Report No. 19992-AR Argenti na Poor People in a Rich Country (In Two Volumes) Volume 1: A Poverty Report for Argentina March 23, 2000 Poverty Reduction and Economic Management Latin America and Caribbean Region Document of the Wortd Bank Acronyms and Abbreviations ASOMA Apoyo Solidario a Los Mayores (Food Support for the Elderly) ENGH or ENIGH Encuesta Nacional de Gastos de Hogares (National Household Survey of Income and Consumption) EPH Encuesta Permanente de Hogares (Permanent Household Survey [labor force]) FAO Food and Agriculture Organization (UN) FONAVI Fondo Nacional de Vivienda (National Housing Fund) FONCAP Fondo de Capital Social (Social Capital Fund) FOPAR Fondo Participativo de Inversion Social (Participatory Social Investment Fund) GPC Gasto Puiblico Consolidado (Consolidated Public Spending) GPS Gasto Publico Social (Public Social Expenditures) INDEC Institute National de Estadistica y Censos (National Institute of Statistics) IPV Instituto Provincial de la Vivienda (Provincial Housing Program) MBA Metropolitan Buenos Aires MCE Ministry of Culture and Education NBI Necesidades Basicas Insatisfechas (Unsatisfied Basic Needs) PAMI Health Insurance for the Elderly PMI Programa Materno Infantil (Program for Mothers and Infants) POSOCO Politicas Sociales Comunitarias PRANI Programa Alimentario Nutricional Infantil PROEMPLEO Programa de Empleo (Employment Program) PROHUERTA Programa de Huertas (Garden promotion program) PROSONU Programa Social Institucional PSA Programa Social Agropecuario PSE Programa Social Educativo PYMES Pequefia y Medianas Empresas (Small and Medium Enterprises) i POOR PEOPLE IN A RICH COUTNTRY A Poverty Report for Argentina Table of Contents Executive Summary ........................................ iii Volume I: Part I: Overview I. Poverty, Growth and Distribution ....................................... 2 II. Employment, Unemployment and the Informal Sector .................... 19 III. Social Sector Spending: Does It Reduce Poverty? .......................... 27 IV. Reducing Poverty In Argentina: The Way Forward ........................ 39 Part II: Background Analysis Chapter I: The Determinants of Poverty ....................................... 52 Chapter II: Labor Markets, Employment and Poverty .......................... 75 Chapter III: Poverty and the Social Sectors ....................................... 94 Chapter IV: Poverty and Education .................. ..................... 106 Chapter V: Poverty and Health ....................................... 118 Chapter VI: Rural Poverty ....................................... 133 Chapter VII: Urban Development and Inequality ................................. 142 Volume II: Background Papers Task Team This report is a product of a team, led by Norman Hicks, which included World Bank staff members Alexandre Abrantes (health), Sandra Cesilini (poverty perceptions), Michael Cohen (urban),Wendy Cunningham (gender), William Experton (education), Luis Guasch (labor), Haeduck Lee (poverty data and analysis), William Maloney (labor), Vicente Paqueo (social protection), and Tom Wiens (rural). Contributions were also incorporated from Myrna Alexander, Carlos Arango, Omar Arias, Dariush Akhavan, Ana Maria Arriagada, Omar Arias, Mercedes Avellaneda, Katherine Bain,Walter Sosa Escudero, Estanislao Gacitua-Mario, Paul K. Freeman, Gillete Hall, Kihoon Lee, Mariana Marchioni, Pia Peeters, Steven Schonberger, Carlos Velez, Quentin Wodon, and Enrique Zuleta. The peer reviewers were Martin Ravallion, Ariel Fiszbein and Miguel Szekely. Valuable comments and advice were received from Myrna Alexander (country director), Guillermo Perry (chief economist), and Paul Levy (lead economist). Editing done by Marta A. Cervantes-Miguel and Tania Gomez. iii POOR PEOPLE IN A RICH COUNTRY A Poverty Report for Argentina Executive Summary Argentina is a relatively rich country. Yet despite this wealth, it is also a country with a relatively high level ofpoverty. Since 1991, the country has gone through a period of adjustment that has led to a remarkably sharp drop in rate of inflation, the privatization of state-owned industries, and the opening of the economy to foreign commerce. All of these adjustments have affected the poor, particularly through their effects on labor demand. Added to these shocks have been the recurrence of economic crisis, particularly in 1995 and 1998, which have also slowed the growth process. Progress on the economic front produced real gains in terms of reducing poverty and improving welfare. Poverty rates fell from 40% in 1990 to a low of 22% in 1994. However, since 1995 poverty has grown slightly as a percentage of the population, and income distribution has deteriorated. The deterioration of income distribution reflects the fact that while overall growth has been positive, and average per capita income has risen, the gains have gone largely to the more skilled and educated in the labor force, and not to the poor. In addition, unemployment has risen, and unemployment rates are higher among the poor and extreme poor, than for the general population. Many poor are underemployed or in temporary jobs ,or work in highly variable construction activities. Women, particularly poor women, have increasingly entered the labor force as a strategy to maintain household income. The lack offull time and secure employment is seen by the poor as one of their most critical needs. In general, poor families have low levels of education, have a large number of dependents, and are younger than families that are not poor. Large family sizes are the result of much higher fertility rates among poorer women, a factor that tends to perpetuate poverty. They live in areas lacking often in water and sanitation services, roads, and other public amenities, live in areas affected by flooding, and live in overcrowded conditions. The often lack titles to the land they inhabit, and therefore lack the incentive and the collateral to invest in their housing. The distribution of urban services is uneven between urban areas; some seem to do a better job than others in meeting these basic needs. The Government spends about 18% of GDP on social programs, however not all of these programs were designed to reduce poverty. The largest part of government spending is for social insurance, which provides pensions and some unemployment benefits to workers from the formal sector. However, workers in the informal sector, which are more often poor, do not receive these benefits. While workers in the formal sector enjoy benefits and relative job security, workers in the informal sector have neither. Most of the unemployment seems to come from the informal sector, and younger workers are more often unemployed Workers in the informal sector are not necessarily Iv poor; the poor can be found in both the formal and informal sectors, and there is movement in both directions between the these sectors. However, workers in the informal sector are both more prone to job loss and salary reductions, and are relatively unprotected against these events. Generalized social programs, principally in education and health, benefit all groups, and generally the poor benefit more than most. The poor particularly benefit from primary education, in part because they have larger families. Higher education spending, however, is highly regressive. Most of the students in public universities are not poor, and receive essentially afree education. Government programs that are specifically targeted to the poor generally work well, and are well targeted The more general problem is one of coverage. Only about 25% of poor families receive any form of direct public assistance, in the form of cash, food, etc. However, it is estimated that public and private transfers together probably reduce overall poverty by 4 percentage points, and are particularly important for the elderly. Government programs tend to be procyclical, and are reduced during downturns in the economy just when they are most needed And there are several government programs of limited value which could be reduced or redirected (housing, labor training). Shifting demand for labor has put a high premium on education. While rates of return to primary education are extremely low (about 3%o), returns to tertiary education is 29%. Despite these high returns, the poor often do not complete secondary school and are underrepresented in higher education. Repetition rates are high, as are dropouts. Only 24% of those aged 18-24 among the poor have a secondary education. The low quality of education, particularly in poorer areas, and the need to work all work against school completion. Rural areas tend to be ignored in most surveys, in part because Argentina is heavily urban. However, limited information suggests that there is substantial poverty among the rural population, particularly in the Northwest and Northeast. Most of these poor are notfarmers, butfarm and non-farm workers who are often unemployed and lack skills and education. The indigenous people of the rural areas seem to be particularly poor, since they live in remote areas away from public services. (i) Conclusions Future anti poverty efforts need to focus on three broad areas: * First, reforms and policies that will lead to a pattern of growth that will be more rapid overall, andfeature a higher level of employment per unit of output. * Second, improving the access of the poor to basic services, that will both raise their overall welfare and, by improving their human capital, improve the productivity and their ability to compete in an increasingly globalized economy; * Thirdly, reduce the vulnerability of the poor to shocks and losses in income, chiefly by improvements in safety nets to both protect the poor during economic downturns, v and keep them from making short term adjustments that will have negative impacts on their long term ability to reduce their poverty. Generating Labor Intensive Growth. Macro-economic policies that permit rapid and stable economic growth without inflation are an essentialfirst step to a significant decline in poverty. A sustained growth ofper capita income of 1. 8% could reduce poverty by 35% in ten years, provided the benefits go to all parts of the economy. This is more likely to happen ifArgentina 's labor markets operate efficiently. However, Argentina 's labor market is one of the most rigid and regulated in the developing world, preventing wage adjustments from taking place easily. Some of key short term reforms that would facilitate a more orderly operation of the labor market include: a elimination of centralized or sectoral collective bargaining agreements which are automatically extended to all workers in a sector, even if not signed and even when expired; * reduce the high cost of labor by reducing labor taxes, severance payments and moving to a fully-funded unemployment insurance system based on individual accounts; * allow temporary employment that is not subject to payroll taxes, as under the former modalidades promovidas, and * extending programs, such as PYMES, which permit exceptions for small scale enterprises. In the longer term, the critical problem remains that a large part of the labor force in the informal sector lacks any form of pension or unemployment insurance coverage. A major reform of the labor laws that would reduce their present high level of protectiveness should be followed by an extension of at least minimum coverage to smallfirms. Increasing Access to Services. A major effort should be undertaken to raise the level and quality of education available to the poor, and increase their access to secondary and higher education. One of the key problems is that children ofpoor families are more likely to drop out of schoolfor various reasons. The impact of the recent recessions in 1995 and 1999 seems to have actually worsened the situation with enrollment rates for the poor declining A viable strategy in education would include: * greater investments in secondary schools in poor neighborhoods, such as by extending the present Plan Social Educativo; * cash grants to poor families conditional on keeping children in school particularly at the secondary level, in order to offset the economic incentives from school leaving and the effects of unemployment; * the establishment of a system ofpartial cost recovery from students at public universities, who generally tend to be from non-poorfamilies, and the establishment of a nation-wide system of scholarships for students from poor families. * expanding the capacity of the current public university system, both by improvements in operating efficiency and through further investments. While the situation in health is less critical, greater efficiency in the health sector could improve the quality of service available to the poor. Particularly, the Government vi shouldfocus public health care expenditures on those without health insurance, by improving cost recovery from those with insurance and the ability to pay, and by improving the operating efficiency of the public hospital system. While granting more autonomy to public hospitals can improve their efficiency, care needs to be taken to avoid building in incentives that will reduce services to the poor. Eventually, health insurance coverage should be extended to those in the informal sector not presently covered. Existing programs of maternal and child health (PROMIN) need to be expanded, and linked with family planning and reproductive health services for the poor, in order to reduce their currently high rate offertility among the poor. Deficiencies in infrastructure both reduce the productivity of the poor, and limit human resource development. The urban poor live in areas usually devoid of adequate sanitation and safe water, and often without paved roads. Provision of such public services in poor neighborhoods can improve health outcomes. But attention also needs to be focused on building up communities, especially in urban areas, that lack roads, lights, and other services, and do not have legal titles to their land Existing large public sector subsidies for housing (FONA VI) which are not well targeted would be better reallocated to improvements in basic urban infrastructure. The urban poor are particularly vulnerable to problems of crime and violence, and attention needs to be paid to alcohol and drug abuse, and improvements in police protection and access to justice. Reducing Vulnerability. Recent economic "shocks " clearly demonstrate the needfor a strong system of safety nets. The Government needs to: • identify high priority programs that will be protectedfrom budget cuts during a crisis; • undertake a thorough evaluation of existing programs, eliminate weak programs, combine or streamline programs, and put more resources into programs which have proven effective; • identify programs that can be expanded during a crisis to provide emergency employment and income opportunities for the poor; and • take additional steps to improve targeting, so as to reduce leakage to the non-poor. P:\lcopr\poverty\summary-gray.doc POOR PEOPLE IN A RICH COUNTRY A Poverty Report for Argentina PART I. OVERVIEW 2 POOR PEOPLE IN A RICH COUNTRY A Poverty Report for Argentina --Among developing countries, Argentina is relatively rich, with an annual per capita income officially calculated at over $8000 per person. It is also a country that has undertaken major macro economic reforms and improved economic performance substantially in the past decade. Yet, despite this relative wealth and an extensive network of social programs, it is a country with a high degree of poverty, and a high degree of unemployment. The key question is why improved economic performance has not translated into greater improvements in welfare, especially among the poor? And what could be done differently in the future? I. Poverty Growth and Distribution Reform, Growth and Poverty During the past twenty years, Argentina has undergone a major economic restructuring, moving the economy from one with a large public sector, high inflation and high protection of trade and labor, to a more open economy with low inflation, and low levels of trade protection. The role of the public sector has been changed radically, with the privatization of state-owned enterprises in oil, transport and telecommunications, as well as basic infrastructure. In addition, the Government improved fiscal administration, eliminated many distortionary taxes, reformed the civil service, reformed the social security system, undertook a limited amount of labor reforms, and decentralized major social functions. The record of the past twenty years can be divided into two periods (see Table 1). In the first, during the 1980s, the economy was affected by the debt crisis and high fiscal deficits, which led to an inflationary spiral as the Government failed to take sufficient corrective steps. Inflation peaked at over 3,000 percent in 1989. While various stabilization programs were attempted, none of them successfully dealt with the problems of the fiscal deficit, aggravated by the losses of state enterprises and the erosion of the tax base by inflation. During this period, growth was negative and unemployment and poverty gradually climbed, peaking in 1990 in the aftermath of the hyper-inflation which erupted in July 1989. However, this period was also marked by a return to democracy, and the beginning of reforms, particularly in terms of trade liberalization or apertura. Import tariffs were reduced sharply and most quantitative restrictions on irnports were reduced (and largely eliminated by 1993). The second period, beginning with the Menem administration in 1989, a successful stabilization effort was undertaken. Inflation was brought under control through fiscal restraint and deregulation of the economy was begun. At the same time, privatization took the public sector out of production and provided fiscal space for expanding social programs and improving social safety nets. The Convertibility Plan, an important part of the stabilization effort, was introduced in 1991, linked the peso with the U.S. dollar on a one-to-one basis, and eliminating deficit financing through money creation. 3 By the late 1990s, growth had been restored, the economy had almost doubled in size, and inflation had fallen to less than 1% per year, a remarkable achievement in contrast to the earlier hyperinflation period. However, the control of inflation, deregulated and competitive markets and fiscal discipline seems to be associated with higher levels of unemployment, perhaps because employers were unable to use inflation as a means of lowering real wages. In addition, apertura, begun in the 1980s, increased competition from imports, requiring adjustments in the tradable sectors', and resulted in the introduction of new technologies that may have made some parts of the labor force obsolete or redundant. Thus, while the 1990s saw a decline in poverty and unemployment from the peak of 1990, these rates have never returned to the low levels of the early 1980s. Table 1: Long Term Trends in Poverty, Employment, Growth and Inflation (1980-1998) Year Poverty Rate Unemployment Inflation Growth of real Per Capita GNP Buenos Aires (% of labor force) (% increase) GDP(%) (1980=1100) 1980 8.0 2.0 100.8 4.5 100.0 1985 16.0 6.1 672.2 -2.0 79.5 1988 33.1 6.3 343.0 -1.9 83.6 1989 38.1 7.6 3079.5 -6.9 73.3 1990 41.2 7.5 2314.0 -1.8 74.3 1991 26.4 6.5 171.7 10.5 82.2 1992 18.7 7.0 24.9 9.9 90.0 1993 16.9 9.6 10.6 5.7 94.3 1994 17.0 11.5 4.2 5.9 100.3 1995 22.6 17.5 3.4 -2.7 94.5 1996 25.5 17.2 0.2 5.5 97.6 1997 25.2 14.9 0.5 8.1 104.1 1998 24.9 12.8 0.9 4.0 108.4 Note: Unemployment for 1980 refers to Greater Buenos Aires only. GDP growth for 1985 is for period 1980-85. Source: poverty rates: INDEC from EPH data; GNP from World Bank. The stabilization program was clearly successful in restoring growth, as well as controlling inflation. By 1998, per capita GDP had increased by about 30% over its 1988 levels, and the average overall GDP growth rate was 5.8% (1990-98). These macroeconormic events clearly impacted on poverty. Poverty rates (for Buenos Aires) rose quickly from a low of 8% in 1980 to a high of 41% during the hyperinflation of 1989, as fixed wages were not adjusted quickly enough to keep up with price changes. However, during the stabilization period of 1990-93, poverty rates did not fall back to their previous lows. Instead, urban poverty fell to 22%, and then rose as the economy stagnated under the impact of the Tequila crisis (see Table 2).2 Thus the 1990s can be divided into two sub-periods: 1990-94, when stabilization restored growth and reduced poverty and unemployment, and the period from 1995 to 1998, when growth slowed under the impact of the Tequila and Russia/Brazil crises, and unemployment and poverty remained high. As a result, urban poverty in 1998 was relatively high at 29%, with 7% of I Imports went from about 6% of GDP in 1988 (goods and non-factor services), to 11% in 1998. 2 For the period of the 1980s, poverty rates refer to only Buenos Aires. 4 the population living in extreme poverty3. In absolute numbers, given that there are about 32 million people living in urban areas in Argentina in 1998, this means that about 9 million are living in poverty and about 2 million in extreme poverty. Among children, poverty rates are even higher, with 45% of Argentina's children living in poverty. Table 2: Rates of Poverty, Extreme Poverty, and Poverty Gaps (urban areas only) Poverty Extreme Poverty Poverty Year Headcount Poverty Gap Depth (FGT2) 1990 41.4 11.3 16.4 8.8 1991 30.4 5.9 11.2 6.0 1992 24.1 4.4 7.8 3.7 1993 21.8 4.3 7.4 3.6 1994 21.6 3.7 7.2 3.4 1995 27.2 6.1 9.9 5.1 1996 30.0 7.3 11.3 6.0 1997 29.4 6.8 11.1 5.8 1998 29.4 7.1 11.2 5.9 Source: Calculated from EPH surveys, average of May and October. In this table the poverty rate includes the extreme poor. Poverty Estimates and Methodology Trends for poverty are based on the use of a poverty line, and the standard headcount and poverty gap measures. For this report, we use the Government's official poverty line calculated based on the 1986/87 Income and Expenditure Survey, updated using price indices for its food and non food components. This poverty line is equal to about $160 per male adult, per month, in 1998. The extreme poverty line, or indigence line, is based on the food consumption portion of the poverty line, and is equal to $69 per month in 1998. These lines applied to the Permanent Household Survey (EPH) which is undertaken semiannually by INDEC, and which covers approximately 26 urban areas in the country, or about 70% of the total urban population. However, data before 1990 refer only to Greater Buenos Aires, since data for other urban areas are not available on a consistent basis, or non existent in early years. These estimates are deficient in a number of ways. First, they exclude the rural population which would tend to understate poverty. Since there are no price indices outside of Buenos Aires, price adjustments over time for other urban areas are based on the Buenos Aires index, which may not be appropriate They are based on income poverty, not consumption, since there are no time series on household consumption data, and the existing survey in 1997 on consumption does not appear to give reliable consumption estimates. Finally, no attempt has been made to adjust the data for the apparent unreporting of incomes between the EPH and the national accounts. (For a more detailed discussion of data sources, estimation methods, and results see H. Lee, Background Paper No. 1.) 3 It should be noted, however, that the poverty line used here is $160 per month per adult, somewhat higher than that used in other Latin American countries. Estimates for Buenos Aires for 1999 indicate that poverty rose by about 7% in that urban area. 5 Figure 1: Changes in Per Capita Income, 1990-1998 1400 1200 1000 - 1st Decile O 5th Decile O- 80 10th Decile 600 - + 8th Decile 10 400 - _ 0 200 - _ X 0- *, *, * *1 * * 0 * 1990 1991 1992 1993 1994 1995 1996 1997 1998 Year Who Benefited from Growth? While growth resumed in the 1990s, the poor benefited relatively little in terms of income gains from this resumption in growth although they did gain in terms of social services and various other social programs. In other words, as per average capita income rose, the distribution of income worsened and the incomes of the poorest 20% actually declined. Table 3. Income Distribution (urban areas), 1990-98 (percent) 1990 1991 1992 1993 1994 1995 1996 1997 1998 Gini Coefficient 0.46 0.46 0.45 0.46 0.46 0.47 0.48 0.48 0.49 Upper 20% 50.90 51.40 50.40 50.90 51.10 52.20 53.00 53.20 54.25 Lower 20% 4.55 4.50 4.85 4.50 4.55 4.30 4.05 4.00 3.80 Upper 20/Lower 20 11.19 11.42 10.39 11.31 11.23 12.14 13.09 13.30 14.28 As shown in Table 3, the share of income received by the lowest 20% has declined since 1990, so that in 1998 they received less than 4% of total GDP. The upper 20% increased their share at the same time from 51% to 54%. In fact it is the 10 percent richest group that has gained the most, from 34 % to 37%. The share of the following twenty percent was constant, and the shares for the rest of the income groups fell by at least 10 percent. Furthermore, not only have the shares of the lower groups declined, but there has been virtual stagnation of real incomes or actual declines in incomes for the lower income groups. Given the changes in per capita income, the real income received by each group shows stagnation for the lowest groups (see Fig. 1), and rising income only for the highest deciles of the income distribution. Furthermore, since 1995, there has been a decline in average incomes for almost all groups, except for the highest 20% (see Figure 1). 6 We can estimate what the effect of changing income distribution has been on the poverty rate. If income distribution had remained as it was in 1990, we estimate, the poverty rate would have fallen more dramatically; and instead of reaching 29% in 1998 it would have been only 25% (for methodology, see Part II, Chapter 1). It is important to note that these calculations are based on urban poverty only and throughout the report these are the figures used. Given that poverty is probably higher in rural areas the correct national numbers for both income distribution and poverty would be slightly worse. On the other hand, since the country is about 89% urban, including the rural population would not make a major change in the conclusions.4 Table 4: Growth of Output and Employment, 1990-98 Sector Growth Rate of Growth Rate Growth Employment Standard Standard Employment, of Rate of elasticity, Deviation Deviation of 1980-90 Employment, Value 1990-98 of Growth Employment (Buenos Aires 1990-98 Added, only) 1990-98 Agriculture 8.5 1.6 2.1 .80 5.1 6.6 Manufacturing -0.7 -1.0 4.2 -.23 5.2 6.2 Construction -7.3 11.2 9.6 1.17 15.2 13.0 Trade -1.1 4.2 4.8 .87 6.1 6.8 Finance 4.2 2.9 6.3 .47 6.5 5.8 Services 3.8 1.2 4.2 .28 4.9 3.9 Total 0.9 2.2 4.5 .48 4.8 2.7 Note: Growth rates are based on end points; standard deviations are from annual averages. The factors causing a worsening of income distribution and fall in incomes are complex, but an important one has been the change in labor demand. During the 1980s, a declining GDP resulted in only small amounts of employment growth (about .9% per annum) and rising unemployment as the labor force continued to expand (see Table 4). However, even with the resumption in growth in 1990, employment growth has been sluggish. While value added grew on average by 4.5% during 1990-98, employment grew by only 2.2% for an overall elasticity of only .48. In manufacturing, employment declined despite positive growth in output, suggesting that new investments and modernization tended to be capital, rather than labor, intensive. The result has been very large increases in productivity in the manufacturing sector. (Such tendencies, however, may reflect the temporary adjustments to the new regime of openness, and are similar to those experienced by Mexico and Chile during their transition periods.) In addition, the low employment creation in services in particular-which might normally have been a source of employment for the unskilled-is striking and is partly the result of the privatization of public enterprises and the shedding of excess employees. One of the more dynamic sectors-which likely benefits the poor-has been the reactivation of the 4 Argentina badly needs a comprehensive household survey that would include rural areas, and provide a consumption based poverty estimates over time. It also needs to have price indices for urban areas outside of Buenos Aires, and for rural areas. The present work plan of INDEC includes launching a more comprehensive living standards survey, and to expand the coverage of the consumer price index. 