61294 A Break with History: Fifteen Years of Inequality Reduction in Latin America APRIL 2011 A Break with History: Fifteen Years of Inequality Reduction in Latin America APRIL 2011 Contents Executive Summary ····················································································································································································································· 1 1. Household income inequality in Latin America is on the decline······························································································ 3 2. Understanding today's inequality ··········································································································································································· 8 2.1 Labor markets ·····················································································································································································································8 2.2 Pensions and Monetary Transfers ····································································································································································13 2.3 Other factors ·····················································································································································································································14 3. Insights into tomorrow's inequality: is the decline sustainable? ···························································································· 17 Statistical Annex ·························································································································································································································· 22 lac poverty and labor brief Cover photograph: © istockphoto/Anastasia Khan Crédits: Design: Alejandro Espinosa/sonideas.com This brief was produced by the Poverty and Gender Unit from the Poverty Reduction and Economic Management Team (LCSPP) in the Latin America and Caribbean Region. The work was led by Joao Pedro Azevedo, Louise Cord and Carolina Díaz-Bonilla. The team included Bernardo Atuesta Montes, Andrés Castañeda Aguilar, María Eugenia Dávalos, Will Durbin, Gabriel Facchini and Carlos Sandoval. The team benefited from comments from Amparo Ballivian, Maurice D. Kugler, José Antonio Cuesta Leiva, Luis Felipe López Calva and Jamele Rigolini, and from key inputs from Francisco Ferreira, Sergei Soares, Miguel Foguel, and Leopoldo Tornarolli. The views and interpretations expressed here are the sole responsibility of the authors and should not be attributed to the World Bank, the Board of Executive Directors or the countries which they represent. The numbers presented in this brief are based on a regional microdata harmonization effort (called SEDLAC) to increase cross-country comparability of selected variables from the official household surveys. Given this objective, often the indicators constructed through this process are different from the official statistics, reported by Governments and National Offices of Statistics. Such differences should not be interpreted in any way as an indicator of methodological superiority since both numbers depart from different, yet equally important and valid, objectives: regional comparability and best possible national representation. The welfare aggregates used in this study are income based. v Contents Boxes Box 1. Caveats of income measures ················································································································································································· 7 Box 2. Other measures of inequality for Latin America ································································································································· 16 Figures Figure 1. Household Income Gini coefficient in Latin America, 1995-2009 ····················································································· 3 Figure 2. Annualized percent changes in household income inequality by subregional averages ···························· 3 Figure 3. Income share of top 10 percent of the income distribution in LAC ················································································· 4 Figure 4. Income Polarization Index ··············································································································································································· 4 Figure 5. Decomposition of Latin America household income inequality into "within" and "between" countries inequality (percent shares)·············································································· 5 Figure 6. Annualized Percentage Changes in Gini coefficient of BRIC countries, circa 1995-circa 2005··················· 5 Figure 7. Income Gini coefficient in LAC and OECD countries (2006 OECD, 2009 LAC) ·························································· 6 Figure 8. Decomposition of Latin American household income inequality, by share attributable to each source of income, 1995 and 2009············································································································· 8 Figure 9. Labor income Gini (individual hourly wages)···································································································································· 9 Figure 10. Decomposition of the Labor Income (individual hourly wages) Gini coefficient: quantity, price and residuals effects, 1995-2009 ········································································································ 10 Figure 11. Average returns to education, experience and other factors (1995-2009) ·························································· 11 Figure 12. Mean and Standard deviation of education and experience (1995-2009)··························································· 11 Figure 13. Absolute change in years of education (25-64 years) of top and bottom quintiles, 1995-2009········· 12 Figure 14. Share of monetary transfers (private and public) and pensions in total income in Latin America ·· 13 Figure 15. Marginal effect on inequality of each income source (elasticities) ··········································································· 14 lac poverty and labor brief Figure 16. Percentage change in dependency ratio, 1995-2009 ·········································································································· 15 Figure 17. Female labor force participation, 1995-2009, percentage point change······························································ 15 Figure 18. Change in the Educational Mobility Index for Teenagers (13-19), 1995-2009 ·················································· 17 Figure 19. Poverty rate ($2.5/day PPP) by age ······················································································································································ 19 Figure 20. Labor Income Poverty Index (LIPI) ­ 2.5 USD/PPP ··················································································································· 20 Figure B2.1. Urban and Rural Poverty, Poverty Gap and Poverty Gap Squared (country averages) ························· 16 Tables Table 1. Typology of Changes in the Gini coefficient of Labor Income (individual hourly wages), between 2000 and 2009········································································································································ 12 Table 2. Income inequality before and after taxation ···································································································································· 18 Table 3. Number of Latin American countries with subsidies ·················································································································· 18 Table B2.1. Regional household per capita income inequality by different measures ························································ 16 vi ANNEX Figures Figure A1. Ratio of hourly wages (prime-age males) between high and low educated workers, 1995-2009 ··· 23 Figure A2. Absolute change in years of education (25-64 years), 1995-2009·············································································· 24 Figure A3. FGT[2]% ­ 2.5 USD/PPP ················································································································································································· 24 Figure A4. FGT[0]% ­ 2.5 USD/PPP (adjusted for adult equivalence and household economies of scale) ·········· 25 Figure A5. FGT[2]% ­ 2.5 USD/PPP (adjusted for adult equivalence and household economies of scale) ·········· 26 Figure A6. Changes in the HOI (circa 1998-circa 2008) ·································································································································· 28 Figure A7. Mexico Labor Income Poverty Index (LIPI) ­ 2.5 USD/PPP ································································································ 28 Tables Table A1. Gini coefficient (household income) in Latin America ·········································································································· 22 Table A2. Population in the sample ··············································································································································································· 