 23.7 Primary .. 81.0 65.0 57.3 57.3 14.6 Secondary .. 94.8 71.2 71.6 71.9 6.2 Higher . 96.7 86.5 86.5 86.3 2.5 aPru 19086 76,. 112.0 59,9 70.9 64.7 65$5 None 124.1 175.0 .. .. In.PtmuF 85.0 1290 .. .. .. . Comp, ?riMrAy 42.0 56.0 .. ... SOConat + 22.0 22.0 ,, .. ., Trinidad &Tob. 1987 31.1 34.4 .. 44.0 87.0 87.1 6.9 95.9 Inc. Primary 27.5 32.6 . 43.0 84.8 75.8 11.8 95.5 Conp. Primary 24.7 28.9 . 42.1 85.8 86.5 7.6 96.0 lnt. Secondary 34.8 36.8 45.5 87.7 87.3 6.5 96.3 Adv. Secondary 61.0 61.0 . 48.2 92.8 91.6 1.5 93.7 Source: Westinghouse Demographic and Health Surveys, except El Salvador (1988) which was conducted by the Salvadoran Demographic Association with the assistance of the U.S. Centerfor Disease Control. Notes: a Number of infants, per thousand live births, who die before reaching one year of age. b Percent of children immunized. For Bolivia, Brazil, Colombia (1990), Guatemala, Paraguay, Peru (1991/92),figures arefor the 12-23 month age group. For Colombia (1986), Dominican Republic, El Salvador (1988), El Salvador, (1985), Peru (1986), Trinidad and Tobago, figures arefor the under 5 years age group. ' Refers to 0-59 month age group in most cases, although the age group variesfor afew countries. d Percentage of births occurring in hospitals, clinics, or other professional medical establishments. " Brazilian malnutritionfigure isfor the Northeast only. Social Indicators 91 Table 5.3: Intra-Country Variability of Infant Mortality Country/Region Mortality Country/Region Mortality Rate Rate Bolivia (1986) 95.8 Mexico (1977-87) 56.2 Altiplano 95.7 Northwest 47.8 Valles 105.6 North-central 39.6 Llanos 83.5 Northeast 73.5 Brazil (1986) 86.0 Central-east 62.8 Rio de Janiero 46.0 Central-west 43.0 Sao Paulo 61.0 Paraguay (1990) 35.4 Southern 44.0 Greater Asunci6n 28.1 Mid-east 55.0 North 41.6 Northeast 142.0 South Central 33.6 North-Central-West 57.0 East 40.8 Colombia (1990) 26.9 Peru (1992) 64.0 Atlantic 22.9 Loreto 89.0 East 27.9 A.A. Caceres 63.0 Central 24.2 Arequipa 36.0 Pacific 39.5 Chavin 58.0 Bogota 22.6 Grau 79.0 Guatemala (1987) 79.1 Inka 103.0 G}uatemala 72.3Mariategui 101.0 Guatemala 729.3 Libertadores 73.0 Central 119.3 Northeast 61.0 Southwest 72.3 La Libertad 52.0 Northwest 75.2 San Martin 85.0 North 484.1 Ucayali 84.0 Southeast 84.9 Lina 30.0 - continued 92 Powrtvy and Income Distribution in Latin America: The Story of the 1980s Table 5.3: Intra-Country Variability of Infant Mortality Rate (continued) Mortality Mortality Country/Region Rate Country/Region Rate Venezuela (1986) 21.4 Ecuador (1989) 44.2 Distrito Federal 16.6 Azuay 44.4 Anzoategui 18.8 Bolivar 47.5 Apure 21.1 Canar 40.0 Aragua 23.9 Carchi 51.9 Barinas 31.4 Cotopaxi 66.6 Bolivar 29.1 Chimborazo 56.4 Carabobo 23.3 El Oro 26.5 Cojedes 26.3 Esmeraldas 52.8 Falcon 24.1 Guayas 51.0 Guarico 27.8 Imbabura 52.1 Lara 28.9 Loja 27.6 Merida 29.9 Los Rios 61.2 Miranda 26.9 Manabi 23.6 Monagas 25.5 Morona Santiago 34.2 Nueva Esparta 23.1 Napo 34.0 Portuguesa 35.5 Pastaza 34.0 Sucre 22.2 Pichincha 41.6 Tachira 22.6 Tunguragau 56.1 Trujillo 37.1 Zaamora Chinchipe 28.7 Yaracuy 18.7 Galapagos 17.8 Zulia 29.1 Sucumbios 42.3 T.F. Amazonas 53.3 T.F. D. Amacuro 33.7 Notes: All data come from the Westinghouse DHS surveys except Brazil, Ecuador, and Venezuela. Brazil: IMR are for Infant classifi cation of 2 years and under. IBGE, Perfil Estat(stico de Criancas e mdes no Brasil: Aspectos de Saude e Nutricdo da Criancas no Brasil 1989. Brazil: IBGE, 1992. Ecuador: INEC, Yearly Vital Statistics, Ecuador: INEC, 1989. Mexico: See 1987 Mexico DHSfor complete breakdown of regions. Many states were left out of regional tabulation. Venezuela: World Bank, Venezuela Poverty Study: From Generalized Subsidies to Targeted Programs. Report No. 9114-VE. Washington, D.C., 1991. Social Indicators 93 Mothers' Education: The educational level of the mother is strongly correlated with infant mortality. In each of the DHS surveys, IMR was 3-5 times higher overall for mothers with no education than those with some university education. Bolivia, Brazil, Ecuador and Peru stand out as countries where the IMR was greater than 105 infant deaths per 1000 live births for mothers with no education (Figure 5.2).71 Within country variation due to mothers' education was highest in Peru where mothers with no education experienced IMR 5.6 times higher than mothers with post-secondary schooling. This disparity was severe in Colombia (5.3 times) and Ecuador (4.8 times) as well. However, a low IMR for women with some university education is, to a degree, reflective of critical hidden variables such as income level. In the context of improving the IMR for groups suffering the worst conditions, it is much more useful to examine the effect of basic schooling. On average, mothers with only 14 years of education often experienced infant mortality rates 30 percent lower than those mothers without education. The drop in IMR for women with 1-4 years of education was greater than 30 percent in Brazil, Colombia, Ecuador and Peru. This observation underscores the importance of basic literacy training. Figure 5.2: Infant Mortality by Mother's Education Mortality per 1000 births 140 120 100 80 60 40 200 iiii1ih lllf lZ 20 Colombia(1 990) Ecuador(1 987) Peru(1 986) M None 03 Primary C Secondary 3 Higher Source: [DHS survey] 71 Brazil is not included in Figure 5.2 because of different divisions of educational level. 94 Povery and Income Distribution in Latin America: The Story of the 1980s Urban vs. Rural: Some of the greatest differences in IM can be seen in the urban vs. rural companson. Generally, urban areas experience lower levels of IM than do rural areas. In Mexico rural figures were up to 255 percent higher than urban (28.6 vs. 78.8 infant deaths per thousand live births for the smallest population areas). Other notable disparities were Bolivia, Brazil, El Salvador (around 40 percent higher for rural), Ecuador (50 percent higher), and Peru (90 percent higher). Ethicity: IMR varies across ethnic/linguistic groups. Recent data for IMR by ethnic groups were only available for Bolivia and Guatemala. In Bolivia, IMR for the indigenous population was 35 percent higher than for the Spanish speaking population, (116.6 vs. 86.5 respectively). In Guatemala, the DHS survey reveals that the indigenous population experiences IMR below that of the Ladino population (76.4 vs. 84.8 respectively), but this rate was reversed for mortalities under 5 years of age where the Ladino population does 20 percent better than indigenous groups (142.0 for indigenous vs. 119.6 for Ladino).72 Child Global Malnutrition In the LAC region an estimated 10 million preschool children suffer from moderate to severe levels of malnutrition as defined by weight-for-age.73 The consequences of malnutrition vary from among such conditions as stunted growth, frequent and more severe illness, and irreversible mental retardation in the case of protein-calorie malnutition. As a result malnutriton tkes its toll on longevity and productivity, and its effects can be trans-generational since malnourished mothers are more likely to have premature or malnourished infants. In general the picture for malnutrition is similar to that for infant mortality in pattern and variation. Child malnutrition, like infant mortality, decreased in many LAC countries during the 1980s regardless of deteriorating or improving income equality or absolute poverty levels. However, there are signs that improvements in nutritional status have slowed during the last decade as economic problems intensified. It is estimated that the total undernourished population will increase from 55 million in 1983-85 to 62 million by the end of the century.74 The greatest incidence of child malnutrition is found in a small number of countries with large populations. Brazil and Mexico together contain two-thirds of the malnourished children in the region, an estimated 6,608,800. Bolivia, Colombia, and Guatemala, account for 72 See country tables in Annex 14 for exact figures. 73 Philip Musgrove, Feeding Latin America's Children: An Analytical Survey of Food Programs. Regional Study No. 11. World Bank, Washington, D.C., 1991. 74 Proceedings of the 21st Regional Conference of Latin America and the Caribbean, FAO, July 1990. Social Indicators 95 Box 5.1: Malnutrition in Guatemala In 1987, Guatemala exhibited a high degree of variation for child global malnutrition. The most striking differences existed between regions and educational levels of the mother. Guatemalan child malnutrition was 22.4 percent of the population between 3 and 36 months of age in the Guatemala City region, and around 40 percent in many of the predominately indigenous regions (Central, Southwest and Northwest). For mothers with no education, child malnutrition was greater than 40 percent. This figure dropped to 30 percent for mothers who received several years of primary education. Malnutrition rates varied across urban/rural settings. Malnutrition in rural areas was more than 40 percent higher than urban areas. Indigenous children suffered malnutrition rates around 40 percent higher than Ladino children. Child Malnutrition Guatemala, 1987 Urban Rural . ,- * -l IndigeandousO Ladino Guatemala North Northeast Southeast Central Southwest Northwest No Educabton Primary Secondary Post-Secondary 0 10 20 30 40 50 Percent of Children Malnourished C Urban/Rural E IEthnicity 0 Region U Mother's Education Suuro. Demographic and Health Survey. Guatemala, 1987. 96 Poverty and Income Distribution in Latin America: 7he Story of the 1980s another significant share of the regions malnourished children, an additional 1,750,600.75 Intra-Country Regional Variation: Again, aggregate figures hide large intra-country variations, most notably in Guatemala (see Box 5.1), Paraguay and Peru. In Paraguay, the northem region had a malnutrition level that was three times that of Gran Asunci6n. And in Peru, Lima had a malnutrition level of less than 3 percent of the child population while the poor states of Loreto, Inka and San Martin had levels of 20, 20 and 16 percent, respectively. In 1989, the number of Brazilian children moderately and severel6y malnourished was less than 1.9 percent in the South, and almost 10 percent in the Northeast.7 In Honduras child malnutrition affected 37 percent of total population under 5 years of age, but reached 55 percent in some of the poorest rural areas. Malnutrition was a factor in 60 percent of infant deaths in Honduras.77 Malnutrition in Mexico was nearly four times higher in rural southern Mexico than in northern Mexico. In the states of Chiapas and Oaxaca, about 70 percent of children under 5 suffer from some degree of malnutrition.79 In Nicaragua child malnutrition was estimated at 10-20 percent in the country, except Managua which had a malnutrition rate of 7.9 percent.79 Mother's Education: Table 5.1 clearly illustrates the dramatic impact which mother's education plays on child malnutrition. For countries with complete data, there was without exception a decrease in child malnutrition as mothers' level of education increased. Bolivia experienced the smallest variation in malnutrition associated with mother's education - a 350 percent difference in malnutrition between mothers with no education and those with some post-secondary education. Guatemala and Paraguay experienced 400 percent reductions across similar educational increases. The most dramatic disparity due to mother's education occurred in Peru where malnutrition dropped 950 percent between mothers with no education and those with post-secondary (23.7 percent vs. 2.5 percent respectively), and Colombia where child malnutrition was 0 percent for mothers sampled from the highest educational bracket. For mothers with only 1-4 years of education child malnutrition dropped by 30 percent when compared to mothers with no education in Bolivia, Brazil (malnutrition data for Northeast only), Guatemala and Peru. 75 Musgrove, 1991, Op. cit. 76 World Bank, Addressing Nutritional Problems in Brazil. Sector Report No. 8881-BR. Washington, D.C., 1990. 77 World Bank, Republic of Honduras: Second Social Investment Fund Project. Staff Appraisal No. 10451-HO. Washington, D.C., 1992. 78 World Bank, Mexico Basic Health Care Project. Staff Appraisal Report No. 8927-ME. Washington, D.C., 1990. 79 World Bank, Republic of Nicaragua Social Sector Issues & Recommendations Report. Report No.10671-NI. Washington, D.C., 1992. Social Indicators 97 Urban vs. Rural: Like infant mortality, there was a high degree of variation in malnutrition between urban and rural regions. In all cases for which survey data were available, rural areas had child malnutrition levels which were at least 40 percent higher than their urban counterparts. This was most notable in Paraguay (rural was double urban) and Peru (rumral was almost triple urban). Ethnicity: Ethnicity is also associated with malnutrition levels. Indigenous children suffer higher levels of malnutrition than their non-indigenous counterparts. In Bolivia malnutrition was twice as high among the Indian-speaking population than among the Spanish-speaking population. In Guatemala malnutrition was over 40 percent higher for the indigenous population than for the Ladino population. Access to Matemal Health Care at Birth Maternal and child health care (MCH) programs have an important role to play in lowering infant mortality rates. (See Box 5.2.) MCH programs cover a wide variety of services including pre- and post-natal care, nutrition programs for both mothers and children, and food subsidy programs, among others. MCH programs can be targeted at a specific illness or at improving the general health of mothers and children.'0 Tables 5.1 and 5.2 enumerate the percentage of births occurring in hospitals, clinics, or other professional medical establishments. The current state of MCH programs in the LAC region is inadequate. A recent study conducted by the Pan American Health Organization (PAHO) on MCH programs in Latin America concluded that MCH services were unsatisfactory in 80 percent of the 1,700 laboratory 80 Human Resources in Latin America and the Caribbean: Priorities and Action. Washington, D.C.: World Bank, 1993. 98 Poverty and Income Distribution in Latin America: The Story of the 1980s services and hospitals sampled. The main reason cited for the poor performance of Latin America's MCH services was poor program design.81 Mother's Education: Mother's educational level is correlated with some of the highest variations occurring in access to maternal care. All countries with available data experienced sharp increases in maternal care at birth with increases in mother's educational level. The most extreme case was Bolivia where delivery in a professional medical environment for mothers with post-secondary education was 11 times greater than mothers with no education. Urban vs. Rural: The prevalence of maternal and infant health care at birth differed between urban and rural regions. In Bolivia the urban percentage of access to professional health care at birth was 3 times the rural percentage (58.2 percent vs. 18.3 percent respectively). These large disparities persist throughout for countries in the survey for which data on maternal health care was available with the notable exception of Trinidad & Tobago where there was a negligible difference between urban and rural areas. Immunizations Data for the percentage of children immunized with BCG, measles, polio (3 doses) and DPT (3 doses), are shown in Table 5.2. At the aggregate level, Brazil and Trinidad & Tobago performed better than other countries in the region. Brazil led in measles immunization with 83.6 percent of children having received inoculation and second in polio3 with 84.4 percent of children. Guatemala, at the other extreme, had generally low percentages of child immunization; few vaccines were received by much more than one-half of the child population. Mother's Education: Again, mother's education played an important role. With few exceptions, child immunization levels increased with mother's education. However, in the 1986 sample for Colombia and the 1987 for Guatemala, immunizations showed a considerable decrease between secondary and post-secondary educational achievement, marking a reversal of the expected trend. Two possible explanations may be (i) sample design error or (ii) those who rely on public health care, i.e. often the less monetarily and educationally advantaged, may be receiving more complete health coverage in certain health areas, specifically preventive health care, than those relying on private health care. Urban vs. Rural: Generally urban areas had higher percentages of child immunizations. In some cases rural areas performed better, most notably Trinidad & Tobago where rural polio3 inoculations were about 50 percent higher than urban (88.0 percent vs. 58.6 percent for urban). SI Segunda Reunion Subregional Andina Sobre Salud Maternoinfantil. Buletin de la Oficina Sanitaria Panamericana 107, no. 3(September 1989). Social Indicators 99 Box 52: Maternal Health, Infant Health, and Mortality Access to maternal health care is highly correlated to infant mortality. In Colombia infant mortality experienced high variation when levels of maternal health care were compared. For those women who did not receive qualified professional pre-natal health care or assistance at birth, IMR averaged a very poor 187.5. For those women who received either one, but not both types of health care, IMR dropped to 21.9, and those who received both experienced an IMR of only 14.5. Variability in infant mortality rates was also high with varying levels of infant health care. In the northeast region of Brazil, IMR dropped dramatically between the years 1983 and 1987, reflecting an augmentation of medical services to the area, including a 31.4 percent increase in DPT3 vaccinations and oral rehydration programs between 1980 and 1985. Regional Infant Mortality Brazil 1980-88 Infant Mortality Rate 120 Northeast 100 80 -*.,,___ 60 40 - South .... .. .. .. .. .. . 