7 construction industry during the 1990s. It is likely that this has been a significant sources of employment for the unskilled, albeit on a temporary basis. Table 5. Income from Primary Job by Skill Level Year Professional Skilled Unskilled (monthly income in 1998 constant prices, pesos) 1990 1176.7 570.2 366.2 1992 1483.6 685.2 438.9 1994 1715.8 725.4 424.4 1996 1661.9 632.5 354.5 1998 1794.1 644.8 356.2 Growth, 1990-94 45.8% 27.2% 15.8% Growth, 1994-98 4.6% -11.1% -16.1% Growth, 1990-98 52.5% 13.1% -2.7% Source: INDEC/EPH / annual data are the averages of two surveys While the poor, typically unskilled, have not benefited from growth, the growing demand for more highly skilled workers has increased the wages of those with secondary and university level education. As shown in table 5, the incomes of professionals grew by a total of 53% over the 1990-98 period, while skilled wages rose 13% and those of the unskilled declined by 3%. Furthermore, looking at the data only since 1994, we find substantial declines in the real wages of both skilled and unskilled workers, while professional workers have made small gains. What is not clear is whether the change since 1994 represent a permanent trend reflecting international comparative advantage, or reflect temporary changes in labor demand and lags in adjustment. Table 6: Returns to Education and Predicted Earnings by Education Level Years of Returns to Education Schooling (% increase in income) 1992 1998 2 6.8 5.0 4 7.2 6.1 5 7.3 6.7 6 7.5 7.2 8 7.9 8.3 10 8.3 9.4 12 8.6 10.4 14 9.0 11.5 16 9.4 12.6 18 9.8 13.7 sources: INDEC, Encuesta Permanente de Hogares, October of 1992 and 1998. The indicated returns are for cumulative education through the year indicated, not the marginal returns. Since professionals and skilled persons tend to be located in the upper levels of the income distribution, an increase in their incomes relative to others naturally tends to worsen the distribution. The same effect can be seen from a measure of the returns to education. As Table 6 shows, estimates reveal that individual workers with higher education attainment were reaping higher returns for their investment in 1998 than in 1992. For those with only two years of schooling, the returns to education dropped from 6.8 to 5.0%. And in general, for those with up to eight years of schooling returns to 8 education have been falling between 1992 and 1998, while those with more than 8 years of education have seen rising wages. At the extreme, for those with 18 years of schooling returns increased from 9.8% to 13.7%. An analysis, which decomposes the changes in the Gini coefficient, finds that about 40% of the change in the Gini is due to the differences in returns to education. Who are the Poor? Who are the poor, and how do they differ from the non-poor? Overall, the poor: - have significantly larger families (4.6 vs. 3. - have younger families with a much higher dependency ratios (3.0 vs. 1.4) - have much higher unemployment rates (twice the rate of the non-poor), - have fewer years of schooling (about 25% less), and - are more likely to work in the informal sector. As shown in Table 7, the differences are clearly even more significant in these respects for the indigent (defined as those lacking sufficient money to afford a basic food basket). These indigent poor represent about 7% of the population (1998) and are those at the extreme margins of poverty. Most indicators are worse for the indigent, compared to the poor: they are younger, have more children, have more people depend on one wage-earner, and particularly important have much higher unemployment rate (37%), and lower average hours worked (35, compared to 42 for the poor). The latter seems to represent their inability to secure full time work. Moreover, given their low level of education-about 7 years on average-and the declining returns to workers with that level of education, they have poor prospects for increasing their incomes in the current labor market. While female headed households are often thought of as poor, this is not borne out by the data. In addition, available evidence suggests that the poor are not, as is often alleged, the result of migration from outside of Argentina (for details see Part II, Chapter 1). 5 Larger families can utilize scale economies to maintain welfare even at lower per capita income levels. However, the nature of scale economies is unknown. In order to reduce the bias against large families, however, poverty rates calculated in this report are done on the basis of converting family members into adult equivalents. 9 Table 7: Characteristics of the Poor, 1998 Indigent Poor Non-Poor Family size 5.6 4.6 3.1 Family size, adult equiv. 4.3 3.7 2.5 Female Headed (%) 26.8 21.0 37.7 Average Age (years) 20.9 25.1 34.7 Dependency Rate (per worker) 4.1 3.0 1.4 Labor Force Participation (% of 15-64) 54.8 56.4 67.3 Years of Schooling 6.9 7.9 10.5 Informal employment (%) 46.4 44.6 36.2 Unemployment Rate (%) 36.5 22.6 8.9 Hours Worked 34.9 41.7 45.1 Source: EPH 1998. Demography of Poverty As indicated above, the poorest households in Argentina tend to be younger, larger and have more children. The average age of the lower income quintile household is just 25 years, while the mean age of the upper income quintile households is 41 years. The average size of the lower income quintile household is 5.1 people and only four percent are single person households. Sixty percent of these households have children aged 0-19 years and only 1 percent of these households have no children. Table 8: Poverty Rate by Age Group and Sex (Urban Argentina), 1998 Total po ulation Female population Age group % indigent % poor % indigent % poor Total 7.1 29.4 7.2 30.1 0-4 12.2 43.2 11.5 43.0 5-14 13.0 45.3 12.5 44.4 Subtotal, 0-14 12.7 44.6 12.1 43.9 15-24 7.0 30.6 7.1 31.1 25-39 5.6 26.0 5.1 25.1 40-64 4.5 21.0 4.8 21.7 Subtotal, 15-64 5.6 25.4 5.6 25.6 65 & older 1.4 13.2 1.7 15.2 Source: Average of EPH for May and October, 1998, all urban areas. Because poor families are both larger and younger than average, children are more apt to be in poverty than adults. As shown in Table 8, 45% of all children (aged 0-14) were living in poverty in 1998, compared to a 25% poverty rate for adults. Poverty among older adults is below average. Only 13% of those 65 and older were found to be below the poverty line. The lower rate of poverty among older people probably reflects the fact that the formal sector, which covers about 60% of the labor force, ensures that retired workers receive adequate pensions. An obviously important factor in explaining differences between the non-poor and the poor is the size of family and the number of children. Population growth in Argentina is relatively low, at 1.3% per year. Total fertility rate was 2.7 per 1,000 women aged 15- 10 45 in 1995, which is about the average for the Region, but are high when compared with other middle income countries such as Chile and Uruguay and higher than one would expect from the country economic and educational situation. Birth and fertility rates are clearly higher among lower income groups. While birth rates are generally not available by income groups, they can be approximated from the Social Development Survey (see Table 9). These estimates show a crude birth rate that varies from 45 for the lowest quintile, to 14 for the upper quintile. Likewise, the fertility rate for the lowest quintile is about 4.2 per thousand, compared to 1.3 per thousand women in the richest, quintile. Unfortunately, there is no way to estimate death rates, or the rate of population growth by quintile. However, from these estimates we can see that there will be a natural tendency for poorer families to increase faster than richer families, which is likely to perpetuate a cycle of poverty, and a gradual worsening of income distribution. Given these demographics, programs that provide maternal and child care, as well as reproductive services, will be naturally tend to benefit the poor. Table 9: Average family age, size and composition by household income quintile Per capita Household Income Quinliles _ _ _ __ _ 1 2 3 4 5 Average Age 25 31 35 37 41 Average family size 5.1 4.2 3.8 3.5 3.1 Percent of households with no children 1 1 3 3 10 Percent of households w/ children aged 0-19 66 59 53 48 36 Estimated Crude Birth Rate(births per 1000)6 45 32 23 18 14 Estimated Fertility Rate(births per 1000 women 4.2 2.9 2.0 1.7 1.3 aged 1544)' ___ Source: SIEMPRO, Social Development Survey (EDS), 1997. Table 10: Social Indicators and Deviations from Expected Values Indicator Argentina LAC average 1985 1997 Predicted Deviations* 1996 Values Life Expectancy (years) 71 73 74 -.6 70 Infant Mortality(per 1000 27 ** 22 14 54.8 33 births) Enrollment, Secondary 70 77 84 -8.6 53 Schools(%) Adult Literacy (%) 95 96 91 5.0 85 Access to Safe Water (%) 55 65 87 -25.5 76 Notes: * per cent differences from expected values, based on what would be expected for Argentina's level of GNP per capita. See Part 1, Chapter 1, for details. ** 1987 data. 6 This estimate is made by taking the percentage of children in 1997 aged 0-2 for each quintile, dividing by 2 (to equal one year's births), and multiplying by 10 (to convert from per 100 to per 1000), to obtain a rough estimate of the crude birth rate per 1000 population. This measure is essentially the crude birth rate less infant mortality, since deaths during the first year would be excluded, and the average age of children 0-2 would be one year. 7 Calculated with the same method as the crude birth rate, except the denominator is the number of women aged 15-64. 11 Social Indicators: Another View of Poverty Poverty is a multi-dimensional phenomenon and is not only the absences of money or material goods. It can reflect various shortcomings or negative aspects in life-including bad health, lack of education, malnutrition and other factors such as poor housing, violence and inability to participate in political processes for example. While Argentina exhibits social indicators that are above average when compared to other countries in the region (see Table 10), such good performance is to be expected given its high per capita income. If adjustments are made for the level of per capita income, in fact Argentina's social indicators are at or below average. For instance, infant mortality is about 55% higher than would be expected for a country of Argentina's income, even though it is below the LAC average. Likewise, secondary school enrollment is 9% below what should be expected, and access to safe water is 24% below the expected level. On the other hand, Argentina performs at or slightly above the expected levels for life expectancy and literacy. Voices of the Poor How do the poor see their poverty? The above statistics obtained from a national household survey have been confirmed by a separate survey on perceptions. A survey undertaken for this report of 1,200 households in 29 cities8 asked people if they considered themselves poor, 26% responded positively. When those who self-identified as poor were asked the source of their poverty, the most important responses were: not having work 31% having a low salary 18 not being able to pay for basic needs 14 difficult retirement 7 labor problems 7 What is clear from this is that the poor give heavy weight to employment, higher than that given to wages per se or being able to pay for basic needs: they see poverty as a problem of employment, underemployment and low wages, and less as not being able to access basic needs as health, education and sanitation. In the eyes of the poor, being unemployed or not have steady work implies a loss of dignity, as well as a loss of income. These results were confirmed and amplified through a series of focus group discussions. These discussions revealed the sense of hopelessness among the extremely poor concerning their ability to improve their situations. The lack of steady work and the problems of violence and insecurity were considered the most important problems facing the urban poor. Lack of employment leads to problems of alcohol and drug abuse, crime and violence in the home and in the community. In rural areas, attention focused on the s See S. Cesilini and E. Zuleta, Background Paper No. 2. 12 lack of land, credit, and access to markets. Both rural and urban poor cited problems derived from the lack of adequate infrastructure: water, sanitation, roads, lighting. A common and sweeping complaint by the poor consulted for this report, and consistent with other public opinion polls in Argentina, is the lack of credibility of public institutions in meeting their needs. Institutions are valued by their ability to meet the poor's immediate needs and few public institutions are seen as doing that. Government at all levels are seen to be corrupt, unresponsive and inefficient. Yet, a strong role for the Government is considered by these same people. There is another important feature of the above survey results: with the exception of the indigent, who do see their level of poverty very clearly, 59 percent of those classified as poor by their socio-economic levels actually did not think of themselves as poor while some of the middle-class groups and even the upper class thought of themselves as poor. This seems to suggest that poverty is very much a relative concept, dependent on social and cultural considerations. Lack of work also signifies a shift in gender relations with poor women able to earn more than their partners in domestic service. This is also seen to be the source of tensions within family and domestic violence appears to be increasing. Youth also stand out as a particularly vulnerable population. Poor youth report high incidences of drug and alcohol abuse and a general feeling of despair as they are faced with a future of insecurity marked by unemployment and violence. The police are identified in rural and urban areas as playing a prominent role in promoting violence and corruption rather than offering public security. While the poor are generally wary of the agendas of any external organizations, they point to their own informal networks and contacts as mechanisms that promote cohesiveness and unity and, perhaps, provide a haven in an increasingly vulnerable context. Poverty and Infrastructure. The poor state of basic services is a major concern among both rural and urban populations consulted for this report. In part, this can be explained by the fact that, unlike many other countries, the poor of Argentina possess a higher standard of living, as evidenced by their assets and living conditions. For instance, of the urban poor9: > 78% own a refrigerator > 85% own a television (31% have cable); > 59% own a washing machine > 19% own a car, and > only 4% lack electricity. Yet, the reality is that most live in difficult circumstances in poor neighborhoods that lack basic infrastructure services. Many do not have formal titles to the land they occupy, and their houses are constructed informally and illegally, often in areas lacking adequate drainage or near garbage dumps or other marginal areas. If we focus not on the 9 Data here are from the Social Development Survey of SIEMPRO (1997); "poor" here is defined on the basis of structural poverty (NBI). 13 poor, but the lower 20%10o, we find that crowding is high; over 50% live in houses with two or more persons per room; 17% live in houses with 3 or more persons per room, defined as "critical" crowding (see Table 11). While only 4% lack electricity, most of these connections are probably illegal. Table 11: Housing and Infrastructure Indicators, 1997 (percent of households) Indicator: Lower 20% Upper 20% Poor (NBI) Non-poor Crowding: -more than 2 persons/room 51.0 2.8 71.8 15.8 - more than 3 persons/room 17.2 0.0 46.8 . I Without access to safe water 8.3 .2 14.0 1.7 Without access to running water in 34.2 .8 54.4 6.3 the house Without adequate sanitation 47.0 5.1 68.0 15.6 Without electricity 4.4 .3 6.7 .5 With insecure tenancy of land 44.6 11.5 58.0 18.8 Located in flood area 28.5 11.9 33.6 16.9 Source: SIEMPRO, Social Development Survey (EDS), 1997. Note: Poverty defined as having one or more deficient basic needs. NBI is an index of unsatisfied basic needs, based on having any one of five deficiencies. For definition, see Part II, Chapter 3. Access to other public services is more critical; 47% lack adequate sanitation, and 42% lack running water in the house. In fact, the access to safe water is one of the main social criteria in which Argentina falls well behind where it should be in terms of international comparisons. Water drawn from wells and other sources is often contaminated by sewerage and chemical pollution, and even piped water is often Box: Privatization and the Poor As part of the reforms undertaken in the 1990's, a major program of privatization of public services was undertaken. This program included telephone services, transport, banks and productive facilities. It is often alleged that such privatization impact adversely on the poor, since they often result in higher prices for basic services which were previously subsidized, even if they the quality of services being provided. In Fig. 2, the prices of basic infrastructure services are compared to the general price index (CPI). We can see that most prices did not even keep up with inflation during the 1990s. Only after 1997 do the prices of urban transport and telecommunications exceed the general price index. However, an examination of household budgets reveals a large increase in consumption of basic infrastructure services. Between 1986 and 1996, families in the lowest quintile of the income distribution increased their expenditures on infrastructure services from 9% to 17% of their total budget, at a time when monthly household expenditures in real terms were declining. There are two possible causes for this phenomenon: better services offered after privatization may have induced more spending; or, privatization also improved the collection rate for fees charged, without necessarily raising prices. Source: N. Hicks and H. Lee, "The Impact of Privatization on the Poor: the Case of Argentina." For similar conclusions, see F, Navajas, "El Impacto Distributivo de los Cambios en Precios Relativos en la Argentina." FIEL, 1999. 10 The focus shifts here to the lower 20%, because the poverty definition being used is the NBI, which itself is based on housing quality indicators. 14 contaminated. About 45% live on land with insecure tenancy, that is without legal title or a formal arrangement. The lack of clear titles reduces the ability of the poor to finance housing improvements and reduces the incentive for the poor to upgrade the quality of their housing. In addition, almost twice as many poor as non-poor live in areas that are prone to flooding. By any standard, Argentina is highly vulnerable to economic loss from flooding. Within the developing countries of the world, Argentina falls within the top seven in terms of size of risk compared GDP. In Latin America, Argentina has the highest exposure to direct flood loss. This further increases the vulnerability of the poor to catastrophic losses. Fig.2: Prices of Public Services, Buenos Aires (1988=100) 350000 - 300000 4-*CPI 250000 = = Electrcity 200000 -* - san services, gas 150000 - X pub~~~4 transport --*-communications 50000 0 "p "p90z~99959>99 9 99 Comparing data on the NBI "1across different cities, we see that some cities are much more homogeneous in terms of poverty levels and urban services-some of these show that there is uniformly poor service, and very high levels of the total population not served by sanitation, for example. This is the case of Greater Buenos Aires (excluding the City) and the large cities of Cordoba and Rosario, plus cities in the poorest Northeast and Northwest of the country (e.g. Posadas, Jujuy, and Resistencia). Other cities, with relatively good levels of basic services (as measured by access to sanitation) show very high variations between unsatisfied basic needs, indicating that the poor typically do not have such services. On investment itself, data assembled for this report (for the City of Buenos Aires) show that, within a single jurisdiction, there is wide variation between school districts in terms of investments made in streets, sidewalks, drainage and other basic local infrastructure. Richer areas tend to receive more investments than poorer areas. As a result of Argentina's privatization program, public services for electricity, gas, water and sanitation, telephones and major trunk and access roads are in the hands of private concessions. Urban bus service, at least in the case of the City of Buenos Aires, is also privately run. Thus, an interesting question is whether or not privatization of public services has hurt the poor's access to these public services. Available evidence seems to I The NBI is a measure of poverty based on a household being deficient in any of five indicators; including crowding, access to water, sanitation, school attendance. For exact definition, see Part II, Chapter 3. 15 show that this is not the case; price adjustments have generally kept pace with inflation. Poor consumers are spending more on these services (see box) but whether they are spending more because service is better and they now have access or because collection is more efficient is not clear. It is clear, however, that the poor are able and willing to pay for urban services and that the total bill being paid for these services is affordable. The degree to which the poor benefit from privatization overall depends greatly on the type of regulation. A recent simulation study finds that privatization lowered the costs of utilities by 41% on average. However, with ineffective regulation, these gains would go largely to the rich. With effective regulation, the poor would receive more of the benefits12, on the assumption that regulation would provide subsidies for low-level consumption. Regional Poverty Argentina is a vast country with many differences among the provinces in terms of endowments, production, institutional capacity, income levels and social indicators. Thus, it is important to look at differences in poverty by region and not just national averages. As shown in Table 12, there are considerable differences. Poverty rates are substantially higher in the North West (46% in 1998) and North East (49%), followed by the mountainous, arid regions of the Cuyo (36%). Conversely, the City and Province of Buenos Aires and the five provinces which make up the resource rich region of Patagonia have poverty rates lower than average (24% and 22% respectively). In fact, the City of Buenos Aires shares a standard of living comparable to many OECD countries. In the middle, we find the region of Pampeana, mainly, the provinces in the agricultural plain, whose poverty rate is close to that of the country as a whole. In terms of income distribution, the regions outside of Buenos Aires tend to be slightly more equal, with Cuyo, Pampaena and the Northwest having somewhat more equal distributions. Argentina's economic transformation in the 1990s seems to have had differing impact on poverty depending on the region. The 1990s can be divided into two sub periods: 1990-94, when poverty decreased as the economy reactivated, and 1994-98, when poverty increased as a result of Tequila crisis and the increasingly competitive environment. In the first period, overall poverty rates fell by 48%, but poverty in the two poorest regions fell by about half of that. In other words, the poorest areas seem to benefit less from the general economic growth of the initial period than the richer provinces. However, when poverty rose during the 1994-98 period, the increases in poverty were greater in areas outside of the Northeast and Northwest, particularly in Buenos Aires, Cuyo and Pampeana. It seems probable that the areas outside of Buenos Aires are more dependent on agriculture, and less vulnerable to external shocks and financial crises. It may also be the case that these regions have been slower to react to 12 Omar Chisari, Antonio Estache and Carlos Romero, "Winners and Losers from the Privatization and Regulation of Utilities: Lessons from a General Equilibrium Model of Argentina" 7he World Bank Economic Review, 13(May 1999),357-378. 16 the new economic incentives which favor Argentina's comparative advantage in natural resources.'3 Table 12: Povery Rates by Region, 1990-98 (% poor of urban population) Greater All Year Buenos Aires North West North East Cuyo Pampeana Patagonia Areas 1990 41.2 54.4 55.7 48.1 33.7 26.7 41.5 1992 18.7 43.1 44.6 30.4 22.6 18.3 24.2 1994 17.0 41.6 40.3 26.1 19.8 17.1 21.6 1996 25.5 48.3 47.5 36.6 28.0 20.9 30.1 1998 24.9 46.0 48.8 36.0 27.4 22.4 29.4 % change: 1990-94 -58.7% -23.6% -27.6% 45.8% -41.2% -35.8% -48.1% 1994-98 46.5% 10.7% 21.0% 38.2% 38.4% 31.0% 36.2% Gini (1996) .484 .455 .477 .452 .434 .462 .483 Lowest 4.1 4.5 4.1 4.6 4.6 4.3 4.0 Poor, 1998 2.9 1.3 1.0 .7 2.6 .3 8.6 (millin)I Source: INDEC/EPH, various years. See Background Paper No. I for details. The Forgotten Poor of the Rural Areas. Behind the variation by region, there are also significant differences between rural and urban populations which have a bearing on poverty levels and local conditions. As mentioned above, most government income, employment and consumption surveys ignore the rural areas. Recent Government surveys for two provinces --Salta in the Northwest and Misiones in the Northeast--give us some insight into the state of the rural poor. The weighted average rural poverty rate for these two provinces is about 75%, with a rate of indigency of 35%14. Since Salta and Misiones are both from the poorest areas of Argentina, these results are not likely representative of the country as a whole. However if they were typical, their rural poverty rate would have the effect of combining with the urban poverty rate of 29% in 1998 to produce a national poverty rate of 34%. Regardless, these statistics are indicative of the very critical situation faced by the rural poor, including the indigenous who live predominantly in rural areas.'5 Even with this little information on the rural poor, we can see that the characteristics of the rural poor are both similar and different than the urban poor. One of the surprising results is that the rural poor are less likely to be farmers when compared to other non-poor rural population; they rather are more likely to be wage workers, self- employed or unemployed. Unemployment is very high in rural areas, about 19% overall 13 One needs to be cautious about interpreting these trends, however, since we have only 8 years of data for urban areas only, and are excluding the rural population. 14 The poverty line used here is based on a line that reflects rural prices and consumption patterns. See Part II, Chapter VII. 15 There are no national statistics which identify indigenous peoples in Argentina; thus little systematic analysis can be made of their particular situation. However, the conjecture is that indigenous peoples would face a situation comparable or worse than that of the rural population as a whole. 17 and about 31% if one includes the "discouraged workers" who have given up job search. The poor are more likely to be illiterate, to have dropped out of school earlier, and to show lower gains from the schooling they have received. In general, the rural poor-not being mainly farmers-- are net consumers of food, so that lower food prices raise their welfare, even though it lowers the welfare of food producers. The poor who are farmers have less land per capita, but have much higher productivity per hectare through the use of intensive production. They generally lack clear titles to their land, have little or no savings, and are less likely to have borrowed money. Higher farm productivity seems to be related mostly to the level of farm inputs, not capital investment. Since the poor lack access to credit, they tend to use fewer inputs, and have lower productivity per unit of labor. One of the very critical features of the rural poor is average household size: the non-poor farm households have some 3.7 members on average while the poor farm households have an average of 5.5 and the indigent, 7.4 members per household. These figures are slightly lower for the non-farm rural poor but show the same tendency. Household size for the rural poor are significantly greater than the average size of households among the urban poor at all income levels. A factor linked to large family size is migration to urban centers from the rural areas. Migration is mainly for work, and typically is to the nearest urban center in the same province as the gains from migrating to Greater Buenos Aires are marginal. Addressing the basic needs of the rural poor is generally more difficult due to the low population densities in most of rural Argentina and the existence of isolated, disperse rural populations. Overcoming the generally higher transactions costs of identifying and providing services to the rural poor requires even greater reliance on civil society structures. For example, the limited surveys indicate that membership in cooperatives or farmer's organizations is associated with greater likelihood of receiving technical assistance. Of particular concern is the apparent high rate of poverty among rural indigenous groups, which are often located in remote areas that are not reached by public services. Growth, Income Distribution, and Poverty As demonstrated in previous sections, Argentina's success in reducing poverty, despite admirable rates of economic growth as compared to the past, has been made more difficult by the rising gap in wages between the skilled and unskilled and the declining employment opportunities for the latter. The clearest demonstration of this is the worsening of income distribution, especially in the past five years. And the worst combination of these factors is faced by the young with low levels of education and little work experience who also have significant family responsibilities and children to care for. Future efforts in poverty reduction will thus be clearly related to overall growth in income, as well as changes in income distribution. For the latter, investment in human capital is paramount. 