22 Table A3. The circa criteria ···································································································································································································· 23 Table A4. Typology of Changes in the Gini coefficient of Labor Income (individual hourly wages) by group, between 2000 and 2009 ················································································································ 27 Table A5. Subsidies in Latin America ············································································································································································ 29 lac poverty and labor brief vii Executive Summary In 2004 the World Bank released a regional report titled "Inequality in Latin America: Breaking with History?"1 Analyzing data from the early 1990s to the early 2000s, a period in which many countries in the region were experiencing increasing inequality, this study raised the question of whether Latin America could reverse its historical pattern of high and persistent inequality. The report concluded that although not easy, breaking with history was more than ever possible in the region. An additional decade of data presented in this brief shows that it was possible, mainly due to changes in labor markets (including a reduction in educational inequality and the skill premia, and greater female labor force participation), a higher incidence of government transfers, and additional factors such as demographic changes. Household income inequality in Latin America is on the decline Latin America is today a more equal region than 15 years ago. From 1995 to 2009, income inequality fell across a range of inequality measures; for example, the Gini and Theil declined 7 and 13.8 percent, respectively. Greater reduction of inequality within countries (as opposed to a convergence of income levels across countries) has driven the fall in inequality, as the share of income held by the top decile has declined in favor of the middle deciles. The improved income distribution has been witnessed throughout the region, but has been most notable in the countries of the Southern Cone with Brazil being a critical force behind the overall improvement in equal- ity. These trends in Latin America have occurred at a time when other developing economies, such as India and China, are experiencing increasing inequality. Nevertheless, household income inequality remains high in the region, and the least unequal country in Latin America (Uruguay) is still more unequal than the most unequal non-LAC country in the OECD (Portugal). Understanding today's inequality: labor markets, pensions, transfers, and other factors The decrease in inequality is driven mostly by improvements in labor income, particularly a reduction in lac poverty and labor brief skill premiums, reflecting improved access to education as well as other factors. Despite driving the decline in inequality, labor incomes continue to account for the majority of inequality in the region, reflecting their dominant share of total income. Pensions play a regressive role in the distribution of income in the region, although they clearly have an important role in keeping the elderly from falling into poverty, especially in the Southern Cone countries (i.e. Argentina, Brazil, Chile and Uruguay). Public and private monetary transfers, while small in scope, have contributed to the decline in inequality due to their growing size and better targeting. Between 1995 and 2009, transfers became relatively more important as a share of total household income rising from 1.8 percent to 3.1 percent. It is important to note that transfers, at least public monetary transfers (in particular the conditional cash transfers programs that have expanded in the region during this period) were not meant to be an instrument for fiscal redistribution, but were aimed more at reducing chronic poverty and promoting human capital development. 1 De Ferranti, David; Guillermo E. Perry, Francisco H. G. Ferreira and Michael Walton (2004) "Inequality in Latin America, Breaking with History?" Washington, DC: The World Bank. 1 Among other factors, the demographic transition (due to falling fertility rates) has the potential to continue reducing inequality in the next few years, although it might also pose important fiscal implications as the region must learn to cope with a larger elderly population (as life expectancy in the region is also expanding) and a smaller workforce. In addition, the increasing female labor force participation found across the region could continue to be a force behind declining inequality should the trend continue. Insights into tomorrow's inequality: is the decline sustainable? Household income inequality in Latin America has declined over the last 15 years, showing that the region can make a dent into its high levels of inequality. But is the decline sustainable and are there policy areas that can be tapped to support future accelerations? Five policy priorities are important if the current trends are to be sustained. First, labor income is likely to remain the most important determinant of inequality in the region. Going forward, it is important to continue to monitor the trends on the skill premia in the region, to better understand what factors are driving its change, and to what extent each one is contributing to the declining trend. Second, despite significant improvements, access to education still needs to expand in the region and the quality of education needs to be improved with special attention on schooling for low-income children. Third, the region needs to strengthen the role of fiscal policy in promoting equality of opportunities and to avoid generalized subsidies that are costly and not targeted. Fourth, arresting childhood poverty and malnutrition will be key to reducing poverty and ensuring that inequality continues to decline as poor children are able to take advantage of and fully exploit their increased access to opportunities. One out of every four children in the region lives below the extreme poverty line, suggesting that halting childhood poverty and malnutrition is an unmet goal even in Latin America. Fifth, the increasing frequency of exogenous shocks and climatic variability calls for a renewed effort to strengthen safety net systems and implement high frequency data collection efforts. lac poverty and labor brief 2 A Break with History: Fifteen Years of Inequality Reduction in Latin America 1. Household income inequality in Latin America is on the decline Household income inequality has been falling in Latin America and the Caribbean since 1995, with a more pronounced reduction in the most recent decade. In 2009, the Gini coefficient of the region for total household per capita income was 7 percent lower than in 1995, decreasing from 0.574 to 0.534 (Figure 1 and Table A1). A similar declining trend is captured by the Theil index of inequality, which drops by about 0.093 points between 1995 and 2009, from 0.677 to 0.584. The reduction in inequality accelerated substantially in the past decade (a 0.7 percent annual decrease between 2000 and 2009), compared to what was achieved between 1995 and 2000 (-0.3 percent) (Figure 2). These estimates of inequality in Latin America (LAC) cover 14 countries in the region, ac- counting for around 75 percent of the total population (Table A2), and use the circa criteria defined in Table A3. Although the decline in household income inequality has been widespread in the past decade, the magnitude of the change has been heterogeneous across countries and subregions in Latin America2. Comparisons by subregion suggest that the more pronounced fall in inequality after 2000 has been driven by countries in the Southern Cone region, with an average yearly change in inequality of -1.0 percent (Figure 2). In Figure 1. Household Income Gini coefficient in Latin America, 1995-2009 0.60 0.574 0.567 0.546 0.55 Gini coe cient 0.534 0.50 0.45 0.40 Circa Circa Circa Circa 1995 2000 2005 2009 Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank).Note: Calculated using total household income per capita. For details on the years (circa criteria) by country refer to Table A3. Figure 2. Annualized percent changes in household income inequality by subregional averages lac poverty and labor brief 0.0% Yearly % change in Gini index 0.0% 0.0% -0.2% -0.2% -0.1% -0.3% -0.4% -0.4% 0.6% -0.6% -0.7% -0.8% -1.0% -1.0% -1.2% 1995-2000 2000-2009 Southern Cone Andean Region Central America, DR and Mexico LAC Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using country population-weighted averages. For details on the years (circa criteria) by country refer to Table A3. 2 Central America, Mexico and DR region includes: Dominican Republic, Costa Rica, El Salvador, Honduras, Mexico, Panama; Southern Cone includes: Argentina, Brazil, Chile, Paraguay, and Uruguay; Andean region includes: Bolivia, Colombia and Peru. Subregional averages are calculated as population-weighted averages. 