20 - o 1 I I I I l 1980 1981 1982 1983 1984 1985 1986 1987 1988 Source Colombia DHS Survey and IBGE, Perfil Estafistiso de Cnanzas e maes no Brasil: Aspectos de Saude Nutnrio de Cnangas no Brasil 1989. Brazil: IBGE, 1992 100 PoE and Income Distribution in Latin America: The Story of the 1980s Education It is clear that mother's educational level plays a vital role in determining levels of health indicators. The critical influence of mother's education emphasizes a need to examine education as a whole within the region. Education varies greatly in LAC between and within countries. Guatemala, Honduras, Bolivia and northeast Brazil mark some of the worst areas with regard to education. Illiteracy, years of schooling and repetition and dropout rates are the measures of education examined below. Figures are based on independent sources and the household surveys which provided further opportunity to examine education indicators by urbanlrural division and income quintile. Illiteracy Though absolute illiteracy has declined in LAC over the 1980s from 44.3 million of those over 15 years of age in 1980 to 42.8 million in 1987, it is still a serious problem which characterizes 15 percent of the population. It is even more critical when considering the problem is concentrated in five countries: Brazil, Dominican Republic, El Salvador, Guatemala and Honduras. ILUiteracy is also concentrated among indigenous groups in four other countries: Bolivia, Ecuador, Mexico and Peru. Women and those over 40 years of age tend to be disproportionately affected throughout the region. In light of the importance of mother's education on a number of health indicators, this disproportionate representation of female illiteracy suggests widespread repercussions throughout the health sector. In 1990 in Guatemala illiteracy reached 52.9 percent of the female population. Regarding those over 40, one-half of absolute illiteracy is now accounted for by this older age group, despite great strides made towards reducing illiteracy among the young population. 82 According to the most recent data available, Guatemala and Honduras had the highest percentages of illiteracy with 40 and 25 percent of the population over 15 years of age, respectively.83 The lowest percentage was Paraguay where 2.2 percent of the population was illiterate (see Table 5.4). 82 E. Schiefelbein and J. C. Tedesco, Primary Schooling and Illiteracy in Latin America and the Caribbean: 1980 - 1987. The Major Project in the Field of Education in the Latin American and Caribbean Region, Bulletin No. 20, December 1989. 83 For Bolivia, Chile, Costa Rica, Panama, Paraguay and Uruguay, illiteracy rates, as defined by the household survey analysis, are the percent of persons with no schooling. The no schooling variable will give an upper bound for illiteracy rate since people with no schooling could have achieved literacy through other means such as the alphabet school. Social Indicators 101 Figure 5.3: Illiteracy Prevalence (Percentage of Illiterate Population) IBRD 28450 PERCENTAGE OF ILLiTERATE POPULATION ~ ~ ~ ~ omincanATLNTI No data shownl alv u shw n h,ma ant rpyna h or t ofTeW rl Ban Ip oERaCENTAGE nF I o IEAi b onaPUlAT Ove : 5 SOUCe:Socal ndi Leos ofa Deeomn5 192%isels ol ak F-1 No data shown PAOFIC 1996 OCEAN.1iH, o,, CA This~~ ~ ~ ~ ma wa pmme by "B lz mamic io Uni ofroWr ak Souce:SoialInicaor ofDeelomet ( 92)isEllesd, WorlBank.a^ ,_ .~~~~~~~~~~~~ECME19 102 Povery and Income Distribution in Latin America: The Story of the 1980s Illiteracy varied considerably between urban and rural regions. In Brazil illiteracy ranged from 11 percent in the urban South/South-Central to five times that in the rural Northeast. Illiteracy rates in Costa Rica for 1984 were over three times higher in rural areas than in urban. 84 In 1981, Peru's illiteracy rate ranged from 4.5 percent in Lima to more than 45 percent in the poorest departments, marking a tenfold difference between the two regions.85 In 1989, Venezuela experienced illiteracy four times greater in the Andean region than in the Caracas metropolitan am.e86 The correlation between income and illiteracy is very strong. Individuals in the top income quintile of urban areas stood out in terms of low illiteracy rates. Even in Guatemala and Honduras, where overall illiteracy rates are high, the top income quintiles had low rates - comparable to other top income quintiles in LAC. (See Table 5.4) 84 World Bank, Costa Rica - Basic Education Rehabilitation Project. Staff Appraisal Report No. 9893-CR. Washington, D.C., 1991. 85 H. Fernandez and J. Rosales, Educacion una Mirada Hacia Adentro. Peru: IPP, 1990. 86 World Bank, 1991, MScit. Social Indicators 103 Table 5.4: Iliteracy by Income Quintile (% of 15+ Age Group) Per Capita Income Quintile Country Year Total 1 2 3 4 5 Argentina (Gran Buenos Aires) 19f9 2.3 4.4 3.5 2.5 1.4 0*.6 Bolivia" (Urban) 1989 6.6 10.5 9.6 6.4 4.6 3.2 Brazil 1989 18.8 43.0 30.2 18.2 9.6 3.5 utban 134 35.6 25.0 151 9.4 $3 Rural 36.4 45.4 40.0 29.8 17.1 10.3 Chile 1989 6.8 11.5 8.8 7.2 5.0 3.2 Urban 4.9 8.8 6.7 5.6 3.8 1.8 Rural 17.7 19.7 18.6 19.1 18.6 12.1 Colombia (Urban) 1989 3.2 6.2 4.2 3.4 2.4 0.9 Costa Rica! 1989 7.0 13.6 8.5 7.5 4.4 2.7 Urban 4.4 10.4 6.9 5.6 3.0 2.1 Rural 9.1 14.8 9.2 8.8 5.8 4.1 Guatemala 1989 39.7 64.0 53.7 42.7 29.7 13.2 UTr @20.3 48.7 3846 29.6 20.6 9.0 Rural 51.8 63S9 57.3 48.6 38.4 25.9 Honduras 1989 25.4 40.6 36.6 30.7 20.5 8.1 Urban 10.8 27.1 24.8 19.8 11.3 4.9 Rural 33.5 41.6 38.3 34.7 28.2 14.8 panama 1989 5.6 13.2 7.0 5.3 3.3 1.4 UrbM 2.6 6.4 4.1 3.5 21 1.2 Rurat 10.2 15.2 9.5 7.9 6.7 3.2 Paraguay" 1990 2.2 3.6 2.5 3.6 1.4 0.5 Urban 2.1 3.1 2.1 4.0 1.4 0.6 Rural 3.2 6.5 5.9 0.0 1.7 0.0 Peru 1990 9.1 20.3 11.8 7.8 S'. 3.6 Utban 6.1 12.0 9,5 7.0 4.3 3.1 Rura 22.2 26.0 21.1 14.2 14.6 16.5 Uruguaya (Urban) 1989 3.5 5.0 6.0 4.1 2.4 1.0 Venezuela 1989 8.9 18.3 12.4 8.8 6.3 2.9 Utban 6.) 12.4 8.9 6.8 4.7 2,2 Rurat 25.4 30.1 25.9 21.5 21.1 16.5 Note: a Figures reflect the illiteracy rate as reflected in household surveys, exceptforfive countriesfor which thefigures reflect the percentage of the populktion with no schooling. Thesefive countries are Bolivia, Chile, Costa Rica, Panama and Paraguay. For these countries, the no schooling variable will give an upper boundfor illiteracy. 104 Powert and Income Distribution in Latin America. The Story of the 1980s Ye= Of Schooling Annex Table 14.1 shows the distribution of average number of years of schooling attained by persons over 18 years. In all countries, the urban residents are substantially better schooled. At the aggregate level, Peru had the highest average at 9.4 years and Guatemala had the lowest at 3.2 years. In predominantly rural countries such as Guatemala and Honduras, the average schooling gap between urban and rural areas reached 3.5 years. In these two countries, the gap between the top and bottom income quintiles reached its most extreme manifestation at approximately 6 years difference. With the exception of Guatemala and Honduras, the top income quntile had about twice as much schooling as the bottom quintile. In Guatemala and Honduras this multiple was about four. (see Figure 5.4) Figure 5.4: Educational Attainment by Income Quintile Schooling Years 12 1 0 8 6. ,. 1~~~~~~~~~~~~~~~X 4 i 2 0 0 Brazil Chile Guatemala Honduras Paraguay Peru Income Quintile: 1 (bottom 1/5) 0 5 (top 1/5) Social Indicators 105 Grade Repetition and Dropout Rates Among the most serious problems facing education in the world today is grade repetition and school dropout. The LAC region experiences the highest grade repetition and school dropout rates in the world. Within the region one-half of all children who enroll in the first grade never complete the fourth grade. Twenty-nine percent of primary students are repeaters each year; and 42 percent are repeating first grade. As a result, grade repetition has become the main issue in LAC primary education, both in improving education and reaching suitable levels of quality.87 Like other indicators, repetition and dropout rates exhibit regional variation. In Costa Rica dropout rates are three times higher in Lim6n, Cartago and Canas than in metropolitan San Jose while the percentage of students completing sixth grade is 33 percent lower in these regions. In Honduras in 1987, only 46 percent of those who entered first grade reached sixth grade. This figure was only 23 percent for rural areas.89 In 1988 school repetition rates in Mexico were more than twice as high in poor states (Oaxaca, Chiapas, Guerrero, Campeche) than in the Federal District, and dropout rates were 4 or 5 times as high in the poorer states. In Chiapas, less than one-third of students completed primary school, while nearly 80 percent did so in the Federal District. A recent study of grade repetition, based on the household data, estimated various probabilities of grade repetition based on a number of individual variables. In Bolivia it was found that the percentage chance of grade repetition did not vary appreciably for gender or income. Among the factors that did increase chances for grade repetition were being of an indigenous group, enrolled in a public school, having a poorly educated mother and being from a female- headed household.90 (see Figure 5.5) 87 Harry Patrinos and George Psacharopoulos, Socioeconomic and Ethnic Determinants of Grade Repetition in Bolivia and Guatemala. Washington, D.C.: World Bank, 1992. 88 World Bank, 1991, opcit. 89 World Bank, Republic of Honduras - Second Social Investment Fund. Staff Appraisal Report No. 10451-HO. Washington, D.C., 1992. 90 It was noted that houses without a male head may be suffering more difficult economic consequences thus explaining the worsening grade repetition rates for children of female headed households. For further discussion see Patrinos and Psacharopoulos, 1992, op. cit. 106 Powrty and Income Distribution in Latin America. The Story of the 1980s. Figure 5.5: Grade Repetition in Bolivia Repetdion probabtldy (%) 50 40.8 10 Non8-indig Indig Public Private Bottom20 Top20 Pim- Prm Sec+ Ethnic group School type Family income quintile Mother's education An "education deficit" index was created to measure the shortfall of schooling for individuals between the ages of seven and seventeen (Annex Table 14.3).91 This index measures the shortfall of education with regards to an expected complete and uninterrupted enrollment. Guatemala and Paraguay had the greatest education deficit value of .59, indicating that 59 percent of expected education has been foregone. Chile had the lowest at .12. The greatest disparities between top and bottom quintiles were in Chile and Venezuela where the bottom quintiles had deficit values three times greater than the top quintiles. In most countries the education shortfall of the bottom income quintile was about twice that of the top income quintile. 91 The education deficit index is defined as follows: for individuals in the age group between seven and seventeen, a variable is defined, AGE-6, which is the potential years of schooling ignoring the possibility of skipping the grade and not accounting for enrollment starting or ending outside of the seven to seventeen year age band. By subtracting the years of schooling from the above, AGE-6-S should be the short-fall of schooling from what it would have been had one continued their education up to his/her age. This is termed the education deficit or education gap. In order to have a consistent measure across different age groups, the education deficit is normalized by dividing by AGE-6, which yields the education deficit index, Education Deficit Index = (AGE-6-S)/(AGE-6). By subtracting this index from one, an opposite measure of the education deficit index is created, i.e., the education attainment index. Social Indicators 107 Demographic and Employment Indicators This section is based on the household surveys, which provide a unique opportunity to examine timely data across and within income quintiles. Each heading is based on a corresponding table in Annex 14 and briefly describes observable trends and correlations. Indigenous Population Indigenous peoples represent 8 percent of Latin America's population, an estimated 37 million people across 589 tribes. Bolivia, Guatemala, Peru and Ecuador are among countries whose indigenous populations comprise more than 40 percent of the entire national population. In Bolivia over 70 percent of the population identifies itself as indigenous. In absolute numbers, Mexico contains the largest population of indigenous peoples at 12 million. The smallest percentage of indigenous people is in Uruguay where indigenous people represent less than 1 percent of the population or fewer than 500 people.92 92 Jordan Pando, Roberto. Desarrollo en Poblaciones Indi2enas de America Latina y El Caribe. Mexico, D.F.: Instituto Indegenista y FAO, 1990. 108 Poverty and Income Distribution in Lotin America: The Story of the 1 980s Figure 5.6: Indigenous Population of Latin America IBRD 28449 ATLANTIC OCEAN Guatemal u S a33 El Solvo > ragua . a na h Guiana Ecuador- W.t I H I I I I 1970 1975 1980 1985 1990 GNP per capita (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 3.3 -- 2.0 4.1 for basic social services in GDP Net primary enrollment -- 47.3 50.4 54.4 60.5 65.4 -- Under-five mortality -- -- -- 94.1 66.7 50.3 Child malnutrition -- -- -- -- 43.5 33.5 -- Life expectancy at birth - male -- -- -- -- 54.5 56.8 59.7 - female -- -- -- -- 58.4 61.3 64.4 Total fertility rate -- 6.7 6.5 6.4 6.2 5.9 5.6 Maternal mortality -- -- -- -- 9.1 11.2 9.2 Percentage of Urban Population -- 34.3 35.7 37.1 32.7 32.7 32.7 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All General Characteristics Rural area (h) 89.6 80.3 68.3 50.6 25.0 62.8 Mean household size 6.6 5.9 5.6 5.0 4.1 5.4 Indigenous population (%) 66.9 49.2 35.3 21.9 8.0 36.3 Female-headed household (%) 11.2 14.0 15.3 17.5 20.0 15.7 Dependency ratio 30.6 6.4 3.8 3.1 2.5 4.3 258 ANNEX 15 - continued Social Indicators, Guatemala Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All Education Years of schooling among 1.0 1.5 2.3 3.5 7.0 3.2 adults aged 18 and older Illiteracy rate among persons 64.0 53.7 42.7 29.7 13.2 39.7 aged 15 and older (%) Education deficit index 0.75 0.67 0.60 0.48 0.35 0.59 (aged between 10 and 18) Labor Market Employment category - employees (%) 8.4 36.7 56.2 62.8 69.1 49.0 - self-employed (%) 54.6 39.2 29.8 27.0 21.9 33.2 - owners of enterprises (%) 0.1 0.4 0.5 1.1 4.4 1.5 - others (%) 37.0 23.7 13.5 9.2 4.7 16.3 Among employees - public sector (%) 3.1 3.3 6.1 13.0 28.3 15.1 - private sector (%) 96.9 96.7 93.9 87.0 71.7 84.9 Informal sector (%) 96.3 79.4 62.7 55.2 43.6 65.3 Labor force (%) 43.2 44.8 49.0 53.4 61.3 50.3 unemployment rate (%) 1.0 1.3 2.2 2.5 2.0 1.9 Labor income index (mean= 100) 26.8 45.1 59.7 80.8 167.9 100.0 Housing Characteristics Home ownership (%) 88.3 78.6 69.0 64.1 62.8 72.2 Rooms per capita 0.35 0.43 0.49 0.62 1.07 0.60 Access and Uses of Public Services Access to electricity (%) 16.1 29.0 47.1 65.7 86.1 49.8 Access to running water (%) 46.9 52.6 61.0 70.1 86.8 64.0 259 ANNEX 15 - continued Social Indicators, Honduras Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 US$) - -- 770 780 950 810 660 Social Indicators, Honduras 1200 1000n 600 1 ll M ii 1[ 400 F RIRFF 1970 1975 1980 1985 1990 GNP per capta (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 6.9 -- -- -- for basic social services in GDP Net primary enrollment -- 68.2 78.4 77.9 80.8 91.0 -- Under-five mortality -- -- -- -- 27.1 -- Child malnutrition -- -- -- -- 20.8 20.6 - Life expectancy at birth - male -- -- -- -- 60.4 60.0 61.9 - female -- -- -- -- 59.6 64.0 66.1 Total fertility rate -- 7.4 7.4 6.9 6.3 5.8 5.4 Maternal mortality -- -- -- 9.4 - -- Percentage of Urban Population -- 25.7 28.9 32.3 35.9 39.7 42.8 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All General