18 Table 13: Estimated Impact of Changes in Per Capita Income and Distribution on Poverty Percent Change in Povert Headcount: %changein Gini total change in per capita income % +10% I +20%0 | +30% 0 -13.8 -27.7 -41.5 +5% -10.0 -23.8 -37.7 +10% -6.2 -20.0 -33.8 Estimated Poverty Headcount: change in Gini total change m per capita income c +10% +20% +30% 0 25.3 21.3 17.2 +5% 26.5 22.4 18.3 +10% 27.6 23.5 19.5 Using pooled cross section and time series data from the EPH for 1990-98, we can estimate the elasticities of poverty rates with respect to changes in per capita income and the Gini coefficient. 16 The estimated elasticity for poverty with respect to growth is 1.38; a 10% increase in per capita income will reduce the poverty rate by 13.8%, provided there is no change in the Gini. Table 13 shows alternate combinations of changes in per capita income and the Gini. For instance, a 20% growth in per capita income would imply a growth rate over 10 years of about 1.8% per annum. This would reduce the poverty rate by 28%, if there is no change in the Gini, but by only 20% if the Gini were to increase by 10%. If we start with a poverty rate of 29%, then the forecast poverty rate would vary between 21% and 24%. This simulation somewhat mechanical, but illustrates both the importance of growth in income in any plan to reduce poverty, as well as the importance of how gains in national income are distributed among the population. However, it also assumes that the elasticities of the Gini and poverty with respect to growth are constant over the cycle. Recent work by DeJanvry and Sadoulet17 suggest that these elasticities can have important asymmetries: they tend to be higher in a downturn than in a recovery. Poverty is likely to be reduced faster if, for the same average growth rate, there is a greater degree of stability in the rate of growth. 16 The estimated equation regresses the change in the log of the poverty rate (dlnPR) against the change log of per capita income (dlnPCY) and change in the log of the Gini (dlnGINI), with dummy variables used to capture the fixed effects of different regions. The estimated equation (without the dummy variables) is: dlnPR = -0.004 -1.383dlnPCY + 0.768dlnGINI R2=.64 (-0.454) (-9.635)* (3.670)* n=102 * significant 1% level. The equation is estimated using semi-annual data for the period 1990-98 for six urban areas. 17 Alain de Janvry and Elisabeth Sadoulet, "How Effective Has Aggregate Income Growth Been in Reducing Poverty and Inequality in Latin America?" processed, Nov 9, 1999. 19 II. Employment, Unemployment and the Informal Sector Poverty and Jobs Employment is a key factor, but not the only factor, in determining poverty. Estimates show that a person is 17% more likely to be poor if the head of the household is unemployed. Poverty trends correlate with unemployment and wage trends, as shown in Figure 3. Figure 3: Poverty, Unemployment, Income 45 14.0 _ 40 12.0 E 8 E - -s..~~~~~ 10.0 -0 30- -6:- 25 8.0 0 20 6.0 -W 4.0 8 10- CU 4--.- ~~~~~~~~~2.0 (0 CO 5 CO 0 0.0 90 91 92 93 94 95 96 97 98 |_Unem ployment Poverty - Household Income ReaI Wa e The labor market in Argentina is divided roughly into two parts; about 55% of employment is "formal" in the sense that workers are covered by social security and labor legislation protection, and usually have union membership and health care coverage. The remaining 45% can be divided into two groups: the self-employed and the informal salaried workers. The range of conditions for these two groups is wide: the former may consist of well off professionals who operate outside the formal corporate system in order to avoid taxes and social charges or they may be relatively unskilled who work on a contract basis or on their own. The latter may be working for formally constituted companies (the so-called workers en negro, contracted informally in order to avoid paying taxes and social charges) or for informal enterprises which are typically very small, "mom and pop" operations. While 45% of the poor work in the informal sector, only 36% of the non-poor work there. In fact, the majority (about 65%) of.the workers in the informal sector are not poor. Thus, being in the informal sector does not automatically mean poverty, nor does being formal automatically mean one is out of poverty. 18 See S. Cesilini and E. Zuleta, Background Paper No.2 20 As shown in Fig 4, the size of the formal sector has fallen since 1988, when it was over 60% of the economy, partly as a result of the downsizing of the public sector and shedding of excess employment by privatized public utilities. Real wages of the informal salaried workers are lower than those in the formal sector, but since they generally escape the payment of social security and other wage taxes, their cash income may be the same or higher. Informal self employed workers have average wages even higher than those in the formal sector. What is clear (Table 14) is that the unemployment rates are substantially higher among the poor (23%), and even higher for the indigent poor (37%) as compared to the non-poor ( 9%). Unemployment spells, however, are similar between poor and non-poor. Most unemployment does not last more than six months; 44% last less than two months. Thus, it appears that the unemployed have been able to find work, albeit that work may be of a temporary nature, in construction for example, and not sustained in time. At the same time, participation rates are lower among the poor, and particularly among poor women. Given that the poor have larger families, and are generally younger, poor women seem more likely to be unable to participate in the labor market given their child raising responsibilities. In addition, given their lower levels of education and skills, the opportunity cost of staying at home is also lower. Figure 4: Sectoral Participation and Median Real Wages in Argentina, 1988-1997 (includes small firm owners in sel-employed) 70.0% 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 14 13 60.0% 12 50.0% 0 _ 40.0% O 30.0% 20.0 4 2 0.0% ~~~~~~~~~~~~~~~~~~~~~~0 1988 1991 1993 -1995 1997 1% formal % informal ~% self-employed | Dformal wages A informalwages maself employed wages Furthermore, the total hours worked by the employed poor is lower, reflecting an inability to obtain full-time work or, to a lesser extent, a preference for part-time work. Defining underemployment as those working less than 35 hours per week and seeking employment, there is a clear tendency for underemployment to rise over time (see Fig 5). In Buenos Aires, underemployment has risen from 5% to 14% of total employment since 1980. About two-thirds of part-time workers report that they are seeking more work. The preoccupation expressed by the poor about finding work is therefore collaborated by these statistics which show frequent spells of unemployment, combined with increasing underemployment for those who have found work. 21 Table 14: Labor Market Characteristics of Urban Population by Poverty Group, 1998 Poverty Group Indigent Poor Moderate Non-Poor Total .________________________________ Poor Labor Force Participation (Age 15-64) 54.8 56.4 67.3 64.5 Female Participation 37.4 35.5 53.4 49.0 Unemployment Rate 36.5 22.6 8.9 12.6 Among the Unemployed: Unemployment Spell (%) 100.0 100.0 100.0 100.0 Under 2 months 45.5 46.8 42.8 44.3 2 - 6 months 33.6 29.4 26.3 28.2 7 months- I year 13.3 18.8 23.0 20.5 More than 1 year 7.6 5.0 7.9 7.0 Among the Employed Number of Jobs Held 1.03 1.04 1.10 1.09 Total Hours Worked per Week 34.9 41.7 45.1 44.2 Female Hours Worked 26.8 31.9 37.0 36.1 % Employed in the Informal Sector 46.4 44.6 36.2 38.2 %ofPart-time Workers 43.9 26.8 17.0 19.4 Sources: INDEC, Encuesta Permanente de Hogares, average of May and October of 1998, except for unemployment spell, which is October 1997. Figure 5: nderemployment as a Share of Total Employment, 1974 - 1997 14 12 10 8 4 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 Source: Permanent Household Survey. Gran Buenos Aires, October 1997. Ministerio de Economria y Obras Pzublicas Underdemployment defined here as workers working 30 hours or less per week. Panel data which traces workers over two year periods during the overall 1993-97 period shows that there is considerable movement between job categories over time. Workers do move between the informal and formal sectors, as well as from the formal to the informal. Over two years, only 48% of male informal salaried workers remained in that sector; 12% moved to the formal sector, 18% became self-employed and 14% were found to be unemployed. For formal workers, by contrast, 80% were found to remain in the formal sector after two years, 8 percent became unemployed. Of unemployed men, 32% moved to the informal sector, either as salaried workers or as self employed, and 16% moved to the formal sector. For women, the results are similar but with some important differences. For women who were unemployed at the beginning of the period, 22 only 28% found work, 22% remained unemployed, but 50% left the labor force entirely (were not seeking work).19 The stability of workers in the formal sector is probably current labor laws which provide considerable job security for formal sector workers, at the detriment of greater job mobility. There is also considerable stability among informal self-employed, but considerable movement among informal salaried workers, who appear to be most at risk. In fact, it is not surprising that a major finding of this study is that the bulk of unemployment originates from the informal sector since those workers lack the protection of formal workers and there is no cost to employers associated with lay-offs. On the other hand, the fact that it is the informal workers who suffer the most from unemployment is disturbing since they are precisely those who are not protected by severance payments or unemployment insurance. As shown in Table 15, of the paid workers moving to unemployment, 38% come from the informal salaried sector, even though this sector constitutes only about 19% of total workers.20 Table 15: Sector of Origin of the Unemployed Sector of Origin All Paid Informal Self-Employed 18 38 Salaried Informal 12 26 Formal Salaried 17 36 Previously Unemployed 34 -- Previously Out of Labor Force 6 -- School Graduates 13 -- Total 100% 100% Unemployment hits the young hardest. The unemployment rates for workers in their teens and twenties is twice as high as for those in their 40s and 50s. While this is an important problem, one needs to recognize that the majority of the unemployed are young, and may not be household heads. Thus, the impact of unemployment on poverty per se will depend more on whether or not there are other older wage-earners in the household. (Admittedly, the large number of youths who are not pursuing studies or working may constitute another set of social problems associated with the use of drugs and violence.) High rates of unemployment, particularly among male workers, seem to have led to a rise in participation rates, particularly among women Since 1988, labor force participation rates overall have risen by about three percentage points. Across all economic groups, women have increased their participation in the labor force over time (see Table 16). Today, about 49% of adult women in Argentina work compared to 37% in 1980. Guasch estimates that this alone could account for 25% of the increase in 19 For details of the transition matrix, see Part. II, Chapter 2. 20 It also appears that some employers have both formal and informal workers, with the informal or workers en negro, providing flexibility in times of labor force reductions, without the need to pay severance payments. 23 unemployment.21 Since women tend to have lower education levels than men, the growing participation of women would mean an increase in the supply of unskilled workers, and may account for their stagnant real wages over the last ten years. Table 16: Female Labor Force Participation Rates (urban areas) Indigent Poor Non-Poor Total 1985 21.1 20.4 39.9 37.4 1990 31.8 32.0 49.5 42.9 1995 36.4 37.4 52.3 48.8 1998 37.4 35.5 53.4 49.0 Source: EPH, various years. 1985 data is for Buenos Aires only. Among the poor and indigent, however, the increase has been even more pronounced. Poor women have increased their participation from 20% to 36%, while indigent women have increased from 21% to 37%, although still well below the rate of participation for non-poor women. Female participation also appears to be procyclical, going up in crisis years (1988, 1990, 1995), particularly for the poor. Increased female participation is clearly one coping mechanism the poor can use to offset the adverse shocks of male unemployment and/or declining real wages. Another factor explaining high unemployment among the poor relates to the sectoral composition of their employment. The poor, and particularly the indigent poor, are heavily located in the construction industry which is more volatile than other industries, and therefore more subject to short term unemployment and uncertainty. About 18% of the poor and 23% of the indigent work in construction, compared to 7% of the non-poor (see Table 17). The variance of employment growth in construction is over four times higher than the average variance for overall employment. Table 17: Employment Distribution by Sector, 1998 (urban, %_of total) percent distribution of mployment standard standard sector Indigent Poor Non-Poor deviation of deviation of Agriculture 1.9 1.4 1.8 5.1 6.6 Manufacturing 9.6 15.6 15.7 5.2 6.2 Construction 23.4 18.2 6.8 15.2 13.0 Trade 17.0 18.4 19.1 __ 6.1 __6.8__ Finance 3.8 3.7 104 6.5 _ 5_.8_ Services 44.3 42.7 46.2 4.9 _ 3.9 X Total lOo 100 loO 4.8 2.7 Source: EPH. Services includes "unknown". Standard deviations are for period 1990-98, and are deviations from mean growth rates for employment and value added by sector. But how important is reducing unemployment toward the elimination of poverty? Since the poor are often unskilled and undereducated, and have large families to support, even full employment would leave a substantial number of people below the poverty line. We can estimate this by looking at the relationship between income and labor force 21 J.L Guasch, "Argentina Labor Market in the New Millennium", World Bank, August, 1999. 24 characteristics of the employed workers, and then assuming that unemployed workers with similar characteristics would have the same earnings were they to be employed. The results of this simulation (Table 18) show that, in May 1998, full employment (0% unemployment) would have only reduced the overall poverty rate from 29% to 23%, and the rate of extreme poverty from 7% to 4% (about a 40% decline in extreme poverty). Table 18: Impact of Full Employment on Poverty Headcount, 1992-1998 l_ From: From: Poverty Headcount Being An Indigent to: Being a Moderately Poor to: Year Actual Staying Becoming Escaping Actual Staying Escaping Actual Full Indigent Indigent Moderately poverty Poverty Moderately poverty Employment Rate poor Rate Poor 1992 4.4 3.3 0.7 0.4 19.7 16.9 2.9 24.1 20.9 1993 4.3 2.9 0.8 0.6 17.5 14.1 3.4 21.8 17.8 1994 3.7 2.3 0.8 0.6 17.9 13.8 4.1 21.6 16.9 1995 6.1 3.1 1.9 1.1 21.1 14.7 6.4 27.2 19.6 1996 7.3 3.9 2.1 1.3 22.7 16.2 6.4 30.0 22.2 1997 6.8 3.8 2.1 1.0 22.6 17.0 5.8 29.4 22.8 1998 7.1 4.3 1.9 0.9 22.3 17.2 5.1 29.4 23.3 Source: Calculated from the EPH surveys, average of May and October. For details see Background Paper No.1 These estimates are rough, but indicate that employment can affect poverty levels. Even if unemployment were to be reduced by 50%, the poverty level Table 19: Real Wage Growth (%) would fall from 29% to 26%, and (Male Workers, Greater Buenos Aires) extreme poverty would fall by about 1988-93 1993-97 20%, both significant changes. Formal Sector 24.2 .6 However, it is also important to note Informal -Salaried 4. .6 InformalSalaried64.4 -12.6 that even with full employment, 0 poverty would remain. The low wages Sorce: E Sel s paid to unskilled workers (see Table 4) suggest that even when employed such workers would remain well below the poverty line.22 These estimates also assume that increases in employment occur through increases in demand. If more employment is achieved by increased labor market flexibility, which lowers real wages, then poverty effect would be considerably reduced. Not only would some of the existing poor not escape poverty, but some of the non-poor might fall below the poverty line. The impact of labor reform, could in the short run, increase poverty. However, for labor market flexibility to reduce poverty, one must assume that these reforms increase incentives to invest over time, and thus increase the demand for labor. Wage Growth Poverty is as much, if not more, a function of wages as of unemployment. As noted above, real wage growth has been highest for skilled and highly educated workers while wages for unskilled workers has fallen in real terms since 1990 (Table 4). Technological 22 For instance, an unskilled worker receiving 356 pesos per month, and having a family of five whose spouse did not work, would have an average per capita income in terms of adult equivalents of 101 pesos, less than the poverty line of 160 pesos. 25 progress, easier access to capital, and trade openness have put unskilled workers as a disadvantage, particularly since 1993. There are important wage differences between the informal sector and the formal sector. In fact, informal salaried workers gained more than formal salaried workers in the period 1988-93, but lost more during the period 1993-97 reflecting the flexibility in that segment of the labor market (see Table 19). For instance, between 1993 and 1997, real informal sector salaries increased 13% while formal sector salaries remained unchanged; in the prior period informal sector salaries rose almost three times faster. Informal workers are, once again, at more risk of sudden declines in income. This is another sign that existing labor laws provide stability in income levels to formal sector worker, both in good times and bad. Within all sectors younger workers have done better in terms of wage growth than older workers, particularly salaried workers (however, they probably are starting at a lower level of wages). In some cases, younger workers with less education may have had faster wage growth than those with more education. But those who lost the most seem to be long-tenured informal salaried workers with primary or secondary education along with those with secondary education who were just starting out in the labor force. In general, the data of Table 20 suggest that older workers may possess obsolete skills and may not be competitive with younger, better educated workers. Table 20: Experience, Education and Wage Growth, Buenos Aires, 1988-97 Formal Informal, Informal, Self Salaried Employed Primary Ed 1-5 years exp. 32.8 51.1 22.4 > 5 yrs. exp. 22.2 .8 28.3 Secondary Ed. 1-5 years exp. 20.2 35.7 54.9 > 5 yrs. exp. 6.3 -6.0 14.3 Source: EPH Surveys. Labor Market Reform and Poverty There is good reason to believe that current labor legislation is biased against the absorption of labor by the formal sector and instills rigidities in the market's ability to adjust. More importantly, current labor legislation appears to create a great incentive for employers to take on workers--in negro -- without all of the privileges of formal sector employment, forcing workers to accept employment without any form of social protection. Conversely, it may also promote self-employment in the informal sector by those with high skills in order to avoid mandatory social charges, including contributions to health insurance and pensions. For both groups-whether they are relatively high- income self-employed or low paid informal workers-there are high social risks attached to not being part of the formal social security system. At the same, the fact that such as 26 large share of the population falls outside of the social security system creates pressures on public spending to compensate.23 Centralized or sectoral collective bargaining agreements limit firms' abilities to adapt to competitive conditions. The ergo omnes clause extends these agreement to all workers and firms in the sectors, even if not signed. Ninety percent of present contracts are "ultra-active" remaining in force even after expiration and lock in work arrangements that may no longer be appropriate to market conditions. High non-wage benefits drive labor costs up relative to the costs of capital. In addition to a 17% wage tax paid by employees, employers provided legally mandated benefits and fiscal contributions that raise labor costs to them by 48%.24 Restrictions on firing discourage firms from hiring additional workers. The effects of this rigidity are clearly seen in the relative stagnation of wages and employment in the formal, large business sector which by and large respects the existing labor agreements. Recent experience with temporary reforms in the 1990s supports the view that these contracts are binding. Partial reform lowered payroll taxes somewhat, improved the pension system, and provided for temporary contracts that were exempt from payroll tax and severance payments. Employment under these modalidades promovidas reached nearly 5% of total employed workers by 1998, and accounted for 40% of the new jobs created between 1996-98. However, the modalidades promovidas were abolished in 1998, which may further add to unemployment in 1999. The labor law changes in 1998 also shortened probation periods to one month (from six months), and further centralized collective bargaining arrangements, moves which reduced overall flexibility. However, recent draft legislation has been sent to the Argentine Congress that would reverse these reversals; probation periods would be reextended, ultra-activity would be gradually phased out, and power of centralized collective bargaining would be reduced. In addition, labor taxes on new employees would be reduced. Overall, the existing labor legislation is ill-suited to a new competitive environment, makes adjustments to adverse shocks difficult and discourages firms from taking on additional workers. Programs such as PYMES (proposed in 1994) which gives firms of less than fifty employees a special regime for collective bargaining and facilitating labor contracting are important for giving emerging firms the flexibility to adapt and grow. Another possible reform would be to move from the current system of severance payments to an unemployment insurance system based on individualize accounts. These could be portable, capitalized and fully funded, and would both reduce labor costs and eliminate the perverse incentives in most unemployment schemes that discourage employment.25 23 This is part of a wider problem facing Argentina which needs to be addressed through labor and tax policies as well as improved enforcement. 24 L. Guasch. Argentina Labor Market in the New Millennium, World Bank, Report No. 19996 AR, January, 2000, p.97. 25 For a fuller discussion of how such a system would work, see J. L. Guasch, Argentina Labor Markets in the New Millenium, August, 1999 (World Bank). 27 III. Social Spending: Does it Reduce Poverty? Who pays, who gets Argentina has one of the highest levels of per person social spending in Latin America26. Argentina spends more per person than countries such as Chile, Costa Rica and Uruguay. Yet, despite substantial spending poverty persists and social indicators indicate a mixed performance. Social spending has risen in importance over time, from about 15% of GDP in the early 1980s, to about 18% today. Overall, social spending has three main components: social insurance (57% of total), social sector programs that are universally available to the public (36%), and programs specifically targeted to the poor (7%). Thus the amount of explicit targeting to the poor is relatively small, although the poor benefit in varying degrees from all social programs. The net impact of social programs and the tax system which pays for them is to improve the welfare of the poor. As shown in Table 21, the lowest 20% in the income distribution receive 22% of total social sector spending, and pay 7% of all taxes27. People in the upper quintile pay 47% of taxes and receive only 19% of social sector spending. While social sector spending, both universal and targeted, definitely benefits the poor, social insurance spending is regressive, giving more benefit to the upper quintiles. Likewise, the tax system appears regressive, since the poor carry a large tax burden than the rich: the 7% of the taxes paid by the poor constitutes a larger share of their income-only 4 percent of the total-- than the 47% paid by the rich who have more than 53 percent of total income. This situation may be explained by the fact that little tax revenue is generated by income taxes, and a significant share is generated by consumption taxes (VAT and special taxes on gasoline, etc) which fall heavily on the poor28. The net effect of the system taxes and expenditures seems to benefit the poor. Recent estimates by Llach and Montoya29 suggest that the total net benefit to the lowest quintile is equal to an increase of 55% in their total income. However, these estimates depend heavily on the values given to such free public services such as health and education, and estimates of their consumption by the poor. 26 Government of Argentina, Direcci6n Nacional de programaci6n del Gasto Social - Secretaria de Programaci6n Econ6mica y Regional Caracterizaci6ny.Evoluci6n del Gasto Ptublico Social en el Perfodo 1980-199 7, Buenos Aires, 1999 27 These estimates are from L. Gasparini, "Incidencia Distributiva del Gasto Publico" FIEL, April 1999. 28 In addition, there is estimated to be a substantial amount of tax evasion and underreporting of income by the upper income groups, because of the underreporting of non-wage income (see box, Part II, Chapter 1 "Adjustments for Underreporting") 29 Juan Jose Llach and Silvia Montoya, En Pos de la Equidad, IERAL, June, 1999. Their study also suggests that the tax system is roughly proportional. See Part. II, Chapter 3 for details. 28 Table 21 :Total Social Expenditures and Taxes by Quintiles, Urban Argent na 1996 Quintile: I II III IV V Total Expenditures: (percent of total) Social Sectors 29.8 18.8 21.7 16.8 13.0 100 Social hisurance 9.9 20.6 19.5 23.6 26.5 100 Total Social Expenditures 21.8 19.5 20.8 19.5 18.4 100 Tax Distribution 7.1 10.7 14.9 20.1 47.2 100 hicome Shares 4.0 8.4 13.2 21.2 53.2 100 Source: L. Gasparini, FIEL, 1999 and 1998, income shares EPH May and October. Who Takes the Cuts? Targeted programs generally should expand during an economic crisis, as Table 22: Elasticities of Government Spending with emplomen and pe Respect to Changes in GDP, unemployment and poverty By Type of Spending, 1980-97 increase. At the same time, we Spending Type Elasticity know that declines in Total Govemment .95govement revenues often Total Social Sector 1.28 produce reductions in budgets Social Security .84 for social and other spending. Universal 1.87 What has been the case in Targeted 1.86 Agnia Source: Staff calculations based on Ministry of Economy Argentina? data (DNPGS, 1998). For the period 1980-97, calculated elasticities (Table 22) reveal that all kinds of government spending are procyclical; they rise and fall with GDP changes. But the situation is even worse for the social sectors; while the elasticity of total spending with respect to GDP is .95, for the social sectors is 1.3. Thus, social spending seems to be more sensitive to the business cycle than overall spending. One would hope that targeted programs would be part of a social safety net that could expand during a crisis. Instead of being counter cyclical, however, they seem to fluctuate by a factor greater than even total social sector spending. Thus, a five percent drop in GDP seems to be associated with a 9 percent drop in targeted social sector spending. The effect is even greater if one considers that a drop in GDP is normally associated with an increase in the number of people below the poverty line. Thus, while poverty is going up, targeted programs for the poor tend to decline. Do the Poor Benefit? Incidence vs. Coverage of Social Programs Argentina has numerous universal and targeted social assistance programs that seek to improve the quality of life of the general population as well as the poor, vulnerable, and disadvantaged population. They include food, nutrition, health, employment, training, education, shelter, clothing, income subsidies, and other emergency programs. Outside of the universal program of education and health, and social security, many programs are targeted directly to poor families and most of them have been put in place or expanded since 1995, as a result of the Tequila crisis. Others are aimed at 29 strengthening the ability of communities to address various social and economic issues. Unfortunately, there are only a few evaluation studies that examine program impact and cost-effectiveness. The 1997 Social Development Survey (EDS), financed with World Bank assistance, provides information on cash assistance, food and other direct benefit programs received by households (social welfare assistance). It indicates that social welfare assistance appears to be strongly pro-poor on the whole; 65% of benefits go to the poor; 47 percent goes to the poorest 20% (see Table 23). In contrast, the highest quintile gets only 2.4 percent of the benefits. Thus, while there is some scope for reducing "leakage" of benefits to the non-poor, overall targeting is respectable. Table 23: Benefit Incidence of Government Social Assistance Income Distribution of % households receiving assistance quintiles public benefits (%)a Public Public and private sources First 47.2 29.5 45.0 Second 29.0 18.8 31.4 Third 15.0 9.8 22.0 Fourth 6.4 4.9 16.1 Fifth 2.4 1.4 9.8 total 100.0 12.8 24.8 Poor 64.6 25.5 39.8 Non-poor 35.4 7.2 18.1 Total 100.0 12.8 24.8 Source: SIEMPRO, Social Development Survey (EDS), 1997, Includes food, cash and other forms of in kin Note: (a) Weighted by provincial government spending on social welfare assistance. Includes only urban population Within certain groups, public transfers, when combined with private transfers, can be very important in reducing poverty. Without public and private transfers, the poverty rate would be about 4 percentage points higher than it is. However, for household heads over the age of 65, the absence of transfers, mainly social security, would increase their poverty rate by 21%, and for female headed households by 33%. Clearly, the present system of public pensions provides a very considerable safety net for the aged albeit one which provides benefits to all the retired regardless of income level.30 More serious problems emerge in terms of coverage. About 75 percent of the poor do not receive any public assistance and over half got neither public nor private help. This would indicate that a substantial percentage of the poor are outside Argentina's public and private social safety nets. Appendix A provides an overview of the identified vulnerable populations and the extent to which each population is uncovered. 