3 fact, Brazil, Argentina and Paraguay, representing around 50 percent of the population in the sample, achieved some of the fastest reductions in inequality in Latin America during this period (Table A1). The reduction in inequality has been driven mostly by an expansion of the middle class, not by improvements in the share of income of the extreme poor. A stagnant share of total income for the lowest decile (remaining between 0.9-1.0 in fifteen years), coupled with a declining share for the top ten percent of the distribution (Figure 3), suggests that those in the middle of the distribution are capturing a larger share of total income (however, see Box 1 for caveats on income measures, in particular those referring to incomes at the top). In addition, income polarization (the share of income held by either the extreme rich or poor) has fallen across the board, especially in the Southern Cone, resulting in an overall decrease of the polarization index of around 1 percent per year for Latin America (Figure 4)3. Figure 3. Income share of top 10 percent of the income distribution in LAC 48 45.7 46 44.8 % of total income 44 42.9 42 41.6 40 38 1995 2000 2005 2009 Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank). Notes: For details on the years (circa criteria) by country refer to Table A3. Figure 4. Income Polarization Index3 1.0% % yearly change in Wolfson polarization index 0.6% 0.5% 0.0% -0.1% -0.3% -0.3% -0.5% lac poverty and labor brief -0.5% -0.5% -0.8% -1.0% -1.0% -1.0% -1.5% -1.4% -1.4% -2-0% -1.8% 1995-2000 2000-2009 1995-2009 Southern Cone Andean Region Central America, DR and Mexico LAC Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using country population-weighted averages. Notes: For details on the years (circa criteria) by country refer to Table A3. 3 The Wolfson polarization index used here is based on the notion that a distribution is more polarized than another one (based on a notion of bi-modal distributions) if it shows a lower number of individuals or households around the median. Thus the Wolfson measure looks at the mass in the middle of the distribution. 4 Figure 5. Decomposition of Latin America household income inequality into "within" and "between" countries inequality (percent shares) 1.0 0.9 0.8 0.7 0.6 95% 97% 97% 96% 0.5 0.4 0.3 0.2 0.1 5% 3% 3% 4% 0.0 1995 2000 2005 2009 Between Within Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). Notes: The shares refer to a decomposition based on the Theil index. For details on the years (circa criteria) by country refer to Table A3. A more equal distribution of income within countries is the dominant force behind the drop in household income inequality, as opposed to convergence of income across countries. Almost all (96 percent) of total inequality in Latin America is explained by within-country inequality in 2009 (Figure 5), a consistent phenom- enon for the past 15 years. Moreover, 94.5 percent of the decline in inequality between 1995 and 2009 can be attributed to a reduction in inequality within countries. In contrast to the recent Latin American trends, Asia is witnessing rising inequality, pushed up by China and India ­ where income has traditionally been more equal. Focusing on the BRIC countries (Brazil, Russia, India and China) for the period between (circa) 1995 and (circa) 2005, Latin America is, on average, reducing inequality faster than these countries, for a decline in the Gini coefficient of 0.5 percent per year. Brazil had the Figure 6. Annualized Percentage Changes in Gini coefficient of BRIC countries, circa 1995-circa 2005 6% 6% Annualized % change in Gini coe cient 5% lac poverty and labor brief 4% 3% 2% 1% -0.30% 0% -0.39% -0.48% -0.50% -1% China India Russia Brazil LAC Source: LAC: author's calculation with data from SEDLAC (CEDLAS and The World Bank); Other countries: UNU-WIDER World Income Inequality Database (WIID). For China, the years are 1995 and 2003; for India 1992 and 2004, for Russia 1995 and 2005. 5 Figure 7. Income Gini coefficient in LAC and OECD countries (2006 OECD, 2009 LAC) Bolivia Colombia Honduras Brazil LAC Panama Chile LAC Paraguay countries Mexico in 2009 Costa Rica Dominican Republic Peru El Salvador Argentina Uruguay Portugal Greece Poland Estonia United Kingdom Italy Ireland Spain Norway Luxembourg Belgium A set of Germany OECD countries France in 2006 Hungary Netherlands lac poverty and labor brief Iceland Finland Slovak Republic Austria Czech Republic Slovenia Denmark Sweden 0 10 20 30 40 50 60 70 Gini coe cient Source: LAC: author's calculations with data from SEDLAC (CEDLAS and The World Bank); Other countries: UNU-WIDER World Income Inequality Database (WIID). Sample of OECD countries, measuring the Gini coefficient in terms of income, are Sweden, Denmark, Slovenia, Czech Republic, Austria, Slovak Republic, Finland, Iceland, Netherlands, Hungary, France, Germany, Belgium, Luxembourg, Norway, Spain, Ireland, Italy, United Kingdom, Estonia, Poland, Greece and Portugal. 6 fastest inequality decline (0.48 percent annual rate) followed by Russia (0.39 percent), whereas inequality sharply increased in China and modestly increased in India (Figure 6). However, inequality remains very high in Latin America, with levels significantly above other middle in- come countries. Figure 7 compares the income Gini coefficient of Latin American countries in 2009 with a set of OECD countries in 2006 (latest year available in UNU- Wider dataset). Without exception, all countries in Latin America have a higher level of inequality ranging from around 0.42 to 0.57 in 2009, compared to a set of OECD countries whose maximum inequality level reached 0.38 in 2006. In other words, the least unequal country in Latin America (Uruguay) is still more unequal than the most unequal non-LAC country in the OECD (Portugal). Box 1. Caveats on income measures In many developing countries, welfare measures, including inequality, are estimated using consumption rather than income data, since it is believed that consumption is less volatile and better captures well-being than income (e.g. Deaton and Zaidi, 2002). Few countries in the Latin American region routinely conduct national household surveys with consumption/expenditures-based questionnaires, while all of them include questions on individual and household income (Methodological Guide - SEDLAC). Hence, as nearly all distributional studies in the region, this brief uses income data to measure household welfare, but some drawbacks need to be kept in mind. Consumption measurement is not without its own challenges, as the recent work from Kathleen Beagle and others (2011) suggests significant discrepancies on consumption-based welfare measures depending on different methodological choices. Total household income can be misreported in a household survey, especially for the lower and higher ends of the income distribution. Incomes at the top of the distribution are usually underreported or not captured at all. This happens because these groups could be deliberately omitting some sources of income, and/or because data on capital income (a large fraction of income of the rich) is barely or not at all collected by surveys. On the other hand, income from farming activities in poor rural households can also be difficult to measure given that, for example, farm animals or home production are not easily valued. In addition, the income of those engaged in multiple jobs or illegal activities is also usually misreported in the survey. Finally, income measures could exclude or minimize wealth, and fail to reflect the value of public and private transfers, as well as the insurance value of government programs. The bias of the above points on inequality measures and on comparability across countries is ambiguous. Omit- lac poverty and labor brief ting income from the top of the distribution can result in a downward bias in inequality, but failing to fully account for incomes at the bottom of the distribution (e.g., farming income, small amounts of income from many sources) can push inequality upwards. For the US, consumption inequality is less pronounced than income inequality and changes in consumption inequality differ considerably from changes in income inequality (Meyer et al. 2010). Moreover, differences in the process of data collection (e.g., survey design, supervision) and in underreporting over time can affect comparabil- ity across years and countries. In order to address these measurement and comparability issues in the LAC region it is of critical importance to have the National Offices of Statistics working on the documentation of the magnitude of this challenge, in particular by triangulating information from anonymized tax records with the household surveys. Sources: (i) Methodological Guide - SEDLAC (CEDLAS and The World Bank); (ii) Khandker, Shahidur and Jonathan Haughton (2009) "Handbook on Poverty and Inequality". The World Bank; (iii) Meyer, Bruce D. and James X. Sullivan (2010) "Consumption and Income Inequality in the U.S. Since the 1960s". Deaton and Zaidi (2002) "Guidelines for Constructing Consumption Aggregates for Welfare Analysis", The World Bank. Beegle, K et al (2011) Methods of Household Consumption Measurement through Surveys: Experimental Results from Tanzania. Policy Research Working Paper 5501. Washington, DC. 7 2. Understanding today's inequality 2.1 Labor markets Labor income accounts for the highest share of total household income and contributes the most to Latin America's inequality. In 2009, labor income accounted for around 74 percent of total income, compared to 12 percent for pensions and 3 percent for monetary transfers (private and public). The share of labor income decreased by a total of 3.4 percentage points (or 4.4 percent) since 1995, as both transfers and pensions increased as a share of household income. Decomposing inequality by income source shows that almost 73 percent of total inequality in 2009 can be attributed to labor income inequality. However, while labor