Characteristics Rural area (%) 93.7 87.0 72.6 53.5 33.1 68.0 Mean household size 6.6 6.1 5.8 5.3 4.3 5.5 Female-headed household (%) 19.2 16.5 19.1 18.1 17.1 17.9 Age structure of population - children aged 10 and under (%) 45.0 41.2 37.1 31.9 24.8 36.0 - 11-64 (%) 47.4 56.2 59.6 64.8 72.3 61.1 - 65 and older (%) 2.6 2.6 3.3 3.3 2.9 2.9 Dependency ratio 22.4 10.7 7.1 5.7 3.6 6.9 260 ANNEX 15 - continued Social Indicators, Honduras Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All Education Years of schooling among 2.1 2.4 3.1 4.4 7.5 4.2 adults aged 18 and older Illiteracy rate among persons 40.6 36.6 30.7 20.5 8.1 25.4 aged 15 and older (%) Education deficit index 0.67 0.62 0.57 0.48 0.31 0.55 Labor Market Employment category - employees (%) 16.9 31.9 42.7 48.6 59.0 42.9 - self-employed (%) 56.2 46.8 40.3 37.2 23.3 38.3 - owners of enterprises (%) 0.5 0.5 0.6 1.2 4.7 1.8 - others (%) 26.4 20.8 16.4 13.0 13.1 17.0 Among employees - public sector (%) 3.3 4.5 10.6 19.2 39.1 22.5 - private sector (%) 96.7 95.5 89.4 80.8 60.9 77.5 Informal sector (%) 92.0 83.2 70.8 60.5 37.3 64.4 Labor force (%) 41.7 44.3 48.1 51.3 59.9 49.7 Unemployment rate (%) 2.7 2.6 4.2 4.7 2.4 3.3 Second job (%) 12.8 12.6 10.2 6.9 5.6 9.0 Labor income index (mean= 100) 21.6 33.6 48.0 73.5 172.2 100.0 Housing Characteristics Home ownership (%) 85.1 82.4 78 5 75.6 68.7 77.3 Rooms per capita 0.38 0.42 0.52 0.62 1.04 0.63 Access and Uses of Public Services Access to electricity 8.5 16.6 32.8 56.3 80.5 42.6 Access to public water supply (%) 20.1 24.4 37.1 52.5 70.7 43.6 261 ANNEX 15 - condnued Social Indicators, Mexico Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 US$) -- -- 1,310 1,520 1,920 1,790 1,730 Social Indicators, Mexico 2500 2000 15000 sooo _ Eli VI 0 1970 1975 1980 1985 1990 GNP per capta (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 6.7 6.4 5.8 5.5 for basic social services in GDP Net primary enrollment -- 66.6 77.8 83.3 92.6 100.5 -- Under-five mortality -- -- -- -- 37.2 27.5 -- Child malnutrition -- -- -- -- -- -- 13.7 Life expectancy at birth - male -- -- -- -- 62.6 64.2 65.7 - female -- -- -- -- 68.2 70.6 72.3 Total fertility rate -- 6.7 6.5 5.5 4.5 3.8 3.3 Maternal mortality -- -- -- -- 9.4 6.4 -- Percentage of Urban Population -- 54.9 59.0 62.8 66.4 69.6 72.6 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All General Characteristics Mean household size 6.5 5.6 5.1 4.6 3.8 4.9 Female-headed household (%) 10.5 11.8 14.4 16.0 17.8 14.6 Age structure of population - children aged 10 and under (%) 38.5 31.8 26.0 22.7 19.4 27.7 - 11-64 (%) 57.7 63.8 68.7 72.5 74.9 67.5 - 65 and older (%) 3.8 4.4 5.3 4.8 5.7 4.8 262 ANNEX 15 - continued Social Indicators, Mexico Quintile Characteristics, 1989 Per CaRita Income Ouintile 1 2 3 4 5 All Dependency ratio 8.7 5.5 4.5 3.6 3.2 4.5 Education Years of schooling among 3.0 4.4 5.6 7.0 9.00 6.1 adults aged 18 and older Illiteracy rate among persons 27.1 16.4 10.2 5.7 2.7 11.3 aged 15 and older (%) Education deficit index 0.45 0.34 0.27 0.23 0.19 0.32 Labor Market Employment category - employees (%) 41.4 63.7 69.9 75.3 73.9 66.7 - self-employed (%) 36.0 25.9 22.5 18.2 16.1 22.6 - owners of enterprises (%) 1.6 1.6 1.4 2.6 5.8 2.8 - others (%) 21.1 8.8 6.2 3.9 4.2 7.9 Labor force (%) 45.3 45.5 46.7 51.6 54.6 49.0 Unemployment rate (%) 7.3 7.6 6.1 5.8 5.7 6.4 Second job (%) 25.2 15.5 11.4 8.4 10.1 13.2 Labor income index (mean= 100) 16.1 44.7 73.3 117.2 251.4 100.0 Housing Characteristics Home ownership (%) 86.5 77.9 75.4 72.8 73.3 76.4 Rooms per capita 0.35 0.46 0.59 0.75 1.19 0.72 Ownership of Durable Goods Television (%) 39.5 67.4 83.3 89.4 94.3 78.0 Refrigerator (%) 10.9 30.3 55.9 73.4 88.4 56.8 Washer (%) 4.1 15.1 31.6 46.8 64.6 36.4 Access and Uses of Public Services Electricity for lighting (%) 66.2 86.3 95.1 96.9 99.0 90.6 Access to public water (%) 50.2 70.8 82.9 90.9 95.0 80.7 Access to public sewer (%) 14.2 35.2 54.6 69.4 83.2 55.7 Connection to telephone (%) 0.7 3.2 10.3 27.2 48.6 21.2 263 ANNEX 15 - continued Social Indicators, Panama Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 US$) -- -- 1,750 1,940 2,030 2,150 1,660 Social Indicators, Panama 2500 2000 - X - 1 U 1500 0I soo 111I I_ I _II II I 1970 1975 1980 1985 1990 GNP per capita (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- -- 11.9 14.4 16.1 for basic social services in GDP Net primary enrollment -- 82.6 74.1 87.4 88.5 88.6 -- Under-five mortality -- -- -- -- 24.0 23.5 21.1 Child malnutrition -- -- -- -- 15.1 -- -- Life expectancy at birth -male -- -- -- -- 67.6 69.2 70.1 - female -- -- -- -- 70.9 72.9 74.1 Total fertility rate -- 5.7 5.2 4.4 3.7 3.3 3.0 Maternal mortality -- -- -- -- 7.2 5.7 3.8 Percentage of Urban Population -- 44.4 47.6 49.1 50.5 52.4 54.3 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All General Characteristics Rural area (%) 76.4 52.2 40.4 26.1 11.7 41.4 Mean household size 4.8 4.8 4.4 4.0 3.3 4.2 Female-headed household (%) 25.2 27.6 25.4 25.1 21.2 24.6 Age structure of population - children aged 10 and under (%) 30.5 29.1 23.5 19.4 15.5 23.6 - 11-64 (%) 63.6 66.4 71.1 74.0 75.9 70.2 - 65 and older (%) 5.9 4.5 5.4 6.6 8.6 6.2 Dependency ratio 14.9 7.0 4.8 3.7 2.6 4.7 264 ANNEX 15 - continued Social Indicators, Panama Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 AU Education Years of schooling among 5.1 7.0 8.0 9.2 11.7 8.5 adults aged 18 and older No schooling among persons 13.2 7.0 5.3 3.3 1.4 5.6 aged 15 and older (%) Education deficit index 0.31 0.26 0.19 0.14 0.14 0.22 Labor Market Employment category - employees (%) 20.3 59.8 69.0 77.1 86.8 67.2 - self-employed (%) 56.8 34.9 27.2 20.0 8.6 25.8 - owners of enterprises (%) 0.9 0.8 1.3 1.6 3.9 2.0 - others (%) 22.0 4.5 2.5 1.3 0.7 5.0 Among employees - public sector (%) 11.0 19.5 32.1 42.1 49.9 38.8 - private sector (%) 89.0 80.5 67.9 57.9 50.1 61.2 Informal sector (%) 92.5 60.8 46.5 32.7 22.7 45.6 Labor force (%) 54.5 53.0 56.6 58.9 64.8 58.0 Unemployment rate (%) 20.4 25.5 21.1 16.9 7.3 17.2 Labor income index (mean= 100) 18.5 43.8 61.2 84.5 163.3 100.0 265 ANNEX 15 - continued Social Indicators, Paraguay Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 USS) -- -- 570 690 1,070 950 1,040 Social Indicators, Paraguay 1200 1000 200 1970 1975 1960 1998 1990 GNP per capita (in 1987 USS) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 3.8 3.3 4.6 2.6 for basic social services in GDP Net primary enrollment -- 80.0 88.4 88.8 95.2 95.2 -- Under-five mortality -- -- -- -- 66.8 49.4 -- Child malnutrition -- -- -- -- -- -- 4.2 Life expectancy at birth -male -- -- -- -- 64.1 64.4 64.8 - female -- -- -- -- 68.1 68.6 69.0 Total fertility rate -- 6.6 6.0 5.3 5.0 4.8 4.6 Matermal mortality -- -- -- -- 36.5 28.3 19.4 Percentage of Urban Population -- 36.2 37.1 39.0 41.7 44.4 47.5 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 AU General Characteristics Rural area (%) 12.9 12.0 9.2 8.2 4.0 9.3 Mean household size 5.8 5.4 5.0 4.4 3.9 4.8 Monolingual Guarani-speakers (%) 16.8 12.1 9.0 4.6 2.2 8.9 Female-headed household (%) 21.3 20.1 21.4 17.7 18.1 19.5 Age structure of population - children aged 10 and under (%) 36.2 28.9 22.0 19.1 15.1 24.3 - 11-64 (%) 58.8 67.4 71.9 74.5 78.4 70.2 - 65 and older (%) 5.0 3.7 6.1 6.4 6.5 5.5 Dependency ratio 7.3 4.7 3.7 3.4 3.7 4.2 266 ANNEX 15 - continued Social Indicators, Paraguay Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All Education Years of schooling among 6.3 7.4 8.0 9.4 11.2 8.7 adults aged 18 and older No schooling among persons 3.6 2.5 3.6 1.4 0.5 2.2 aged 15 and older (%) Education deficit index 0.63 0.61 0.57 0.55 0.55 0.59 Labor Market Employment category - employees (%) 55.7 60.4 64.5 63.2 51.0 58.9 - self-employed (%) 30.2 23.5 21.0 21.2 19.6 22.2 - owners of enterprises (%) 4.2 6.8 5.8 6.3 16.2 8.6 - others (%) 9.8 9.2 8.8 9.3 13.2 10.3 Among employees - public sector (%) 32.7 33.3 33.1 38.5 37.8 36.2 - private sector (%) 67.3 66.7 66.9 61.5 62.2 63.8 informal sector (%) 67.1 59.2 52.4 50.1 49.7 54.2 Labor force (%) 61.6 65.5 69.1 70.0 75.5 68.9 Unemployment rate (%) 17.5 8.9 4.8 3.3 1.9 6.3 Labor income index (mean= 100) 60.2 73.3 81.0 103.4 156.2 100.0 Housing Characteristics Home ownership (%) 75.6 78.8 70.3 73.0 73.1 74.0 Rooms per capita 0.62 0.72 0.91 1.17 1.74 1.09 Access and Uses of Public Services Electricity for lighting (%) 94.5 96.6 96.9 97.7 99.2 97.2 Access to public water (%) 53.7 59.8 70.3 78.6 88.8 72.1 Access to public sewer (%) 10.4 24.6 29.7 43.7 62.2 36.7 267 ANNEX 15 - continued Social Indicators, Peru Basic Indieators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 US$) -- -- 1,270 1,340 1,350 1,120 940 Social Indicators, Penu 1670 17 1 1400_ 1200 1 i l t 800 _ l . 600 31- o 1970 1975 1980 t985 1990 GNP per capita (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 5.0 3.3 3.9 0.2 for basic social services in GDP Net primary enrollment -- 70.5 77.7 84.1 86.5 97.5 -- Under-five mortality -- -- -- -- 45.2 39.4 -- Child malnutrition -- -- -- -- -- 14.4 13.2 Life expectancy at birth -male -- -- -- -- 55.2 56.8 59.5 - female -- -- -- -- 58.8 60.5 63.4 Total fertility rate -- 6.7 6.0 5.3 4.7 4.2 3.8 Maternal mortality -- -- -- -- 10.8 8.9 -- !ercentage of Urban Population -- 51.9 57.4 61.4 64.5 67.4 70.2 Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All General Characteristics Rural area (%) 60.0 19.7 10.7 6.5 4.1 20.2 Mean household size 5.5 5.6 5.5 5.2 4.5 5.2 Indigenous population(%) 35.4 8.9 5.6 3.6 2.1 11.1 Female-headed household (%) 15.8 17.4 19.3 14.7 13.0 15.9 Age structure of population - children aged 10 and under (%) 32.4 27.5 24.1 19.4 16.7 24.0 - 11-64 (%) 63.2 67.6 72.1 77.1 79.2 71.9 - 65 and older (%) 4.4 4.9 3.8 3.5 4.0 4.1 Dependency ratio 15.8 8.3 6.0 4.5 3.8 6.0 268 ANNEX 15 continued Social Indicators, Peru Quintile Characteristics, 1989 Per Canita Income OuintileA 1 2 3 4 5 All Education Years of schooling among 7.1 8.5 9.3 9.7 11.1 9.4 adults aged 18 and older Illiteracy rate among persons 20.3 11.8 7.8 5.0 3.6 9.1 aged 15 and older (%) Education deficit index 0.32 0.24 0.23 0.20 0.16 0.24 Labor Market Employment category - employees (%) 19.0 39.9 46.3 52.7 56.0 44.3 - self-employed (%) 21.1 40.2 44.1 42.6 40.2 38.1 - employed in farm (%) 58.9 18.0 7.6 3.4 3.0 16.2 - others (%) 1.0 1.9 2.0 1.3 0.8 1.3 Among employees - public sector (%) 28.1 36.4 34.7 32.1 33.3 33.3 - private sector (%) 71.9 63.6 65.3 67.9 66.7 66.7 Access to social security (%) 7.3 23.0 27.0 35.5 41.6 28.4 Second job (%) 8.1 8.6 6.1 5.9 5.1 6.6 Labor force (%) 65.7 56.6 59.7 64.0 66.7 62.7 Unemployment rate (%) 2.6 6.4 5.2 4.1 3.1 4.2 Labor income index (mean= 100) 33.5 54.5 68.2 86.8 166.9 100.0 Housing Characteristics Home ownership (X) 79.8 79.1 73.0 73.0 64.7 73.5 Rooms per capita 0.59 0.67 0.69 0.79 0.98 0.76 Access and Uses of Public Services Electricity for lighting (%) 45.6 78.9 88.8 93.6 95.8 81.2 Access to public water (%) 48.3 72.7 80.9 80.2 89.8 75.0 Access to public sewer (%) 26.7 53.9 71.3 75.6 83.8 63.2 269 ANNEX 15 - continued Social Indicators, Uruguay Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 US$) -- - 2,150 2,210 2,670 2,020 2,460 Social Indicators. Uruguay 2500( 2000 1 _1 " 1970 1975 1980 1985 1990 GNP per capita (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- 13.9 13.6 11.4 15.7 for basic social services in GDP Net primary enrollment -- 85.4 82.8 81.2 83.0 91.0 -- Under-five mortality --- - -- 38.6 30.3 24.9 Child malnutrition - 7.4 -- Life expectancy at birth - male -- -- -- -- 66.4 67.8 68.8 - female -- -- -- -- 73.2 74.3 75.3 Total fertility rate -- 2.4 2.9 2.9 2.7 2.5 2.3 Maternal mortality -- -- -- -- 5.0 4.3 2.8 Percentage of Urban Population -- 81.1 82.1 83.0 83.8 84.6 85.3 Quintile Characteristics, 1989 Per Capita Income Qguintile 1 2 3 4 5 All General Characteristics Mean household size 4.5 3.5 3.2 3.0 2.7 3.3 Female-headed household (%) 25.0 25.2 24.0 24.9 22.6 24.2 Age structure of population - children aged 10 and under (%) 29.3 19.4 15.2 12.8 9.8 17.3 - 11-64 (%) 63.8 64.0 68.2 72.9 76.7 69.1 - 65 and older (%) 6.9 16.6 16.6 14.3 13.5 13.6 Dependency ratio 5.7 4.1 3.3 2.8 2.6 3.4 270 ANNEX 15 - continued Social Indicators, Uruguay Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All Education Years of schooling among 6.1 6.4 7.2 8.4 10.4 7.9 adults aged 18 and older No schooling among persons 5.0 6.0 4.1 2.4 1.0 3.5 aged 15 and older (%) Education deficit index 0.28 0.20 0.15 0.14 0.12 0.20 Labor Market Employment category - employees (%) 67.2 76.4 76.4 76.0 69.3 73.1 - self-employed (%) 26.5 19.9 17.6 16.0 16.2 18.4 - owners of enterprises (%) 0.9 1.3 2.4 4.2 9.1 4.3 - others (%) 5.4 2.4 3.6 3.8 5.4 4.2 Among employees - public sector (%) 18.6 26.2 30.2 33.3 31.8 29.3 - private sector (%) 81.4 73.8 69.8 66.7 68.2 70.7 Informal sector (%) 18.5 14.3 14.9 15.7 15.5 15.6 Labor force (%) 52.5 49.1 54.7 61.6 66.4 57.4 Unemployment rate (%) 16.7 11.1 8.4 5.6 2.8 8.0 Second job (%) 7.0 8.3 9.0 10.5 17.1 11.2 Labor income index (mean= 100) 48.3 66.3 81.3 98.9 160.0 100.0 Housing Characteristics Home ownership (%) 58.7 64.8 67.8 68.9 72.6 67.3 Rooms per capita 0.89 1.18 1.28 1.36 1.69 1.32 Access and Uses of Public Services Electricity for lighting (%) 89.4 95.9 98.6 99.2 99.7 97.2 Access to public water supply (%) 79.4 89.9 93.3 95.7 98.0 92.3 271 ANNEX 15 - continued Social Indicators, Venezuela Basic Indicators 1960 1965 1970 1975 1980 1985 1990 GNP per capita (1987 USS) -- -- -- 3,110 3,360 2,790 2,920 Social Indicators, Venezuela 4000 3000 2500 I _m i _ t75 1980 1985 1990 GNP per capita (in 1987 US$) Social Indicators 1960 1965 1970 1975 1980 1985 1990 Share of public expenditures -- -- -- -- 8.0 7.2 -- for basic social services in GDP Net primary enrollment -- 82.6 88.9 92.6 103.2 105.6 -- Under-five mortality -- -- -- -- 33.8 27.5 -- Child malnutrition -- -- -- -- 10.2 -- - Life expectancy at birth -male - -- -- -- 64.9 66.0 66.7 - female -- -- -- -- 70.7 72.1 72.8 Total fertility rate -- 6.1 5.3 4.7 4.2 3.9 3.6 Maternal mortality -- -- -- -- 6.5 5.8 -- Percentage of Urban Population -- 69.8 72.4 77.8 88.3 81.8 83.7 Quintile Characteristics, 1989 Per Cavita Income 24intile 1 2 3 4 5 ARl General Characteristics Rural area (%) 34.5 20.9 13.9 9.0 4.6 16.6 Mean household size 6.8 6.0 5.4 4.9 3.7 5.2 Female-headed household (%) 24.0 19.3 18.7 19.0 15.3 18.7 Age structure of population - children aged 10 and under (%) 38.7 34.4 29.7 23.7 18.2 28.9 - 11-64 (%) 58.4 62.9 67.4 72.9 78.4 68.0 - 65 and older (%) 2.9 2.7 2.9 3.4 3.4 3.1 Dependency ratio 9.7 6.6 4.7 3.6 2.7 4.5 272 ANNEX 15 - continued Social Indicators, Venezuela Quintile Characteristics, 1989 Per Capita Income Ouintile 1 2 3 4 5 All Education Years of schooling among 6.3 6.8 7.4 8.1 9.9 8.0 adults aged 18 and older Illiteracy rate among persons 18.3 12.4 8.8 6.3 2.9 8.9 aged 15 and older (%) Education deficit index 0.30 0.24 0.20 0.16 0.10 0.21 Labor Market Employment category - employees (%) 53.3 62.1 69.8 70.7 70.7 67.1 - self-employed (%) 34.7 28.2 22.5 20.4 16.1 22.4 - owners of enterprises (%) 2.7 4.5 5.0 6.9 12.0 7.3 - others (%) 9.2 5.1 2.7 2.0 1.3 3.3 Among employees - public sector (%) 14.6 22.6 26.5 30.5 35.4 28.7 - private sector (%) 85.4 77.4 73.5 69.5 64.6 71.3 Informal sector (%) 62.5 49.4 40.0 35.5 29.8 40.0 Labor force (%) 39.7 43.2 48.1 55.6 66.0 51.4 Unemployment rate (%) 6.9 6.0 5.3 4.0 1.9 4.7 Labor income index (mean= 100) 48.9 69.9 82.6 94.1 140.6 100.0 Housing Characteristics Home ownership (%) 82.8 80.0 76.4 73.9 68.8 75.3 Rooms per capita 0.43 0.49 0.55 0.66 0.94 0.66 Access and Uses of Public Services Electricity for lighting (%) 92.6 95.2 97.5 98.1 98.8 96.9 Access to piped running water (%) 81.5 88.0 92.2 94.5 96.9 91.6 ANNEX 16 Bibliography on Poverty and Income Distribution in Latin America TABLE OF CONTENTS Latin America ............ 275 Argentina ............. 283 Bolivia ............ 285 Brazil ............ 285 Chile ............ 292 Colombia ............ 294 Costa Rica ............. 297 Dominican Republic ............ 298 Ecuador ............ 298 El Salvador ............. 300 Guatemala ............ 300 Guyana ............. 300 Honduras ............ 300 Jamaica ............. 301 Mexico ............. 301 Nicaragua ............. 303 Panama ............. 303 Paraguay ............. 303 Peru ............ 304 Uruguay ............. 306 Venezuela ............. 