30 The present system of public pensions is the remainder of the pay-as-you-go system and thus covers the needs of all those retired from the formal sector who had made contributions previously as well as some provision for those who did not make contributions in the past. Thus, it is not a targeted social program per se and pays out according to the previous level of salary (on which past contributions were supposedly made). The transition to a fully funded private system will continue for some 40 + years until all those over age 45 years are no longer living 30 Box: TRABAJAR Argentina's TRABAJAR is an excellent example of a well targeted social safety net program. The program provides temporary employment at below-market wages to poor and unemployed workers. TRABAJAR projects include minor social and economic infrastructure projects, such as the repair or remodeling of schools, based on projects proposed by local governments. The national government pays for the unskilled labor component, about one-third of the cost of the proposed projects; the rest is financed by local government. Data from TRABAJAR II indicate a coverage rate of about of about 1-2% of the total labor force, but 20 to 30 percent of the unemployed poor. By paying below market wages, the program attracts only those poor and with few prospects of employment. Experience thus far shows that self-targeting is effective: evaluation of the program indicates that on average 40 percent of its resources have reached those in the bottom 5 percent of the income distribution, while 75 percent have benefited the bottom 20 percent. Moreover, targeting of poor communities has also been effective. Nevertheless, TRABAJAR's experience also shows that rationing of resources among competing municipalities can lead to possible political interference. Finally, experience shows that temporary public sector employment is not a long-term substitute for sustainable job creation in the private sector and that this kind of program is not a substitute for sustained job creation in the private sector and only acts as a temporary palative for those who may be seeking full-time, steady employment Source: J. Jalan and M. Ravallion, "Income Gains from Workfare and their Distribution: Estimates from Argentina's Trabajar Progran", World Bank (processed), June 1998. Examination of specific programs provides details. For instance, food and nutrition programs for young children and senior citizens are one of the most important thematic areas in the inventory of social assistance programs. The most vulnerable group to long- term negative effects of malnutrition are the children aged 0-2 in the lowest quintile. Of these, only 44% receive benefits from public nutrition programs, even though they receive 70% of the benefits. Some 1 million children under the age of five are not being served by early childhood and nutrition programs. Again, targeting is good, but coverage is low. In some cases, where programs have emphasized targeting, results have been impressive. An evaluation of the TRABAJAR program which is financed, particularly, by the World Bank, estimated that 75% of the benefits of the program went to people in the bottom 20%, and 40% of the benefits reached the lowest 5%, an impressive targeting record by international experience (see box). For other programs, improvements in targeting are still possible, such as by using the existing SISFAM system of proxy means testing developed by SDS (see Part II, Chapter 3). One critical problem is the proliferation of safety net programs, even though the total amount of federal funding allocated to targeted programs is small. Total federal expenditures on targeted programs are about $2.2 billion ($60 per Argentine per year, 3.8% of national social sector expenditures, and 0.68% of GDP). This small budget finances a complex array of over 60 programs -- which comprises the current federal safety-net (see Appendix A). The result is a system with overlapping objectives and 31 target groups, higher administrative costs than needed (see below), but also dispersion of efforts, and thus low overall impact. The sheer number of programs suggests a surplus of administrative effort which could be rationalized, and hence made far more cost effective, by collapsing the number of programs into a far simpler structure. Because these programs are managed by a large number of agencies, they utilize different targeting mechanisms, selection criteria, size of transfers, etc., which lead to further inefficiencies in the use of resources. Moreover, there is not a umique system to identify beneficiaries across programs, which not only leads to inefficiency, but opens up to the possibility of people receiving multiple from several programs. One major reform worth considering is the merger of existing programs into a single cash grant. While this may not be appropriate for all programs, it could greatly reduce administrative costs, such as involved in food handling and storage, and provide a more even distribution of benefits among the poor. EDUCATION: THE KEY To LONG TERM POVERTY REDUCTION While social assistance programs can mitigate poverty in the short run, and protect the poor against major variations in their incomes, long-term poverty reduction will depend both on rapid economic growth in the economy combined with substantial investments in human capital. The most important type of human development investment remains education. Argentina's education system is one of the most advanced in the region. However, with its high per capita income, comparisons with other Latin American countries are not appropriate. In fact, for a country of its wealth, Argentina underperforms in certain key indicators, including secondary school enrollment. Nevertheless, it can boast a literacy rate of 97% for the adult population, universal primary enrollment, and 9.7 mean years of education of the adult population. Who Benefits from Public Education? Overall, the poor benefit from public education spending more than the non-poor, in part because higher income families tend to send their children to private schools. The poorest 20% of the households receive about 25% of total education benefits; the richest about 12%. However, the lowest quintile also potentially has about 25% of the students, and the upper quintile only 15%; thus, public spending on education is not progressive. Within the sector, the poor tend to benefit more from primary education; secondary education tends to benefit slightly more the middle class. The major distortion comes in tertiary or university education; the lowest quintile receives only 10% of the benefits, while the richest quintile receives 34% (see Table 24). Af the same time private enrollment is low at the university level. While private enrollments equaled 25% of total enrollment at the primary and secondary level in 1996, it was only 14% at the university level. Providing free university education to all groups basically produces a major subsidy to the richest and stands out as a major contradiction in terms of equitable social policies. 32 Table 24: Distribution of Public Education Expenditure Benefits by Level of Education and Income Status (Percent), 1997 Total Primary Secondary Tertiary Quintile: First 25.5 32.6 20.9 10.4 Second 23.4 25.1 24.0 12.1 Third 20.8 19.1 23.4 17.4 Fourth 18.0 15.1 19.3 26.0 Fifth 12.3 8.0 12.3 34.1 Total 100 100 100 100 Poor 40.4 47.2 37.9 17.6 Non-poor 59.6 52.8 62.1 82.4 Source: SIEMPRO, Social Development Survey (EDS), 1997. Problems of Quality Argentina faces major problems of education quality, reflected in high repetition and dropout rates, and low student achievement. The secondary graduation rate in Argentina is 52%, compared to an average rate for the OECD countries of 80%. Of 100 students entering primary school, 84 will enter the seventh grade, 76 will enter the ninth grade, 40 will enter the last year of secondary school, 35 will enroll in university and only seven will graduate. The results are clearly differentiated by income groups; dropout rates increase sharply for the poor after 7 years of mandatory education (see Fig. 6). Only 24% of students in the lowest quintile complete secondary school, compared to 76% for the upper quintile; 37% of the youth aged 14 to 18 are out of school. Fig. 5. Education Attainment Profile, Age 20-24,1997 120.0% - 0 100.0% 0; a - C 80.0% -Quintile 5 2: 60.0% ^' ° 60.0%- \: ~ * Mean , 40.0% X 20.0% 0.0% 0 1 2 3 4 5 6 7 8 9 10 1112 13 14 15 16 Highest Grade Attained In addition, children from low-income families are more likely to be older than other primary school children due to higher incidences of late entrance and repetition among poorer children. Repetition rates are about 25% for the lowest quintile (repeaters of at least one grade) compared to 4% for the richest. Thirty two percent of the children from the poorest income quintile are delayed in prirnary level compared to 9.3% for the richest quintile. Dropping out of schooL An analysis of the returns to education suggests that the rate of return to completing primary education is only 2.5%, while for secondary 33 education it is about 10% and for tertiary education about 29%31. Why do many children dropout of school early if the returns to education are high? The 1997 Social Development Survey focused on this issue. It found that about half of the dropouts considered "personal motives" a key factor, of which 77 percent say lack of interest is the main cause. A common interpretation is that school is boring. Yet, as mentioned, only a few of the respondents blame schools for their decision to drop out and only a small percentage find education useless. About 43% gave "economic difficulties" as a reason, of which seventy-three percent reported that they left school because they needed to work. As might be expected, economic difficulty is a much more salient problem for the low income families. (50% for the lowest quintile vs. 29% for the upper quintile.) An examination of actual school enrollment behavior of 15-19 year olds further reveals that parental education and income tend to have significant positive effects on schooling. Second, distance to public secondary school remains a significant cost that discourages some children from school participation. In this regard, children in the first quintile generally live further from school than those in the fifth quintile. Third, the unemployment rate among household heads is negatively related to enrollment. This suggests that in areas where the unemployment rate among household heads is high, children's human capital formation suffers. Further analysis reveals that the rate of return to education is probably lower for individuals with poor parents, as measured by the education of the head of the family. Students whose parents have only primary education have an incremental rate of return from a year of secondary education that is 60 percent lower that those whose parents have tertiary education. The reasons for this are multiple and may have to do with differences in the quality of education received by the poor as well as less conducive home environments. There may be less emphasis on educational attainment, more pressure to seek work instead of study and less opportunity to study because of crowded household conditions. It may be that the poor children, discouraged by what they perceive as fewer opportunities or outright discrimination,32 may cease to see the value in further education. The fact that the poor have lower rates of return from the same years of education might explain recent changes in enrollment rates at the secondary and tertiary levels. At the secondary level, enrollment rates between 1992 and 1997 have decreased, while enrollment rates for other deciles have increased (see Table 25). For higher education, the shift is even more pronounced. Enrollment rates for those in the lower 40% all declined, while those for the upper 60% increased. Hence, the richer groups are responding to higher returns to education, while the poorer groups are not, potentially perpetuating the gap between rich and poor. 31 These calculations come from the 1997-Social Development Survey (EDS), and are the marginal returns from additional schooling, The results shown in Table 4 give average returns for a given number of years of schooling. This data combines returns from both private and public education. 32 One of the conclusions of the Voices of the Poor exercise in Argentina was that the poor experience discrimination on the basis of where they live, with certain barrios connoting that the persons who live there are likely to be of a lower social standing. 34 Table 25: Enrollment Rates in Secondary and Higher Schools by Decile Secondary Higher Decile 1992 1997 1992 1997 1 70.5 62.1 22.7 8.7 2 74.9 68.9 20.5 10.7 3 80.3 83.7 17.7 13.3 4 80.0 85.4 22.3 16.1 5 82.5 103.6 20.8 22.5 6 76.1 95.9 21.9 28.0 7 92.2 105.3 27.4 37.8 8 96.4 102.4 38.0 53.8 9 95.2 117.1 54.4 75.5 10 106.2 115.9 80.8 82.0 Source: Maria Echart, " Educaci6n y Distribuci6n Del Ingreso", FIEL, 1999. Efforts to Improve the System The Federal Government uses two main instruments to reduce inequities in the education system: the Pacto Federal Educativo and the Plan Social Educativo. A principal objective of the Pacto Federal Educativo, which spent $ 398 million between 1995 and 1998, is to help the provinces meet the physical demands of extending mandatory education from 7 to 10 years. This change was introduced in the 1994 Education Law and is presently being implemented by the provinces, which the World Bank support. Because most children who are not presently enrolled in secondary education come from underprivileged households, this major effort to expand physical capacity is expected to benefit poor children and youth in particular. The Plan Social Educativo (PSE) was created in 1993 to directly address inequities in the education system by targeting resources to schools serving underprivileged children. The PSE serves 11,820 Inicial, primary and secondary schools and approximately 6 million children, or 75% of total enrollment at these levels. Since its creation, the PSE has disbursed a total of US$703 million. In 1997, the PSE represented approximately 20% of all federal education expenditures and approximately 2% of total expenditures in the sector.33 Most of this investment has been directed to Inicial and primary schools. The PSE comprises three kinds of programs; improvements in school infrastructure; quality improvements in schools including programs targeted at indigenous students, rural schools, adult and special education; and a national program of student scholarships to keep children from poor families enrolled in secondary school The adequacy and effectiveness of PSE have not yet been rigorously examined, although feedback has been positive. Indeed, it appears that PSE is strongly pro-poor. The 1997 EDS reveals that 60 percent of the PSE beneficiaries come from the first quintile and 40 percent from NBI families (Table 26). Moreover, the two lowest income quintiles account for over 80 percent of the beneficiaries. Once again, targeting is quite 33 See note 4 about the way these percentages were calculated. 35 good. In terms of coverage, it can be seen that only about half of the children in the poorest quintile have received assistance from PSE. Table 26: Coverage Rate and Distribution of PSE Beneficiaries by Income Class Total Per capita income quintile Poverty 1 2 3 4 5 NBI No NBI Coverage rate (a) 42.0 51.8 36.6 33.9 23.8 25.3 55.7 36.2 Distribution of beneficiaries 100 59.9 21.3 11.3 5.5 2.0 39.5 60.5 (a) percent of children enrolled in public school. Source: 1997 EDS. SIEMPRO Tabulation. Does it Pay to Invest in Education? While we cannot evaluate the PSE program directly, it is possible to make some indirect estimates of the returns on investment in secondary schools. Based on estimates derived from regional data, we estimate that a 10% increase in public secondary expenditure would produce a 3.5% increase in the lifetime wage of a 20-29 year old with a complete secondary education. Thus, an expenditure of 800 pesos per student would produce an annual return of about 200 pesos per year, or an internal rate of return of about 25%, clearly a productive investment and one that would reduce poverty in the longer term. Health Status and Poverty Overall health indicators of Argentina are good when compared with those of other countries in the Region, and have improved markedly over the past ten years. In 1995, life expectancy stood at 76 years, 12 more than in 1960, the crude death rate at 8 per 1,000 population, infant mortality at 18 per 1000 births in 1998, down over 30 percent from 1987, and maternal mortality ratio at 14 per 10,000 live births in 1995, down by 14 percent from 1990.34 These statistics, however, mask wide variation by region, and within provinces by particular locations. Moreover, they are worse than those of other middle income countries, and lower than one would expect for a country at Argentina's level of development For example neighboring Chile, with lower income than Argentina, had in 1995 an infant mortality rate of 12 per 1,000 live births and maternal mortality ratio of 6.5 per 10,000 live births. As elsewhere, the poor in Argentina carry a disproportionate part of the total burden of disease and are disproportionately affected by a number of diseases. Diseases which affect disproportionately the poor are mostly avoidable communicable diseases, such as tuberculosis, syphilis, diarrhea, tetanus, and Chagas disease, as well as maternal, perinatal and nutritional conditions. Of particular note for the poor are health problems of women: cervical cancer, hypertension associated with pregnancy, abortions, and maternal mortality. Most of these could be controlled with universal access to a comprehensive low-cost package of quality public health and personal health services. Theses 34 Bos, E, Hon, V., Maeda, A. et al: Health Nutrition and Population Indicators, a statistical handbook. (Washington, DC: The World Bank, 1998) 36 differences in health status are also consistent with environmental conditions, most notably access to safe water and sanitation, behaviors, and health care utilization which can be improved with concerted effort. Health insurance coverage. In general, the poor are not covered by health insurance. Thirty-seven percent of the general population and 62 percent of the poor are not covered by health insurance, their health care needs being met by a network of public hospitals, health centers and health posts. The other 63 percent of the population is covered by mandatory, provincial or union-run insurance schemes (Obras Sociales) for formal sector workers and retirees, or voluntary private insurance (Prepagas or Mutuales), their health care needs being mostly met by the private sector. Health insurance covers more of those in the higher income groups, but also covers a significant proportion of lower income people: 28 percent of those in the lower income quintile and 33 percent of those with urnmet basic needs are covered by an Obra Social. Of those not covered by any health insurance plan, the majority are domestic workers and other informal sector workers. Sixty-two percent of the unemployed also lack any health insurance coverage, since when a worker becomes unemployed, he or she loses coverage. It is also interesting to note that 10 percent of those in the highest income quintile choose not to insure against health care expenditures, bearing their relatively low health risk and assuming that they will use the public hospital in case they incur in catastrophic health care expenditures. Quality of care. There seems to be no discrimination against the poor in terms of admitting them to a public hospital when they have need. However, the poor tend to have to go farther to reach a public hospital, waiting times are long, public hospitals are open only a few hours a day, and there are few health posts to meet the basic health needs which means that the poor have to wait in hospitals for even the simplest of care. In addition, the coverage of public facilities varies greatly by province and region with a concentration of services in the richer, urban areas. Formosa, a poor province in the Northeast, has only I doctor per 1000 inhabitants, as compared to 9 per 1000 in the City of Buenos Aires. Generally, the poorer provinces have fewer private health care providers and less public facilities as well. But once in the hospital the poor do not get the same quality of care as those in the upper income groups: over 25 percent of women coming from the lowest household income quintile are delivered by non-doctors, while virtually no woman from the upper quintile has such experience. Over 30 percent of women coming from the lowest household income quintile have no post delivery follow up, while only two percent of women from the upper income group miss such follow up. In ambulatory care the same pattern appear to be true: 74 percent of patients coming from the lowest household income quintile had to wait more than 24 hours for a consultation, while only 45 percent of those in the upper quintile had such delay. Of particular importance for the poor is the quality and coverage of maternal and child care services. Using prenatal care as a tracer for the quality of public health care, we see that those with lower incomes are more likely not to receive good quality care. Some 25 percent of the extremely poor receive fewer than five prenatal exams. The 37 same association is true for immunizations where there is a strong association between income level and correct immunizations. Out-of-Pocket Health Expenditures Coverage against high pharmaceutical expenditures is a problem for poor families, as drugs tend to be expensive in relation to their income and the public system offers very inadequate coverage: 43 percent of the poor in the lowest family income group quintile had to pay pharmaceuticals out-of-pocket and eight percent could not afford the prescriptions written by their doctors. According to the National Survey of Household Expenditures, the poorest quintile spent about 5% of their total budget on medicines, medical services and supplies. The Poor and Health Care Financing35 Total public health expenditures are sufficient to finance an acceptable package of public health and personal health benefits for the entire population. In 1997, Argentina invested about US$12,900 million public monies in the health sector, which correspond to about, 15 percent of total public expenditures and 22.3 percent of total social public expenditures. Although health expenditures have increased by more than 50 percent from 1990 through 1997, they decreased as a proportion of the GDP (from 4.6 percent in 1995 to 4.0 in 1997) and as a proportion of total social expenditures (from 23.6 percent in 1995 to 22.8 percent in 1997). Table 27: Source of health sector finance 1997 Federal Province Municipal Work-related contributions Percent 8% 27 % 7% 58 % Source: The Government of Argentina: National Directorate for Social Expenditure Programming Over the nineties, the relative responsibility of different government levels for public health sector finance has changed: the role of the Federal government has declined, that of the Provinces has remained constant and that of municipalities has grown in importance. Over the same period and in the area of work related contributions, the relative weight of the national union-based health insurance, the Obras Sociales, has declined, while expenditures related with the insurance for the elderly and pensioners (PANI) and those related to provincial health insurance schemes have increased their share of the total public health expenditures. In 1996, provinces invested on average about 11 percent of their total public expenditures in the health sector. However the relative weight of the health sector in the total public expenditures varies widely, ranging from less than 5 percent total public expenditures in Formosa, to a high of 27 percent in the city of Buenos Aires. The investment each province makes in its public sector, both in pesos per capita and as a 35 The Government of Argentina: "Caracterizaci6n y evoluci6n del gasto publico social en el periodo 1980 -1997." (Buenos Aires, AR: Secretaria de Programaci6n Econ6mica y Regional, mimeography, October 1998). 38 proportion of their geographical product is highly correlated with the respective income per capita (R= 0.7 and R= 0.6, respectively). Thus, public health services behave as luxury goods: the richer the province, the more it puts into its public health system. Public resources available to care for the poor are about half what one would have in an equitable distribution of public health monies. Care for the poor is financed from provincial budgets which constitute about twenty seven percent of total public health sector financing. Provincial budgets have to: a) cover most public health services for the entire population; b) subsidize the care provided to the insured population at public hospitals, given that such care goes mostly unpaid by the respective insurance carriers; and c) provide personal health services for the uninsured population. Data suggest that about 8 to 10 percent of provincial health budgets goes to finance general public health services, and that 30 percent of public health services are provided free of charge to insured patients, which means that only 60 percent of provincial public health expenditures, or 18 percent of total public health expenditures, are available to pay for personal health care for the 37 percent of the population who is uninsured, about on half of what one would expect in a equitable situation. Improving health protection for the poor. The Government, with earlier Bank assistance, has embarked on several major national level reforms of the health insurance system to improve its efficiency and equity.36 The next big push in Argentina's health system reform should be to continue consolidation of the insurance system and improve health care services for the poor without insurance. As mentioned, the provincial public health system by providing virtually free services acts as a safety net for the poor. The problem, however, is that because of inefficiency and lack of focus it is unable to ensure that the poor get good quality services and needed medicines. This problem is indicated by the extent to which poor families are disadvantaged in terms of the length of time they have to wait to be attended, the amount of unfilled medical prescriptions, the unrealized recommended diagnostic analyses and the lower prevalence of reproductive and prenatal care. Under the current system, government spending is used to maintain public hospitals and their staff regardless of quantity and quality of services rendered and the ability to pay of the clients being served. The consequence is that there is no incentive from the financing system to encourage provision of good quality service, to be productive in terms of the number of clients served and to improve subsidized services to the poor. The hospitals received their budget on the basis of past allocation and some allowance for inflation. The experience of Tucuman, Salta, San Juan and Rio Negro highlights the above issue. Recent study of these provinces reveals that public health facilities are 36 Essentially, the reforms have sought to rationalize the process and rules of the game governing the obligatory national Obras Sociales. One of the noteworthy reforms is providing members of these Obras some choice and establish some competition among them. Furthermore, it sought to address funding for catastrophic illnesses and raising the minimum health package for members of low income national Obras Sociales. 