income reduced its contribution to inequality between 1995 and 2009, pensions gained importance as a disequalizing factor (Figure 8). Figure 8. Decomposition of Latin American household income inequality, by share attributable to each source of income, 1995 and 2009 100% 13.7% 12.8% 12.0% 11.7% 90% 1.4% 1.8% 1.8% 1.6% 80% 12.8% 9.2% 12.1% 13.9% 70% 60% 50% 40% 30% 75.8% 73.3% 73.5% 72.9% 20% 10% 0 1995 2000 2005 2009 lac poverty and labor brief Labor income Pensions Transfers Other Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank). Notes: Decomposition applies the methodology by Lerman & Yitzhaki (1985). For details on the years (circa criteria) by country refer to Table A3. Labor income inequality has declined in Latin America since 1995. Figure 9 shows a persistent decline in labor income inequality, more pronounced between 2000 and 2005 (3.2 percent total reduction in that period). In the case of labor income, 9 out of the 14 countries in the sample experienced a fall in inequality in the past fifteen years, with the largest declines taking place in the Southern Cone countries, particularly Chile (around 10 percent overall reduction from 1995 to 2009), Brazil (9 percent) and Paraguay (5 percent). Why is labor income inequality declining? The literature focuses on two key determinants of earnings and thus of labor income inequality: "quantity" effects (which capture the composition and dispersion of skills) and "price" effects (returns to observable characteristics or the skill premium). Looking at regional and country trends in 8 Figure 9. Labor income Gini (individual hourly wages) 0.60 0.562 0.556 Labor Income Gini coe cient 0.55 0.538 0.532 0.50 0.45 0.40 1995 2000 2005 2009 Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank). Notes: For details on the years (circa criteria) by country refer to Table A3. these outcomes can shed some light on the evolution of labor income inequality, measured in hourly wages and at the individual level. Changes in labor income inequality in the region are mostly driven by a decline in the skills premium, known as a "price" effect. Using a variation of the methodology proposed by Juhn, Murphy and Pierce which enables a counterfactual interpretation of the changes of inequality observed in the period4, changes in labor income inequality (individual hourly wages) over time are decomposed into a price effect, a quantity effect5, and residuals (i.e., other factors that affect labor income, such as characteristics of labor markets and macro policies). The price effect measures the returns to experience and education. The quantity effect, in turn, measures how changes in the dispersion of experience and years of education affect income inequality; more specifically, it aims to understand how much labor income inequality is explained by the dispersion of workers' skills, measured by years of education, experience, and the interaction of both (as opposed to the differences in the premium given to workers with the same type of characteristics). Figure 10 presents the observed total changes in the Gini, as well as the changes attributed to each component from 1995 to 2009. Results are interpreted as deviations from the year 2000. The driving factor behind labor income inequality declines, all else equal, appears to be the falling returns to skills, also known as the "price effect" (Figure 10). This finding is consistent with previous work exploring the lac poverty and labor brief fall in inequality in Latin America.6 In contrast, the quantity effect was slightly inequality increasing. 4 Chinhui Juhn, Kevin M. Murphy and Brooks Pierce (1993), "Wage Inequality and the Rise in Returns to Skill."; Journal of Political Economy, 101(3), pp. 410-42 and Foguel, Miguel Nathan and João Pedro Azevedo (2007), "Decomposição da desigualdade de rendimentos do trabalho no Brasil: 1995 -2005", in Desigualdade de Renda no Brasil: uma análise da queda recente, Ricardo Paes de Barros, Miguel Nathan Foguel and Gabriel Ulyssea eds., vol. 2, ch. 27, Brasília: IPEA. 5 Including years of education, experience (defined as four categories 0-10 years; 11-20 years; 21-30 years and over 30 years), and an interaction term. Years of experience is defined as age minus years of education minus six. 6 See for example, López-Calva, Luis Felipe and Nora Claudia Lustig (eds) (2010) "Declining Inequality in Latin America, A Decade of Progress?" Brookings Institution Press and United Nations Development Programme, c. 253pp. 9 Figure 10. Decomposition of the Labor Income (individual hourly wages) Gini coefficient: quantity, price and residuals effects, 1995-2009 a. Observed b. Quantities 0.05 0.05 Changes in the Gini from 2000 0.04 0.04 0.03 0.03 0.02 0.02 0.01 0.01 0 0 -0.01 -0.01 -0.02 -0.02 -0.03 -0.03 -0.04 -0.04 -0.05 -0.05 1995 2000 2005 2009 1995 2000 2005 2009 c. Prices d. Other Factors 0.05 0.05 0.04 0.04 0.03 0.03 0.02 0.02 0.01 0.01 0 0 -0.01 -0.01 -0.02 -0.02 -0.03 -0.03 -0.04 -0.04 -0.05 -0.05 1995 2000 2005 2009 1995 2000 2005 2009 Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using labor population-weighted averages. The decomposition follows Foguel and Azevedo (2007). The countries included in this analysis are: Argentina, Bolivia (only Urban), Brazil, Chile, Colombia, Costa Rica, Dominican Republic, El Salvador, Honduras, Mexico, Panama, Paraguay, Peru, and Uruguay (only Montevideo- urban). For details on the years (circa criteria) by country refer to Table A3. The unbundling of the "price effect" suggests that returns declined for all three components (i.e. education, experience and other factors) during the period of analysis, and that the pace of reduction accelerated after 2005. Figure 11 also suggests that the decline in the premia for education was the largest of the three, with an almost 25 percent reduction from 1995 to 2009. This reduction reflects the shrinking of the earnings gap between skill groups, captured in a reduction of the ratio of hourly wages between high and low-educates workers (Figure A1). It is important to emphasize that although returns are on average decreasing, the patterns are heterogeneous among countries in the region. lac poverty and labor brief The decline in returns to skills is related to an across-the-board expansion in educational attainment follow- ing decades of investments in education. From 1995 to 2009, there was an increase of 28 percent in the average years of education in the region (Figure 12.a). The increase was larger for those in the bottom income quintiles (Figure 13), with the largest gain for countries in the Southern Cone (almost 2 extra years of education since 1995 ­ Figure A.2). In particular, Brazil was the highest achiever in the period increasing from an average of 5.6 years of education in 1995 to 7.8 in 2009. An increase in school attainment has resulted in a widespread decline in the share of workers with only primary education, falling as much as 25 percentage points in Brazil, 14 points in Mexico and 13 points in Argentina from 1995 to 2009. Interestingly, the increase in educational attainment in Latin America has been accompanied by a convergence of educational attainment among individuals within and between the countries, reflected by a lower standard deviation over this period (Figure 12.a), and thus strengthening the evidence on a potential equalizing effect of education. 10 Figure 11. Average returns to education, experience and other factors (1995-2009) 115 110 105 100 95 Index 2000=100 90 85 80 75 1995 2000 2005 2009 Education Experience Other Factors Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages. The aggregation of the returns follows Foguel and Azevedo (2007). For details on the years (circa criteria) by country refer to Table A3. However, although educational inequality has decreased, the opposite trend in experience dominated result- ing in an inequality-increasing "quantity effect". In other words, even if the dispersion in educational level has fallen in the region (Figure 12.a), an increasing disparity of experience levels (i.e., higher standard deviation on the experience component over time, Figure 12.b) due to an aging population in the region offset the educational convergence within countries. Figure 12. Mean and Standard deviation of education and experience (1995-2009) a. Education b. Experience 120 102.0 101.5 115 101.0 110 100.5 Index (2000=100) Index (2000=100) 100 105 99.5 100 99.0 98.5 95 98.0 lac poverty and labor brief 90 97.5 1995 2000 2005 2009 1995 2000 2005 2009 SD Mean SD Mean Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages. For details on the years (circa criteria) by country refer to Table A3. Disaggregating by country reveals significant consistency across the region in the inequality-reducing nature of the "price effect" (i.e., a declining skill premium), but also shows significant heterogeneity in the impact of the quantity effect (Table 1). While the quantity effect was inequality-enhancing in most of the countries of the region, an exception occurred among the three largest countries of the Southern Cone (Argentina, Brazil and Chile) and Bolivia where it was inequality-reducing. It is important to note that, in some cases, changes in the three components partially offset each other, thus minimizing the net impact on inequality. A closer look shows that the overall directional impacts of the price and quantity effects were not affected by gender or sector of employment of the worker (formal or informal). More specifically, the 11 Figure 13. Absolute change in years of education (25-64 