306 LATIN AMERICA Albanez, Teresa et al., "Economic Decline and Child Survival: The Plight of Latin America in the Eighties," UNICEF, Innocenti Occasional Papers no. 1, Florence: 1989. Altimir, Oscar, "La dimension de la pobreza en America Latina," CEPAL, Cuadernos de la CEPAL no. 27, Santiago: 1979. Altimir, Oscar, "Poverty, Income Distribution and Child Welfare in Latin America: A Comparison of Pre- and Post-Recession Data," World Development 12 (Special Issue), 261-82, 1984. Altimir, Oscar, "The Extent of Poverty in Latin America," The World Bank, Staff Working Paper no. 522, Washington: 1982. Altimir, Oscar, "Income Distribution Statistics in Latin America and their Reliability," Review of Income and Wealth 33(2):111-155, 1987. Altimir, Oscar and Mufloz, Oscar, "Distribuci6n del ingreso en America Latina," CLASCO and CIEPLAN, Buenos Aires: El Cid Editor, 1979. Altimir, Oscar and Pinera, S., "Analisis de descomposiciones de las desigualdades de ingreso en America Latina," in 0. Muffoz, La Distribuci6n del Ingreso en America Latina, Buenos Aires: El Cid Editor, 1979. Altimir, Oscar and Sourrouille, Juan V., "Measuring Levels of Living in Latin America: An Overview of the Main Problems," The World Bank, Development Research Center, LSMS Working Paper no. 3, Washington: 1980. Annis, Sheldon and Hakim, Peter, Direct to the Poor: Grassroots Development in Latin America, Boulder: L. Rienner, 1988. Ascher, William, Scheming for the Poor: The Politics of Redistribution in Latin America, Cambridge, MA: Harvard University Press, 1984. Behrman, Jere and Wolfe, Barbara, "Labor Force Participation and Earnings Determinants for Women in the Special Conditions of Developing Countries," Journal of Development Economics 15:259- 288, 1984. Berger, Marguerite and Buvinic, Mayra (eds.), La muier en el sector informal. Trabalo femenino y microempresa en America Latina, Quito: ILDIS-Editorial Nueva Sociedad, 1988. Berry, Albert, "Poverty and Inequality in Latin America," Latin American Research Review 22(2):202-214, 1987. Berry, Albert, "Economic Performance, Income Distribution, and Poverty in Latin America: The Experience of the 1980s," Background Paper for World Development Report 1990, World Bank, Washington (mimeo). 276 ANNEX 16 - continued Berry, Albert, "The Effects of Stabilization and Adjustment on Poverty and Income Distribution: Aspects of the Latin American Experience," Background Paper for World Development Report 1990, World Bank, Washington (mimeo). Bourguignon, Francois, "Optimal Poverty Reduction, Adjustment and Growth: An Applied Framework," The World Bank, Latin America and the Caribbean Region, Human Resources Division, Washington: 1989 (mimeo). Buvinic, Mayra, "Women and Poverty in Latin America and the Caribbean: A Primer for Policy Makers," International Center for Research on Women, Washington: 1990. Carbonetto, Daniel et al., El sector informal urbano en los pafses andinos, Quito: LEPESIU-ILDIS, 1985. Cardoso, Eliana and Helwege, A., "Below the Line: Poverty in Latin America," World Development 20, no. 1(1992): 19-37. Carnoy, Martin and Lobo, Jose, Can Educational Policy Equalize Income Distribution in Latin America?, Saxon House for the IIEP, 1979. Carrizosa, S.M. and Musgrove, Philip, "Ingreso, desigualdad y pobreza en America Latina," ECIEL, Estudio , Rio de Janeiro: 1982 (mimeo). Cauas L., Jorge and Selowsky, Marcelo, "Potential Distributive Effects of Nationalization Policies," The World Bank, Staff Working Paper no. 178, Washington: 1974. Centro Latinoamericano de Administraci6n para el Desarrollo (CLAD), "Desarrollo, transformaci6n y equidad: la superaci6n de la pobreza," Selecci6n de documentos claves 5:9-56, 1988. CEPAL, "Determinacidn de las necesidades de energfa y protefnas para la poblaci6n de nueve pafses latinoamericanos," Santiago: 1988. CEPAL, Nota de la secretarfa, "La dinamica del deterioro social en America Latina y el Caribe en los affos ochenta," CEPAL, Estudios e Informes de la CEPAL, Santiago: 1989. CEPAL, "La pobreza en America Latina: dimensiones y polfticas," CEPAL, Estudios e Informes de la CEPAL no. 54, Santiago: 1985. CEPAL, "Magnitud de la pobreza en ocho pafses de America Latina en 1986," CEPAL, Santiago: 1988. CEPAL, "Magnitud de la pobreza en America Latina en los afhos ochenta," CEPAL, Santiago: 1990 (mimeo). CEPAL, Nota de la secretarfa, "Notas preliminares sobre la situaci6n social y los gastos sociales de pafses seleccionados de America Latina y el Caribe," CEPAL, Estudios e Informes de la CEPAL, Santiago: 1989. 277 ANNEX 16 - continued CEPAL, Nota de la secretarfa, "Opciones y falsos dilemas para los afios noventa: lo nuevo y lo viejo en polftica social en America Latina," CEPAL, Santiago: 1989. CEPAL, Nota de la secretarfa, "Polftica macroeconomica y pobreza," CEPAL, Santiago: 1989. CEPAL,"Estudios sobre la distribuci6n del ingreso en Am6rica Latina," United Nations, Economic and Social Counsel, 1967. CEPAL, "Distribuci6n regional del producto interno bruto sectorial en los pafses de America Latina," 1st ed., Cuadernos Estadfsticos de la CEPAL 6, Santiago: 1981. CEPAL, "Estructura del gasto de consumo de los hogares segdn finalidad del gasto, por grupos de ingreso," 1st ed., Cuadernos Estadfsticos de la CEPAL 0251-9437(8), Santiago: 1984. CEPAL, "Development and Change: Strategies for Vanquishing Poverty," Estudios e Informes de la CEPAL 69, Santiago: 1988. Cline, William R., Potential Effects of Income Redistribution on Economic Growth: Latin American Ca, New York: Praeger Publishers, 1972. Corbo, Vittorio and Stelcner, M., "Earnings Determination and Labor Markets," Journal of Development Economics 12:251-266. Cornelius, Wayne A. and Felicity M. Trueblood, Urbanization and Inequality: The Political Economy of Urban and Rural Development in Latin America, (Latin American Urban Research 5), California: Sage Publications 1975. Cornia, G., Jolly, R. and Stewart, F. (eds.), Adjustment with a Human Face, Oxford: Clarenton Press, 1988. [Ajuste con rostro humano: estudio de pafses, New York: UNICEF, 1987.] Cornia, Giovanni Andrea, "Economic Decline and Child Survival: The Plight of Latin America in the 1980s," International Child Development Centre, UNICEF, Florence: 1988. de Janvry, A. and Sadoulet, E. S, "Investment Strategies to Combat Rural Poverty: A Proposal for Latin America," Department of Agricultural and Resource Economics Working Paper No. 459, University of California at Berkeley, 1988. de Janvry, A. and Sadoulet, E. S., "Rural Development in Latin America: Rethinking Poverty Reduction to Growth," for The World Bank IFPRI Poverty Research Conference, October 1989. de Janvry, A. Sadoulet, E.S., and Wilcox, Linda, "Rural Labour in Latin America," International Labour Office, World Employment Programme, Research Working Paper no. 79, 1986. Dornbusch, Rudiger and Edwards, Sebastian, "Macroeconomic Populism in Latin America," Cambridge, MA: National Bureau of Economic Research, 1989. 278 ANNEX 16 - continued Duke University Center for International Development Research, "Poverty, Conflict and Hope: A Turning Point in Central America," report of the International Commission for Central American Recovery and Development, 1989. Edwards, Alejandra Cox, "Wage Trends in Latin America," World Bank, Latin American Technical Department, 1991. Fallon, Peter R. and Riveros, Lufs A., "Adjustment and the Labor Market," The World Bank, Country Economics Department, PPR Working Paper no. 214, 1989. FAO, Regional Office for Latin America and the Caribbean: "Rural Poverty in Latin America and the Caribbean: An Essay of Diagnosis," FAO, Rome: 1984. FAO, "Rural Poverty in Latin America and the Caribbean," FAO, Rome: 1988. Fields, Gary S., "Changes in Poverty and Inequality in Developing Countries," The World Bank Research Observer 4(2): 167-85, 1989. Fields, Gary, "Crecimiento, distribucion del ingreso y pobreza en America Latina: algunos hechos estilizados," IDB, Washington: 1992 (mimeo). Fields, Gary, "Poverty, Inequality, and Economic Growth," in G. Psacharopoulos (ed.), Essays on Povert.= Equity and Growth, Oxford: Pergamon Press, 1991. Fields, Gary, "Poverty, Income Distribution and Child Welfare in Latin America: A Comparison of Pre- and Post-recession Data," World Development, 1984. Figueiredo, J., R. Frenkel, P. Meller and G. Rozenwurcel, "Empleo y salarios en America Latina," ECIEL, Rio de Janeiro: 1985. Figueroa, Adolfo and Richard Weisskoff, "Viewing Social Pyramids: Income Distribution in Latin America," Yale University, Economic Growth Center Discussion Paper no. 204, New Haven: 1974. Foxley, Alejandro, Income Distribution in Latin America, Cambridge, New York: Cambridge University Press, 1976. Gilbert, Alan and Peter M. Ward, Housing. the State. and the Poor: Policy and Practice in Three Latin American Cities (Mexico City. Mexico: Bogota. Colombia: and Valencia. Venezuela), New York: Cambridge University Press, 1985. Grosh, Margaret, "From Platitudes to Practice: Targeting Social Programs in Latin America" I & II, The World Bank, Technical Department, Latin America and the Caribbean Region, Human Resources Division, Report no. 10720-LAC, yellow copy, Washington: 1992. Harrison, Paul, Inside the Third World: The Anatomy of Poverty (Africa. Asia. Latin America), Harvester Press, 1980. 279 ANNEX 16 - continued Helleiner, Gerald, "Stabilization Policies and the Poor," University of Toronto, Department of Economics Working Pager no. B.9, Toronto: 1985. ILO, Programa Regional del Empleo para America Latina y el Caribe, "Meeting the Social Debt," Santiago: 1988. ILO, Programa Regional del Empleo para America Latina y el Caribe, "Profiles of Rural Poverty: A Popularized Version of Poverty and Landlessness in Rural Asia, with Additional Material on Africa and Latin America," International Labour Office, Geneva: 1979. Kanbur, R., "Malnutrition and Poverty in Latin America," University of Warwick, Economic Research Poper no. 278, Coventry: 1987. Keith, Nelson W. and Keityh, Novella Zett, New Perspectives on Social Class and Socioeconomic Development in the Periphery, New York: Greenwood Press, 1988. Lassen, Cheryl A., "Landlessness and Rural Poverty in Latin America: Conditions, Trends and Policies Affecting Income and Employment," Cornell University, Center for International Studies, Rural Development Committee, Ithaca: 1980. Latorre, Carmen Luz and Yonemura, Akio, "Formation of Urban Low Income Class and Education: Chile and Mexico," Institute of Developing Economies, JRP Series no. 59, Tokyo: 1986. Looney, Robert E., Income Distribution Policies and Economic Growth in Semi-Industrialized Countries: A Comparative Study of Iran, Mexico. Brazil. and South Korea, New York: Praeger Publishers, 1975. L6pez, Ramdn E. and Riveros, Luis A., "Macroeconomic Adjustment and Labor Market Structure in Four Latin American Countries: An Econometric Study," The World Bank, Washington: 1989. Mackenzie, G.A., "Social Security Issues in Developing Countries: The Latin American Experience," International Monetary Fund, Fiscal Affairs Department, IMF Working Paper WP/88/21, Washington: 1988. Mesa-Lago, Carmelo, "Social Security and Extreme Poverty in Latin America," Journal of Development Economics 12:83-110, 1983. Molina, Sergio, "Poverty: Description and Analysis of Policies for Overcoming It," CEPAL Review 18:87-110, 1982. Mufloz G., and Altimir, Oscar, Distribuci6n del ingreso en America Latina, Consejo Latinoamericano de Ciencias Sociales and Corporaci6n de Investigaciones Econ6micas para America Latina, Buenos Aires: El Cid Editor, 1979. Morley, Samuel A., "Macroconditions and Poverty in Latin America," IDB, Washington: 1992 (mimeo). 280 ANNEX 16 - continued Musgrove, Philip, "Desigualdad en la distribucidn del ingreso en diez ciudades sudamericanas: descomposicion e interpretaci6n del coeficiente de Gini," Cuadernos de Economfa 23:201-27, Chile: 1986. Musgrove, Philip, "Feeding Latin America's Children," The World Bank, Technical Department, Latin America and the Caribbean Region, Human Resources Division, Washington: 1991. Musgrove, Philip, "Food Needs and Absolute Poverty in Urban South America," Review of Income and Wealih 1:63-83, 1985. Musgrove, Philip, "The ECIEL Study of Household Income and Consumption in Urban Latin America: An Analytical History," The World Bank, Development Research Center, LSMS Working Paper no. 12, Washington: 1982. Paldam, Martin and Riveros, Lufs, "The Causal Role of Minimum Wages in Six Latin American Labor Markets," The World Bank, Discussion Paper, Washington: 1987. Peek, P., "Rural Poverty in Central America: Dimensions, Causes and Policy Alternatives," ILO, World Employment Programme, Research Working Paper, Geneva: 1984. Petrei, A. Humberto, "El gasto piblico social, sus aspectos distributivos: un examen comparativo para cinco pafses de America Latina," ECIEL, Serie Documentos no. 6, 1987. Pfeffermann, Guy Pierre, "Poverty in Latin America: The Impact of Depression," The World Bank, Washington: 1986. Pfeffermann, Guy Pierre, "Economic Crisis and the Poor in Some Latin American Countries," Finance and Development 32-35, 1987. Pfeffermann, Guy Pierre, "Public Expenditure in Latin America: Effects on Poverty," The World Bank, Discussion Paper no. 5, Washington: 1987. Pfeffermann, Guy Pierre and Griffin, C., Nutrition and Health Programs in Latin America: TargetinF Social Expenditures, The World Bank in association with the International Center for Economic Growth, 1989. Pinto de la Piedra, Matilde, "El componente social del ajuste econ6mico en America Latina," for "Seminario de alto nivel: C6mo recuperar el progreso social en America Latina," sponsored by UNICEF, ILPTES, and IDE, Chile:1988. PNUD, "Magnitud y Evolucion de la Pobreza en America Latina," Proyecto Regional para la Superacion de la Pobreza, Comercio Exterior, 380-392. Pollitt, Ernesto, Halperin, Robert, and Eskenasy, Patricia, Poverty and Malnutrition in Latin America: Early Childhood Intervention Programs: A Report to the Ford Foundation, New York: Praeger, 1980. 281 ANNEX 16 - continued Portes, Alejandro, Castells, Manuel and Benton, Lauren A., The Informal Economv. Studies in Advanced and Less Developed Countries, Baltimore: The Johns Hopkins University Press, 1989. Portes, Alejandro and John Walton, Urban Latin America: The Political Condition from Above and Below,' Austin: University of Texas Press (Pan-American Series), 1970. PREALC, "Pobreza y mercado de trabajo en cuatro pafses: Costa Rica, Venezuela, Chile y Perd," PREALC, Working Paper no. 309, Santiago:1987. PREALC, "Interrelaciones entre poblaci6n y desarrollo. Bases para polfticas de poblacion en el Istmo Centroamericano," PREALC, Working Paper no. 339, Santiago: 1989. Psacharopoulos, George, "Poverty Alleviation in Latin America," Finance and Development, 27(1):17-19, 1990. Psacharopoulos, George (ed.), Essays on Poverty. Equity. and Growth, Oxford: Pergamon Press, 1991. Psacharopoulos, George, "Recovering Growth with Equity in Latin America," The World Bank Internal Discussion Paper no. 33, Latin America and the Caribbean Region, 1989. Psacharopoulos, G., Lee, H. and Wood, W., "Poverty and Income Distribution in Latin America and the Caribbean: An Update", The World Bank, Latin American Technical Department, 1992 (mimeo). Redclift, M., "Urban Bias and Rural Poverty: A Latin American Perspective," Journal of Development Studies 20:123-138, 1984. Riveros, Lufs A., "Recession, Adjustment and the Performance of Urban Labor Markets in Latin America," The World Bank, Macroeconomic Adjustment and Growth Division, Washington: Washington: 1989. Robbins, Donald, "Real Minimum Wages and Adjustment in Six Latin American Countries," The World Bank, Latin American Technical Department, Human Resources Division, 1989. Rodgers, Gerry (ed.), Urban Poverty and the Labor Market, Geneva: International Labour Office, 1989. Rosenthal, Gert, "Some Thoughts on Poverty and Recession in Latin America," Journal of Interamerican Studies and World Affairs, 1989. Sachs, Jeffrey D., Social Conflict and Populist Policies in Latin America, NBER Working Paper No. 2897, Cambridge MA: 1989. Selowsky, Marcelo, "Balancing Trickle Down and Basic Needs Strategies: Income Distribution Issues in Large Middle-Income Countries with Special Reference to Latin America," The World Bank, Staff Working Paper No. 335, Washington: 1979. 282 ANNEX 16 - continued Shaull, Millard Richard, Heralds of a New Reformation: The Poor of South and North America, Maryknoll, NY: Orbis Books, 1984. Sheahan, John, Patterns of Development in Latin America: Poverty. Repression. and Economic Strate=, Princeton, NJ: Princeton University Press, 1987. Tockman, Victor E., "Adjustment and Employment in Latin America: The Current Challenges,' International Labor Review 125:5, 1986. Tockman, Victor E., "Crisis, ajuste econ6mico y costo social," Trimestre econ6mico 53(1), 1986. Tockman, Victor E., "Distribucidn del ingreso, tecnologfa y empleo: analisis del sector industrial en el Ecuador, Perd y Venezuela," Instituto Latinoamericano de Planificaci6n Econ6mica y Social, Santiago: 1975. Tockman, Victor E., "El sector informal: quince afnos despuds," El trimestre econ6mico 54(3), Mdxico: 1987. Twomey, Michael J., "Devaluations and Income Distribution in Latin America," Southern Economic -oumnal 49(3):804-21, 1983. UNDP, "Conferencia regional sobre la pobreza en America Latina y el Caribe," PNUD Projecto RLA/86/004, Bogota: 1988. UNICEF, "Urban Basic Services: Reaching Children and Women of the Urban Poor: Report," New York: United Nations, 1982. United Nations, "Update on the Nutritional Situation: Recent Trends in Nutrition in 33 Countries," 1989. United Nations, Statistical Office, "National Accounts Statistics: Compendium of Income Distribution Statistics," Series M(79), New York: 1985. van de Walle, Dominique, "Poverty and Inequality in Latin America and the Caribbean During the 70s and 80s: An Overview of the Evidence," The World Bank, Technical Department, Latin America and the Caribbean Region, Human Resources Division, A View from LATHR no. 22, Washington: 1991. van Ginneken, Wouter and Park, Jong-goo, Generating Internationally Comparable Income Distribution Estimates, International Labour Office, Geneva: 1984. Weisskoff, Richard, "Income Distribution and Economic Growth in Puerto Rico, Argentina, and Mexico," Yale University, Economic Growth Center, Discussion Paper, no. 93, New Haven: 1970. Weisskoff, Richard, "Equity, Efficiency and Social Welfare: A Comparison of Latin American Areas," Yale University, Economic Growth Center, Discussion Paper, no. 212, New Haven: 1970. 