39 underutilized. They spend too much on salaries and too little on medicines, equipment and other materials. They not only are overstaffed; their staff mix also lacks balance (they have too many administrative and general support staff). The problem is that incentives to focus on the poor, perform efficiently and deliver good quality services are not embedded in the way public health facilities are financed. Moreover, their frontline managers do not have the necessary authority to manage effectively and efficiently. To address the above issue, these provinces have decided to adopt, with World Bank support, a health sector reform strategy that would concentrate provincial government health expenditures on the uninsured poor and change the payment mechanism. The hope is to both improve health care protection of the poor and raise the quality of services rendered to them by creating "choice and voice" mechanisms.37 IV. Reducing Poverty in Argentina: The Way Forward It is clear, that despite an impressive record of stabilization and recovery, Argentina has a relatively high level of poverty. Recent trends indicate a widening gap between rich and poor, and a rising level of poverty, despite economic growth. While this may be the natural consequence of major structural shifts (privatization, liberalization, increased competition and deregulation) which is still going on, nevertheless the results, particularly since 1995, have been disappointing. While a fairly extensive system of social programs has had some positive impact in offsetting poverty, but has not produced the kinds of results one might have expected. In this situation, what can be done? In general, we can see that the poor are poorly educated, often unemployed or employed in low salaried jobs, and frequently work in the informal sector, but subject to a high degree of vulnerability in terms of possible job losses and income declines. While most of the poor are in the urban areas, the poor in the rural areas are also heavily dependent on non-farm employment and earnings. Clearly, employment and wages are key to poverty reduction. In this situation, and following the strategy laid out in the Bank's World Development Report 1990 and subsequent thinking, we can conceptualize a strategy built around three main pillars: * First, reforms and policies that will lead to a pattern of growth that will be both more rapid and more stable than in the past, and feature a higher level of employment per unit of output. * Second, improving the access of the poor to basic services, that will both raise the level of their human capital of the poor, and give them opportunities to improve the productivity and their ability to compete in an increasingly globalized economy; * Thirdly, reduce the vulnerability of the poor to shocks and losses in income, chiefly by improvements in safety nets to both protect the poor during 37 For a description of the design of the strategy and its rationale, see World Bank (1997), Argentina: Improving Health and Education (Salta, Tucuman and San Juan). 40 economic downturns, and keep them from making short term adjustments that will have negative impacts on their long term ability to reduce their poverty. Generating Labor Intensive Growth. Macro-economic policies that permit rapid and stable economic growth without inflation are an essential first step to a significant decline in poverty. A sustained growth of per capita income of 1.8% could reduce poverty by 35% in ten years, provided the benefits go to all parts of the economy. This would bring the urban poverty rate down to about 19 percent, comparable to the poverty rate in Chile (22 percent in 1997) and Uruguay (19 percent in 1998). However, it is also important that this growth be stable, as well as high. Macro-management, particularly monetary and fiscal policies, should be used to not only generate a high rate of growth, but also to stabilize the economy, avoiding large swings in output, employment and poverty. There are a number of necessary conditions for sustained growth in Argentina. The first is lowering the perception of country risk which at present constitutes a drag on all economic activity via the increase in the cost of capital to both the private and public sectors. The lowering of the cost of capital, among other things, would make construction even more attractive, generating low skilled employment. In order to do that, however, fiscal discipline and compliance with the new Law of Fiscal Responsibility, which requires a fiscal surplus by 2003, are essential along with a gradual reduction in public debt and the promotion of exports. Continued deregulation of factor markets and simplification of administrative processes are also elements of a growth strategy for Argentina. Given that most other factor markets have been deregulated, it is increasingly important that labor markets de jure are modernized to provide a level playing field for labor. At present, Argentina's labor market is one of the most rigid and regulated in the developing world, preventing adjustments in the formal sector from taking place easily. While de facto there may have been deregulation via extraordinary negotiations between unions and large companies or the shift to informal sector employment and employment in negro for smaller firms, the continued existence of outdated labor laws constitutes a future uncertainty and imperils the security of both firms and workers. The high incidence of evasion both of paying taxes and contributing to health insurance, unemployment insurance, and pensions constitutes a very serious fiscal and social problem for the Government. Greater labor flexibility in the formal sector, in contrast to the informal sector which by definition is flexible, would have two beneficial effects. First, such a move would likely attract new investments on the part of large companies with commensurate growth in employment albeit still at relatively high capital-labor ratios. One estimate of this is to reduce unemployment by about 3 percentage points38. While this may not be significant in terms of immediate poverty reduction, as much of this new job creation would be for the highly skilled, there would be secondary income effects on services and construction which would likely be of a more labor intensive nature. 38J.L. Guash, Estimating the Benefits of Labor Reform in Argentina, World Bank Report No. 15643 AR, (May, 1996) 41 The second and perhaps more important impact would be on small and medium scale enterprises which tend to be more labor intensive. Some 75% of all employment is in firms with 40 or fewer employees, and many of these firms operate largely in negro which has the negative effect of restricting their access to credit:39 Understandably, firms which do not keep accounts or pay taxes or only reveal a limited part of their business activity- to the authorities are not able to get credit from the banking system. Thus, their growth potential is limited to alternative sources of financing, including internally generated funds and becoming part of the formal sector would allow access to the banking system. A collorary to this is that the employees of these firms, many of whom earn low wages, would be engaged on a formal basis and benefit from social security and pension benefits. This would be a major step forward in protecting these workers, given that the SMEs are the most volatile part of the business community and informal salaried workers are the most likely to suffer from unemployment. Some of key reforms that would facilitate a more orderly operation of the labor market include: * elimination of centralized or sectoral collective bargaining agreements which are automatically extended to all workers in a sector, even if not signed and even when expired. This would have the particular advantage of being able to tailor labor agreements to regional conditions and to promote investment in that are adapted to the comparative advantage of different parts of the country; - reduce the high barriers to labor mobility (hiring, firing, changes from one employer to another) by moving to a fully-funded unemployment insurance system based on individual accounts. This would have the advantage of allowing firms in effect to capitalize these costs so that they are not incurred at a point in time when the firm is typically experiencing financial difficulties. It would reduce the uncertainty about the eventual costs of engaging new employees; * allow temporary employment that is not subject to full payroll taxes, as under the former modalidades promovidas, but still offer employees basic coverage to social services; * extending adopting programs, such as PYMES, which permit exceptions to present labor legislation for small and medium scale enterprises; and * as new agreements are negotiated, vastly simplifying and reducing the array of employer contributions and employee deductions that are part of the collective bargaining agreement, keeping the core programs (pension and health 39 Clarin, Feb. 7, 2000, p. 12. 42 insurance), at least for smaller firms which cannot afford to make these contributions40 A critical problem remains that a large part of the labor force in the informal sector lacks any form of pension or unemployment insurance coverage. While modifications of the present labor legislation may provide more incentive for firms to join the formal sector, there is still likely to be a significant number of firms which will not see this to their financial advantage even if this allows them to access banking credit. A concerted effort must be made with the tax authorities to reduce tax evasion as a complement to efforts to reduce labor costs. There also may be resistance from the employee side. If the standard deductions from the employee's pay-cheque--about to 20 percent of payroll-may too not be attractive to the lowest paid workers.41 Thus, they too may prefer to remain in the informal sector, working in negro, rather than have their take-home pay reduced. In the case of unskilled workers, with an average monthly wage of about $350, a drop of $70 in disposable income may be too large to absorb even if there are clear benefits in the long- term to pension and health insurance coverage. In these cases, from the point of view of risk management, a reasonable option may in fact be a public subsidy to help cover the costs of basic health and pension services for the lowest paid workers. Labor training remains an unknown quantity. While the government supports and large number of programs, and spends significant sums on these programs, little is known of their effectiveness in raising productivity or reducing unemployment. Studies from other countries generally are skeptical about the benefits of these programs, and a major evaluation of these programs would be a high priority. If anything, the current situation in Argentina, with very high rates of unemployment for the young, would point to an increased focus on technical training, apprentice programs, and mixed school-work programs which combine efforts to reduce drop-out rates among poor youths and provide needed work experience for them to be able to enter the labor market full time. At the same time, these programs would have the advantage of making secondary school and technical training more responsive to the needs of the private sector and the emerging demand for skills. For those workers hardest hit by wage declines since 1993-the long- tenured, informal salaried workers with primary or secondary education-there may be little effective remediation possible other than reactivation of the economy, particularly the construction industry. Increasing Access and Quality of Services. Since productivity is clearly linked to education, effort should be intensified to raise the level and quality of education available 40 For example, in the case of the collective agreement for the Commerce Union, which covers some 900,000 employees (only 600,000 of which are in the formal sector), the total of employer and employee contributions amounts to 59 percent of base salary. Of this, 26 percent is for severance pay. Other contributions-such as union contributions, life insurance, complementary pension funds, family assistance - amount to 17.6 percent. 41 Evasion by all workers and the self-employed is a wide-spread problem. Only 47 percent of formal sector workers contribute to the privately run pension scheme, down from 60 percent when the system was first introduced. While some continue to contribute to the public pay-as-you-go scheme, a significant number evade. 43 to the poor. While the system as a whole suffers from problems of quality and of relevance, the poor are particularly disadvantaged as they do not have the option of joining the private system. Moreover, it is likely that public schools in poor areas are of less physical quality, more crowded, and less well equipped than schools in the wealthier areas. Finally, poor children are often more likely to need extra attention, in part due to their poor access to early childhood education, fewer available resources in the home, and greater pressure to generate income and/or care for other household members. One of the key problems is that children of poor families are more likely to drop out of school for various reasons. The impact of the recent recessions in 1995/96 and - 98/99 seems to have actually worsened the situation with enrollment rates for the poor declining. As a result, the poor have less access than the rich to higher levels of education, which increasingly exhibit the highest returns. A viable strategy in education would include: * expansion of early childhood development centres and pre-school for the high risk poor; * greater investments in secondary schools in poor neighborhoods, such as by extending the present Plan Social Educativo and by ensuring equitable-if not progressive-funding of schools on a per student basis, rather than by school as is currently the case; * expansion of the hours of schooling in high risk schools from the present four hours per day to full days (6-8 hours) including extra-curricular and remedial programs42 * direct provision of books and texts for poor children so that they may work extra on their assignments (at present texts and books are only available during school hours); * cash grants to poor families conditional on keeping children in school particularly at the secondary level, in order to offset the economic incentives from school leaving and the effects of unemployment; * accommodation of the special needs of young mothers, who would otherwise drop out from school, either through in-house nursery services or other means for them to continue to study; * diversification of technical offerings and creation of new technical and job-related programs, jointly with the private sector, which would provide poor students with early employment experience and encourage drop-outs to return to school; 42 Extension of the school day is something which Argentina should generalize throughout the system as well as extend the school year. 44 * the establishment of a system of partial cost recovery from students at public universities, who generally tend to be from non-poor families, and the establishment of a nation-wide system of scholarships for students from poor families; and * expanding the capacity of the current public university system, both by improvements in operating efficiency and through further investments, as well as developing an integrated system of community colleges. As noted above, these efforts to improve education for the poor should be accompanied by system wide improvements to education quality, including reform of teacher statutes, gradual upgrading of teacher qualification, greater school autonomy and accountability for results. While the situation in health is less critical as most agree that basic services are provided, there are three main avenues which would greatly help improve the situation of the poor and provide greater equity and quality in the system. These efforts would go hand-in-hand with programs to instill greater efficiency in the health sector as a whole. These actions are as follows: * Reorientation of the public health care system to those without health insurance, by improving cost recovery from those with insurance and the ability to pay, and by improving the operating efficiency of the public hospital system. While granting more autonomy to public hospitals can improve their efficiency, care needs to be taken to avoid building in incentives that will reduce services to the poor. - Eventually, health insurance coverage should be extended to those in the informal sector not presently covered. The public health insurance system for the elderly poor (PAMI) could be improved by opening PAMI to competition by allowing freer entry and exit to other health insurance institutions. Likewise, allowing people to switch between Obras Sociales would encourage competition and reduce costs. In addition to improve financial management and accountability in both institutions. * Reinforcement of public health programs which address the particular needs of the poor, particularly women and children. Existing programs of maternal and child health (PROMIN) need to be expanded, and linked with family planning and reproductive health services for the poor, in order to reduce their currently high rate of fertility among the poor, the incidence of early pregnancies and maternal mortality rates As well, existing nutrition programs need to be reoriented away from using them as a proxy to income supplements to genuine support for nutritional deficiencies of the very young and lactating mothers. More generally, these programs need to be better integrated into the primary health care system and early childhood development programs. 45 * Help in controlling out-of-pocket expenditures. At present, since drugs are not provided by the public hospital system and the obra social insurance schemes demand co-payments, it is important to address such costs in whatever system is providing care for the poor. At 5 percent of income on average they could be covered under the basic health care package for the poor. * Greater emphasis on preventative health expenditures, including vaccinations, nutrition and health education, and family planning. Deficiencies in infrastructure both reduce the productivity of the poor, and limit human resource development. The urban poor live in areas usually devoid of adequate sanitation and safe water, and often without paved roads. Provision of such public services in poor neighborhoods can improve health outcomes. But attention also needs to be focused on building up communities, especially in urban areas, that lack roads, lights, and other services, and do not have legal titles to their land. Existing large public sector subsidies for housing (FONAVI) which are not well targeted would be better reallocated to improvements in basic urban infrastructure. Other major infrastructure programs- such as the one for Greater Buenos Aires, Cono Urbano-need to be redirected to ensure basic coverable for poor areas. The urban poor are particularly vulnerable to problems of crime and violence, and attention needs to be paid to alcohol and drug abuse, and improvements in police protection and access to justice. The core features of the reoriented FOVANI and Cono Urbano programs would be as follows: * Promotion of private credit schemes for new housing construction, using the private banking sector, facilitated by the creation of a secondary mortgage market. These schemes could be open to various income levels with a subsidy according to need. * Where appropriate, slum-upgrading programs which would regularize land title and provide basic services accompanied by self-help housing improvement programs. * Targeted subsidies for the poor to connect to the privately operated water and sewerage systems. The subsidy would be for connection charges only with the consumers expected to pay for consumption but according to a reduced tariff schedule. * Financing of municipal level urban investments in drainage, flood control, and roads targeted to poor neighborhoods. The problems of crime, social exclusion, alcohol and drug abuse, violence in the home, etc, need to be addressed at the local level, through community based programs which are locally managed. This is an area in which non-governmental organizations, community based groups and others in civil society can play a great role and there are many local initiatives underway. 46 Nevertheless, more concerted efforts need to be made to improve systems of information to help guide these programs. Statistics are poor or not available on many of these questions or are of doubtful reliability, such as the reporting of crimes. A very helpful first step is to undertake a national survey of victimization which can assist policy makers, community leaders and program managers to make more informated decisions on the nature and magnitude of the problems that exist and the trends. Reducing Vulnerability. Recent economic "shocks" clearly demonstrate the need for a strong system of safety nets. There is a more common acceptance that the global trends to liberalized markets and rapid technological change have increased uncertainties and may make employment more volatile. In the case of Argentina, the dependence of the country on external capital makes it highly exposed to the perceptions of the markets and to external shocks. While, in the past, there was high job security in the public sector or protected private firms that guarantee of life-long employment is not longer there. At the same, the demands of the market place for skills have increased leaving behind a part of society which cannot compete. All of these factors heighten the need for greater management of social risks. The Government has many good and well targeted programs, such as Trabajar, which have proven effective in reaching down to the poorest segments of the population. Nutrition programs are generally well targeted (albeit of unproven effectiveness to address malnutrition), as is the Plan Social Educativo. However, often these programs reach only a relatively small percentage of the eligible poor. In addition, they often do not expand to meet needs during a crisis. Rather, they tend to be cut back just when they are needed most. Overall, there has been inadequate evaluation of the impact of various programs, or their usefulness. To address these concerns, the Government needs to: * carefully evaluate existing activities and identify high priority programs that will be protected from budget cuts during a crisis. This should be an integral part of the planning and budget processes and explicitly covered by the Law of Fiscal Responsibility in how the fiscal cushion should be used; * eliminate weak programs and duplication of efforts, and put more resources into programs which have proven effective. This calls for a major effort on evaluation of program impacts and to reorder federal-provincial responsibilities for social programs. Most programs should be decentralized and managed at the municipal level. This would facilitate greater integration of social programs; * ensure support those programs that can be expanded during a crisis to provide emergency employment and income opportunities for the poor; and * take additional steps to improve targeting, so as to reduce leakage to the non- poor. This could be done by extending the present SISFAM system of beneficiary identification to a larger number of social programs, especially 47 those outside of the Ministry of Social Development, and expanded use of SINTyS to detect overlap and misuse among beneficiaries. It is also recognized that the issues of the working poor and unemployment are only partially addressed by the above programs; it is evident for unskilled workers, especially those with a large number of dependent, the present level of average monthly wage is insufficient to bring that household above the poverty line. Moreover, without unemployment insurance for informal sector workers, individuals will still confront periods of income loss. Affordability At present, Argentina spends a considerable amount of money on social programs only a small portion of which are directly targeted to the poor. The bulk of the monies are allocated to universal programs of health, education and pensions. Nevertheless, some of the recommended changes can be accommodated through program changes. For instance, in the case of health spending, where the analysis carried out for this report indicates that the redirection of public health spending to the uninsured, combined with cost-recovery from those with insurance, would be adequate and leave funds for expanding public health programs. In the case of secondary education, greater funding would likely be needed, even if efficiency gains being sought under on-going administrative reforms are realized and redirected to new programs. This would also be the case for expansion of early childhood programs and pre-school and for expansion of the school scholarship/family subsidy program. Savings, however, could be generated from present job training programs. In the case of several targeted social programs, more analysis would be needed to determine if savings can be generated in order to free up monies for new activities. There are likely to be considered gains from consolidation of nutrition and even reducing the kind of assistance presently provided (moving away from food basket hand-outs to particular fortified foods for target groups). There are two obvious cases in which a major change in program orientation and cost-recovery are warranted: this is the case of FONAVI and Cono Urbano in order to fund the proposed program of urban infrastructure and housing for the poor and the other is eliminating free public university in order to fund scholarships and student loans for the poor. Universal income support programs are likely to be too costly to consider seriously. The options that can be considered includes generalizing or expanding the income subsidy for poor households proposed for those with school age children. One way to finance such a targeted cash grant program would be to reduce number of present social protection programs (over 60), and combine into a single cash grant program, that could be made conditional on school attendance or visits to health clinics. Such a consolidation would reduce administrative costs, eliminate duplication, and could improve targeting 48 Another group with particular needs are the aged not presently covered by the Government's pension scheme. About 55% of the elderly poor (200,000 people) do not receive any form of pension support. Extending the existing non-contributory pension system to all of the elderly poor would cost $252 million per year; extending coverage to just the indigent poor would cost $30 million per year. Likewise there is very little in the way of ongoing programs for the rural poor, with the exception of programs such as PROINDER. Although they are a small part of the overall poverty question, the degree of poverty among the rural populations, and indigenous groups, as evidenced in the survey of two provinces indicates that more serious attention should be provided to that group. A careful examination of the entire budget needs to be made to prioritize spending decisions. Establishing priorities between the social sectors and other needs cannot be done by this report, but efforts at an overall public expenditure survey are already underway. In addition, the Bank is carrying out a more intensive review of social protection programs, designed to give more concrete advice concerning improved targeting and administrative savings. With additional information on program effectiveness and target populations, the Government will be able to make a considered choice on priorities: for example the degree of coverage of particular groups-covering only the indigent and not the rest of the poor, or focusing on those with school age children and not those poor who have few dependents, and so on. Another area of choice is in terms of "who implements": there are choices among the different levels of government, or between the public and private sector or non-governmental organizations. These choices will affect not only the cost of new and expanded programs but potentially their effectiveness. When these choices on funding and program support are decided, the Government will need to follow through on these choices in terms of institutional capacity to deliver. At present, existing institutional capacities are of mixed quality, mandates often overlap and internal rivalries inhibit effective collaboration. Moreover, as mandates change so necessarily existing institutional make-ups change: this means reducing drastically federal level present as programs are decentralized. Poverty Monitoring Such informed choices, however, depend on good information and Argentina still does not have a good system of surveys or poverty maps that can be used to do poverty analysis and program targeting. Existing information is scattered among three different sources: INDEC's Permanent Household Survey (EPH) provides income data and labor force information, but no information on consumption, social indicators or the use of public services. Consumption data are available only in the National Survey of Expenditures (ENGHO), carried out every ten years, and social information has to be obtained from special surveys, such as the 1997 Survey of Social Development (EDS). Most surveys do not cover the rural areas, which have the highest rates of poverty, and particularly exclude the indigenous, about which little is known. Furthermore, price 49 indices are only available for Buenos Aires, making the deflation of income estimates outside of Buenos Aires very difficult, and as a result most studies of poverty until now have only concentrated on the capital city, further biasing the results. The Government is in the process of redefining the poverty line and revising the Index of Unsatisfied Basic Needs. A more comprehensive National Survey of Living Conditions will be introduced as a supplement to the existing EPH. This survey will include more questions concerning social conditions and welfare, benefits from government programs, and include questions on consumption as well as income. There are plans to produce a national consumer price index, but this should be supplemented with regional price indices as well. Thus, while progress is being made, further efforts are still needed to improve the quality, coverage and timeliness of poverty related data. 50 Appendix A. Types of Risks by Group, and Number of Poor Uncovered Age Group/ MAIN Ris Leading Indicator Value Number of Indigence and Indicators of a) Indigent poor Poverty Rates Selected Risks Poor uncovered43 * poorest qui ntile 0-5 Years - Stunted development - malnutrition - pre-school /ECD 22%* 1,000,000 12% indigent program 43% poor coverage 6-14 Years - Poor education - late entry 8%* 13% indigent quality (low human - grade repetition 27%* 45% poor capital development) 15-24 Years - low human capital - Secondary school 62%* 400,000 development (education enrollment 7% indigent quality/attainment) repetition 33%* 31% poor - Unemployment/low - unemployment wages - inactivity (violence, - inactivity substance abuse, etc.) 25-64 Years -- Low income - unemployment 36% 23% 1,000,000 5% indigent - below poverty 23% poor eamings (under- employment) Over 65 Years - Low income - Pension coverage 47% 55% 200,000 1.4% indigent rate 13% poor General - Poor health care - health insurance 35%* 6,000,000 Population coverage - Poor housing /lack of - running water 66%* 800,000 hhds. 7% indigent basic infrastructure --sewerage 53%* 1,000,000 " 29% poor - in flood-prone 28%* 600,000 " area P:\Acopr\poverty\Part-I-gray.doc 43 See Annex I for notes on calculation methods. POOR PEOPLE IN A RICH COUNTRY A Poverty Report for Argentina PART II. BACKGROUND ANALYSIS 52 POOR PEOPLE IN A RICH COUNTRY A Poverty Report for Argentina CHAPTER I: THE DETERMINANTS OF POVERTY Introduction Argentina is a relatively rich country, with an annual per capita income officially calculated at $8970 per person (1998, World Bank Atlas). Yet, despite this relative wealth it is a country with a surprising high degree of poverty. According to estimates calculated for this report, the overall poverty rate in 1998 was 29%, and rates in the poorer Northeast and Northwest regions exceed 50%. The overall objective of this report is to look at the reasons such high rates of poverty continue to exist, and how resources and policies might be adjusted to affect a more rapid reduction in poverty in the future. It is important to view the situation with regard to poverty in its economic and social context. Since 1990, the country has undergone a fundamental economic restructuring. The result has been a major move away from a high inflation economy with heavy state intervention and protection, to a more open, private sector oriented economy, and one characterized by low rates of inflation. During this process, there has been a monetary reform, linking the peso to the U.S. dollar, a fiscal reforrn, which has improved tax administration, and a major public sector overall, including civil service reform, privatization of state enterprises1, social security reform and fiscal restructuring. In addition, trade reform reduced import tariffs and eliminated export taxes, as well as quantitative controls on imports. Added to the impact of these adjustments was the impact of the Tequila crisis in 1995, which produced a decline in GDP and a sharp rise in unemployment However, the overall has been a period since 1991 of sharply lower inflation, and substantially higher rates of economic growth. Inflation fell from a peak of over 1,300 percent in 1990, to less than one percent during the years 1996-98. Real economic growth, on the other hand, increased from an average rate of -1.1% during the 1980s, to 5.8% during the 1991-98 period. Yet the period of economic restructuring produced both benefits and costs, many of which impacted on the poor. Privatization and public sector restructuring increased unemployment, and raised prices of some basic goods and services. On the other hand, privatization improved the quality of basic services, and economic growth raised substantially per capita consumption levels. Data and Coverage Trends for poverty are based on the use of a poverty line, and the standard headcount and poverty gap measures. For this report, we use the Government's official poverty line calculated based on the 1986/87 Income and Expenditure Survey, updated ' It is often alleged that privatization has impacted adversely on the poor. Analysis of trends in public service prices does not support this hypothesis (see Chapter VII). 53 using price indices for its food and non food components (see Box 1). This poverty line is equal to about' $160 per male adult, per month, in 1998.2 To calculate poverty, household composition is converted into male adult equivalents using standard conversion factors. The extreme poverty line, or indigence line, is based on the food consumption portion of the poverty line, and is equal to $69 per month in 1998. These lines applied to the Permanent Household Survey (EPH) which is undertaken semiannually by INDEC, and which covers approximately 28 urban areas in the country, or about 70% of the total urban population. However, data before 1990 refer only to Greater Buenos Aires, since data for other urban areas is not available on a consistent basis, or is non existent in early years. Since the rural population is excluded from these estimates, they tend to understate overall poverty levels, since what little evidence we have suggests that poverty is higher in rural areas. However, since Argentina is 89% urban, including the rural population would not seriously change the results4. Box: Adjustments for Underreporting We know that the total income reported in the households surveys, such as the EPH, fall short of the estimates of disposable income derived from the national accounts. The differences are often attributed to the underreporting of income in the surveys. Adjustments for underreporting are difficult because detailed breakdowns of the income side of the national accounts are not made every year, and there is no evidence how underreporting varies by income group. We do know that there is more underreporting of entrepreneurial and interest income than wages, and one common technique is to assume equal adjustments for each type of income. This approach tends to result in larger adjustments for the upper quintiles who have most of this income. A recent study by Llach and Montoya, uses this approach, based on data from 1993 for Buenos Aires. It estimates an underreporting of income of 87%, overall. Revised estimates of poverty for Buenos Aires based on this methodology change their estimate of the poverty rate from 28% to 18% in 1998. However, such adjustments also worsen income distribution; the share of the lowest 20% falls in their estimates from 3.6% to 3.1%. However, to accept these estimates one must assume that the differences in found in 1993 are valid to adjust 1998 figures, that underreporting by type of income is constant across income groups, and that there are no errors in the calculation of disposable income in the national accounts. Because of the tenuous nature of such adjustments, we have not introduced adjustments for underreporting in the poverty estimates given here. source: J. Llach and S. Montoya, En Pos de la Equidad, Buenos Aires, April 1999. 2 The poverty line is actually $160 for Greater Buenos Aires, and slightly lower for other regions (see H. Lee, Background Paper No. 1). One should note that the poverty line being used here, roughly $5 per day, is somewhat higher than that used in most LAC countries, which is closer to $2 per day. On the basis of $2 per day, the poverty rate in 1996 would be about 15%. See World Bank, Poverty and Policy in Latin America and Caribbean, Latin America and Caribbean Regional Study, 2000(forthcoming). 3 This report also makes use of the INDEC's National Survey of Household Expenditures for 1996/97 (ENIGH), which covers the full urban population, and the INDEC/SIEMPRO's Survey of Social Development (8/97) which has a similar coverage. Unfortunately, neither survey have comparable time series data. Estimates of poverty for 96/97 using the ENIGH, however, are comparable to those obtained using the EPH data. Annual poverty rates are an average of two surveys, May and October. 4 Rural poverty estimates for Salta and Misiones, two of the poorest provinces in the country, give estimates of poverty of about 75% (1998). If urban poverty is 29%, and rural poverty throughout the country is 75% (an extreme estimate), the total poverty rate would rise only to 36%. The term "urban" refers to population living in towns of 5,000 or more. 54 Trends in Poverty For the period before 1990, poverty estimates have to be based on data that cover only Greater Buenos Aires. The gradual deterioration of the economy, marked by increasing inflation, slow and even negative growth, and rising unemployment had a telling effect. Poverty rates in Buenos Aires rose from 8% in 1980 to 41% in 1990, perhaps not surprising given a 26% drop in per capita income (see Table 1.1), and a doubling of the unemployment rate. Table 1.1: Long Term Trends in Poverty, Employment, Growth and Inflation (1980-1998) Poverty Rate Poverty Rate Unemployment Inflation Per Capita GNP Growth of real year Buenos Aires Urban Areas (% of labor (% (1980=100) GDP(%) force) increase) 1980 8.0 -- 2.6 100.8 100.0 4.5 1985 16.0 -- 6.1 672.2 79.5 -2.0 1988 33.1 -- 6.3 343.0 83.6 -1.9 1989 38.1 -- 7.6 3079.5 73.3 -6.9 1990 41.2 41.5 7.6 2314.0 74.3 -1.8 1991 26.4 30.5 6.1 171.7 82.2 10.5 1992 18.7 24.2 6.9 24.9 90.0 9.9 1993 16.9 21.8 9.6 10.6 94.3 5.7 1994 17.0 21.6 11.0 4.2 100.3 5.9 1995 22.6 27.2 17.3 3.4 94.5 -2.7 1996 25.5 30.1 16.5 0.2 97.6 5.5 1997 25.2 29.4 14.7 0.5 104.1 8.1 1998 24.9 29.3 13.5 0.9 108.4 4.0 Note: Unemployment prior to 1990 refers to Greater Buenos Aires only; GDP growth for 1985 is compound growth rate 1980-85. Source: INDEC. Poverty Rates calculated from EPH data by Bank staff, GNP from World Bank Since 1990, there has been a substantial drop in poverty, with the overall headcount ratio for all urban areas dropping from 42% in 1990, to 29% in 1998 (see Table 1.2). Extreme poverty falls from 10% to 7%. However, this longer period can be subdivided into two sub periods. During 1990-94 both poverty and extreme poverty declined, with poverty rate reaching 22% in 1994. The second sub-period, 1994-98, poverty increased from 22% to 29%, and extreme poverty from 4% to 7%.The other standard poverty measures, (FGT), which include the poverty gap and the poverty depth (poverty gap squared), show a similar trend. 55 Table 1.2: Rates of Poverty, Extreme Poverty, and Poverty Gaps (urban areas only) Poverty Extreme Poverty Poverty Income Income Year Headcoun poverty gap depth gap gap/GDP t 1990 41.4 11.3 16.4 8.8 39.6 10.1% 1991 30.4 5.9 11.2 6.0 36.8 5.8% 1992 24.1 4.4 7.8 3.7 32.4 3.7% 1993 21.8 4.3 7.4 3.6 33.9 3.3% 1994 21.6 3.7 7.2 3.4 33.3 3.0% 1995 27.2 6.1 9.9 5.1 36.4 4.4% 1996 30.0 7.3 11.3 6.0 37.7 4.9% 1997 29.4 6.8 11.1 5.8 37.8 4.5% 1998 29.4 7.1 11.2 5.9 38.1 4.5% Source: Calculated from EPH surveys, average of May and October. Poverty depth is squared poverty gap, or FGT2. Given that there were about 32.2 million people in urban Argentina in 1998, this means 9.4 million living in poverty, and about 2.3 million living in extreme poverty. The poverty gap index in 1998 was 11.2%, and the increase in income needed by the average poor person (the income gap) was about 38% (see Table 1.1). Calculating the income gap as a share of GDP, we see that about 4.5% of GDP would be sufficient to eradicate poverty (bring everyone up to the poverty line. These poverty estimates could be lower, if one were to adjust the numbers for what appears to be underreporting of income in the household surveys. However, because of the methodological difficulties that this entails, this report does not introduce any adjustments for underreporting (see Box). Poverty Trends and Employment As shown in Figure 1.1 the poverty rate drops sharply from a peak of 42% in 1990 to about 22% in 1994, rises after the Tequila crisis to about 30% in 1996, and stays roughly at that level despite a resumption of growth. The numbers in severe poverty show a similar, although more stable trend. In short, the resumption of growth and decline in inflation after 1991 seem to have had a material effect on poverty through 1994. Since then, poverty seems to persist despite a return to the higher growth rates experienced earlier in the decade.5 The causes of this are not entirely clear.6 Figure 1.1: GDP Grov and Poverty 50 40 - 30 * a._ rae headcot % 11990 1991 1992 1993 1994 1 1996 1997 1998 --Irdgeny, srae -10 Of 5 The trends for the poverty gap and the FGT2 index are similar; see Background Paper No. 1 for details. 6 This phenomenon is not unique to Argentina, but seems to be common throughout Latin America. See Juan Luis Londono and Miguel Szekely, "Persistent Poverty and Excess Inequality: Latin America, 1970- 1995", Inter-American Development Bank, Working Paper Series 357 (October, 1997). 56 One possible cause is found in the link between poverty and unemployment. In Figure 1.2 we can see that unemployment shows a trend similar to that of poverty. In this case, however, unemployment rates rise after 1991, reaching a peak of 17% in 1995, and decline to 13% in 1998. The presence of persistent unemployment may be linked to a period of price stability. It is possible that in the past, employers relied on high inflation rates to reduce real wages, and were able to bargain for nominal wage increases that did not match inflation. In the present, low inflation environment, nominal wages are rigid, and inflation is not available to reduce real wages, so employers resort to layoffs. Figure 1.2: Unemployment and Poverty 45 0 355 *~30 25 ==-_ WN M 9)20 ____ ^ 15 -l-- Poverty, 10I headcount % 10i X 7(ndigency, 5 ~~~~~~~~~~~~~~~~~~~~rate of 0 = -1--unemploymen 1990 1991 1992 1993 1994 1995 1996 1997 1998 t(urban) A Profile of Poverty Who are the poor, and how do they differ from the non-poor? As summarized in Table 1.3, the poor: - have significantly larger families (4.6 vs. 3.1), 7 - have younger families with a much higher dependency Ratio (3.0 vs. 1.4),8 - Have much higher unemployment rates (twice the rate of the non-poor), - Have fewer years of schooling (about 25% less), and - Are more likely to work in the informal sector. The situation is clearly even more critical in these respects for the indigent. While most indicators are worse for the indigent, compared to the poor, particularly important is the much higher unemployment rate (37%), and lower average hours worked (35, compared to 42 for the poor). The latter seems to represent their inability to secure full time work 7 Although the gap is smaller if looked at in terms of adult equivalence, and no allowance is made here for scale economies within the family. s The dependency ratio is the number of non-workers in the family to the number of workers. 57 Table 1.3 Characteristics of the Poor, 1998 Indigent Poor Non-Poor Family size 5.6 4.6 3.1 Family size, adult equiv. 4.3 3.7 2.5 Female Headed (%) 26.8 21.0 37.7 Average Age (years) 20.9 25.1 34.7 Dependency Rate (per worker) 4.1 3.0 1.4 Labor Force Participation (% of adults) 54.8 56.4 67.3 Years of Schooling 6.9 7.9 10.5 Informal employment (%) 46.4 44.6 36.2 Unemployment Rate (%) 36.5 22.6 8.9 Hours Worked 34.9 41.7 45.1 Source: EPH, 1998 There does not appear to be much of a relationship between female-headed households and poverty. Surprisingly, the poor have a lower rate of labor force participation (among those aged 15-64), and work fewer hours per week. Among the indigent, the unemployment rate is a surprising 37%, which suggests that many of those classified as "indigent" are temporarily in this situation because of loss of employment. Chapter II explores these questions in more detail. Age and Poverty. Because poor families are both larger and younger than average, children are more apt to be in poverty than adults. As shown in Table 1.3a, 45% of all children (aged 0-14) were living in poverty in 1998, compared to a 25% poverty rate for adults. Poverty among older adults is below average. Only 13% of those 65 and older were found to be below the poverty line. The lower rate of poverty among older people probably reflects the fact that the formal sector, which covers about 60% of the labor force ensures that retired workers receive adequate pensions, and this is supplemented by a system of non-contributory pensions for certain poor groups outside the formal sector. Poverty among women is not significantly different from the general population, with the possible exception of the 65 and over group, which is slightly poorer Table 1.3a Pove Rate by Age Group and Se, (Urban Argentina, 1998) Total po pulation Female p pulation age group % indigent % poor % indigent % poor total 7.1 29.4 7.2 30.1 0-4 12.2 43.2 11.5 43.0 5-14 13.0 45.3 12.5 44.4 subtotal, 0-14 12.7 44.6 12.1 43.9 15-24 7.0 30.6 7.1 31.1 25-39 5.6 26.0 5.1 25.1 40-64 4.5 21.0 4.8 21.7 subtotal, 15-64 5.6 25.4 5.6 25.6 65 & older 1.4 13.2 1.7 15.2 Source: average of EPH for May and October, 1998, all urban areas. Probit Analysis: Of these characteristics, which are the most important in terrns of defining poverty? Using probit analysis, the characteristics of poor and non-poor families were analyzed (see Background Paper No. 1, Table 18), to estimate the probabilities of being poor given certain key characteristics. Table 1.4 reports some of these interesting findings. 58 Table 1.4 Probit Regression Results on the Probability of Being Poor Characteristics Probability of individual being poor Households: female headed -.015* households size larger than 7 .047 number of children .124 number of employed -.143 Individuals: age 65 and over -.035 retired -.161 female -.012 Household Head: Illiterate .217 secondary school completed -.183 higher education completed -.253 Unemployed .171 skilledjob -.055 professional job -.263 *Not significant at the 95% level; all other coefficients are significant at 95% or better. Probit regressions based on EPH data for 1998 In terms of relative importance, the education variables dominate. For instance, completion of secondary school by the household head reduces the chances of an individual being in poverty by 18%, and higher education by 25%. Illiteracy of the household head raises the chances of being poor by 22%. Unemployment is also an important factor, with a probability of 17%, but having a household head with a skilled job is of relatively little importance. Again, coming from a female-headed household is not a significant factor, and being older or retired actually reduces the chances of being in poverty. Also, place of birth is not important. Migrants from outside of Argentina are no more likely to be in poverty than people born in Argentina (see box). However, this Migration and Poverty It is often aledged that poverty in Argentina is a result of migration of poor people from other countries, principally Bolivia and Paraguay. However, an analysis of the 1997 EDS shows that this is not the case. Looking at poverty in terms of place of birth, we find that only 4.1% of the poor come from a country other than Argentina, compared to 4.8% for the non-poor. For the indigent poor, the ratio is even lower at 2.6%. In fact, most of the poor are living in the same place where they were born. About 70% live in the same locality where they were born, and 84% in the same province (see Table). Distribution of Population by Place of Birth (% of total) l______________ Indigent Poor Non-Poor Total Same Locality 69.4 70.2 60.7 69.3 Same Province 16.9 13.7 19.1 17.8 Other Province 11.2 12.0 15.4 14.4 Other Country 2.6 4.1 4.8 4.5 Total 100.0 100.0 100.0 100.0 J Source: 1997 Social Development Survey (EDS) 59 analysis also excludes the indigenous population, for which no data is collected. Anectdotal evidence suggests that indigenous people tend to have higher levels of poverty than the general population, and often live in remote rural areas where service delivery is difficult (see box). Indigenous People and Poverty Government sources estimate that there are about 1 million indigenous people in Argentina, while indigenous groups put the total at closer to 1.5 million. Because of long standing prejudice against indigenous people, many of their descendants do not recognize themselves as indigenous. The constitutional reforms of 1994 recognized the preexistence of indigenous groups, and gave them important rights with regards to land, education, economic development, and cultural preservation. As a result, there has been a resurgence of indenties and the formation of new tribal communities, which have been registered in the National Registry of Indigenous Communities. Thus, the Census of 2000, which will include a question on indigenous background, might find substantially more indigenous people in Argentina than previously estimated. Poverty appears high among indigenous groups. While these groups occupy about 2 million hectares of land, much of it is in areas with difficult access, such as in the Andean mountains and in the chaco. Indigenous people are also found in National Parks, and in border areas close to Paraguay and Bolivia. Those living in mountainous areas subsist by raising animals, through the sale of artesian artifacts, and tourism. Those in the chaco live close to rivers, and subsist on fishing, hunting, and the collection of fruits, and supplement their incomes by working on larger estates during peak harvest season. The lack of titles to the land they occupy puts these communities at risk of being dislocated. While most communities have a primary school, learning is made difficult by the lack of bilingual education in most schools. Access to secondary schools is limited to those that can afford to travel and live in urban centers, and the number of government scholarships to assist in secondary and tertiary education are limited. hi terms of health, indigenous people suffer from high degrees of malnutrition, particularly among children, as a result of normally eating but once a day. Respiratory diseases are also common, including tuberculosis. High fertility patterns, with the average indigenous family having about seven children, further impacts on matemal and child mortality. - source: based on materials supplied by Mercedes Avellaneda Trends in Income Distribution Another factor irnportant in understanding the failure of poverty to decline in the 1 990s is the change in the distribution of income. It is obvious that, given a rapid growth in overall income, and little change in the incomes of those in poverty, that income distribution must have worsened. The implication is that the non-poor benefited to a far greater extent than the poor in the recovery of the economy. As can be seen in Table 1.4, there is a slow but steady deterioration in income distribution during the period, by any measure used. The Gini coefficient rises from .46 in 1990 to .49 in 1998. The share of income received by the upper 20% rises from 51% to 54%, while the share of the lower 20% falls from 4.6% to 3.8%. The ratio of the two rises from 11 to 14. These trends do not seem to have been affected by the decline in growth in 1995. The trends in income distribution are clearer if we look a per capita income for various income levels. In Figure 1.3, we plot per capita incomes over time for the first, fifth and tenth decile. Incomes for the first and fifth declines were virtually stagnant of the period; only those in the top deciles advanced. 60 Table 1.4 Changes Income Distribution, 1990-98 (1998 Pesos) Per Capita Household hicome byHDecile 1990 1991 1992 1993 1994 1995 1996 1997 1998 Average Income 1 42.6 49.5 51.3 50.4 52.3 42.1 39.2 40.7 39.2 2 75.0 84.4 90.0 92.4 90.1 79.3 74.4 76.8 74.6 3 98.7 112.1 120.8 125.2 121.8 108.5 103.4 107.5 105.2 4 124.4 141.2 153.8 159.7 155.7 138.4 132.7 139.5 137.4 5 156.0 172.9 185.3 198.1 194.7 169.8 165.5 176.6 173.8 6 190.5 207.4 228.9 243.6 237.9 212.6 207.3 217.3 216.3 7 231.0 256.1 282.4 301.8 294.3 261.7 260.0 273.9 276.0 8 292.1 325.2 359.2 387.4 378.7 337.4 338.4 353.7 364.0 9 398.5 445.2 498.3 528.9 517.0 476.4 476.2 499.0 515.1 10 839.4 974.7 1,027.0 1055.6 1,090.1 1,054.1 1,023.8 1,071.4 1,144.1 Overall 244.8 276.9 299.7 314.3 313.3 288.0 282.1 295.7 304.6 Income Share 1 1.7 1.8 1.7 1.6 1.7 1.5 1.4 1.4 1.3 2 3.1 3.1 3.0 2.9 2.9 2.8 2.6 2.6 2.5 3 4.0 4.1 4.0 4.0 3.9 3.8 3.7 3.6 3.5 4 5.1 5.1 5.1 5.1 5.0 4.8 4.7 4.7 4.5 5 6.4 6.2 6.2 6.3 6.2 5.9 5.9 6.0 5.7 6 7.8 7.6 7.6 7.7 7.6 7.4 7.3 7.3 7.1 7 9.4 9.2 9.4 9.6 9.4 9.1 9.2 9.3 9.1 8 11.9 11.7 12.0 12.3 12.1 11.7 12.0 12.0 12.0 9 16.3 16.1 16.6 16.8 16.5 16.5 16.9 16.9 16.9 10 34.3 35.2 34.3 33.6 34.8 36.6 36.3 36.2 37.6 Gini Coefficient 0.451 0.456 0.453 0.453 0.461 0.479 0.483 0.482 0.498 Share of Income 50.6 51.3 50.9 50.4 51.3 53.1 53.2 53.1 54.5 Claimed by Top 20% Share of Income 4.8 4.9 4.7 4.5 4.5 4.2 4.0 4.0 3.7 Claimed by Bottom 20% To Bottom 20% Ratio 10.4 10.5 10.8 11.2 11.4 12.6 13.3 13.3 14.7 (C=A/B) I I_I_I_I_I_I Sources: INDEC, Encuesta Permanente de Hogares, October of 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, and 1998. Household per capita incomes are from all sources including wages and salaries from primary and secondary jobs, as well as retirement, capital and other transfer incomes. 61 Figure 1.3: Changes in Per Capita Income 1990-1998 I1400 - . 1200- X 1000- $ 00 __> 800-~ __-. -5th Decile _ 600 - 1-- -t-10th Decile g 400 U 200- [L 0 Year While it is not completely clear what has caused the worsening of income distribution, one important factor has been the change in labor demand since apertura. Growing demand for more highly skilled workers has increased the wages of those with secondary and university level education. As shown in table 1.6, below, the incomes of professionals have grown by about 53% over the 1990-98 period, while skilled wages have risen only 13%, and unskilled declined 3%. Furthermore, looking at the data only since 1994, we find virtually substantial declines in the real wages of skilled and unskilled workers, while professional workers have continued to show gains. It is not clear whether this is a trend, however, or a temporary reaction to international recession and shifts in labor demand. Table 1.6 Income from Primary Job by Skill Level (monthly income in 1998 constant prices, pesos) Year Professional Skilled Unskilled 1990 1176.7 570.2 366.2 1992 1483.6 .685.2 438.9 1994 1715.8 725.4 424.4 1996 1661.9 632.5 354.5 1998 1794.1 644.8 356.2 Growth 1990-98 52.5% 13.1% -2.7% Source: INDEC/EPH / annual data are the averages of two surveys. Nevertheless, since professionals and skilled persons tend to be located in the upper level of the income distribution, an increase in their incomes relative to others naturally tends to worsen the distribution. The same effect can be seen from a measure of the 62 returns to education. Estimating Mincerian earnings functions for the period 1992 and 1998 showed that the results were highly significant.9 As Table 1.7 shows, a simple simulation exercise reveals that individual workers with higher education attainment were reaping higher returns for their investment in 1998 than in 1992. For those with only two years of schooling, the retums to education dropped from 6.8 to 5.0%. However, those with 18 years of schooling, the returns increased from 9.8% to 13.7% Table 1.7: Returns to Education and Predicted Earnings by Education Level Years of Returns to Education Predicted Earnings Schooling (% increase in income) (1998 pesos/month) 1992 1998 - 1992 1998 1 6.6 4.5 378.9 339.9 2 6.8 5.0 405.1 356.5 3 7.0 5.6 434.0 376.0 4 7.2 6.1 465.7 398.6 5 7.3 6.7 500.7 424.9 6 7.5 7.2 539.3 455.4 7 7.7 7.7 582.0 490.7 8 7.9 8.3 629.3 531.6 9 8.1 8.8 681.6 579.1 10 8.3 9.4 739.7 634.1 11 8.5 9.9 804.2 698.1 12 8.6 10.4 876.0 772.8 13 8.8 11.0 956.0 860.0 14 9.0 11.5 1045.2 962.2 IS 9.2 12.0 1144.8 1082.4 16 9.4 12.6 1256.3 1224.2 17 9.6 13.1 1381.2 1392.0 18 9.8 13.7 1521.4 1591.4 Sources: INDEC, Encuesta Permanente de Hogares, October of 1992 and 1998. The indicated returns are for cumulative education through the year indicated, not the marginal returns. The increase in demand for persons with college education should encourage more enrollments at the university level and higher graduation rates from secondary schools, more production of graduates, and eventually a closing of the wage gap. 9 The estimated Mincerian earnings functions are given by: October 1992: log Earnings = 5.155 + 0.064 S + 0.00093 s2 + 0.051 EX - 0.00078 EX2 - 0.423 FEM (t-ratios) (167.2) (12.6) (3.9) (43.4) (-31.8) (-47.0) + 0.543 EMP - 0.057 SELF + E (Regional Dummies) (23.4) (-5.4) adj. R2=0.389, n=20,511 October 1998: log Earnings = 5.016 + 0.040 S + 0.00269 s2 + 0.053 EX - 0.00078 EX2 - 0.477 FEM (t-ratios) (156.5) (7.6) (11.3) (43.3) (-30.5) (-51.1) + 0.565 EMP - 0.223 SELF + I (Regional Dummies) (23.0) (-20.0) adj. R2 = 0.351, n = 24,183 where S = years of schooling, EX = years of experience, FEM = female dummy variable, EMP employer, and SELF = self-employed. 