years) of top and bottom quintiles, 1995-2009 2.5 Bottom quintile 2.1 Bottom quintile Change in years of education 2.0 Top Bottom 1.8 quintile Bottom Top quintile Top 1.5 quintile quintile 1.4 quintile 1.5 1.4 Top 1.3 1.2 quintile 1.0 0.9 0.5 0.0 Southern Cone Andean Region Central America, LAC DR and Mexico Change in years of education for bottom quintile (left) and top quintile (right) Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages for the population from 25 up to 65 years of age. For details on the years (circa criteria) by country refer to Table A3. Table 1. Typology of Changes in the Gini coefficient of Labor Income (individual hourly wages), between 2000 and 2009 Inequality-reducing Inequality-increasing Argentina Colombia Bolivia Costa Rica Brazil Dominican Republic Chile El Salvador Quantity effect Honduras Mexico Uruguay Panama Paraguay Peru Argentina Costa Rica Bolivia Uruguay Brazil Honduras Chile Peru Colombia Price effect Dominican Republic lac poverty and labor brief El Salvador Mexico Panama Paraguay Argentina Chile Bolivia Honduras Brazil Mexico Colombia Paraguay Other Factors Costa Rica Peru Dominican Republic Uruguay El Salvador Panama Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank). The decomposition follows Foguel and Azevedo (2007). The countries included in this analysis are: Argentina, Bolivia (only Urban), Brazil, Chile, Colombia, Costa Rica, Dominican Republic, El Salvador, Honduras, Mexico, Panama, Paraguay, Peru, and Uruguay (only Montevideo-urban). 12 inequality-reducing "price effect" is consistent across these groups, while the "quantity effect" shows more heterogeneity (i.e., only for formal workers is the quantity effect inequality-reducing between 2000 and 2009). A further disaggregation of this typology by male and female, and individuals in formal or informal employment is presented in Table A4. The increased supply of skilled labor is a factor behind the fall in returns to education and inequality re- duction, but the demand for labor also plays a role. For example, although the supply of skilled workers also expanded during the 1980s and 1990s, returns to education increased driven by a higher demand for skilled labor (i.e., the demand effect associated with trade liberalization and technological change). In the 2000s, the expansion in education (supply-side) pushed returns down and this fall was only partially offset by an increased demand for skilled labor as in previous decades. Using Tinbergen's idea of a race between technology and education7 suggests that, this time around, education won in Latin America. Nonetheless, it is difficult to unambiguously determine how much of the decline in returns is due to supply-side factors (i.e., more skilled workers) or demand-side factors (e.g., changes in sectoral composition and labor demand for highly educated labor)8. This is likely to be heterogeneous across countries. In Argentina, for example, the shrinking wage gap between skilled and unskilled workers could be related to factors such as the commodity boom of the last decade, the exchange rate devaluation, and the role of labor unions, all of which pushed up the demand for unskilled labor relative to skilled labor9. 2.2 Pensions and Monetary Transfers Transfers (private and public) have become relatively more important as a share of total household income in Latin America. While the share of pensions 10 rose from 8.4 percent to 12.1 percent from 1995 to 2009,the contribu- tion of monetary transfers to total income increased by around 75 percent in the period, reaching 3.1 percent of total income in 2009 (Figure 14).One component captured in our measure of transfers is private transfers from remittanc- es- a key source of income for the region, estimated around $58 billion for 2011 and increasing in the coming years.11 Although we cannot systematically disentangle the impact of public vs. private transfers on inequality with the Figure 14. Share of monetary transfers (private and public) and pensions in total income in Latin America 14% 11.4% 12.1% 12% 10.8% 10% 8.4% 8% 6% lac poverty and labor brief 3.1% 3.1% 4% 2.6% 1.8% 2% 0% 1995 2000 2005 2009 Transfers Pensions Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). Notes: For details on the years (circa criteria) by country refer to Table A3. 7 Tinbergen, Jan (1975) "Income Distribution: Analyses and Policies". Amsterdam: North-Holland. 8 López-Calva, Luis Felipe and Nora Claudia Lustig (eds) (2010) "Declining Inequality in Latin America, A Decade of Progress?" Brookings Institution Press and United Nations Development Programme, c. 253pp. 9 Idem 10 Pensions include both contributory and non-contributory pensions. 11 World Bank (2010) "Remittances to Developing Countries Resilient in the Recent Crisis" Press Release No:2011/168/DEC, The World Bank. 13 available data, some studies suggest that international remittances have a limited redistributive role in the short and medium-term. Nonetheless, remittances can be inequality-reducing in the long-term as migrant networks strengthen and are made available to the unskilled12. On the other hand, in terms of public transfers, an in depth study of inequality in four Latin American countries finds that public transfers became increasingly equalizing after 2000, particularly from targeted spending through programs like Bolsa Familia in Brazil and Progresa/ Oportunidades in Mexico13. Pensions have increased household income inequality and continue to do so, while transfers, although still small in scope, have contributed to the decline in inequality. As previously seen in Figure 8, pensions have increased their contribution to total income inequality, increasing from 9.2 percent in 1995 to 14 percent in 2009. Computing an elasticity of inequality to each source of income over time suggests that pensions have become increasingly regressive (Figure 15). In particular, in 2009, a 1 percent increase in pensions increased total inequality more than proportionately by 1.8 percent. Conversely, transfers (public and private) have the highest inequality-reducing marginal effect with an elasticity of -1.5 percent. Figure 15. Marginal effect on inequality of each income source (elasticities) 2.0% 1.8% 1.4% 1.5% 1.2% 1.0% 0.8% 0.5% 0.0% -0.5% -0.4% -0.7% -1.0% -1.5% -1.3% -1.5% -2.0% Labor income Pensions Transfers Other 1995 2000 2005 2009 Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages. Notes: The marginal effect of changes on each income sources on overall inequality were estimates using the method proposed by Lerman, R. I. & Yitzhaki, S. (1985), Income inequality effects by income source: A new approach and applications to the United States, Review of Economics and Statistics, 67(1): 151-156. For details on the years (circa criteria) by country refer to Table A3. lac poverty and labor brief 2.3 Other factors Additional factors, such as socio-demographic changes, have had an impact on household income inequal- ity14. In particular, a decline in fertility rates in the region has resulted in an overall reduction in dependency ratios, more pronounced for the lowest income quintile (Figure 16). As household income is spread among fewer individuals, particularly in the poorest households in the region, there is room for increasing investments and standards of living, thus potentially reducing poverty and inequality. 12 Cornia, Giovanni Andrea (2010) "Income Distribution under Latin America's New Left Regimes". Journal of Human Development, 11(1); Mckenzie , David and Hillel Rapoport (2007) "Network effects and the dynamics of migration and inequality: Theory and evidence from Mexico" Journal of Development Economics, 84 (1): 1-24. 13 López-Calva, Luis Felipe and Nora Claudia Lustig (eds) (2010) "Declining Inequality in Latin America, A Decade of Progress?" Brookings Institution Press and United Nations Development Programme, c. 253pp. 14 Besides the additional factors discussed in this section, recent studies refer to the role of social-democratic political regimes in the region during the past decade, and how the policies put in place by them had a more pronounced redistributive effect. See for example: Cornia, Giovanni Andrea (2010) "Income Distribution under Latin America's New Left Regimes". Journal of Human Development, 11(1); Birdsall, Nancy; Nora Lustig and Darryl McLeod (2011) "Declining Inequality in Latin America: Some Economics, Some Politics". Center for Global Development. 