283 ANNEX 16 - continued World Bank, "Targeted Programs for the Poor During Structural Adjustment: A Summary of a Symposium on Poverty and Adjustment," Washington: 1988. World Bank, World Development Report. Poverty, New York: Oxford University Press, 1990. World Bank, World Development Report 1991, New York: Oxford University Press, 1991. Yamada, Gustavo, "Adjustment and the Social Sectors in Latin America," The World Bank, Technical Department, Latin America and the Caribbean Region, Human Resources Division, Washington: 1989. ARGENTINA Aguirre Dubarry, Alejandro J., "Los procesos inflacionarios como instrumento de la polftica econ6mica," Boletfn de la Direccidn General Impositiva 65:9-18, Argentina: 1986. Altimir, Oscar, "Estimaciones de la distribuci6n del ingreso en la Argentina, 1953-1980," Desarrollo Economico 25:521-66, Argentina: 1986. Antelo, Roberto, "Medida de la desigualdad en la distribucion," Banco Central de la Repiblica Argentina, Centro de Estudios Monetarios y Bancarios, Buenos Aires: 1978. Beccaria, Luis A., and Riquelme, Graciela C., El gasto social en educaci6n y la distribuci6n del ingreso, Colecci6n FLASCO, Serie Documentos e Informes de Investigaci6n no. 41, Buenos Aires: 1985. Beccaria, Luis A. and Riquelme, Graciela C., Efectos distributivos del gasto pdblico en la educaci6n piblica y privada, Colecci6n FLASCO, Buenos Aires: 1985. Beccaria, Luis A. and Martfnez, E. La influencia de la educaci6n en la distribuci6n del ingreso. Un analisis exploratorio, Buenos Aires: Instituto Nacional de Estadfstica y Censos, 1985. CEPAL, "La distribuci6n personal del ingreso en el gran Buenos Aires en el perfodo 1974-1983," Documento de Trabaio no. 23, Santiago: 1986. CEPAL, "Economic Development and Income Distribution in Argentina," New York: 1969. CEPAL, "Argentina: Canasta basica de alimentos y determinacidn de las lfneas de indigencia y de pobreza," Santiago: 1988. CEPAL, "Antecedentes estadfsticos de la distribuci6n del ingreso, Argentina 1953-1982," Seri Distribuci6n del Ingreso no. 5, Santiago: 1987. Dieguez, Hector, "Social Consequences of the Economic Crisis, Argentina," report prepared for the World Bank, Buenos Aires: 1986. 284 ANNEX 16 - continued Dieguez, Hector and Petrecolla, A., "Distribuci6n de ingresos en el gran Buenos Aires," in P. Musgrove (ed.), Ingreso desigualdad y Dobreza en America Latina, ECIEL, Rio de Janeiro: 1982. FIDE, "La distribuciondel ingreso entre 1974 y 1982," Coyuntura y Desarrollo 60:33-44, Buenos Aires: 1983. Fiszbein, Ariel, "Essays on Labor Markets and Income Inequality in Less Developed Countries," Ph.D. Dissertation, University of California, Berkeley: 1991. Gertel, H., de Sanctis, M. and Pereyra, L., "Educaci6n y distribucidn de ingresos en la ciudad de C6rdoba," Anales de la XXI Reuni6n de la Asociaci6n de Econdmica Polftica, Cordoba, Argentina: 1987. INDEC (Instituto Nacional de Estadfstica y Censos), La pobreza urbana en la Argentina, Buenos Aires: 1990. Mann, Arthur J., "Multivariate Analysis of Argentina's Urban Income Distribution, 1974-1975 and 1978," Economics Letters 10:251-5, the Netherlands: 1982. Marshall, Adriana, "Income Distribution, the Domestic Market and Growth in Argentina," Labour and Society, 13:79-103, International Institute for Labour Studies, 1988. Minujin, Alberto and Vinocur, Pablo, "LQuienes son los Pobres del Gran Buenos Aires?," Comercio Exterior 42(4):393-401, Mexico, 1992. Morley, Samuel A. and Alvarez, C., "Recession and the Growth of Poverty in Argentina," IDB, Washington: 1991 (mimeo). Orsatti, Alvaro, "La nueva distribuci6n funcional del ingreso en la Argentina," Desarrollo econ6mico 23:315-37, 1983. Orsatti, Alvaro, and Mann, Arthur, J., "Desigualdades regionales e ingresos familiares en la Argentina," Desarrollo econ6mico 26:289-314, 1986. Palomino, Hector, "Reflexiones sobre la evoluci6n de las clases medias en la Argentina," El Bimestre Polftico y Econ6mico, 42-43, 1989. Petrei, A. Humberto, "El gasto piblico social y sus efectos distributivos: un examen comparativo de cinco pafses de America Latina," ECIEL, Rio de Janeiro: 1987. Rodrfguez, Carlos A., "Inflaci6n, salario real y tipo real de cambio," Centro de Estudios Macroecon6micos de Argentina, Documento de Trabajo no. 41:1-21, 1984. Vinocur, Pablo and L6pez, N., "Evolucidn de la pobreza en la Argentina, (1970-1984)," for the Inter- American Dialogue, Buenos Aires: 1990. 285 ANNEX 16 - continued Welna, David, "Housing Solutions for Buenos Aires' Invisible Poor," Grassroots Development, 1988. World Bank, "Argentina, Social Sectors in Crisis," a World Bank Country Study, Washington: 1988. BOLIVIA Boltvinik, Julio, "Necesidades blsicas y pobreza, conceptos y metodos de medici6n" in Calidad de vida y pobreza urbana, Centro Latinoamericano de Encuestas Sociales, 1990. LeBaron, Allen, Brown, Bruce and Ortiz, Rail, "Estimates of the Distribution of Urban and Rural Family Incomes in Bolivia," Consortium for International Development, La Paz: 1976. Newman, John, Jorgensen, S. and Pradhan, M., "How Did Workers Benefit from Bolivia's Emergency Social Fund?", The World Bank, Washington: 1990. World Bank, "Country Assessment of Women's Role in Development: Proposed Bank Approach and Plan of Action," R.N.8064-BO, Latin America and the Caribbean Regional Office, Country Operations Division 1, 1989. BRAZIL Andrade, Thompson Almeida and Lodder, Celsius A., "Sistema Urbano e Cidades Medias no Brasil," IPEA/INPES, Rio de Janeiro: 1979. Araquem da Silva, Ednaldo, "A RelacZo Salario-lucro no Brasil: Analise de Insumo-Produto, 1970 e 1975," Revista Brasileira de Economia 42:3-12, 1988. Bacha, Edmar Lisboa and Klein, Herbert S., "A Transicao Incompleta: Brasil desde 1945," Paz e Terra 2, Rio de Janeiro: 1986. Bacha, Edmar Lisboa and Taylor, L., "Brazilian Income Distribution in the 1960s: Facts, Models, Results and the Controversy," Journal of Development Studies 14(3):272-96, 1978. Behrens, Alfredo, "Energy and Output Implications of Income Redistribution in Brazil," Energy Economics 6:110-16, UK: 1984. Behrman, J.R. and Birdsall, Nancy, "Quality of Schooling: Quantity Alone Is Misleading," American Economic Review 73:5, 1983. Behrman, J.R. and Birdsall, Nancy, "The Equity-Productivity Tradeoff: Public School Resources in Brazil," European Economic Review, 1988. 286 ANNEX 16 - continued Birdsall, Nancy, "Public Inputs and Child Schooling in Brazil," Journal of Development Economics 18:67-86, 1985. ["Comment," in Are World Population Trends a Problem?, B. Wattenburg and K. Zinsmeister (eds.), Washington: American Enterprise Institute, 1985.] Birdsall, Nancy and Fox, M.L., 'Why Males Earn More: Location and Training of Brazilian Schoolteachers," Economic Development and Cultural Chan2e 33(3), 1985. Birdsall, Nancy and Meesook, O.A., "Children's Education and the Intergenerational Transmission of Inequality: A Simulation," Economics of Education Review 5(3), 1986. Birdsall, Nancy and C. Griffin, "Poverty and Rapid Population Growth in Poor Countries," Journalof Policy Modeling, 1988. Bonelli, Regis and da Cunha, Paulo Vieira, "Crescimento Econ6mico, Padrao de Consumo e Distribuigao da Renda no Brasil: Uma Abordarem Multissetorial para o Perfodo 1970/75," Pesguisa Planejamento Economico 11:703-56, Brasil: 1981. Bonelli, Regis and da Cunha, Paulo Vieira, "Distribuicao de Renda e Padr6es Crescimento: Um Modelo Dinamico daEconomia Brasileira," Pesguisa ePlanejamento Econ6mico 13:91-154, Brasil: 1983. CEPAL, "Antecedentes estadfsticos de la distribuci6n del ingreso, Brasil 1960-1983," Serie Distribucion del Ingreso no. 2, Santiago: 1986. CEPAL, "Brasil: Canastas basicas de alimentos y determinaci6n de las lfneas de indigencia y de pobreza," Santiago: 1989. Clements, Benedict J., "Foreign Trade Strategies, Employment, and Income Distribution in Brazil," New York: Praeger, 1988. Clements, Benedict J. and Kim, Kwan S., "Comercio Exterior e Distribuigao de Renda: 0 Caso Brasileiro," Pesquisa e Planejamento Economico 18:17-36, Brasil: 1988. da Silva, Ednaldo Araquem, "Pregos e Distribuicao da renda no Brasil: Uma Analise de Insumo-Produto 1975," Pesguisa e Planejamento Economico 18:361-78, Brasil: 1988. da Silva, Fernando Antonio Rezende and Mahar, Dennis J., Sadde e Previdencia Social: Uma Anglise Econ6mica," IPEA/INPES, Rio de Janeiro: 1974. de Almeida, Anna Luiza Ozorio, "Distribuiqao de Renda e Emprego em Servigos," IPEA/INPES, Rio de Janeiro: 1976. de Carvalho, Jose Alberto and Wood, Charles H., "Crescimento Populacional e Distribuigao da Renda Familiar: 0 Caso Brasileiro," Estudos EconOmicos 11:5-25, Brasil: 1981. de Carvalho, Jose Alberto and Wood, Charles H., "Mortality, Income Distribution and Rural-Urban Residence in Brazil," Population and Development Review 4:405-20, Brazil: 1978. 287 ANNEX 16 - continued del Campo, Carlos Patricio, "Is Brazil Sliding Toward the Extreme Left?: Notes on the Land Reform Program," New York: The Society, 1986. Denslow, David, Jr. and Tyler, William G., "Perspectives on Poverty and Income Inequality in Brazil: An Analysis of the Changes During the 1970s," The World Bank, Staff Working Paper no. 601, Washington: 1983. Denslow, David, Jr. and Tyler, William G., "Perspectives on Poverty and Income Inequality in Brazil," World Development 12:1019-28, UK: 1984. do Valle, Silva N., "Os Deserdos do Milagre," Laboratorio de Computacio Cientffica, Rio de Janeiro: 1987 (mimeo). Drobny, Andres and Wells, John, "Wages, Minimum Wages, and Income Distribution in Brazil: Results from the ConstructionIndustry," Journal of Development Economics 13:305-30, the Netherlands: 1983. Fields, G., "Who Benefits from Economic Development? A Reexamination of Brazilian Growth in the 1960s," The American Economic Review 67(4):570-82, 1977. Fields, Gary, "Who Paid the Bill? Adjustment and Poverty in Brazil 1980-95, The World Bank, SIaff Workinf Paper no. 648, Washington: 1991. Fishlow, A., "Brazilian Size Distribution of Income," The American Economic Review 62(2):391-402, 1972. Fox, M. Louise, "Income Distribution in Post-1964 Brazil; New Results," World Bank Reprint Series, 1983, reprinted from Journal of Economic History 43(1) 1983. Fox, M. Louise, "Poverty Alleviation in Brazil, 1970-87," The World Bank, Latin America and the Caribbean Regional Office, Country Department I, Discussion Paper no. IDB-072, 1990. Fox, M. Louise and Morley, Samuel A., "Who Paid the Bill? Adjustment and Poverty in Brazil, 1980-1985," The World Bank, Washington: 1990. Fox, M. Louise and Morley, Samuel A., "Who Carried the Burden of Brazilian Adjustment in the Eighties?" The World Bank, Washington: 1991 (mimeo). Hakkert, Ralph, "Who Benefits from Economic Development? The Brazilian Income Distribution Controversy Revisited," Boletfn de Estudios Latinoamericanos y del Caribe, 36:83-103, the Netherlands: 1984. Hicks, James F. and Vetter, David Michael, "Identifying the Urban Poor in Brazil," The World Bank, Staff Working Paper no. 565, Washington: 1983. 288 ANNEX 16 - continued Hoffman, Rudolfo, "Evolugao da Distribuicao da Renda no Brasil, Entre Pessoas e entre Famflias, 1979-1986," in Guilherme Luis Sedlacek and R. de Barros (eds.), Mercado de Trabalho e Distribuiclo da Renda: Uma Coletanea, IPEA/INPES, Rio de Janeiro: 1989. Hoffman, Rudolfo and Kassouf, Ana Lucia, "Modernizacao e Desigualdade na Agricultura Brasileira," Revista Brasileira de Econ6mica 43:273-303, Brazil: 1989. Homem de Melo, Fernando, "New Technologies and Income Distribution: The Case of Brazil," CERES. FAO Review on Agriculture and Development, 19:15-19, 1986. Iten, Oswald, "Self-Help in the Slums of Brazil," Swiss Review of World Affairs 38:14-19, Switzerland, 1988. Jallade, Jean Pierre, "Basic Education and Income Inequality in Brazil: The Long-term View," The World Bank, Staff Working Paper no. 268, Washington: 1977. Juve, Luis Chiodo, "Food Intervention Planning: Distribution in Brazil's Favelas," Food Policy 15:227-38, Brazil: 1990. Knight, Peter T. and Moran, Ricardo, "Bringing the Poor into the Growti Process: The Case of Brazil," Finance and Development 18(4):22-25, 1981. Knight, P., Mahar, D. and Moran, R., "Brazil, Human Resources Special Report," World Bank, a Couniy Study, 1979. Langoni, Carlos Geraldo, Distribuiclo da Renda e Desenvolvimento Economico do Brasil, Rio de Janeiro: Editora Expressao e Cultura, 1973. Langoni, Carlos Geraldo, "Income Distribution and Economic Development in Brazil, "Conjuntura Economica 27, Rio de Janeiro: BNH Information Office, 1973. Langoni, Carlos Geraldo, "Review of Income Distribution Data: Brazil," in C.R. Frank and R.C. Webb (eds.), Income Distribution and Growth in Less-Developed Countries: 114-32, The Brookings Institute, Washington: 1975. Lluch, Constantino, "Pobreza e Concentragao da Renda no Brasil," Pesguisa e Planejamento Economico 11:757-81, Brazil: 1981. Lluch, Constantino, "Sobre Medic6es da Renda a Partir dos Censos e das Contas Nacionais no Brasil," Pesguisa e Planejamento Economico 12:133-48, Brasil: 1982. Locatelli, Ronaldo Lamounier, "Efeitos Macroecon6micos de uma Redistribuicao da Renda: Um Estudo para o Brasil," Pesguisa e Planejamento Economico 15:139-70, Brasil: 1985. Lodder, Celsius A., "Distribuigao da Renda nas Areas Metropolitanas," IPEA/INPES, Rio de Janeiro: 1976. 289 ANNEX 16 - continued Macedo, Roberto,"Salgrio Mfnimo e Pobreza no Nordeste," Banco do Nordeste do Brasil 13:241-82, Brasil: 1982. Macedo, Roberto, "Brazilian Children and the Economic Crisis: Evidence from the State of Sao Paulo," World Development 12(3):203-21, 1984. Macedo, Roberto, "Brazilian Children and the Economic Crisis: Evidence from the State of Sao Paulo," in G. Cornia, R. Jolly and F. Stewart (eds.), Adjustment with a Human Face:28-56, Oxford: Clarenton Press, 1988. Maddison, A. and Associates, "The Political Economy of Poverty, Equity and Growth: Brazil and Mexico" 1989 (mimeo). Maia Gomes, Gustavo, et al., "Polfticas Recessivas, Distribuicao da Renda e os Mercados Regionais de Trabalho no Brasil: 1983-1984," Pensamento lberoamericano 10:261-82, Spain: 1986. Martine, George, "Formaci6n de la familia y marginalidad urbana en Rio de Janeiro," Centro Latinoamericano de Demograffa (CELADE), Series E no. 16, Santiago: 1975. Mata, Milton da, "Concentracao da Renda, Desemprego e Pobreza no Brasil," IPEA/INPES, Rio de Janeiro: 1979. Morley, Samuel A., Labor Markets and Inequitable Growth: The Case of Authoritarian Capitalism in Brazil, Cambridge (UK) and New York: Cambridge University Press, 1982. Morley, Samuel A. and Smith, Gordon Whitford, "The Effect of Changes in the Distribution of Income on Labor, Foreign Investment, and Growth in Brazil," Rice University, Program of Development Studies, Houston: 1971. Morley, Samuel A. and Williamson, Jeffrey, G., "The Impact of Demand on Labor Absorption and the Distribution of Earnings: The Case of Brazil," Rice University, Program of Development Studies, Houston: 1973. Musgrove, Philip, "Por una mejor alimentaci6n: evaluaci6n de programas destinados a mejorar el consumo alimentario y el estado nutricional de familias pobres en Brasil," PAHO, Washington: 1989. Musgrove, Philip and Osmil Galindo, "Do the Poor Pay More? Retail Food Prices in Northeast Brazil," Economic Development and Cultural Change 37:91-109, 1988. Mota, Roberto and Scott, Parry, "Sobrevivencia e Fontes de Renda: Estrategias das Famflias de Baixa Renda no Recife," Fundacao Joaquim Nabuco, Serie Populaclo e Emprego no. 16, Recife: Editora Massangana, 1983. Norris, William P., "The Social Networks of Impoverished Brazilian Women: Work Patterns and Household Structure in an Urban Squatter Settlement," Michigan State University, Office of Women in International Development, East Lansing: 1985. 