63 However, this does not appear to be happening, as college enrollments are stagnating, and drop-out rates from secondary school remain high, particularly among the poor (see Chapter 4). One can estimate what would have happened to the poverty rate if there had been no deterioration income distribution, by decomposing the changes in poverty into changes in overall income, and changes in distribution. These estimates show that between October 1990 and 1998, income growth alone would have reduced the poverty rate by 14 percentage points to 25.3%, instead of the estimated 29.9%.This potential change was partially offset by the negative effect of the change in distribution, which alone accounted for a 4.7 percentage point increase in the poverty rate.10 This rather technical calculation hides somewhat the reality of what was happening; the pattern of the changes in demand and labor supply did not benefit the poorest elements of society. In a sense, they were excluded from the growth and recovery process. Table 1.8: Decomposition of Poverty Changes, 1990-1998 (rate of poverty, % of population) 1990 1991 1992 1993 1994 1995 1996 1997 1998 Component Oct Oct. Oct. Oct. Oct. Oct. Oct. Oct. Oct Poverty Rates 39.4 27.1 23.3 21.7 22.8 28.4 30.7 29.0 29.9 -- Income effect -11.1 -15.3 -16.8 -16.4 -12.5 -11.1 -13.3 -14.1 Poverty rate w/o distribution effect 39.4 28.3 24.1 22.6 23.0 26.9 28.3 26.1 25.3 -- Distribution effect -1.3 -0.8 -0.8 -0.2 1.5 2.4 2.9 4.7 -- Overall effect -12.3 -16.1 -17.7 -16.6 -11.0 -8.7 -10.4 -9.5 Source: based on EPH data and Bank staff estimates However, we can also ask another key question. How much of the change in the Gini is attributed to the changes in the returns to education, and how much to other factors? Between 1992 and 1998, the Gini changed by 4.0 points. Of that change, 1.5 points or about 38%, can be attributed to changes in the returns to education. However, that leaves 2/3rds of the changes unidentified and attributable to other factors, such as experience, hours worked, family size and others which are also influencing income distribution. ' Growth and Poverty Future efforts in poverty reduction will be clearly related to overall growth in income, as well as changes in income distribution. Using pooled cross section and time series data from the EPH for 1990-98 by regions, we can estimate the elasticities of poverty rates with respect to changes in per capita income and the Gini coefficient. 12 the estimated i° The lower Gini could also have the effect of producing higher growth rates, ftirther reducing poverty. See K. Deininger and L. Squire, "New Ways of Looking at Old Issues: Inequality and Growth", Journal of Development Economics, Vol. 57(1998) 259-287. " See Background Paper No. 1. 12 The estimated equation regresses the change in the log of the poverty rate (dlnPR) against the change log of per capita income (dhnPCY) and change in the log of the Gini (dlnGINI), with dummy variables used to capture the fixed effects of different regions. The estimated equation (without the dummy variables) is: dInPR=-0.004 -1.383dlnPCY+0.768dlnGINI R2=.64 (-0.454) (-9.635)* (3.670)* n=102 * significant 1% level. The equation is estimated using semi-annual data for the period 1990-98 for six urban areas. 64 Table 1.8a: Estimated Impact of Changes in Per elasticity for poverty with respect to Capita Income and Distribution on Poverty: growth is 1.38; a 10% increase in per capita income will reduce the poverty Percent Change in Poverty Headcount: rate by 13.8%, provided there is no % change total'change in per capita income change in the Gini. Table 1.8a shows in Gini +10% +20% +30% alterate combinations of changes in -13.8 L-27 7 -1.5 atraecmmtoso hne 05% -10.0 -23 77 -415 per capita income and the Gini. For +10% -6_.2 -20 3.8 instance, a 20% growth in per capita income would imply a growth rate Estimated Poverty Headcount: over 10 years of about 1.8% per change in total change in per capita income annum. This would reduce the Gini +10% +20% +30% poverty rate by 28%, if there is no 0 25.3 21.3 17.2 change in the Gini, but by only 20% +5% 26.5 22.4 18.3 _ if the Gini were to increase by 10%. +10% 27.6 23.5 19.5 _ If we start with a poverty rate of 29.4%, then the forecast poverty rate would vary between 21.3% and 23.5%. This simulation somewhat mechanical, but illustrates both the importance of growth in income in any plan to reduce poverty, as well as the importance of how gains in national income are distributed among the population. However, it also assumes that the elasticities of the Gini and poverty with respect to growth are constant over the cycle. Recent work by DeJanvry and Sadoulet13 suggest that these elasticities can have important asymmetries: they tend to be higher in a downturn than in a recovery. Poverty is likely to be reduced faster if, for the same average growth rate, there is a greater degree of stability in the rate of growth. Regional Poverty There are also important regional differences in poverty. As shown in Table 1.9, poverty rates are substantially higher in areas in the north, particularly in the North West (46% in 1998) and North East (49%). Conversely Buenos Aires and Patagonia are lower than the national average. The period 1990-98 can be divided into two sub periods: 1990-94, when poverty decreased, and 1994-98, when it increased. In the first period, overall poverty rates fell by 48%, but poverty in the two poorest regions fell by about half of that. In other words, the poorest areas seem to benefit less from general poverty reduction than the richer areas. However, when poverty rose during the 1994-98 period, the increases in poverty were greater in areas, outside of the Northeast and Northwest, particularly in Buenos Aires, Cuyo and Pampeana. Thus, it appears that the poorer areas are more stabile with regard to poverty in times of economic changes, and the impact of macroeconomic crises seem to occur in the richer areas. It seems possible that the areas outside of Buenos Aires are more dependent on agriculture, and therefore less susceptible to economic and financial shocks. 13 Alain de Janvry and Elisabeth Sadoulet, "How Effective Has Aggregate Income Growth Been in Reducing Poverty and Inequality in Latin America9" processed, Nov 9, 1999. 65 Table 1.9: Poverty Rates by Region, 1990-98 (% of population poor) Greater All Urban Year Buenos Aires North West North East Cuyo Pampeana Patagonia Areas 1990 41.2 54.4 55.7 48.1 33.7 26.7 41.5 1992 18.7 43.1 44.6 30.4 22.6 18.3 24.2 1994 17.0 41.6 40.3 26.1 19.8 17.1 21.6 1996 25.5 48.3 47.5 36.6 28.0 20.9 30.1 1998 24.9 46.0 48.8 36.0 27.4 22.4 29.4 % change. 1990-94 -58.7% -23.6% -27.6% -45.8% 41.2% -35.8% -48.1% 1994-98 46.5% 10.7% 21.0% 38.2% 38.4% 31.0% 36.2% Population 11.8 2.9 2.0 1.9 9.4 1.3 29.4 Gini (i996) .484 .455 .477 .452 .434 .462 .483 Lowest 20% 4.1 4.5 4.1 4.6 4.6 4.3 4.0 lgbare. - Poor, 1998 2.9 1.3 1.0 .7 2.6 .3 8.6 Source: INDEC/EPH, various years. Social Indicators: Another View of Poverty Poverty is a multi-dimensional phenomenon, and is not only an absence of money or material goods. Poverty can reflect various shortcomings or negative aspects of life, including bad health, lack of education, malnutrition and other social aspects, as well as other factors, such as crime and violence, absence of legal protections, inability to participate in the political process, etc. One common alternate measure of poverty is to examine social indicators. Such social indicators often are seen as measuring a person's ability to meet their basic needs, and are often combined into a composite index of basic needs (see box) Some important indicators are listed in Table 1.10 for Argentina. These indicators all clearly show progress in the period 1985-96. Life expectancy has increased, infant mortality has declined, secondary school enrollment is up, etc. If we compare Argentina to the average for all of Latin America and the Caribbean (LAC), it is clear that Argentina is well ahead on every front. However, such a comparison is wrong, since Argentina has much higher per capita income than the average country in LAC. In the column market "deviations", Table 1.9 shows the deviation from an expected value calculated from a regression estimate of what would be expected for a country of its per capital income14 What this column shows is that Argentina "underperforms" in certain important aspects. For instance, its infant mortality is 55% higher than it should be, and its secondary school enrollment is about 9% lower than expected. Access to safe water is 26% lower than expected. Only in terms of life expectancy and adult literacy does it appear to be at levels appropriate for its income level. Again, this suggests that 14 Deviations are the percentage differences (residuals) from an equation which predicts social indicators on the basis of per capita income. for 1996, using a world wide sample of countries. In general, the estimates are made using the log of per capita GNP, where GNP has been adjusted for purchasing parity (PPP). For details see N. Hicks and P. Peeters, "Poverty In Latin America: A Survey", World Bank, 1998 66 Argentina, despite being wealthy has relatively high levels of poverty, and is not providing adequate social services to its poor. Table 1.10 Social Indicators for Argentina 1980-96 ARGENTtNA AvCge Indicator: 1985 1997 Predicted deviations* 1996 Life Expectancy (years) 71 73 74 -.6 70 Infant Mortality(per 1000 27** 22 14 54.8 33 births) Enrollnent, Secondary 70 77 84 -8.6 53 Schools(%) Adult Literacy (%) 95 96 91 5.0 85 Access to Safe Water (%) 55 65 87 -25.5 76 *per cent differences from expected values, see footnote for source. **1987 data. Perceptions of Poverty by the Poor and Non-Poor Another way of looking at poverty is through the eyes of the poor themselves. The background work for this report included a survey of opinions and attitudes relative to poverty, based on a survey 1,200 households in 29 cities, which was undertaken in early 199915. The survey was based on a series of open-ended questions, which attempted to identify how society defines poverty, the perceived changes in the last five years, and the roles of institutions in reducing poverty. In addition, focus group discussions were used to add depth to some of the observations note in the survey (see box ). When asked whether they considered themselves poor, 26% responded positively. This figure is remarkably close the to 29% rate reported early based on the standard poverty line analysis. A similar question in the Encuesta Desarrollo Social (1997) undertaken by the Social Development Secretariat found that 32% of the respondents considered their family to be poor, a slightly different concept. The responses from the survey can be classified by the socio-economic level of the respondentsI on a high-medium-low basis. Those at the "low" socio-economic level correspond roughly to what we would consider "poor". What is of interest is that 59% of these "poor" do not consider themselves as "poore. Thus, even though the averages come out roughly the same, there are substantial differences between the self-perception of poverty and the traditional poverty line approach to measuring poverty. Of additional interest is the fact that 17% of the middle class and 2% of the "rich" consider themselves to be "poor". 15 See S. Cesilini and E. Zuleta, Background Paper No.2 16 The Socio-Economic Level (NES) is an index developed by the Argentina Marketing Association similar to the NBI. It combines information on the educational level and occupation of the household head, along with possession and use of various goods and services, including the possession of an automobile. 67 When those who self-identified as poor were asked the source of their poverty, the most important responses were: not having work 31% having a low salary 18 not being able to pay for basic needs 14 retirement undignified (.) 7 labor problems 7 What is clear from this is the heavy weight given to employment, higher than that given to wages, or being able to pay for basic needs. Respondents were also asked to evaluate the situation of their household over the last five years, on a scale of better, much better, the same, worse or much worse (see Table 1. 11). Overall, the data here confirm the trends in income distribution. About 18% of the respondents answered "better" or "much better," while 46% chose "worse" or "much worse." When examined by socio-economic level, about 50% of the "poor" responded negatively, while a negative response was received by only 32% of the "rich". Likewise, 31% of the "rich" responded positively, compared to only 14% of the "poor"., indicating what appears to be a growing gap between rich and poor. Table 1.11: Retrospetive Perception of the Household Situation in the Last Five Years TOTAL By SEX Socio-Economic Level Men Women High Middle Low much better 2.3 3.6 1.2 7.3 3.1 1.0 Better 15.3 18.0 13.0 23.5 17.8 12.5 Equal 35.3 38.3 32.5 36.8 34.5 35.4 Worse 37.5 32.2 42.2 28.4 35.0 40.6 much worse 8.4 6.5 10.0 3.3 8.9 8.9 No answer/don't know 1.2 1.4 1.1 0.6 0.8 1.6 Of interest also, are the results by sex. While 39% of the men gave negative answers (worse, much worse), women answered negatively 52% of the time. This would suggest that any worsening of the situation of the poor impacted more on women than on men, or at least, they feel more strongly about the impacts. How does this compare with the EPH survey results? Looking at a comparison of October 1998 with October 1993 (the closest approximation to the five year gap of the perception survey), we find average income of every decile of the income distribution 68 declined, except for the very top decile. Roughly then, the survey results should say that the situation for 90% of the population was "worse off', compared to 46% in the perception survey. Why is there this gap between perceptions and measured income? It may be that respondents place greater weight on the last 2-3 years when average incomes rose, and are not strictly comparing today's situation with their situation five years ago (which would be before the downturn in the economy in 1995). For instance, if one compares the average income by decile between 1996-98, all deciles except the lowest have improved (i.e. 90% of the people are better off). Alternatively, the EPH may not be measuring all sources and types of income. Or this perception may refer to non-income factors flagged in the focus groups such as increased insecurity, lack of trust, increasing disintegration and little faith in external agents leading to increasing despair and a felling of hopelessness ( Voices of the Poor, 1999). Respondents were also asked to identify institutions according to their capacity and competence to respond to the problem of poverty. Each respondent listed three institutions, according to first, second or third place priority. We then constructed a weighted average of these responses to be able to rank the institutions. 17 The results were as follows: National Government 15.7 CARITAS 15.2 Catholic Church 13.7 Provincial Govt. 11.3 Non-Govt. Organizations 8.3 Municipal Govt. 8.1 Political Parties 5.9 International Organizations 4.8 Unions 4.1 Clearly the two most effective organizations in combating poverty are perceived to be the national government and the Catholic Church, including its subsidiary CARITAS. If one were to consider CARITAS and the Catholic Church as one organization, it would clearly outrank the national government. Despite decentralization, the role of provincial governments, and especially municipal governments, is clearly seen as much lower. Even further down the scale are such organizations as unions, political parties and international agencies. The national government, which has a broad range of social programs, is seen by the poor as a relatively more effective organization (although one should note that only 16% of respondents felt is was effective). However, the qualitative work carried out in poor communities tells a quite different story. While the role of government is seen by the poor as vital in combating poverty, public institutions are generally seen as inefficient, unresponsive and corrupt. This is true for both central and local governments and, most of all, for the police. The police are not seen as helping but as part of the problem acting in many cases as perpetrators of violence and public insecurity. Institutions are valued by the poor in terms of their ability to resolve real problems, access, to deliver what they promise and to do so without bureaucracy. While the independence of some non governmental organizations and community based organizations is questioned, participants in the focus groups pointed to neighborhood associations, nursery schools and the Church as the most effective institutions. Primary 17 The weights were: ISt place =3, second place=2, third place =1 69 schools and health posts also came up high as they are accessible and provide a safe and friendly environment. From the focus group discussions one receives a general feeling of hopelessness and despair among the extreme poor (see Box)18. They feel cut off from national government institutions and taken advantage of by politicians. In urban areas, the lack of employment and problems of security and violence are paramount. Growing unemployment has impacted on both men and women. Men seem more likely to be unemployed as women are more easily absorbed into the service sector particularly as domestic helps. Unemployment among men affects their self esteem as they feel devalued in their communities and lose authority and pride within their families. High unemployment amongst men is seen as promoting high levels of alcohol abuse and domestic violence of which women and children are most often the victims. As women in urban areas enter the informal sector and continue to carry the responsibility for domestic chores and childcare, their load is increased. This may explain why more women feel as they are worse off now than they were 5 years ago. Poor women in urban areas, however, often earn more than their male partners and this is often seen as the source of tension within families. Among the young, the use of illicit drugs is a growing problem. Young people have little hope of being able to find a job upon finishing school and despair and resignation to an undesirable future in the countries youth is high. In rural areas, poverty is linked with the lack of access to land, to markets and to credit, as well as employment opportunities. Unemployment is also viewed by the rural poor as contributing to new rural violence which is manifested as robbery, drugs, and an increasing number of rapes. In both rural and urban areas the poor look to their own organizations and community interactions as a survival mechanism. It is not clear whether these informal organizations have increased or decreased as a result of the "crisis". In urban areas particularly social networks are seen as providing some cohesion and feeling of belonging to counteract an increasing social disintegration. Both rural and urban poor point to problems related to poor infrastructure: the lack of water and sanitation facilities, street lights, roads, land titles, and poor housing. Conclusions Despite a resumption of economic growth during the decade of the 1990's, poverty has remained relatively high, and income distribution has worsened. These trends seem to be related to a number of factors, including increased insecurity as both unemployment and violence rise in urban areas and the rural poor find it more difficult to meet their basic needs naming access to land and credit as major problems. Poor families are characterized by low levels of education, high fertility and therefore large families, high levels of unemployment and a higher degree of employment in the informal sector. In the eyes of the poor, employment is a critical factor in their poverty, and although they clearly believe that the national government has an important role in poverty alleviation programs they are disappointed by the level of coverage, unresponsiveness and corruption in public services. In succeeding chapters we examine how labor market conditions interact with poverty (Chapter II), how government social programs affect poverty and income distribution (Chapter III), how programs in education (Chapter IV) 18 For more details of the focus group discussions, see "Argentina: Consultations with the Poor" paper prepared for the global Synthesis Workshop, Sept. 22-23, 1999, Poverty Group, PREM, World Bank. 70 and health (Chapter V) can reduce poverty, and the particular problems of poverty in rural areas (Chapter IV). 71 Box 2 VOICES OF THE POOR: INSIGHTS FROM THE FOCUS GROUPS Perceptions of Well Being * The lack of employment is seen as a key problem in urban and rural areas. The lack of alternative employment and the thought of being "barely able to meet one's needs" at the same time leads to a feeling that the cycle of rural poverty is inevitable. * Argentineans view themselves as being poor if they do not have enough to meet their basic needs. Being able to provide for themselves and their families with food, health care, basic infrastructure and some level of education in a self sufficient manner is considered as the minimal definition of well being. * The only solution perceived by both the rural and urban poor is education. Education is seen as a way of breaking a cycle of poverty and offering hope to the younger generation. However, an education capable of breaking a cycle of poverty is seen to have costs: "A good quality of life for us would be to have a good education. For the young, there is the possibility that we can acquire the minimum resources for them to be able to study. " - In urban areas, crime and insecurity has grown making people feel more vulnerable and creating a sense of insecurity about the future and their children's future. * Growing inequality on the one hand and an increased social pressure to keep up with a society that is becoming more consumer-orientated on the other hand is also given as a reason for feeling worse off: "...the point is, the more you have, the more you need; my kids come from school and tell me that they want certain brand-name shoes, clothing... it's a consumer society... it's a great contradiction" * Low social capital and solidarity leads to perceived loss of collective values and a need to act individualistically. This is also expressed as increasing isolation and despair: "There are ... people that take advantage... you also can see it in the participation, it's not there... common peoples'worries don't interest them. " • In this context and with the increasing diversity in the population with immigrant influxes, lack of unity, social cohesiveness and harnony is expressed as another reason for further concern. * The consequences of this worsening situation are expressed as a general uncertainty and fear about what the future holds and an increasing sense of despair. In this situation, maintaining spiritual values are seen to be the only hope in a daily situation characterized by a fierce pressure to "keep up" and, - when that inevitably fails - increasing family and social violence and a worsening of health and a growth in addictions and psychologically-related disorders * In rural areas, the lack of access to land, credit and other resources to increase productivity is expressed as the principal concem among the rural poor. The support that the government is perceived to give to new large scale producers is also cited as a reason for growing despair. * Geographical isolation is also cited as a reason for the rural poor being left out of development. Local government is seen as not only distant in geographical terms but unresponsive and uninterested, despite the efforts made by citizens to interact with their governments: "Around here, when people need something, they go to the municipality, knock on the door but no one sees them because there isn 't anyone there or they have just gone out ... they '11 come back tomorrow, maybe the day after..." 72 Similar charges are made about providers of public services such as doctors and public servants. In addition, the rural poor feel that they are often treated with prejudice by public servants. Perception of Institutions and development actors * Institutions are valued by their ability to meet the poor's immediate needs. There is a sweeping and worrying lack of credibility in most public institutions with few exceptions. In particular, the government at all levels and the police are singled out as corrupt, unresponsive and ineffective institutions: * Government at all levels is seen to be corrupt, unresponsive and inefficient. While a strong role for government is seen as being central to development, local, provincial and central governments are seen as being unconnected and, after having won their vote, unconcerned with the poor, "The governor is there because we put him there, we were the ones who voted him in, we're the ones who suffered the heat, the cold, we went through so much to vote so they can get there and today they don't remember us... they take their position andforget about us (the peasants). " "the politicians cheat us, it's a corruptness, it's a deception; before they're taught to talk, they're taught how to make the easy money... " * Government development interventions are also perceived as superficial and compensatory- rather than effectively attacking the root cause of poverty: "The government ... doesn't do anything, for those young people in trouble. there reformatories and crowded places instead of giving them a place for education, a profession... " * The police are also singled out as an ineffective, discriminating and corrupt institution. The negative image of the police is expressed as an inability to carry out their function on the one hand and their role as a perpetrator of abuse and violence on the other. 73 Appendix I: Poverty Lines in Argentina Poverty lines are defined as the income/consumption thresholds that would allow a typical Argentine to enjoy a consistent level of welfare, via satisfaction of a minimun level of food and non-food consumption needs. They draw on the concept of absolute poverty that is consistent with an individual's achievement of constant level of material well-being. Current Poverty Lines Current official poverty lines have been in effect about eleven (11) years since 1988. While debates about their revision are under way in Argentina, definitive consensus is yet to emerge. In the meantime, the current poverty line provides a point of departure for constructing consistent poverty lines utilized for this study. Current poverty lines are based on the data from 1985-86 Income and Expenditure Survey for the Metropolitan Buenos Aires (MBA). With the recommendation by the 1981 Consultative Meeting of FAO/OMS/UNU (1985) as the basic frame of reference, the basic food basket for the Argentine poverty lines was estimated by studying the consumption pattern of households in the 2-4 income deciles in the Metropolitan Buenos Aires. The value of the basic food basket that would allow a representative adult male of age 30-59 with moderate activity level to achieve a daily intake of 2,700 Calories, namely the adult equivalent indigent line, was estimated to be 149.2 Australes, with the corresponding poverty line given by 308.9 Australes, per month each in March 1988. Current poverty lines have two major short-comings as a measure of consistent poverty threshold. First, while indexing of the indigent lines by the price of basic food basket appears methodologically sound, allowing accurate tracing of the purchasing power of the extremely poor over time, the indexing of the non- food component is problematic. Since the non-food CPI index includes luxury article as well as basic necessities, the current poverty lines have a tendency to overstate the consumption of non-foods by the poor significantly over time. Second, the current poverty lines are available only for the MBA. Since the 1985- 86 Income and Expenditure Survey was carried out in the MBA alone, there was no statistical base upon which to construct the regional price index to adjust for the cost-of-living differences across regions within Argentina, thereby limiting their usefulness for regional poverty comparison. Consistent Poverty Lines Construction of poverty lines utilized by this study has been motivated by the need for absolute poverty lines that are time-invariant and location-invariant with respect to welfare. They have been elaborated by combining the current official poverty lines with statistical information which has become available recently. In order to ensure consistency of inter-personal welfare comparison over time and at different locations, poverty lines have been elaborated by taking into account: (i) changes over time in prices of consumption items faced by the poor, and (ii) differences in cost of living at different locations. Defining the non-food CPI for the poor In order to rectify the upward bias in poverty lines imparted upon by inclusion of luxury items in the non-food CPI, this study made an independent estimate of the price index of the non-food commodity bundle that would reflect the consumption pattern of the poor, namely the non-food price index for the poor. The methodology is described in detail in what follows. INDEC provided the project team with the set of price indices for 139 commodity sub-groups, 58 sub- groups for food items and 81 for non-foods (40.1% for foods, and 59.9% for non-foods). From non-food items of non-foods, 19 items (19.5% weights) have been identified to have the characteristics of luxury goods and services with higher income elasticity, e.g., formal education (1.77%), domestic service (2.13%), purchase and maintenance of automobile (7.399%), tourism and hotels (1.77%). Construction of has been performed by recalculating the price index of the non-foods using the weights of the 2nd income decile after purging these luxury items from the non-food CPI bundle. 