14 Figure 16. Percentage change in dependency ratio, 1995-2009 Central America, Southern Cone Andean Region DR and Mexico LAC 0.0% -0.5% -1.0% -1.1% -1.5% Top -1.5% quintile -1.6% -1.8% Top -1.8% -1.8% -1.8% Top -2.0% quintile Bottom Bottom Bottom quintile Bottom quintile quintile quintile -2.1% quintile 2.5% Top quintile Annualized % change in the dependency ratio for the bottom quintile (left) and for the top quintile (right) Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using country population-weighted averages. Notes: The dependency ratio is defined as household size over income earners and it is presented here for the bottom equivalized income quintile. For details on the years (circa criteria) by country refer to Table A3. Women's labor force participation has sharply increased in all subregions of Latin America, particularly for low and medium-educated groups. A falling dependency ratio coupled with a considerable increase in female participation in labor markets can be a powerful source of inequality reduction. In 1995, the share of women in the labor force in Latin America was 54 percent, on average; by 2009, it increased to 63 percent (Figure 17), with the highest increase taking place in the Andean region. Moreover, the total increase in labor force participation (males and females) from 1995 to 2009 was larger for individuals with low and medium levels of education (2.1 and 3.0 percentage points, respectively), compared to those with higher education (-0.8 percentage points); this change was most likely driven by women. For the Andean countries, in particular, the change in total labor force participation was significantly higher (7.1 percentage points) for low-educated individuals than for the other educational groups (4.4 for medium-educated and -0.3 for the top) in this period. Figure 17. Female labor force participation, 1995-2009, percentage point change 14.0 12.0 11.5 lac poverty and labor brief 10.0 8.4 Absolute change 7.4 8.0 8.0 6.0 4.0 2.0 0.0 Southern Cone Andean Region Central America, LAC DR and Mexico Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages for the female population from 25 up to 65 years of age. Notes: For details on the years (circa criteria) by country refer to Table A3.. 15 Box 2. Other measures of inequality for Latin America Table B2.1 shows the recent trends for total household per capita income inequality using several complemen- tary measures. The Gini, for example, decreased 7.1 percent from 1995 to 2009, which is relevant since this indicator is significantly less sensitive to changes that occur around the extremes of the income distribution. Other inequality measures like the Theil index, the relative mean deviation or the mean log deviation, present decreasing trends of 13.8 percent, 8.6 percent and 13.1 percent, respectively. In short, the observed reduction in inequality seems to be robust to the choice of indicator. The increases in food and fuel prices in recent years, however, could be jeopardizing some of the attained progress in inequality. Figure B2.1 suggests that after 2007, urban poverty in the region started to increase -both in terms of incidence (FGT[0]) as well as depth and severity (FGT[1], FGT[2])- as the urban poor tend to be net consumers. Thus, it is ever more important to monitor recent trends (i) using various measures of inequality to better understand the dynamics of inequality (i.e., each measure conveys different information; for example, some are more sensitive to changes in the lower tail of the income distribution, while others capture movements in the middle), and (ii) paying closer attention to key population sub- groups (e.g., urban households), as they might be relatively more affected than others and thus drivers of inequality changes. Table B2.1. Regional household per capita income inequality by different measures Indicator 1995 2000 2005 2009 Gini 0.574 0.567 0.546 0.534 Theil 0.677 0.655 0.611 0.584 Relative mean deviation 0.425 0.418 0.400 0.388 Standard deviation of logs 1.091 1.093 1.045 1.048 Mehran Index 0.705 0.699 0.677 0.667 Piesch Index 0.509 0.502 0.481 0.467 Kakwani Index 0.271 0.265 0.247 0.237 Mean log deviation (or generalised entropy measure with sensitivity parameter 0) 0.623 0.611 0.560 0.542 Ratio of income means between 90th percentile and 10th percentile 18.343 18.400 15.978 16.704 Ratio of income means between 80th percentile and 20th percentile 5.676 5.662 5.216 5.103 Figure B2.1. Urban and Rural Poverty, Poverty Gap and Poverty Gap Squared (country averages) Urban Poverty Rural Poverty 25 60 50 lac poverty and labor brief 20 40 15 30 10 20 5 10 0 0 2000 2005 2006 2007 2008 2009 2000 2005 2006 2007 2008 2009 FGT(0) FGT(1) FGT(2) FGT(0) FGT(1) FGT(2) Source: Policy Brief on "Implications of the Rise in Food Prices for Poverty in Latin America" by the Poverty and Gender Unit in LAC, The World Bank (2011). Notes: The figures use SEDLAC harmonized datasets (CEDLAS and the World Bank). The countries used were: Brazil, Colombia, Costa Rica, Ecuador, El Salvador, Honduras, Paraguay, and Peru. The criterion for selecting these countries was the availability of microdata in at least five of the six years covered in this analysis. 16 3. Insights into tomorrow's inequality: is the decline sustainable? Household income inequality in Latin America has declined over the last 15 years, showing that the region can make a dent into its high levels of inequality. But is the decline sustainable and are there policy areas that can be tapped to support future accelerations? This section identifies five policy priorities that will be important in ensuring future reductions in inequality. 1. A more equitable distribution of labor market income has been the main force behind falling inequal- ity, reflecting declining skill premia and returns to education, as well as improved access to education. Going forward, it is important to monitor the trends on the skill premia in the region to better understand what forces are driving the changes. For example, a falling skill premia could reflect a greater supply of edu- cated workers, a weak demand for skilled labor, and issues related to the quality and other non-measurable characteristics of the labor force. In these instances, should the poor face greater constraints to obtain a quality education or skilled employment, the falling skill premia could potentially increase inequality15. 2. Higher levels of education among low-income families have helped reduce the skill premia and inequality in labor markets; however, some evidence suggests that the quality of educational out- comes remain stratified, which if not addressed could undermine continued declines in inequality. Since 1995, educational mobility has increased steadily in all sub-regions of Latin America, signaling that family background has become less important in explaining teenagers' educational attainment and their ability to access better employment opportunities as adults (Figure 18). However, a recent study tests for the intergenerational persistence of inequality using PISA scores and finds that Latin American countries have relatively higher rates of intergenerational persistence of inequality in edu- cational achievements than, for example, countries in Asia.16 Also employing PISA data, the Human Opportunity Index for quality of education is consistently lower for science, mathematics and reading Figure 18. Change in the Educational Mobility Index for Teenagers (13-19), 1995-2009 1.00% 0.83% 0.80% 0.59% Yearly % change 0.60% 0.56% lac poverty and labor brief 0.40% 0.27% 0.20% 0..0% Southern Cone Andean Region Central America, LAC DR and Mexico Source: Author's calculations with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages for the population between 13 and 19 years of age. Notes: The educational mobility index follows the methodology by Andersen (2001), in which the index is defined as 1 minus the proportion of the variance of the school gap (missing years of education) that is explained by family background. A higher index means a higher mobility, as family background is less important in explaining educational attainment. 15 LAC Skills for the 21st century (forthcoming), LAC Chief Economist Office, World Bank. 16 Ferreira, Francisco H. G. and Jérémie Gignoux (2010) "Educational Inequality and its Intergenerational Persistence: International Comparisons". 17 for Latin American countries than countries in Europe and North America.17 The sustainability of the decline in inequality is tightly related to improvements in the quality of education in Latin American countries because the equalizing effect of an expansion in educational attainment is likely to reach a limit if the high disparities in quality of education persist.18 3. Fiscal policy needs to be used more systematically to redress inequality in Latin America. Both limited tax collection and redistributive spending instruments constrain the use of fiscal policy in this regard. A study of the six largest Latin American countries finds that the distributional impact of the fiscal tax system is very limited in the region when compared to European countries. In particular, before direct taxes and transfers, the Gini coefficient of many European countries is not very different from the Latin America levels (e.g., UK with 0.53, Denmark with 0.49). However, in Latin America direct taxes and public transfers have only a modest impact on the Gini (e.g. around 1 to 2 Gini points reduction with transfers), while they reduce inequality by about 15 percentage points on average in Europe, two thirds of which can be attributed to public transfers (Table 2).19 On the spending side, the prevalence of generalized subsidies often offsets the growing expenditures on targeted programs and basic social services and infrastructure. The vast majority of countries in the region are using subsidies on food, fuel and transport as a tool for policymaking (Table 3 and Table A5). In Mexico, targeted subsidies are progressive in the country, but their redistributive effect is more than offset when generalized subsidies are accounted for, resulting in a regressive pattern of overall subsidies.20 Table 2. Income inequality before and after taxation Argentina Brazil Chile Colombia Mexico Peru Average Gini of Gross income 0.49 0.54 0.46 0.53 0.5 0.49 0.5 Gini of Disposable income 0.48 0.54 0.46 0.52 0.49 0.48 0.5 Gini of Post tax income 0.48 0.55 0.46 0.53 0.49 0.49 0.5 Source: Goñi, Edwin; J. Humberto López and Luis Servén (2008) "Fiscal redistribution and income inequality in Latin America". Policy Research Working Paper 4487, The World Bank. Notes: The authors refer to gross