290 ANNEX 16 - continued Oxfam, "An Unnatural Disaster: Drought in North East Brazil," Oxford: 1984. Pastore, Jose, Zylberstajn, Helio, Pagotto, Carmen Silvia and Fundacao Instituto de Pesquisas Economicas, "Mudanga Social e Pobreza no Brasil, 1970-1980: 0 que Ocorreu com a Famflia Brasileira?", Sao Paulo: Livraria Pioneira Editora for Fundacao Instituto de Pesquisas Econ6micas, 1983. Perlman, Janice E., The My of Marginality: Urban Poverty and Politics in Rio de Janeiro, Berkeley: University of California Press, 1976. Pfeffermann, Guy, "The Social Cost of Recession in Brazil," The World Bank, internal paper, Washington: 1986. Pfeffermann, Guy, and Webb, Richard Charles, "The Distribution of Income in Brazil," The World Bank, Staff Working Paper no. 356, Washington: 1979. Pfeffermann, Guy, and Webb, Richard Charles, "Pobreza e Distribuicao da Renda no Brasil: 1960-1980," Revista Brasileira de Economia 37:147-75, Brazil: 1983. Pfeffermann, Guy, and Webb, Richard Charles, "Poverty and Income Distribution in Brazil," Review of Income and Wealth 29:101-24, 1983. Also issued in World Bank Reprint Series no. 259:101-24, 1987. Ramos, L., The Distribution of Earnings in Brazil: 1976-1985, Ph.D. Dissertation, University of California, Berkeley: 1990. Rezende, Gervasio Castro, "Food Production, Income Distribution and Prices: Brazil 1960-80," ILO, Rural Employment Policy Research Programme, Working Paper no. 10-6/WP89: 1-86, Geneva: 1989. Rios, Jose Arthur, "Invisible Economy of Poverty: The Case of Brazil," Mondes en Developpement 12(45):65-77, France: 1984. Roberge, Roger A., "National Urbanization Strategies and Urban Poverty in Brazil: An Analysis of Variations in the Urban Hierarchy," Institute for International Development and Cooperation, University of Ottawa, Working Paper No. 851:1-43, 1985. Rodrfguez, Octavio, "Agricultura, subempleo y distribucion del ingreso; notas sobre el caso brasileflo," Economfa de America Latina 13:63-77, Mexico: 1985. Rossi, Jose W., Indices de Desigualdade da Renda e Medidas de Concentracao Industrial: Aplicacao a Casos Brasileiros, Rio de Janeiro: Zahar Editores, 1982. Rossi, Jose W., "Decomposigao Funcional do indice de Gini com Dados de Renda do Brasil," Revista Brasileira de Economia 37:337-48, Brasil: 1983. 291 ANNEX 16 - continued Rossi, Jose W., "Notas Sobre uma Nova Decomposicao do fndice de Gini," Pesuuisa e Planeiamento Econdmico 15:241-48, Brasil: 1985. Rossi, Jose W., "Notes on a New Functional Form for the Lorenz Curve," Economic Letters 17(1/2):193-7, the Netherlands: 1985. Sadoulet, Elisabeth, "Croissance inegalitaire dans une economie sous-d6veloppee," Geneva: Libr. Droz, 1983. Sahota, Gian S. and Rocca, Carlos Antonio, Income Distribution: Theory. Modeling, and Case Study ot Brazil, 1st ed., Ames, IA: Iowa State University Press, 1985. Sampaio, Yony de Sa Barreto and Irmao, Jose Ferreira, "Emprego e Pobreza Rural: Uma Visao Critica da Teoria e Aplicacao ao Caso de Pernambuco," Curso de Mestrado em Economia-CME/PIMES, Central de Ciencias Sociais Aplicadas, Universidade Federal de Pernambuco, Recife: 1977. Sant'Anna, Anna Maria, Merrick, Thomas William and Mazumdar, Dipak, "Income Distribution and the Economy of the Urban Household: The Case of Bello Horizonte," The World Bank, Staff Working Papers no. 237, 1976. Sedlacek, Guilherme Lufs and de Barros, Ricardo Paes, Mercado de Trabalho e Distribuicao da Renda: Uma Coletanea, Rio de Janeiro: IPEA/INPES, 1989. Taylor, Lance, Models of Growth and Distribution for Brazil, The World Bank, New York: Oxford University Press, 1980. Thomas, V., "Differences in Income, Nutrition and Poverty within Brazil," The World Bank, S_taff Working Paper no. 505, Washington: 1982. Thomas, V., "Differences in Income and Poverty within Brazil," World Development 15:263-73, UK: 1987. Toloso, H.C., "Pobreza no Brasil: Uma avaliacao dos anos 80," in J. P. dos Reis Velloso, ed., A Ouestio Social no Brasil, Sao Paulo: Nobel, 1991. Willumsen, Maria Jose F., "Impact of Brazilian Production Structure on Income Distribution," Florida International University, Department of Economics, Discussion Papers in Economics and Banking no. 32:1-27, 1985. Wood, Charles H. and de Carvalho, Jose Alberto Magno, The Demography of Inequality in Brazil, New York: Cambridge University Press, 1988. Wood, Charles H. and McCracken, Stephen D., "Underdevelopment, Urban Growth and Collective Social Action in Sao Paulo, Brazil," Studies in Third World Societies 29:101-40, 1984. World Bank, "Brazil, Public Spending on Social Programs: Issues and Options," report no. 70860-BR, Washington: 1988. 292 ANNEX 16 - continued World Bank, "Country Assessment of Women's Role in Development in Brazil," report no. 8043-BR (1 & II), Latin America and the Caribbean Region, Brazil Department, Population and Human Resources Operations Division, Washington: 1989. CHLE Castahleda, Tarsicio, Innovative Policies for Structural Reform in Social Sectors: Chile in the 1980s, Chapters 2, 4 and 5. De Kadt, Emanuel Jehuda, Bienestar v pobreza, Centro de Estudios de Planificaci6n Nacional, Universidad Catolica de Chile, Vicerrectoria de Comunicaciones, Santiago: Ediciones Nueva Universidad, 1974. Foxley, Alejandro and Raczynski, D., 'Vulnerable Groups in Recessionary Situations: The Case of Children and the Young in Chile," World Development, 12(3):233-46, 1984. Foxley, Alejandro, Arellano, Eduardo and Pablo, Jose, "The Incidence of Taxation," ILO, World Employment Programme, Research Working Paper WEP 2-23/WP 51, Chile & Geneva: Labour Office, 1977. Foxley, Alejandro, Arellano, Eduardo, and Pablo, Jose, Redistributive Effects of Government Programmes: The Chilean Case, Geneva & Oxford: International Labour Office (New York: Pergamon Press), 1979. Haindl R., Erik and Weber, Carl S., "Impacto redistributivo del gasto social," Universidad de Chile, Departamento de Economfa, Facultad de Ciencias Econdmicas y Administrativas, Documento Serie de Investigaci6n no. 79, Santiago: 1986. Heskia, Isabel, "La distribuci6n del ingreso en Chile," Universidad Cat6lica de Chile, Centro de Estudios de Planificaci6n Nacional, Estudios de Planificaci6n no. 31, Santiago: 1973. Hojman, David E., "Income Distribution and Market Policies: Survival and Renewal of Middle Income Groups in Chile," Inter-American Economic Affairs 36:43-64, 1982. Hojman, David E., "Neoliberal Economic Policies and Infant and Child Mortality: Simulation Analysis of a Chilean Paradox," World Development, 171:93-108, 1989. Legarreta, A., "Factores condicionantes de la mortalidad en la nifnez," in Livingstone and D. Raczynski (eds.), Salud Piblica y Bienestar Social, CEPLAN, Universidad de Chile, 1976. Oficina de Planificaci6n Nacional, Rist, Miguel Kast and Silva, Sergio Molina, "Mapa de extrema pobreza," Odeplan, Universidad Catolica de Chile, Instituto de Economfa, Santiago: 1975. 293 ANNEX 16 - continued Pollack, Molly and Uthoff, Andras, "Pobreza y mercado de trabajo en el gran Santiago," Estudios de Economa 14:139-92, Chile: 1987. Pollack, Molly and Uthoff, Andras, "Pobreza y empleo en Chile: un analisis del perfodo 1969-1987 en el gran Santiago," Economfa de America Latina 18-19:127-52, 1989. PREALC, "Pobreza y mercado de trabajo en el gran Santiago, 1969-1985," PREALC, Working Paper no. 299, Chile: 1987. PREALC, "Desempleo estructural en Chile: un anAlisis macroecon6mico," PREALC, Working Paper no. 302, Chile: 1987. PREALC, "Pobreza y empleo: un analisis del perfodo 1986-1987 en el gran Santiago," PREALC, Workinf Paper no. 348, 1990. Raczynski, Dagmar, "Social Policy, Poverty, and Vulnerable Groups: Children in Chile," in G. Cornia, R. Jolly, and F. Stewart (eds.), Adjustment with a Human Face n, Oxford: Clarendon Press, 1988. Rodriguez, Jorge G., "Notas sobre el modelo de dos sectores," Universidad de Chile, Departamento de Economfa, Facultad de Ciencias Econ6micas y Administrativas, Santiago: 1977. Rodrfguez, Jorge G., "El papel redistributivo del gasto social: Chile 1983," Interamericana de Planificaci6n 19:46-70, Sociedad Interamericana de Planificaci6n, 1985. Rodrfguez, Jorge G. and Instituto Latinoamericano de Estudios Sociales, "Distribuci6n del ingreso y el gasto social en Chile, 1983," 1st ed., Santiago: Editorial Salesiana for ILADES, 1985. Rosende Ramfrez, Francisco, "Elementos para el disefio de un marco analftico en el estudio de la pobreza y distribuci6n del ingreso en Chile," Estudios Piblicos 34:5-52, Chile: 1989. Torche, Aristides L., "Distribuci6n del ingreso y necesidades basicas en Chile," Economfa de Amdrica Latina 18-19:285-305, Chile: 1989. Universidad de Chile, Departamento de Economfa, "Human Capital, Employment and Poverty," Universidad de Chile, Santiago: 1984. Uthoff, Andras W., "Changes in Earnings Inequality and Labour Market Segmentation: Metropolitan Santiago 1969-78," Journal of Development Studies 22:300-26, UK: 1986. Vergara, Pilar, Polfticas hacia la extrema pobreza en Chile, FLACSO, 1990. World Bank, "Pocalizacion del gasto social en Chile," for "Seminario de alto nivel" on "C6mo recuperar el progreso social en America Latina", June 1988. 294 ANNEX 16 - continued COLOMBIA Acevedo C. and Nelly, Marfa, "La pobreza en Colombia: una medida estadfstica," Trimestre Econdmico 53:15-40, M6xico: 1986. Aguilar, Luis Ignacio and Perfetti, Juan Jos6, "Distribuci6n del ingreso y sus determinantes en el sector rural colombiano," Coyuntura Econ6mica 17:123-55, Colombia: 1987. Ayala, Ulpiano O., "Aproximaci6n al plan de lucha contra la pobreza y para la generaci6n de empleo," Debates de Coyuntura Econdmica 5:3-13, Colombia: 1987. Bamberger, Michael and Kaufmann, Daniel, "Patterns of Income Formation and Expenditures among the Urban Poor of Cartagena: Final Report on World Bank Research Project 672-57," The World Bank, Water Supply and Urban Development Department, Operations Policy Staff, Discussio ager no. UDD-63, Washington: 1984. Bamberger, Michael and Parris, Scott, "The Structure of Social Networks in the Zona Sur Oriental of Cartagena," The World Bank, Working Paper no. 1, Washington: 1984. Berry, A., "Some Determinants of Changing Income Distribution in Colombia, 1930-1970," Yale University, Economic Growth Center, New Haven: 1972. Berry, R. Albert and Soligo, Ronald, Economic Policy and Income Distribution in Colombia, Boulder: Westview Press, 1980. Berry, R. Albert, Soligo, Ronald and Urrutia, Miguel, Income Distribution in Colombia, The Economic Growth Center, Yale University, New Haven: 1976. Bonilla C., Elssy and Velez, Eduardo, Mujer y trabalo en el sector rural colombiano, Centro de Estudio sobre Desarrollo Econdmico (CEDE), Instituto SER de Investigaci6n, Bogota: 1987. Caicedo, Elizabeth and Hoyos de Arbelaez, Luz Helena, "Impacto de una redistribuci6n del ingreso sobre la nutricidn humana," Universidad de los Andes, Facultad de Economfa, Centro de Estudios sobre Desarrollo Econ6mico, Bogota: CEDE, 1977. Carrizosa, Mauricio, "Evoluci6n y determinantes de la pobreza en Colombia," in J.A. Ocampo and Mr. Ramfrez (eds.) El problema laboral colombiano I, Controlorfa General de la Repdblica, DNP, Sena, Bogota: 1987. CEPAL, "Antecedentes estadfsticos de la distribuci6n del ingreso, Colombia 1951-1982," Serie Distribuci6n del Ingreso no. 1, Santiago: 1986. CEPAL, "Colombia: Canasta basica de alimentos y determinaci6n de las lfneas de indigencia y de pobreza," Santiago: 1988. 295 ANNEX 16 - continued CEPAL, "Divisi6n de estadfsticos de la distribuci6n del ingreso," 1st ed., Serie Distribuci6n del Ingreso no. 1 Santiago: 1986. Cifuentes, Jorge Ignacio, Stevenson, Rafael and Paredes, L. Ricardo, "Housing and Urban Development in Bogota: The Institutional Backdrop," The World Bank, Water Supply and Urban Development Department, Discussion Paper no. WUDD-51, Washington: 1984. Cook, Christopher J., "Commodity Price Distortions and Intra-agricultural Income Distribution in Colombia," Journal of Developing Areas, 22:219-38, 1988. DANE (Departamento Administrativo Nacional de Estadfstica), "La pobreza en 13 ciudades colombianas," Boletfn de Estadfstica no. 429, UNDP, 1988. Departamento Nacional de Planeacidn, DANE (Departamento Administrativo Nacional de Estadfstica), UNICEF and UNDP, La pobreza en Colombia I, Bogota: 1989. Dfaz Alejandro, Carlos Federico, "Efectos de las exportaciones no tradicionales sobre la distribuci6n del ingreso: el caso colombiano," Universidad Cat6lica de Chile, Centro de Estudios de Planificaci6n Nacional, Documento no. 50, Santiago: 1975. Glaser, Marion, "Use of Labelling in Urban Low Income Housing in the Third World, Case Study of Bogota, Colombia," Development and Change 16:409-28, the Netherlands: 1985. G6mez Buendia, Hernando, "La propuesta del CONPES: receta apropiada en dosis insuficientes?", Debates de Coyuntura Econ6mica 5:15-23, Colombia: 1987. Hanson, James A., "Growth and Distribution in Colombia: Some Recent Analyses," Latin American Research Review 22(l):255-64, 1987. Jallade, Jean Pierre, Public Expenditures on Education and Income Distribution in Colombia, The World Bank, 1974, Rev. Plan. y Des. DNP 8(3):21-38, Washington: 1976. Kauffmann, Daniel and Bamberger, Michael, "Income Transfers and Urban Projects: Research Findings and Policy Issues from Cartagena, Colombia," The World Bank, Water Supply and Urban Development, Operations Policy Staff, Discussion Paper no. UDD-56, Washington: 1985. Kugler, Bernardo, "Influencia de la educaci6n en los ingresos de trabajo: el caso colombiano," Revista de Planeaci6n y Desarrollo, Bogota, 1974. Le6n R., Alejandro A. and Rodrfguez Lufs F. N., "Distribuci6n de los ingresos monetarios del trabajo urbano en Colombia: una aplicaci6n del modelo de descomposici6n de Theil," Revista de Planeaci6n y Desarrollo 15:71-95, Colombia: 1983. Liuksila, Claire, "Colombia: Economic Adjustment and the Poor," IMF Working Paper no. 91-81, Washington: 1991. 296 ANNEX 16 - continued Londofilo, Juan Lufs, "Distribuci6n del ingreso nacional en 1989," Coyuntura Econ6mica Colombia, 19:131-45, 1989. Londofho, Juan Lufs, Income Distribution during the Structural Transformation: Colombia 1938-1988, PhD Dissertation, Harvard University, 1990. L6pez, Cecilia, "Deuda Social en Colombia: Equidad en los 80 y Perspectiva para los 90," Conijuntura Sal, 2:9-25, 1990. L6pez, Cecilia, "Kuznetsian Tales with Attention to Human Capital: Catching up, Accumulation Modes and Sharp Movements of Income Distribution in Colombia," Cambridge MA: 1990. Lora, Eduardo and Ocampo, Jose Antonio, "Polftica macroecon6mica y distribuci6n del ingreso en Colombia: 1980-1990," Coyuntura Econ6mica 16:109-58, Colombia: 1986. Meldau, Elke C., Benefit Incidence: Public Health Expenditures and Income Distribution: A Case Study of Colombia, North Quincy, MA: Christopher Publishing House, 1980. Mohan, Rakesh, "An Anatomy of the Distribution of Urban Income: A Tale of Two Cities in Colombia," The World Bank, Staff Working Paper no. 650, Washington: 1984. Mohan, Rakesh "The People of Bogota: Who They Are, What They Earn, Where They Live," The World Bank, Staff Workinz Paper no. 390, Washington: 1980. Mohan, Rakesh and Hartline, Nancy, "The Poor of Bogota: Who They Are, What They Do, and Where They Live," The World Bank, Staff Working Paper no. 635, Washington: 1984. Mohan, Rakesh and Sabot, Richard, "Educational Expansion and the Inequality of Pay: Colombia 1973-78," Oxford Bulletin of Economics and Statistics 50:175-82, UK: 1988. Moreno, Alvaro Alfonso, "La distribucion del ingreso laboral urbano en Colombia 1976-1988," Desarrollo y Sociedad 24:65-127, Colombia: 1989. Ocampo, Jose Antonio and Lora, E., Country Study: Colombia, WIDER, 1987. PREALC, Colombia: la deuda social en los 80 I&II, Geneva, 1990. Psacharopoulos, George, Arriagada, Ana Maria, and Velez, Eduardo, "Earnings and Education Among the Self-Employed in Colombia," The World Bank, Education and Training Department, Discussion Paper Series no. EDT70, Washington: 1987. Reyes, Alvaro P., "Evoluci6n de la distribucion del ingreso en Colombia," Desarrollo y Sociedad 21:37-51, Colombia: 1988. Selowsky, Marcelo, Who Benefits from Government Expenditure?: A Case Study of Colombia, Oxford University Press for the World Bank, 1979. 297 ANNEX 16 - continued Thirsk, Wayne R., "Income Distribution and Colombian Rural Education," Rice University, Program of Development Studies, Houston: 1974. Thirsk, Wayne R., "Rural Credit and Income Distribution in Colombia," Rice University, Program of Development Studies, Houston: 1974. Urdinola, A. and Carrizosa, M., "The Political Economy of Poverty, Equity and Growth: Colombia," 1989 (mimeo). Uribe Mosquera, Tomas, "Revaluaci6n de la inseguridad alimentaria en Colombia," Coyuntura Econ6mica 17:157-93, Colombia: 1987. Urrutia, Miguel, "Los de arriba y los de abajo: la distribuci6n del ingreso en Colombia en las dIltimas decadas," Fondo editorial, CEREC, 1984. Urrutia, Miguel, Winners and Losers in Colombia's Economic Growth of the 1970s, New York: Oxford University Press for the World Bank, 1985. Urrutia Montoya, Miguel and Berry, R. Albert, "La distribuci6n del ingreso en Colombia," Medellin: La Carreta (Libros de la Carreta; no. 4), 1975. Various authors, "Polfticas generales de lucha contra la pobreza y para la generacidn de empleo," Documento D.N.P., Revista de Planeaci6n y Desarrollo 18:7-24, Colombia: 1986. World Bank, "Colombia: Social Programs for the Alleviation of Poverty," The World Bank, a Country Sujry, Washington: 1990. World Bank, "Country Assessment of Women's Role in Development: Proposed Bank Approach and Plan of Action," Latin America and the Caribbean Region, Department III, Country Operations Division 1, Washington: 1989. COSTA RICA CEPAL, "Antecedentes estadfsticos de la distribuci6n del ingreso en Costa Rica, 1958-1982," Serie Distribuci6n del Ingreso no. 4, Santiago: 1987. Gindling, T.H., "Labor Market Segmentation and the Determination of Wages in the Public, Private - Formal and Informal Sectors in San Jose, Costa Rica," Economic Development and Cultural Change 39:585-605, 1991. Gindling, T.H. and Berry, A., "The Performance of the Labor Market During Recession and Structural Adjustment: Costa Rica in the 1980's," World Development, forthcoming. 298 ANNEX 16 - continued Gonzales-Vega, C. and Cespedes, V.H.. "The Political Economy of Growth, Equity, and Poverty Alleviation: Costa Rica: 1950-1985" 1989 (mimeo). Fields, Gary, "Poverty and Adjustment in Costa Rica,", IDB, Washington: 1992 (mimeo). Peek, P. and Raabe, C., "Rural Equity in Costa Rica: Myth or Reality?', ILO, World Employment Programme, Research Working Paper no. 67, Geneva: 1984. Pinera, Sebastian, Medici6n.. analisis v descripci6n de la pobreza en Costa Rica, CEPAL, 1979. Pollack, Molly, "Household Behavior and Economic Crisis. Costa Rica 1979-1982," PREALC, Working Eadr no. 270, 1985. Pollack, Molly, "Poverty and the Labour Market in Costa Rica," in G. Rodgers ed., Povery. and the Labour Market: Access to Jobs and Incomes in Asian and Latin American Cities, ILO, Geneva: 1990. Sauma, Fait and Trejos, J.D., "Evoluci6n reciente de la distribuci6n del ingreso en Costa Rica; 1977- 1986," IICE, Working PEper no. 132, 1990. Trejos, J.D. and Elizalde, M.L., "Costa Rica: la distribucion del ingreso y el acceso a los programas de caricter social," Universidad de Costa Rica, Instituto de Investigaciones en Ciencias Economicas, San Pedro: 1985 (mimeo). World Bank, "Country Assessment of Women in Development,' Latin America and the Caribbean Region, Department II, Country Operations Division, Washington: 1989. World Bank, Costa Rica: Country Economic Memorandum, 1988 DOMINICAN REPUBLIC ECOCARIBE, "El impacto distributivo de la gesti6n fiscal en la Repiblica Dominicana," IDB, 1992 (mimeo). Musgrove, P., "Distribucidn del ingreso familiar en la Repiblica Dominicana, 1976-1977," El Trimestre Econmico 53(2):341-92, Mexico: 1986. ECUADOR Barreiros, Lidia, "Distribution of Living Standards and Poverty in Rural Ecuador," Institute of Social Studies-PREALC Working Paper no. 15:1-46, The Netherlands: 1984. 299 ANNEX 16 -continued Barreiros, Lidia, "La Pobreza y los Patrones de Consumo de los Hogares en Ecuador," Comercio Exterio 42(4):366-379, Mexico, 1992. Barreiros, Lidia, "Poverty and Household Consumption Patterns in Ecuador," Rev. Ed. Institute of Social Studies-PREALC Working Paper no. 5:1-44, The Netherlands: 1985. Barreiros, Lidia, "Ecuador's Development Profile and Basic Needs Performance," Institute of Social Studies-PREALC, Working Paper no. 24:1-70, The Netherlands, 1985. Barreiros, Lidia and Teekens, Rudolph, "Poverty and Consumption Patterns in Urban Ecuador, 1975," Institute of Social Studies-PREALC, Working Paper no. 5:1-45, The Netherlands: 1984. Bastiaenen, Michiel and Solf, Johannes, "Urban Poverty, Access to Basic Services and Public Policies: A Case Study of the Marginal Suburbs of Guayaquil, Ecuador," Institute of Social Studies- PREALC, Working Paper no. 7:1-134, The Netherlands: 1983. CEPAL, Ecuador: mapa de necesidades basicas insatisfechas, Divisi6n de Estadistica y Proyecciones, 1989. Luzuriaga, Carlos and Zuvekas, Clarence, "Income Distribution and Poverty in Rural Ecuador, 1950- 1979: A Survey of the Literature," Arizona State University, Center for Latin American Studies, 1983. Moser, Caroline O., "The Impact of Recession and Structural Adjustment Policies at the Micro-Level: Low Income Women and Their Households in Guayaquil, Ecuador," London School of Economics, Department of Social Administration, 1989. Moser, Caroline O., "Impact of Recession and Structural Adjustment on Women: Ecuador," Development 1:75-83, Italy: 1989. Santos, Eduardo, "La pobreza en el Ecuador," Revista de la CEPAL 38:121-32, 1989. Teekens, Rudolf and Barreiros, Lidia, "Theory and Policy Design for Basic Needs Planning: A Case Study of Ecuador, et al," Institute of Social Studies, The Netherlands: 1988. World Bank, "Country Assessment of Women's Role in Development: A Proposed Bank Approach," Latin America Technical Department, Human Resources Division, Washington: 1989. World Bank, "Ecuador: A Social Sector Strategy for the Nineties," Latin America and the Caribbean Region, Country Department IV, report no. 8935-EC, Washington: 1990. 300 ANNEX 16 -continued EL SALVADOR Daines, S., "Analysis of Small Farms and Rural Poverty in El Salvador," in El Salvador, Agricultural Sector Assessment, San Salvador: USAID, 1977. Deere, C.D. and Diskin, M., "Rural Poverty in El Salvador: Dimensions, Trends and Causes," ILO, World Employment Programme Research, Working Paper no. 64, Geneva: 1984. Ministerio de Planificaci6n y Coordinaci6n del Desarrollo Econ6mico y Social, Unidad de Investigaciones Muestrales, El Salvador, "Distribuci6n del ingreso por deciles de perceptores," El Salvador: 1978. Valverde, V., et al, "Life Styles and Nutritional Status of Children from Different Ecological Areas of El Salvador," Ecology of Food and Nutrition 9:167-77, 1980. GUATEMALA CEPAL, "Guatemala: Canasta basica de alimentos y determinaci6n de las lineas de indigencia y de pobreza," Santiago: 1988. CEPAL, "Politica macroecon6mica y pobreza," Santiago: 1989. World Bank, "Guatemala: Population, Nutrition and Health Sector Review," gray cover report no. 6183-GU, 1986. GUYANA Standing, Guy and Szal, Richard, "Poverty and Basic Needs: Evidence from Guyana and the Philippines," ILO, World Employment Programme, Geneva: 1979. HONDURAS Peek, P., "Agrarian Structure and Rural Poverty: The Case of Honduras," ILO, World Employment Programme Research, Working Paper No. 68, Geneva: 1984. 301 ANNEX 16 -continued JAMAICA Ahiran, E., "Income Distribution in Jamaica, 1958," Social and Economic Studies 13(3):333-69, 1964. Boyd, D., "The Impact of Adjustment Policies on Vulnerable Groups: The Case of Jamaica, 1973- 1985," in G. Cornia, R. Jolly and F. Stewart (eds.), Adjustment with a Human Face II:125-55, Oxford: Clarendon Press, 1988. Duncan, G. Cameron, International Monetary Fund Stabilization Programs and Income Distributio Costa Rica and Jamaica, Ph.D. Dissertation, The American University, 1987. McLure, J.C.E., "The Incidence of Jamaican Taxes, 1971-72," Institute of Social and Economic Research, Working aper no. 16, University of the West Indies, Mona, Jamaica: 1977. World Bank, "Country Assessment of Women's Role in Development: Proposed Bank Approach and Plan of Action," Latin America and the Caribbean Region, Country Department IIT, Washington: 1989. World Bank and Statistical Institute of Jamaica, "Living Conditions Survey: Jamaica," preliminary report, Kingston: 1988. MEXICO Aspe Armella, Pedro and Sigmund, Paul E., The Political Economy of Income Distribution in Mexico, New York: Holmes and Meier, 1984. Bazdresch, Carlos and Levy, Santiago, "Populism and Economic Policy in Mexico (1970-1982)," for "The Macroeconomics of Income Distribution in Latin America-Conference," of the Inter- American Development Bank and NBER, Washington: May, 1990. Bergsman, Joel, "Income Distribution and Poverty in Mexico," The World Bank, Staff Working Paper no. 395, 1980. CEPAL, "M6xico: Canasta basica de alimentos y determinaci6n de las lineas de indigencia y de pobreza," Santiago: 1988. CEPAL, "Antecedentes estadisticos de la distribuci6n del ingreso, Mexico, 1950-1977," S-ir Distribuci6n del Ingreso no. 7, Santiago: UN, 1988. Cervantes Gonzalez, Jesus Alejandro, "Mdxico: analisis de la distribuci6n del ingreso; aspectos metodol6gicos," Comercial Exterior 32:43-50, Mexico: Banco Nacional de Comercio Exterior, 1982. Cervantes Gonzalez, Jesuis Alejandro, "Inflaci6n y distribuci6n del ingreso y de la riqueza en M6xico," Trimestre Econ6mico 50:2017-40, Mexico: 1983. 302 ANNEX 16 - continued Garcfa Rocha, Adalberto, "Note on Mexican Economic Development and Income Distribution," El Colegio de Mexico, Centro de Estudios Econ6micos,1990. Gregory, Peter, The Myth of Market Failure: Employment and the Labor Market in Mexico, Baltimore: Johns Hopkins University Press, 1986. Hernandez-Laos, Enrique, "La Pobreza en Mexico," Comercio Exterior 42(4):402-41 1, Mexico, 1992. Hernandez-Laos, Enrique, "Medici6n de la incidencia de la pobreza y de la pobreza extreme en Mexico, 1963-1988," Universidad Autdnoma Metropolitana, Mexico: 1989. Hernandez-Laos, Enrique, "Tendencias recientes en la distribuci6n del ingreso en Mexico, 1974-84," Programa de Doctorado en Ciencias Econ6micas, Universidad Aut6noma Metropolitana, Mexico: 1989. Heroles, G.G. and Reyes, Jesds, "Las polfticas financieras y la distribuci6n del ingreso en Mexico," Trimestre Econ6mico 55:649-702, Mexico: 1988. Levy, S., "La Pobreza Estrema en Mexico: Una Propuesta de Polftica (with English summary). Estud. Econ. 6, no. l(Jan-June 1991): 47-89. Levy, Santiago, "Poverty Alleviation in Mexico," World Bank, Staff Working Paper no. 679, Washington: 1991. L6pez Gallardo, Julio, "La distribucion del ingreso en Mexico: estructura y evoluci6n," Trimestre Econ6mico 50:2227-56, Mexico: 1983. Lustig, Nora, "Crisis econ6mica y niveles de vida en Mexico: 1982-1985," Estudios Econ6micos del Colegio de Mexico 2:227-249, Mexico: 1987. Lustig, Nora, "Economic Crisis, Adjustment and Living Standards in Mexico: 1982-1985," World Development 18, no. 10(1990): 1325-1342. Lustig, Nora, Poverty Indices and Poverty Orderings: An Application to Mexico, The Brookings Institution, Washington: 1990. Lustig, Nora, "PRONASOL: As a Poverty Reduction Strategy," in Wayne Cornelius, ed. San Diego: University of California, Center for U.S.-Mexican Studies, forthcoming. Lustig, Nora, The Incidence of Poverty in Mexico, 1984: An Empirical Analysis, The World Bank, Technical Department, Latin America and the Caribbean Region, Human Resources Division, Washington: 1990 (mimeo). 303 ANNEX 16 - continued Lustig, Nora, "The Social Costs of Adjustment," in Mexico: The Remaking of an Economy. Brookings Institution, 1992. Maddison, A. et al., "The Political Economy of Poverty, Equity and Growth: Brazil and Mexico," The World Bank, Washington: 1989 (mimeo). Stark, Oded, Taylor, J. Edward and Yitzhaki, Shlomo, "Labor Migration, Income Inequality and Remittances: A Case Study of Mexico," The World Bank, Discussion Pager no. 283, Washington: 1987. World Bank, "Mexico: A Strategy for Poverty Alleviation: Summary and Conclusions," yellow cover, Washington: 1989. NICARAGUA Peek, P., "Agrarian Reform and Poverty Alleviation: The Recent Experience in Nicaragua," ILO, World Employment Programme, Research Working Paper no. 69, Geneva: 1984. PANAMA Bunge, C., "Basic Needs in Panama: The Health Sector," ILO, World Employment Programme, Research Working Paper, Geneva: 1980. CEPAL, "Panama: Canasta basica de alimentos y determinaci6n de las lfneas de indigencia y de pobreza," Santiago: 1988. Pinnock, R. and Elton, C., "Rural Poverty in Panama: Trends and Structural Causes," ILO, World Employment Programme, Research Working Paper no. 60, Geneva: 1983. PARAGUAY Miranda, Anibal, "Desarrollo y pobreza en Paraguay," Inter-American Foundation, 1982. 304 ANNEX 16 - continued PERU Alarco, German and del Hierro, Patricia, "Perd 1985-1988: de la redistribuci6n a la reconcentraci6n del ingreso," Economfa de America Latina 18-19:153-89, M6xico: 1989. Alberts, Tom, Agrarian Reform and Rural Poverty: A Case Study of Peru, Boulder: Westview Press, 1983. Amat y Le6n, Carlos and Le6n, Hector, "Estructura y niveles de ingreso familiar en el Perd: c6mo financian sus ingresos las familias en las diferentes areas y regiones de residencia en el Perd?" Universidad del Pacffico, Centro de Investigaci6n, Cuadernos de Investigacidn 10:199, Peru: 1979. Amat y Le6n, Carlos and Monroy, L., 'Los cambios en la economfa de las familias de Lima metropolitana: 1972-85," Universidad del Pacffico, Centro de Investigaci6n, Cuadernos de Investigacidn, Perd: 1985. Arnold, Barry C. and Laguna, Leonor, Universidad Nacional Mayor de San Marcos and United States Agency for International Development, 'On Generalized Pareto Distributions with Applications to Income Data," Iowa State University, Department of Economics, Ames: 1977. Banco Central de Reserva del Perti, Subgerencia de Ingreso y Producto, Departamento de Estudios del Sector Social, "Mapa de pobreza del Perd 1981," Lima: 1986. Berry, Albert, "International Trade, Government, and Income Distribution in Peru Since 1870," Latin American Research Review 25(2):31-59, 1990. Brady, Eugene A., "The Distribution of Total Personal Income in Peru," Iowa State University, Department of Economics, Ames, IA: 1968. CEPAL, "Peru: Canasta basica de alimentos y determinaci6n de las lfneas de indigencia y de pobreza, Santiago: 1988. de Soto, Hernando, The Other Path: The Invisible Revolution in the Third World, New York: Harper and Row, 1989. Dietz, Henry A., Povery and Problem-Solving under Military Rule: The Urban Poor in Lima. Peru, University of Texas at Austin, Institute of Latin American Studies, Latin American Monograghs, no. 51, Austin: 1980. Ferroni, Marco A., "The Urban Bias of Peruvian Food Policy: Consequences and Alternatives," Ph.D. Dissertation, Cornell University, Ithaca: 1980. Figueroa, Adolfo, "El problema distributivo en diferentes contextos sociopolfticos y econ6micos, Perd, 1950-1980," Desarrollo Economico 22:163-85, Argentina: 1982. 305 ANNEX 16 - continued Figueroa, Adolfo, "Integracidn de las polfticas de corto y largo plazo," Pontificia Universidad Cat6lica del Perd, Departamento de Economfa, Lima: Publicaciones CISEPA, 1989. Figueroa, L., "Economic Adjustment and Development in Peru: Towards an Alternative Policy," in G., Cornia, R. Jolly and F. Stewart (eds.), Adjustment with a Human Face II:156-83 Oxford: Clarendon Press, 1988. Glewwe, Paul, "The Distribution of Welfare in Peru in 1985-86," The World Bank, Population and Human Resources Department, LSMS Working Paper no. 42, Washington: 1987. Glewwe, Paul, and de Tray, Dennis N., "The Poor in Latin America during Adjustment: A Case Study of Peru," The World Bank, Washington: 1989. Glewwe, Paul and Hall, Gillette, The Social Costs of Avoiding Structural Adjustment: Inequality and Poverty in Lima. Peru. from 1985-86 to 1990, The World Bank and Cambridge University, 1991. Graham, Carol, "The APRA Government and the Urban Poor: The PAIT Programme in Lima's Pueblos J6venes", Duke University, submitted to Journal of Latin American Studies, 1990). Grupo Nacional del Peril, Proyecto Regional para la Superaci6n de la Pobreza, "La pobreza en el Perd," Lima: 1990. Harrell, Marielouise W., Parill6n, Cutberto, and Franklin, Ralph L., "Nutritional Classification Study of Peru: Who and Where Are the Poor?", Food Policy 14:313-29, UK: 1989. Henrfquez Ayin, Narda and Ana, Ponce A., "Lima: poblaci6n, trabajo y polftica", Pontificia Universidad Catolica del Peril, Facultad de Ciencias Sociales, Lima: 1985. Mosley, Paul, "Can the Poor Benefit from Aid Projects? An Empirical Study of the 'Trickle-Down' Hypothesis," University of Bath, School of Humanities and Social Sciences, Papers in Political Economy 18:1-37 Bath: 1983. Rosenhouse Persson, Sandra, "Identifying the Poor: Is 'Headship' a Useful Concept?", The World Bank, LSMS Working Paper no. 58, Washington, 1989. Thomas, Vinod, "The Measurement of Spatial Differences in Poverty: The Case of Peru," The World Bank, Staff Working Paper no. 273, Washington, 1978. Webb, Richard Charles "The Distribution of Income in Peru," in Foxley (ed.), Income Distribution in Latin America: 11-25, Cambridge, New York: Cambridge University Press, 1976. Webb, Richard Charles Government Policy and the Distribution of Income in Peru. 1963-1973, Harvya Economic Studies no. 147, Cambridge, MA: Harvard University Press, 1977. Webb, Richard Charles, "The Political Economy of Poverty, Equity and Growth: Peru 1948-1985," 1989 (mimeo). 