income as market income plus government cash benefits, or total household income before taxes. Disposable income refers to household income after receipt of government cash benefits (e.g. pensions, unemployment insurance, social assistance transfers) and after payment of direct taxes. Finally, post tax income is gross income minus both direct and indirect taxes. Table 3. Number of Latin American countries with subsidies Food Fuel Transport Water Number of countries with subsidies, by type of subsidy 11 16 18 7 As a percent of all countries (%) 50% 73% 82% 32% Note: 22 countries analyzed. lac poverty and labor brief 4. Despite important gains in access to basic goods and services that are critical for children to be able to develop to their full potential, more than one out of every four children in the region remains be- low the extreme poverty line21, suggesting that arresting childhood poverty and malnutrition is an unmet goal even in Latin America. Between 1998 and 2008, the Human Opportunity Index, which measures children's access to basic opportunities (covering education, water, sanitation and electricity), improved by one percent per year on average in the region (Figure A6). However, given the relatively high extreme poverty rate for children under ten, there are concerns that children may not be able to take full advantage of these expanded opportunities. Figure 19 shows the extreme poverty rate by age for 1995 and 2009 calculated with total household income, and with household income excluding 17 Molinas-Vega J., Barros, R., Saavedra, J. and Giugale, M. (2010) "Do Our Children Have a Chance? The 2010 Human Opportunity Report for Latin America and the Caribbean". The World Bank. 18 This conclusion is also supported by López-Calva, Luis Felipe and Nora Claudia Lustig (eds) (2010) "Declining Inequality in Latin America, A Decade of Progress?" Brookings Institution Press and United Nations Development Programme, c. 253pp. 19 Goñi, Edwin; J. Humberto López and Luis Servén (2008) "Fiscal redistribution and income inequality in Latin America". Policy Research Working Paper 4487, The World Bank. 20 Scott, John (2010) "Retos pendientes de la Política Social". Presentation to the Mexican Congress on March 10th, 2010. 21 Using a poverty line closer to the national poverty line of most countries in the region (4 USD/PPP) shows that one out of every two children in the region remains in poverty. 18 pensions and transfers.22 A first noticeable feature is the much higher poverty rate for children ten and under compared to other age groups. This is particularly worrisome as higher poverty incidence among future generations (i.e. children) might undermine their full potential, given that poor outcomes in the first years of life can determine levels of cognitive development, schooling outcomes and even adult earnings. (See Annex Figure A3 to A5.) A second feature is the large poverty-reducing effect of pensions for the elderly (in the absence of pensions and without any compensatory sources of income, poverty would triple for this group), coupled with the low trickle-down effect for children23. While the comparisons between the two years show a growing positive impact of transfers across all groups, in both years they have a disproportionately smaller impact on young children than the elderly. Moreover, it is important to bear in mind the fiscal implication of an aging population on the existing pension systems of Latin America. Figure 19. Poverty rate ($2.5/day PPP) by age 1995 60 50 40 FGT(0) % 30 20 10 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age 2009 60 50 40 FGT(0)% 30 20 10 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 lac poverty and labor brief Age Net e ect of Transfers Net e ect of Pensions Total Income Note: Extreme poverty rate by age calculated under three assumptions: "Total" uses total household income; "Net Effect of Pensions" uses household income excluding income from pensions; and "Net Effect of Transfers" uses household income excluding both income from pensions and transfers. Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). For details on the years (circa criteria) by country refer to Table A3. 22 This exercise assumes that resources are commonly distributed within households. Figures A.4 and A.5 in Annex report a similar exercise adjusted for adult equivalence and household economies of scale. 23 These aggregate trends can be further unpacked to show the importance of demographics on poverty rates by age. Cotlear and Tornarolli suggest that countries in the region can be organized in three groups: "pro-aging" countries where poverty continuously falls by age ­ Argentina, Brazil, Chile and Uruguay; "work-age" countries, in which poverty falls with age until around age 40, and then becomes either constant or slightly increase ­ Bolivia, Dominican Republic, El Salvador, Honduras, Panama, Paraguay and Peru; and the "anti-aging" countries, in which poverty increases steeply with age in a U-Shape ­ Colombia, Costa Rica, and Mexico. 19 5. Repeated shocks may put at risk gains in inequality reduction ­ at least for some of the region's more vulnerable countries, such as Mexico and parts of Central America, underscoring the importance of strengthening safety nets and high frequency social risk monitoring. Following the 2007/2008 spike in food prices, the world is experiencing a new round of prices increases. Based on the experience of 2007/2008 and the more moderate increase in prices that most countries in Latin America are currently experiencing, the latest acceleration of food prices may not lead to a dramatic increase in pov- erty levels in Latin America; however, certain vulnerable groups and countries may experience rising poverty and a negative welfare shock, which could offset falling inequality. The labor income poverty index (LIPI) provides high frequency insight on individual welfare, as it measures poverty in terms of individuals who cannot obtain the basic food basket with their labor income (which as noted above is the majority of the income of most households). With available data up to December 2010 from Brazil, Mexico and Peru, the LIPI shows that in some countries, and particularly for certain population groups (e.g. urban households in Mexico), the current shock is already having adverse impacts on poverty (Figure 20 and A7), which could make future reductions in inequality less likely (the rate of decline in inequality flattened noticeably in the region during the 2008 and 2009 crises).24 On the other hand, the rural poor who are also net sellers, particularly in some of the agricultural exporting countries of the Southern Cone may benefit from higher prices ­ which could reinforce past trends showing strong poverty reduction and inequality declines in that subregion. Figure 20. Labor Income Poverty Index (LIPI) ­ 2.5 USD/PPP 1.20 1.10 LIPI (Q1 2007 = 1) 1.00 0.90 0.80 0.70 0.60 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2007 2008 2009 2010 Mexico Peru Brazil lac poverty and labor brief Source: Policy Brief on "Implications of the Rise in Food Prices for Poverty in Latin America" by the Poverty and Gender Unit in LAC, The World Bank (2011). Notes: The figures use SEDLAC harmonized datasets (CEDLAS and the World Bank). Finally, as the region seeks to address the above mentioned challenges, it can benefit from the demographic dividend; however, it is important to note that this window of opportunity is likely to close around 2020. During the ongoing demographic transition, there is a larger proportion of working-age adults due to earlier high fertility rates. As a result, the region is likely to generate a demographic dividend that can provide resources to be geared towards inequality and poverty-reducing investments. This is a unique time for Latin America, as it benefits from a large workforce and low fertility rates. This favorable scenario is projected to continue until around 2020, when this ratio of workers/non-workers should reach its maximum, before starting to decline again, this time due to the growing proportion of older persons and a relatively smaller workforce. It is important to notice that while such a demographic transition lasted for over a century in developed countries, similar 24 Did Latin America Learn to Shield Its Poor? Poverty and Gender Group, PREM-Latin America and the Caribbean. World Bank, October 2010. 