306 ANNEX 16 - continued Webb, Richard Charles and Adolfo Figueroa, "Distribucidn del ingreso en el Perd," 1st. ed. Instituto de Estudios Peruanos, Lima: 1975. URUGUAY Bension, Alberto and Caumont, Jorge, "Polftica econ6mica y distribucion del ingreso en el Uruguay, 1970-1976," Montevideo: Acali Editorial, 1979. CEPAL, Oficina de Montevideo, Estructura socio-ocupacional y distribuci6n del ingreso en el Uruguay 1984-1988, 1989. CEPAL, Direcci6n General de Estadfstica y Censos, "Pobreza y necesidades bgsicas en el Uruguay: indicadores y resultados preliminares," 1988. CEPAL, "Uruguay: Canasta basica de alimentos y determinaci6n de las lfneas de indigencia y de Pobreza," Santiago: 1988. Favaro, E. and Bension, A., "The Political Economy of Poverty, Equity and Growth: The Uruguayan Case," 1989 (mimeo). Katzman, Ruben, "La heterogeneidad de la pobreza: el caso de Montevideo," Revista de la CEPAL, 37:141-52, CEPAL, 1989. Melgar, Alicia, "La distribuci6n del ingreso en la decada de los afhos ochenta en Uruguay," Economfa de America Latina 18-19:113-26, Mexico: 1989. Portes, Alejandro and Blitzer, Silvia, "Urban Informal Sector in Uruguay: Its Internal Structure, Characteristics, and Effects," World Development 14:727-41, UK: 1986. World Bank, "Uruguay: Population, Health and Nutrition Sector Memorandum," Washington: 1989. VENEZUELA Baptista, Asdrubal and Mommer, Bernard, "Renta petrolera y distribuci6n factorial del ingreso," in Nissen y Mommer (eds.) Adi6s a la Bonanza? Crisis de la distribuci6n del ingreso en Venezuela, Caracas: Editorial Nueva Sociedad for ILDIS-CENDES, 1989. Bourguignon, Frangois, "Optimal Poverty Reduction, Adjustment and Growth: An Applied Framework," The World Bank, Latin America and the Caribbean Region, Technical Department, Human Resources Division, Washington: 1989 (mimeo). 307 ANNEX 16 - continued Bourguignon, Francois and Michel, G., "Short-run Rigidities and Long-run Adjustments in a Computable General Equilibrium Model of Income Distribution and Development," Journal of Development Economics 13:21-43, Netherlands: 1983. [Also in the World Bank Reprint Series no. 286, 1983.1 CEPAL, "Antecedentes estadfsticos de la distribuci6n del ingreso, Venezuela 1957-1985," Serie Distribucidn del Ingreso 6, Santiago: 1988. CEPAL, Una Estimaci6n de la Mafnitud de la Pobreza en Chile. 1987, LC/L.599, Santiago: 1992. CEPAL, "Venezuela: Canasta basica de alimentos y determinaci6n de las lfneas de indigencia y de pobreza," Santiago: 1988. Cline, William, R., "Venezuela: Economic Strategy and Prospects," IDB, Washington: 1991 (mimeo). de Ferran, Lourdese, "Distribuci6n del ingreso: analisis del caso venezolano," Colecci6n de Estudios Econdmicos no. 5, Caracas: Banco Central de Venezuela, 1977. de Ferran, Lourdes, "Participaci6n econ6mica de la mujer y la distribuci6n del ingreso," Colecci6n de Estudios Econ6micos no. 13, Caracas: Banco Central de Venezuela, 1986. Marquez, Jaime and Shack-Marquez, Janice, "Financial Concentration and Development: An Empirical Analysis of the Venezuelan Case," United States Board of Governors of the Federal Reserve System, International Finance division, Discussion Paper no. 300, 1987. Morley, Samuel A. and Alvarez, C., "Poverty and Adjustment in Venezuela," IDB, Washington (mimeo). Nissen, Hans-Peter and Mommer, Bernard, (eds.), Adi6s a la Bonanza? Crisis de la distribuci6n del ingreso en Venezuela, Caracas: Editorial Nueva Sociedad for ILDIS-CENDES, 1989. Vivancos, Francisco, "El shock externo y la economfa informal," in Nissen and Mommer (eds.), Adi6s a la Bonanza? Crisis de la distribuci6n del ingreso en Venezuela, Caracas: Editorial Nueva Sociedad for ILDIS-CENDES, 1989. World Bank, "Public Policy Options for Venezuela," Latin America and Caribbean Region, Washington: 1988. World Bank, "Venezuela Poverty Study: From Generalized Subsidies to Targeted Programs," Washington: 1990 (mimeo). Distributors of COLOMBIA HAITI ITALY NEW ZEALAND ROMANIA SWITZERLAND Distributors of Infoenlace [ida Cu fure Ddfusion [icosa Commissionaria Sansoni SPA E8SCO NZ Ltd Compani De Librarii Bucurestli S A Libraine Payol Service Insltutionnel W orld Bank Carrera6 No 51-21 5, Rue Capoes Via Duca DiCalabna, 1/1 Private Mail Bag 99914 Str Lipscanino 26, sector3 C6tes-de-Monnbenon 30 Aparlado Aereo 34270 C P 257 Casella Postale 552 New Market Bucharest 1002 Lausanne Publications Santafe de Bogota, D C Por-au-Prince 50125 Firenze Auckland Tel (40 1) 613 9645 Tel (41 21) 341-3229 Paces ald credit terms narltfront Tel (57 1) 285-2798 Tel (5091)39260 Tel (55) 645-415 Tel (649) 524-8119 Fax (401)3124000 Fax (41 21) 341-3235 coliatrV/ to countrV Consutlt your Fax (571) 285-2798 Fax (55) 641-257 Fax (64 9) 524-8067 local distribuitor be~fore placing an HONG KONG, MACAO E-mail licosu80thcc 6 RUSSIAN FEDERATION AEOVnDemnEiinTcnqe lorde Gf istnbtttor before PlaCInS art COTE D'IVOIRE Asia 2000 Ltd Ur hop I/www ftbcc iAicosa NIGERIA Isdatelstvo Ch do L Ducez 41 order Center dEdition et de Dffusion Alncanes Sales & Circulalion Depaitment Universiy Press Limned 9a, LolpachnimyPereulok CH1807 Blonay ARGENTINA (CEDA) Seabird House, uit 1101-02 JAMAICA Three Crowns Building Jencho Moscow 101831 Tel (41 21) 943 2673 Oficina del Libro Intemsacional 949B P 541 22-28 Wyncdham Street, Central Ian Randle Publishers Ltd Private Mail Bag 5095 Tel (7 095) 91768749 Fax (41 21) 942 3605 An Cordoba 1877 Abidlan 04 Hong Kong 206 Old Hope Road Ibadan Fax (7 099) 917 92 59 1120 Buenos Aires Tel (225)246510,246511 Tel (852) 2530-1409 Kingslon 6 Tel (23422)41-1356 TANZANIA Tel (54 1) 815-8354 Fax (225) 250567 Fax (852) 2526-1107 Tel 809-927-2085 Fax (234 22)41-2056 SINGAPORE, TAIWAN, Oxfoerd Unrersoy Press E-mail salesdiasia2000 com hk Pus 809-977-0243 MYANMAR, BRUNEI Maktaha Street Fax (54 1)815-8156 CYPRUS URL hnpi/Aww asia2OOO com hk E-mail irpl8colis7co NORWAY Asahgate Publishing Asia Pacdic Pte Ltd PO Box 5299 AUSTRALIA, FUI, PAPUA NEW GUINEA Cen,er ior Appked Research NIC Info ANS 41 Kalang Pudding Road #04-03 Dar es Salaam SOLOMON ISLANDS, VANUATU, AND Cyprus College INDIA JAPAN Book Department Golden Wheel Buildiug Tel (255 51) 29209 WESTERN SAMOA 6, Diogenes Street, Engomi Allied Publishers Ltd Eastem Book Service PO Box 6125 Eterstad Singapore 349316 Fax (255 51) 46822 O A Iolornation Services PO Box 2006 751 Mount Road 3-13 Hongo 3-chome, Bunkyo-ku N-0602 Oslo 6 Tel (65) 741-5196 648 Whdehorse RoadNicorsi Madras - 696 002 Tokyo 1 13 Tel (47 22) 57-3300 Fax (65) 742-9356 THAILAND 48 3erseRod Tel (357 2)44-1730 Tel (91 44)852-3938 Tel (81 3)3818-0861 Fax (4722)68-1901 E-mail F shgote@asianonnned cnm Central Books Distribution Victona Fan (257 2(48-2051 Fax )9t 44) 852-0648 Fan (81 3) 3818-0864 306 Stiom Roand Tel (61)292107777 E-mail not-eEs@ppp bekkoame or p PAKISTAN SLOVENIA Bangkok 10500 FTa (61) 3 9210 7778 CZECH REPUBLIC INDONESIA URL hipJ/www bekkoame orjp/-svt-ebs Mirza Book Agency Gospodarski Vestnik Publishing Group Tel (66 2) 235-5400 Fax (61) 3 9210 7788 National tntomiaton Center Pt Indira Limited 65, Shahrah-e-Quaid-e-Azam Dunalska cesla 5 Fax (962) 237-8321 E-mail setvice@dadirectcurs an URL htop /wwrw dadirect comau prodetna, Konviklska 5 JaLsa Borobudur 20 KENYA Lahore 54000 1000 L4ubilana UCS -113 57 Prague 1 PO Box 181 Aloca Book Service (E A) Ltd Tel (92 42) 735 3601 Tel (386 61) 133 83 47,13212 30 TRINIDAD & TOBAGO, AUSTRIA Tel (422) 2422-9433 Jakana 10320 Quaran House, Mfangano Street Fax (9242)7585283 Fax (386 61) 1338030 AND THE CARRIBBEAN Gerold and Co Fax (42 2) 2422-1484 Tel (62 21) 390-4290 PO Box 45245 E-mail belicd@gvestnik s Systematics Studies Uni Weihburggasse 26 URL hftp lwww siscz/ Fax (62 21) 4214289 Nairobi Oxord University Press 9 Wans Street A-loll Wig,sen Tel (254 2) 223 641 5 Bangalore Town SOUTH AFRICA, BOTSWANA Curepe Tel (431)512-47-31- DENMARK IRAN Fax (2542) 330 272 Shaae Faisal Forsingle Idles Trnidad, West Indies Te (43 1( 512-47-31-29 SainhardsLdfterafur Keatab Sara Co Pubbshers PO BOo 12023 Oxford Unmversily Press Southem Afrnca Tel (899) 602-5654 URL hI43ww geru5d2cuat onl29 Rosenoems Alie 11 Khaled Eslamboli Aye, KOREA, REPUBLIC OF Karachi-75350 P0 Boo 1141 Pan (809) 662-5654 URL httpJAvww geroldr coat online DK-i 870 FPrderiksberg C 6th Street Daelon Trading gCo Ltd Tel (92 21) 446307 Cape Town 80 E-mail tob@tonidad net BANGLADESH Tel (45 31) 351942 KushehDeltaroozNo 8 PO Box34, Youida Fax (92 21)4547840 Tel (27 21) 45-7266 Micro Industnes Development Fax (45 31) 357822 PO Box 15745-733 700 Seoun Bldg E-mail oup@oup khi erum com pk Fax (2721)45-7205 UGANDA Assistance Society (MIDAS) Tehra 44-6 Yousdo-Dong, Yeongchengpo-Ku Gusnro Ltd House staeRoad ty(M16 AS) EGYPT, ARAB REPUBLIC OF Tel (98 21) 8717818,8716104 Seoul Pak Book Corporation For subscription oreers PO Box 9997, Madhvant Buiklog housen,i Rudea Al Ahran Distotbution Agency Fax (98 21) 8712479 Tel (822) 785-1631/4 Azz Chambers 21 lntematonal Subscnption Service Plot 16/4 Jinja Rd Dhaka 1209 Al Galaa Street E-mail ketab-sara@tnedanet ir Fax (82 2) 784-315 Queen's Road PO Box 41095 Kampala Telk 1289 )36Z Dawno Lahore Craghall Tel (296 41) 254 762 Tal (8802)323427 Tel (202) 578-6083 Kowkab Publishers MALAYSIA Tel (9242)6363222,6360885 Johannesburg 2024 Fan (25641)251 468 Pan (8802(811188 Fax (202(578-833 PO Box19575-511 Unuversay ofMalayaCooperatrve Fax (9242)636223B Tel (2711)880-1448 Tehran Bookshop. Lrffned E-mail pbc@bramnetpk Fax (2711) 880-6248 UNITED KINGDOM JELGIUM The Mddle East Observer Tel (9821)258-3723 PO Box 1127 E-mail nsohiscoza Micinfm Ltd ern De annoy 41, Shern Street Fax (9821)258-3723 Jabn Panta Baru PERU PO Box3 Av1du Roel202 Caro 59700 Kuala Lurnpur Editonal Desanollo SA SPAIN Afton, Hampshire GU34 2PG n 5usses 5 169 Tel (202) 393-9732 IRELAND Tel (603)756-5000 Apanaitrh3824 Mundi-Prensa Libros, S A England a (32 2) 5384-841 Fax (20 2) 393-9732 Govemment Supplies Agency Fax (60 3) 7554424 Lima I Castelo 37 Tel (44 1420) 86648 Fax (32 2) 538-Ml Oifig an tSoLbair Tel (51 14) 285380 28001 Madnd Fax (44 1420) 89889 BRAZIL FtNLAND 4-5 Hancouit Road MEXICO Fax (51 14) 286628 Tel (34 1) 431-3399 E-mail wbank@ukmnlo demon e ouk Publicacdes Tecnuicu Intemacionais Uda AkOateemmen Killakauppa Dubln 2 INFOTEC Fax (341) 575-3998 URL http //www microinfo co uk RaPersoto Gomide, 2D P0 Bun 128 Tel (353 1( 661-31 11 An San Femando No 37 PHILIPPINES E-mail librena@imundiprensa en Ran oPo, SP FIN-00101 Helsinki Fax (353 1)475-2670 Col Tonetlo Guerra Intemabonad Booksource Center Inc URL htpi/www mundiprensa es/ VENEZUELA Tel (5511( 259-6644 Tel (35O)912141 14050 Mexico,DF 1127-A AntipolSt Tecni-Ciencia Lbros, S A Fan (5511(258-6990 Fax (3580)1214441 ISRAEL Tel (525)624-2800 Barangay,Venezuela Mundi-PnensaBarcelona CentoCuidadComercialTamanco E-mail posraastertpli uol hr URL httpilwtww kateeimen conv Yozmot Literature Ltd Fax (52 S)624-2822 Mukali City Consel deCent, 391 Nrvel C2 URL h8pJ/wn nul hr ~~~~~~~~~~~~~~~~P0 B3oo 56055 E-mail laotecrrstins ntxTat (63 2)88096651, 6505,86507 08009 Baorebooa Caracas URL httpMwvww uol br FRANCE 3 Yohanan Hasandar Street URL ht1pJ/dn net mx Fax (632) 896 1741 Tel (343)488-3492 Tel (582)9595547,5035,0016 CANADA Wlodd Bank Publications Tel Aviv 61560 Fax (34 3)487-7659 Fan (582) 9595636 Renaul Publishing Co Ltd 66, avenue dflena Tel (972 3) 5285-397 NEPAL POLAND E-mail barcelana@mundiprensa es 5369r C bnoelt Road 75116 Pans Fax (972 3)5285-397 Everest Medal ntemahonal Services (P)Ud Intemational Publishmg Service ZAMBIA 53w9 Cn notk 9oa Tel (331) 40-69-30-56/57 GPO Box 5443 Ul Piekna 31/37 Universty Bookshop, Universty of Zambia Tel (613)t74532665 Fax (33 1)40-69-30-8 ROY Intemational Kathmandu 00-677 Warzawa SRI LANKA, THE MALDIVES Great East Road Campus Pe (613(745-7660 PO Bux 13056 Tel (9771)472152 TTel (482) 628-6089 Lake House Bookshop PO ox32379 E-mail reno61 t@3x7sumen GERMANY Tel Ami 61130 Fax (977 1)224431 Fax (482)621-7255 100, Sir Chitanmpalam Gardiner Mawatha Lusana URL httpJ/ www loxnstn cat-renouf UNO-Vedag Tel (9723)5461423 E-mail books%ips@ikp atm camp1 Co(ombo 2 Tel (260 1)252576 Puppeludorfer Allen 55 Pun (972 3)85461442 NETHERLANDS URL httpJAwww ipscg wawpLiips/export/ Tel (94 1(32105 Fan (260 1) 253 952 CHINA 53115 Bonn E-madl royd@netvisrson neti DeLindeboorltnOr-Publikafies Fax (94 1)432104 Chms Financial & Economic Tel (49 228) 212940 PO Box 202 PORTUGAL E-mail LHL@sninkanet ZIMBABWE Publshig House Fax (49 228) 217492 Patestn3an Authonttk?4die East 7480 AE Haaksbeigen Lwnana Portugal Longian Znmbabwe (Pte )Ltd Os Pu SirDung Jie IndeIx imiabon Servtces Tel (31 53) 5740004 Apartado 2681 SWEDEN Toude Road, ArcEbenne 8,DaFoSiDongJie GREECE POB 19502 Jerusalem Fax (31 53)572-9296 Rua Do Carmo 70-74 Wennergren-Wiliiams AB PO BoxuST125 Beipog Papasotinou S A Tel (9722)6271219 E-mai lindebooxw:brlclnalnersl 1200 Lisbon P0 Box 1305 Southerton Tel (86 10)6333-8257 35, Stociara Str Fax (9722)6271634 URL httpJAww worldnliene-lindeboo T-I (1)3474982 S-171 25 Subsa Harare Fax (86 10)86401-7365 106 82 Athens Faa (1) 347-0264 Tel (4698) 705-97-50 ikIl (263 4) 6216617 Tel (301)364-1828 Fax (468)27-0O-71 Fax (2634) 621670 Fax (301)364-8254 E-mail malawwnise on01Se7 I RECENT WORLD BANK TECHNICAL PAPERS (continued) No 304 Foley, Photovoltaic Applications in Rutral Areas of the Developing World No 305 Johnson, Education and Training of Accountants in Sub-Saharan Anglophone Africa No 306 Muir and Saba, Improving State Enterprise Performance: The Role of Internal and External Incentives No 307 Narayan, Toward Participatory Research No 308 Adamson and others, Energy Use, Air Polluztion, and Environmental Policy in Krakow: Can Economic Incentives Really Help? No. 309 The World Bank/FOA/UNIDO/Industry Fertilizer Working Group, World and Regional Supply and Demand Balances for Nitrogen, Phosphate, and Potash, 1993/94-1999/2000 No 310 Elder and Cooley, editors, Sutstainable Settlement and Development of the Onchocerciasis Control Programme Area. Proceedings of a Ministerial Meeting No. 311 Webster, Riopelle, and Chidzero, World Bank Lendingfor Small Enterprises 1989-1993 No. 312 Benoit, Project Finance at the World Bank: An Overview of Policies and Instruments No. 313 Kapur, Airport Infrastructutre. The Emerging Role of the Private Sector No 314 Vald6s and Schaeffer in collaboration with Ramos, Surveillance of Agrictultutral Price and Trade Policies A Handbookfor Ecuador No 316 Schware and Kimberley, Information Technology and National Trade Facilitation: Making the Most of Global Trade No. 317 Schware and Kimberley, Information Technology and National Trade Facilitation: Guide to Best Practice No 318 Taylor, Boukambou, Dahniya, Ouayogode, Ayling, Abdi Noor, and Toure, Strengthening National Agricultural Research Systems in the Humid and Sub-hutmid Zones of West and Central Africa A Frameworkfor Action No. 320 Srivastava, Lambert, and Vietmeyer, Medicinal Plants: An Expanding Role in Development No. 321 Srivastava, Smith, and Fomo, Biodiversity and Agricultture Implicationsfor Conservation and Development No. 322 Peters, The Ecology and Management of Non-Timber Forest Resolurces No. 323 Pannier, editor, Corporate Governance of Ptublic Enterprises in Transitional Economies No. 324 Cabraal, Cosgrove-Davies, and Schaeffer, Best Practicesfor Photovoltaic Household Electrification Programs No. 325 Bacon, Besant-Jones, and Heidarian, Estimating Construction Costs and Schedueles: Experience with Power Generation Projects in Developing Countries No. 326 Colletta, Balachander, and Liang, The Condition of Young Children in Suib-Saharan Africa: The Convergence of Health, Nutrition, and Early Education No. 327 Valdes and Schaeffer in collaboration with Martin, Surveillance of Agricultural Price and Trade Policies A Handbookfor Paraguay No. 328 De Geyndt, Social Development and Absolutte Poverty in Asia and Latin America No. 329 Mohan, editor, Bibliography of Puiblications: Technical Department, Africa Region, July 1987 to April 1996 No. 332 Pohl, D)ankov, and Anderson, Restructutring Large Industrial Firms in Central and Eastern Europe: An Empirical Analysis No. 333 Jha, Ranson, and Bobadilla, Measuring thze Burden of Disease and the Cost-Effectiveness of Health Interventions A Case Study in Guinea No. 334 Mosse and Sontheimer, Performance Monitoring Indicators Handbook No. 335 Kirmani and Le Moigne, Fostering Riparian Cooperation in International River Basins: The World Bank at Its Best in Development Diplomacy No. 336 Francis, with Akinwumi, Ngwu, Nkom, Odihi, Olomajeye, Okunmadewa and Shehu, State, Community, and Local Development in Nigeria No. 338 Young, Measuring Economic Benefits for Water Investments and Policies No. 339 Andrews and Rashid, The Financing of Pension Systems in Central and Eastern Europe: An Overview of Major Trends and Their Determinants, 1990-1993 No. 340 Rutkowski, Changes in the Wage Structure during Economic Transition In Central and Eastern Eturope No. 341 Goldstein, Preker, Adeyi, and Chellaraj, Trends in Health Stattus, Services, and Finance: The Transition in Central and Eastern Euirope, Volutme I No. 343 Kottelat and Whitten, Freshwater Biodiversity in Asia, with Special Reference to Fish No. 344 Klugman and Schieber with Heleniak and Hon, A Sutrvey of Health Reform in Central Asia No. 345 Industry and Mining Division, Industry and Energy Department, A Mining Strategyfor Latin America and the Caribbean No. 347 Stock and de Veen, Expanding Labor-based Methodsfor Road Works in Africa THE WORLD BANK I'sIS II StL.Lt \.\\. \shI oiim,m D).( 204331 I S Ic1lcph mc: 2112-4771-1234 I;c,simlc: 2o2- +7T-_( 1) I 1 ( :1 2,4423 \\ () RI I)1: \ k (1I)IL' \.kh1-csN: IN\1'1,)\1 1, I V) \\ \SIiII,\(I \I \I ) w1 \\ l kl \ i. \\ is' II tt p: \I \i \% . \ ,, 1.1k. I) , -lo I -,,,;,il: III:. k ,.,'d. I1'.. ' ISBN 0-8213-3831-5