20 changes are occurring faster in Latin America and other developing countries today. France had 115 years to accommodate the doubling of its elderly population from 7 percent to 14 percent; in Latin America this process is happening much more quickly and the adjustment will likewise need to be quicker. Chile is projected to face this change in 26 years, Brazil in 21, and Colombia in 19 years.25 As Latin America enters a new decade, it does so knowing that inequality reduction is possible in the region. However, there are still many challenges to face in order to sustain the changes and continue improving equality, to extend it to countries that have thus far not joined the declining trend and to prepare for changes yet to come. Nonetheless, fifteen years of declining inequality in Latin America suggest that we have achieved a "break with history." lac poverty and labor brief 25 Cotlear, Daniel (Editor) "Population Aging: Is Latin America Ready?" The World Bank: Washington, DC. 21 Statistical Annex Table A1. Gini coefficient (household income) in Latin America Country circa 1995 circa 2000 circa 2005 circa 2009 Latin America 0.574 0.567 0.546 0.534 Argentina 0.481 0.504 0.490 0.451 Brazil 0.592 0.588 0.564 0.537 Bolivia 0.580 0.617 0.576 0.572 Chile 0.548 0.552 0.518 0.519 Colombia 0.554 0.572 0.562 0.560 Costa Rica 0.447 0.458 0.472 0.502 Dominican Republic 0.472 0.519 0.499 0.497 El Salvador 0.496 0.519 0.497 0.466 Honduras 0.552 0.541 0.559 0.553 Mexico 0.547 0.538 0.510 0.505 Panama 0.551 0.565 0.538 0.521 Paraguay 0.584 0.568 0.530 0.507 Peru 0.543 0.487 0.498 0.469 Uruguay 0.423 0.440 0.450 0.424 Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank) Table A2. Population in the sample Country 1995 2000 2005 2009 Argentina 14,075,355 20,415,792 23,141,580 24,475,148 Brazil 144,471,200 160,829,072 177,859,824 184,528,512 Bolivia 6,856,853 7,706,162 9,291,145 9,493,590 Chile 13,841,897 14,349,163 15,657,923 16,483,947 Colombia 34,447,848 33,728,520 33,311,334 41,513,688 Costa Rica 2,602,960 3,024,301 3,915,076 4,318,115 Dominican Rep. 7,376,168 8,227,291 8,942,731 9,406,675 El Salvador 5,365,552 6,060,664 6,743,786 6,056,755 Honduras 5,063,368 5,580,081 5,259,738 7,394,442 Mexico 90,836,712 97,189,008 101,923,144 104,540,704 lac poverty and labor brief Panama 2,258,821 2,838,800 3,134,515 3,359,728 Paraguay 4,770,148 5,004,014 5,698,069 6,197,857 Peru 23,326,216 17,371,980 20,806,196 21,583,876 Uruguay 2,839,713 2,268,053 2,285,602 3,021,443 Total sample pop. 358,132,800 384,592,896 417,970,656 442,374,496 Total population 480,743,996 518,543,254 552,987,439 578,877,469 % sample/total 74.5 74.2 75.6 76.4 Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank) 22 Table A3. The circa criteria Country Circa 1995 Circa 2000 Circa 2005 Circa 2009 Argentina 1995 2000 2005 2009 Bolivia 1997 2000 2005 2007 Brazil 1995 2001 2005 2009 Chile 1996 2000 2006 2009 Colombia 1996 2000 2004 2009 Costa Rica 1995 2000 2005 2009 Dominican Rep. 1996 2000 2005 2008 El Salvador 1995 2000 2005 2008 Honduras 1995 1999 2005 2009 Mexico 1996 2000 2005 2008 Panama 1995 2001 2005 2009 Paraguay 1995 1999 2005 2009 Peru 1997 2000 2005 2009 Uruguay 1995 2000 2005 2009 Average 1995.6 2000.0 2005.0 2008.6 Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank) Figure A1. Ratio of hourly wages (prime-age males) between high and low-educated workers, 1995-2009 6.0 5.5 5.0 4.5 4.0 3.5 3.0 lac poverty and labor brief 1995 2000 2005 2009 Southern Cone Andean Region Central America, DR and Mexico LAC Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank using population-weighted averages for the male population between 25 and 55 years. Notes: Low educated individuals are those with 0 to 8 years of education, while high-educated are those with more than 13 years of education. For details on the years (circa criteria) by country refer to Table A3. 23 Figure A2. Absolute change in years of education (25-64 years), 1995-2009 LAC 1.7 8.4 Andean Region 1.4 Years of educ 8.5 in 2009 2.0 8.5 Southern Cone Central America, 8.3 DR and Mexico 1.3 0.0 0.50 1.00 1.50 2.00 2.50 Absolute Change in Years of Education Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank) using population-weighted averages for the population from 25 up to 65 years of age. For details on the years (circa criteria) by country refer to Table A3. Figure A3. FGT[2]% ­ 2.5 USD/PPP 1995 30 25 20 FGT(2)% 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age 2009 30 25 lac poverty and labor brief 20 FGT(2)% 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age Net e ect of Transfers Net e ect of Pensions Total Income Note: Extreme poverty rate by age calculated under three assumptions: "Total" uses total household income; "Net Effect of Pensions" uses household income excluding income from pensions; and "Net Effect of Transfers" uses household income excluding both income from pensions and transfers. Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). For details on the years (circa criteria) by country refer to Table A3. 24 Figure A4. FGT[0]% ­ 2.5 USD/PPP (adjusted for adult equivalence and household economies of scale) 50 1995 45 40 35 30 FGT(0) % 25 20 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age 2009 50 45 40 35 30 FGT(0) % 25 20 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age Net e ect of Transfers Net e ect of Pensions Total Income lac poverty and labor brief Note: Extreme poverty rate by age calculated under three assumptions: "Total" uses total household income; "Net Effect of Pensions" uses household income excluding income from pensions; and "Net Effect of Transfers" uses household income excluding both income from pensions and transfers. Individual's equivalized household income is defined as total household income divided by (A+0.5*K1 + 0.75*K2)^0.9, where A is the number of Adults; K1 is the number of children under 5 years old;,and, K2 the number of children between 6 and 14. Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). For details on the years (circa criteria) by country refer to Table A3. 25 Figure A5. FGT[2]% ­ 2.5 USD/PPP (adjusted for adult equivalence and household economies of scale) 1995 30 25 20 FGT(2)% 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age 2009 30 25 20 FGT(2)% 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 Age Net e ect of Transfers Net e ect of Pensions Total Income lac poverty and labor brief Note: Extreme poverty rate by age calculated under three assumptions: "Total" uses total household income; "Net Effect of Pensions" uses household income excluding income from pensions; and "Net Effect of Transfers" uses household income excluding both income from pensions and transfers. Individual's equivalized household income is defined as total household income divided by (A+0.5*K1 + 0.75*K2)^0.9, where A is the number of Adults;K1 is the number of children under 5 years old;,and, K2 the number of children between 6 and 14. Source: Author's calculation with data from SEDLAC (CEDLAS and The World Bank). For details on the years (circa criteria) by country refer to Table A3. 26 Table A4. Typology of Changes in the Gini coefficient of Labor Income (individual hourly wages) by group, between 2000 and 2009 Males Females Formal Informal Inequality- Inequality- Inequality- Inequality- Inequality- Inequality- Inequality- Inequality- reducing increasing reducing increasing reducing increasing reducing increasing Argentina Colombia Argentina Bolivia Argentina Costa Rica Argentina Brazil Dominican Bolivia Costa Rica Uruguay Brazil Bolivia Bolivia Costa Rica Republic Dominican Brazil Chile Brazil El Salvador Chile El Salvador Republic Dominican Chile El Salvador Colombia Chile Panama Mexico Republic Honduras Mexico Costa Rica Honduras Peru Honduras Panama Quantity effect Uruguay Dominican Panama Mexico Uruguay Paraguay Republic Paraguay El Salvador Paraguay Peru Peru Honduras Uruguay Mexico Panama Paraguay Peru Argentina Costa Rica Bolivia Argentina Argentina Costa Rica Argentina Bolivia Bolivia El Salvador Brazil Costa Rica Bolivia El Salvador Brazil Costa Rica Brazil Honduras Chile Paraguay Brazil Uruguay Chile Honduras Dominican Chile Uruguay Colombia Uruguay Chile Uruguay Republic Dominican Dominican Price Colombia Peru Peru El Salvador Republic Republic effect Dominican El Salvador Honduras Mexico Republic Mexico Honduras Mexico Panama Panama Mexico Panama Paraguay Paraguay Panama Paraguay Peru Peru Bolivia Argentina Argentina Chile Argentina Chile Bolivia Argentina Dominican Brazil Chile Bolivia Honduras Bolivia Brazil Chile Republic Dominican Colombia Honduras Brazil Mexico Brazil Honduras Costa Rica Republic lac poverty and labor brief Other Costa Rica Paraguay Colombia Panama Costa Rica Mexico El Salvador Honduras Factors Dominican Peru Costa Rica Paraguay El Salvador Panama Mexico Paraguay Republic Dominican El Salvador Uruguay Peru Paraguay Panama Peru Republic Mexico El Salvador Uruguay Peru Uruguay Panama Uruguay 27 Figure A6. Changes in the HOI (circa 1998-circa 2008) Mexico Nicaragua Ecuador Brazil Peru Guatemala Dominican Republic Paraguay Colombia El Salvador LAC average Chile Honduras Costa Rica Uruguay Panama Venezuela Jamaica Argentina 0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 Annual change in percentage points Source: Molinas-Vega J., Barros, R., Saavedra, J. and Giugale, M. (2010) "Do Our Children Have a Chance? The 2010 Human Opportunity Report for Latin America and the Caribbean". The World Bank. For details on the years (circa criteria) by country refer to Table A3. Figure A7. Mexico Labor Income Poverty Index (LIPI) ­ 2.5 USD/PPP 1.30 1.20 1.10 LIPI (Q1 2007 = 1) 1.00 lac poverty and labor brief 0.90 0.80 0.70 0.60 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2005 2006 2007 2008 2009 2010 Rural Urban Total Source: Policy Brief on "Implications of the Rise in Food Prices for Poverty in Latin America" by the Poverty and Gender Unit in LAC, The World Bank (2011). Notes: The figures use SEDLAC harmonized datasets (CEDLAS and the World Bank). 28 Table A5. Subsidies in Latin America Summary Total Food Fuel Transport Water Countries (#) 22 11 16 18 7 Countries (%) 100% 50% 73% 82% 32% Details Possess Subsidy: Country LC Food Fuel Transport Water Costa Rica LC2 El Salvador LC2 x x x Guatemala LC2 x Honduras LC2 x x x Nicaragua LC2 x x x Panama LC2 x x Belize LC3 Dominican Republic LC3 x x x Haiti LC3 x x x Jamaica LC3 x x St. Lucia LC3 x x x St. Vincent and the Grenadines LC3 x Suriname LC3 x Trinidad & Tobago LC3 x x x Brazil LC5 x Bolivia LC6 x x x Chile LC6 x x x Ecuador LC6 x x x x Peru LC6 x Venezuela LC6 x x x x Argentina LC7 x x x x Paraguay LC7 x x x lac poverty and labor brief Uruguay LC7 x 29 www.worldbank.org/lac