74 Regional poverty lines: accounting for cost-of-living difference Elaboration of regional poverty lines is based on an analysis of the 1996-97 National Household Expenditure Survey (ENGHO) which has the national urban coverage (84% of all Argentine population). The reference population group for the national poverty lines falls 10% range between 16-35% of households by income level, which corresponds to consumption of 2,750 Calories of daily food intake by adult male. The indigent line for the MBA was computed by applying the unit price from the CPI data from INDEC to the basic food basket from the survey data. As for regional indigent lines, those data subsets from the national reference group which correspond to respective regions were analyzed and adjusted proportionally to allow each regional reference group to achieve consumption of 2,750 Calories. Once the value of the respective regional basic food basket was obtained, the coefficients of the pair-wise difference of other regional baskets from that of the MBA have been computed, and applied to the MBA' indigent line to obtain the set of indigent lines for five other regions. Difference in the regional cost-of- living implied by the regional data, along with the associated Engel coefficients, provided the set of basic parameters to define regional poverty lines. Putting them together For households in the MBA, the indigent lines and the non-food component of the poverty lines have been updated by the price index of basic food basket and the non-food price index for the poor, respectively. In September 1996, the adult equivalent indigent line was 67.4 Pesos, and the corresponding poverty line was given by 155.6 Pesos, with the food share of 0.433. In some cases, poverty lines are calculated taking into account economies of scale in consumption; that is, larger families need lower levels of per capita consumption to maintain the same welfare level. Using straight per capita income calculations can overstate the poverty rate among large families. We have not introduced scale economies here, but instead adjust family size into adult equivalence (see Background Paper No. 1 for the equivalence scales used). This adjustment overcomes much of the bias against large families found in standard per capita consumption or income calculations. As for poverty lines for other regions, the following procedures have been performed. The indigent line for the NBA in September 1996 served as the welfare anchor for regional poverty lines. With 67.4 Pesos as the anchor, the coefficients of difference of regional indigent lines from the MBA's indigent line in the 1996/97 ENGHO data have been applied to the MBA's indigent line to obtain the corresponding indigent lines for the regions. While the Engel coefficients for the MBA have been estimated by indexing of poverty lines, the food shares in other regions have been estimated using the ENGHO data. Finally, it would be necessary to update regional poverty lines by appropriate regional price indices in order to carry out regional poverty comparison over time. Since no regional price index system exist outside the MBA in Argentina, the same methodology that was applied to indexing of poverty line in the MBA has been applied to obtain a time-series of regional poverty lines. Box Table 1.1 reports the indigent lines and poverty lines for September 1996, and Annex Tables 5-8 presents a full set of time-series data. Appendix Table 1.1: Adult Equivalent Poverty Lines and Per Capita Household Incomes, Sept. 1996 (pesos per month) Indigent Non-Foods Poverty Implicit Per Capita I _ Lines Engel Consumption (September 1996) EPH EDS ENGHO ENGHO Overall 292.4 284.8 292.6 240.2 Metro Buenos Aires 67.4 88.2 155.6 0.433 342.2 367.9 348.8 292.4 North West 66.0 82.5 148.5 0.444 191.6 208.1 197.6 172.6 North East 63.8 75.3 139.1 0.459 183.0 201.5 199.4 159.5 Cuyo 61.0 85.4 146.4 0.417 242.0 258.2 237.3 196.7 Painpeana 64.9 74.6 139.4 0.465 254.3 260.1 280.2 222.5 Patagonia 64.4 86.9 151.3 0.426 339.4 355.1 310.5 236.0 Sources: INDEC, Encuesta Permanente de Hogares, October 1997. SIFMPRO/ INDEC, Encuesta de Desarrollo Social, August 1997. INDEC, Encuesta Nacional de Ingresos y Gastos, February 1996-February 1997. 75 CHAPTER II. LABOR MARKETS, EMPLOYMENT AND POVERTY Introduction There are strong linkages between poverty status, and position in the labor market, the most obvious being through wages, unemployment, and type of employment (formal or informal). Measures of these aggregates have shown sharp movement over the last decade reflecting long term trends, short term macro-disturbances, and adjustments to structural reforms. Figure 2.1 suggests very tight relationship between the principal labor market aggregates and the moderate poverty headcount index. Figure 2.1: Poverty, Unem ploym ent, Income 4 5 -14.0 8^e 40- -12.0 E QC V6L 4,35- \ E 10.0 '25 t_ L 8.0 a, 20 6.0 i0 ~~~~ 15 - ~~~~~~~~~~~~~~~~~~4.0 4 4) 5 4, 0 0.0 90 91 92 93 94 95 96 97 98 - Unemployment -Poverty - X- Household Income ---mReal Wage * Per capita household income, on which the headcount index is based, is highly correlated with the mean wage of the household head across the period. * After 1994, unemployment also suggests a more direct influence on the poverty rate. Table 2.1 suggests that those appearing to be indigent have the highest rates of unemployment with the rate decreasing with income. This is confirmed by probit analysis that suggests that an individual is 17% more likely to be poor if the head of the household is unemployed. 20 Further, analysis of income transition matrices show that both the probability of falling into poverty, and the length of time an individual is found in poverty are positively related to involuntary separation from previous jobs. 21 20 H. Lee, Background Paper No. 1; poverty defined by income per capita being below the poverty line. 21 C. Arango and W. Maloney Background Paper No. 3. Poverty here based on the worker's income being in the first quintile. 76 * Statistically, the link between unemployment and poverty can be deceptive and more fundamentally suggests the need to use consumption data instead of the income data in calculating poverty indices. Table 2.1 suggests that the preponderance of reported (incomplete) spells are very short 45% being under -two months. Income measured during this spell will be low even though consumption is likely to be much higher. Panel data (see Appendix B) traces individual movements among income quintiles over a two year period and suggests considerable movement in and out of poverty. Nineteen percent of those unemployed and earning zero income, and 26% of those in the bottom income quintile in the first period were found to have moved into the top two quintiles in the next two years. That is, a quarter of those at the bottom of the distribution at any time are probably consuming at much higher levels. In fact, if income over the four periods observed over two years is averaged, only a third of those whose first period income would put them in poverty would remain statistically 22 poor. Using a measure of income that averages across the four semesters that each worker is followed leads to a 7% fall in the Gini. Table 2.1: Labor Market Characteristics of Urban Population by Poverty Group, EPH 1998 Poverty Group The Indigent The Poor Moderate The Non- Total Poor Poor Labor Force Participation (Age 15-64) 54.8 56.4 67.3 64.5 Female Participation 37.4 35.5 53.4 49.0 Unemployment Rate 36.5 22.6 8.9 12.6 Among the Unemploved: Unemployment Spell (%) 100.0 100.0 100.0 100.0 Under 2 months 45.5 46.8 42.8 44.3 2 - 6 months 33.6 29.4 26.3 28.2 7 months- I year 13.3 18.8 23.0 20.5 Morethan I year 7.6 5.0 7.9 7.0 Among the Employed Number of Jobs Held 1.03 1.04 1.10 1.09 Total Hours Worked per Week 34.9 41.7 45.1 44.2 % Employed in the Informal Sector 46.4 44.6 36.2 38.2 Sources: INDEC, Encuesta Permanente de Hogares, average of May and October of 1998, except for unemployment spell, which is October 1997. * The link between unemployment and poverty is further complicated by the fact that figure 2.2 suggests that much of the unemployment is concentrated among the young, many of whom may live at home. A rise in youth unemployment may not obviously be linked to household poverty. 22 Based on averaging non-zero income. 77 The fact remains, however, 30 Figure 2.2: Percent of Males Unemployed At Some Time During Two Year Period that whether poverty figures are affected through open unemployment, or through lower 20 incomes, both phenomena are driven by the demand and supply conditions in the market which 10 have proved adverse over the last - < . five years. L 0 16 22 27 32 37 42 47 52 57 65 Looking at the employment of Age the poor by sector, we see that the poor are about as equally distributed between manufacturing, services and trade as the non-poor (see Table 2.2). Since 1990, the employment composition of the poor has shifted much more toward trade and construction, and away from manufacturing and other services. Substantial differences in the employment patern of the poor and non- poor occur in two sectors: construction, which employs 24% of the poor but only 10% of the non poor, and finance, which employs 10% of the non-poor but only 3.5% of the poor. Among the indigent, the tendency to work in construction is even stronger; 32% work in that sector. The concentration of the poor and indigent in the construction sector adds to their vulnerability. As might be expected, the variance of the growth of construction is much higher than other sectors. As shown in Table 2.4 below, the variation (standard deviation) of the growth rate of construction is almost three times higher than the variation for total GDP, and the variance of employment is more than four times greater. However, its is also true that construction has been one of the most dynamic sectors since the stabilization of the economy, and employment growth in this sector has provided jobs for the poor. The employment elasticity of construction, at 1.17 during 1990-98 is substantially higher than the employment elasticity for the economy as a whole. Table 2.2 Employment Distribution by Sector, 1990-98 (% of total) 1990 1998 Sector Indigent Poor Non- Indigent Poor Non-Poor Poor Agriculture 1.5 .8 .8 1.9 1.4 1.8 Manufacturing 16.3 18.3 17.6 9.6 15.6 15.7 Construction 10.9 6.3 2.9 23.4 18.2 6.8 Trade 18.7 15.9 13.4 17.0 18.4 19.1 Finance 3.4 5.9 8.6 3.8 3.7 10.4 Services 49.2 52.8 56.7 44.3 42.7 46.2 Total 100 100 100 100 100 100 Source: EPH. Services includes "unknown". Determinants of the Evolution of Wages and Unemployment Unemployment has shown a secular increase since about 1980 and does not show an obvious tendency back to the levels of 2-4% of the early 1980s. The rate of increase accelerated in about 1993 with a peak 18.4% in the wake of the Tequila crisis. This 78 particular peak, as with the rise to 14.5% occurring after the recent Brazil crisis, reflects temporary disequilibria that has probably delayed the adjustment to the new low inflation/open economy regime. However, the rise since the 1980s and in particular that which began in 1992 suggest longer term demographic and economic factors. * On the supply side, population growth averaged about 1.6%, and participation rates also rose. Even among adults Table 2.3 Female Labor Force aged 15-64, the rate is about 3 Participation Rates percentage points higher in 1998 (urban areas) compared to 1991, in part Indigent Poor Non-Poor Total because of a higher participation rate among women. Female 1985 21.1 20.4 39.9 37.4 participation rates have risen 1988 23.6 28.0 47.6 42.3 significantly for all economic 1990 31.8 32.0 49.5 42.9 groups since 1985 (see 1992 30.6 28.4 48.6 44.6 Table2.3), and have risen even 1994 33.4 31.8 49.8 46.6 more rapidly for those who are 1995 36.4 37.4 52.3 48.8 poor and indigent. Overall, 49% 1996 36.7 37.4 51.0 47.4 of all women in Argentina work, 1997 39.4 36.6 53.2 49.1 compared to 37% in 1985. 1998 37.4 35.5 53.4 49.0 Guasch estimates that this could Source: EPH various years. Data for 1985-88 is account for roughly 25% of the for Buenos Aires only. higher unemployment observed (Guasch, 1998). However, it also the case that employment growth has been distinctly lackluster. Since GDP fell by five percent across the decade of the 1980's the secular rise in unemployment to 9% is perhaps not surprising given normal growth of the workforce. As can be seen in Table 2.4, employment in certain sectors actually declined during the decade, particularly in manufacturing, construction and trade. However, resumption of growth during the 1990s was marked by a generally low elasticity of labor demand. With an overall growth rate between 1990-98 of 4.5 percent per annum, the growth rate of employment was only 2.2, or an elasticity of about .48. It is interesting to note that employment continued to decline in manufacturing, despite positive growth in the sector, suggesting that adjustments being undertaken during the period reflected capital intensive modernization. Overall, it is difficult to know whether this represents a secular trend or simply the transitional costs of adjusting to a new regime. A similar low rate of absorption is clear in Mexico during their trade liberalization, and Chile experienced a mean unemployment rate of 17.1% during its period of reform, 1974-1989, yet from 1986 on generated employment at rapid rates. 79 Table 2.4 Growth of Output and Em loyment, 1990-98 sector Growth Rate of Growth Rate Growth Employment Standard Standard Employment, of Rate of elasticity, Deviation Deviation of 1980-90 Employment, Value 1990-98 of Growth Employment (Buenos Aires 1990-98 Added, only) 1990-98 . Agriculture 8.5 1.6 2.1 .80 5.1 6.6 Manufacturing -0.7 -1.0 4.2 -.23 5.2 6.2 Construction -7.3 11.2 9.6 1.17 15.2 13.0 Trade -1.1 4.2 4.8 .87 6.1 6.8 Finance 4.2 2.9 6.3 .47 6.5 5.8 Services 3.8 1.2 4.2 .28 4.9 3.9 Total 0.9 2.2 4.5 .48 4.8 2.7 Note: Growth rates are based on end points; standard deviations are from annual averages. Several factors mnay claim partial responsibility for this low growth. - Privatization has led to roughly 300, 000 workers being shed from formal sector jobs. - The stabilization program may have had adverse short and long term effects that offset the positive impact on confidence and investment. Inflationary Inertia: In the short run, any inflationary inertia due to incomplete credibility or indexing may lead, with a fixed nominal exchange rate, to a real appreciation, and hence a lower demand for labor in the tradables sector. This may lead to more workers forced into unemployment or sub-employment, perhaps in the informal sector. Decreased Flexibility. The inability to decrease real wages by lagging nominal wages behind inflation makes it difficult for firms to respond to shocks through wages rather than employment (see Gonzalez, 1998). Since firing is extremely costly, firms are likely to be more reluctant to hire new workers. * Trade liberalization has led many private sector firms to close or radically restructure their operations. The direct implication is that, again, many workers are shed and unemployment may rise over the medium term before jobs are created in newly emerging sectors. Over the longer term, however, to the degree that increased competition leads to higher product price elasticities, this also leads to higher own wage labor demand elasticities. This implies both greater employment or wage response to shocks, but also implies that, at given levels of labor costs, less labor will be hired. Technological progress, and easier access to skilled labor augmenting capital have been argued to have increased the relative demand for skilled (educated) relative to unskilled labor. Table 2.5 suggests that, at first glance, something similar is happening in Argentina and that the real wage series in figure 2.1 are understating the true income losses of the more disadvantaged families. In particular, what emerges is that developments in the non-tradable informal sector seem to be driving much of the movement in unemployment and wages. 80 Table 2.5: Real Wage Growth (%) 97/90 97/93 Education Level Primary 2.3 -17.5 Secondary 4.4 -14.2 College 13.5 -8.6 Skill Level Unskilled -4.9 -17.2 Skilled 9.9 -12.4 Professional 40.5 -1.9 All 8.0 -12.2 Table 2.6: Male Worker Transitions Among Sectors of the Labor Market Across 2 Years Argentinean National Urban Employment Survey, 1993-1997 Probability of moving from initial to final sector, Pij, in percent Initial Sector OLF School Unemp Unpaid SE IS FS Other Total Informal Defined as Unprotected Out of Labor Force 1 10 1 6 4 5 0 100 School 4 1'2010110 20 0 3 7 9 0 100 Unemployed 11 4 3t 1 15 17 16 0 100 Unpaid 15 0~~ 8 3 15 8 8 100 Self Employed 4 1 10 0 j86 10 6 2 100 Informal Salaried 4 3 14 0 1 i 12 0 100 Formal Salaried 3 0 8 0 100 Other 0 0 2 0 32 2 4 100 Total (Pj) 11 6 14 0 24 13 30 2 100 A closer look at the Argentine Labor Force Table 2.6 shows both the relative sizes and mobility patterns of the various different categories of male workers across a two year period drawn from seven combined panels spanning 1993-1998. It shows the probability that a worker found in one of the sectors listed in the first column will be found in another sector after two years. The bottom row gives the share of each category in the total labor force (Pj). The table is useful both for categorizing workers, but also for studying the interactions within the labor market and, particularly, in unemployment. (for data on female workers, see Appendix A). Those working are divided into unpaid workers, formal salaried workers and the informal sector which accounts for 55% of paid workers. This sector is in turn divided into own-account or self-employed workers in firms of fewer than 16 employed who are not covered by social benefits, and those working for these firms, informal wage earners. 81 The informal merit special focus for several reasons. * They tend to have lower income that better educated formal workers and as Table 1 suggests, are more heavily represented among the poor. * The sector is often thought to represent "precarious" and unprotected labor. Since, the informal are, by definition, not covered by social security or other social nets and recent government initiatives, such as Proempleo see formalizing workers as a central goal of their mandate. * Dualistic conceptions of the informal sector see it as the disadvantaged sector of a market segmented by labor market rigidities. The formal sector pays high wages and benefits, and the informal is a low wage residual sector employing those rationed out of the formal sector.25 In this view, a large informal sector is suggestive of rigid labor law and its constituents are particularly at risk. * In many countries the informal sector is seen as a providing part of the safety net for unemployed formal sector workers and hence documenting its relationship to the formal sector and to unemployment is important to understanding the kind of safety net to be designed. * Finally, looking at the relative behavior of wages and relative size in the different sectors sheds light on who gained and lost over the last years and hence, which sectors were driving the adverse movements in poverty. Figure 2.3 shows the median wage rates, and the share of the work force in formal work, informal salaried work, and informal self-employment across the 1992-97 period. Establishing segmentation or the relative desirability of the sectors by comparing the levels of wages is made impossible by unobserved job related characteristics such as the value of benefits foregone, taxes evaded, independence, risk carried, informal training received or, for family workers, payments in kind. Hence the higher earnings of the informal self-employed do not imply that formal jobs are inferior residual class, nor do the lower earnings of the informal salaried workers.26 27 Informal workers are the 25 For an early presentation of this view see J. R. Harris, and M. P. Todaro (1970), AMigration, Unemployment, and Development: A Two-Sector Analysis, American Economic Review, 60:1, 126-142., For an overview of theory on informality, see Gary Fields (1990), "Labor Market Modeling and the Urban Informal Sector: Theory and Evidence," in OECD, The Informal Sector Revisited, Paris. 26 In a competitive system, with flexible wages, benefits given to formal sector employees should result in a relative decline in their wages, reflecting the value of their non-wage benefits. Thus, it might be possible to find higher wages among informal workers. However, if workers discount the value of the benefits, and see formal sector employment as costly in terms of taxes, wages in the informal sector could be lower, without indicating segmentation of the labor market. In addition, informal sector work may offer greater flexibility and independence. See W. Maloney (1999), "Self-Employment and Labor Turnover: Cross Sectional Evidence," IBRD Working Paper #2102 27 However, wage differentials may be related to other factors, such as age, experience, and schooling. Using panel data from the 1993-97 period27, we estimate that a move from informal salaried work to formal salaried work (for men) was accompanied by a 7.8% increase in wages, while a move in the other direction 82 youngest of any class by 8 years, many live at home and may have apprenticeship positions, both of which may imply lower wages in addition to being able to avoid taxes. At a more profound level, the inability to quantify these unobservables implies that raw labor incomes are very imperfect measures of relative welfare. What is more informative is a comparison of the relative wage movements and relative sector sizes over time. It is important also to distinguish between informal salaries workers and informal self-employed. From 1988 to 1993, the rise in the size of the self-employed in the informal sector relative to the formal salaried occurs concomitant with a rise in relative returns to self-employment, exactly the reverse of what would be expected if the sector served primarily as a safety net. In fact, in Argentina, as in Mexico, it is probably better to view the informnal self-employed prirnarily as dynamic entrepreneurs who prefer to be self-employed, an interpretation borne out by interviews with Argentine workers. A small survey in Jujuy revealed that 80% of the self-employed had no desire to change jobs and under 18% saw self-employment as a temporary activity 28 before they found a "real" job. In Gran Buenos Aires, a SIEMPRO/FONCAP survey found that while 36% would have preferred to work more hours, only 26% were looking for other work. 29 Both are consistent with an "entrepreneurial view" of the sector. 3 Multinomial logic regressions show evidence of a life cycle pattern where workers, up to a certain age appear to accumulate human and financial capital and open their own business. 31 Figure 2.3: Sectoral Participation and M edian Real W ages in Argentina, 1988-1997 70.0% 14 13 6 0.01% 12 50.0% % - - - - - - '% S m | 2 0.0 % 4 Hm 28~~ ~~~ Miroeteprs Std for7 Juuy Cosloi odst,Fnod aialScaFNA) 10.0%: 2 0.0% 0 1988 1991 1993 1995 1997 30,Mln% F f rmoa I Informal Sal. r t Self-Employed TKio W Formal epomnW Informal e na o r nw Self-Em, I ed resulted in a 16% drop. However, the standard deviations of these medians is so high, that the results are not statistically different from zero. Indeed, results for women are radically different than those for men. 28 Micro-enter-prise Study for Jujuy, Consultoria Nordeste, Fondo de Capital Social (FONCAP). 29 SIEMPRO/FONCAP (1 998) "Perfiles de la Microempresa". 30 W. Maloney (1999) finds that, among the Mexican self employed, 70% report being voluntarily in the sector. 31Transitions from formal employment to informal employment are grow more frequent with age, as might be imagined if they were accumnulating human and financial capital to start their own business. Though it may be argued that this simply represents laid-off older workers being forced into informality, the results suggest that after 24 years of experience, additional experience makes a worker less likely to move. 83 The apparent contradiction of desiring to be "unprotected" is understandable when the tax burden that implicitly finances such benefits are taken into account, paid either explicitly, or simply as lower wages. Inefficiently provided benefits -poor medical services, a social security system seen as bankrupt and unreliable- provide incentives for workers to work off the books. In addition to avoiding non-labor taxes and regulation, informal workers may be avoiding inefficient "protection." Rather than a traditional dualistic view of the relationship between the formal and informal sector, it is probably better to imagine two sectors, one an unregulated (informal) sector concentrated in non-tradeables, and the other a more regulated tradeables sector. Both may constitute desirable destinations for workers and both can experience positive and negative shocks that drive aggregate output and employment. Within this framework, we can see the boom in commerce, transport, restaurants and construction32 occurring in the wake of the stabilization as a positive demand shock to the informal non-tradeables sector, perhaps due to increased optimnism about future income. This view is consistent with movements in the real exchange rate that track the relative self-employed/formal wage quite closely. This also suggests that the appreciation of the peso seen after the stabilization plan was, at least in part, "equilibrium" in nature and not entirely due to inflation inertia. As a general lesson, whether appreciating currencies hurt the poor or not depends greatly on the source of the overvaluation. If inflationary inertia leads to the tradables sector shedding workers into marginal positions in the informal sector, poverty could worsen. If, as seemed to be the case here until 1993, it is driven by a sector intensive in informal workers, poverty indicators may improve, as they did. Post 1993 evolution of wages, unemployment After 1993, sharp movements in wages, unemployment and sectoral composition, suggest a different story and some explanation for the deterioration in the poverty numbers. The procyclical movement Table 2.7: Sector of Origin of the Unemployed of informality is less apparent. Sector of Origin All Paid There is a sharp decline in the ___________ share of the work force in Informal Self-Employed 18 38 informal self-employment by Salaried Informal 12 26 1997, at a time when the Formal Salaried 17 36 premium of informal self Previously Unemployed 34 employment wages over formal Previously Out of Labor Force 6 sector wages was narrowing School Graduates 13 (see Fig. 2.3). As self Total 100% 100 employment became less attractive, workers moved to other sectors. However, at the same time the gap between informal salaried wages and formal sector salaries was widening, and accompanied by a 32 see Guasch 1999 84 large rise in the size of the informal salaried worker sector. Thus, informal salaried employment appears to be countercyclical, as more dualistic models would predict. Most of the rise in unemployment of previously employed workers appears to originate in the informal sector. Table 2.6 suggests that the unemployed do move disproportionately into the informal sector. When standardized, the movements show a value for movement into informal salaried 2.5 times that into formal salaried work. Further, those who leave school move disproportionately into informal salaried work. An interesting finding is that informal workers also enter unemployment at much higher rates than formal salaried workers. In fact, the informal are the largest contributors to unemployment. Table 2.7 inverts one row of Table 2.6 and tabulates the sector of origin of those found unemployed. The vast majority of those previously employed in paid work who move into unemployment are the informal, 64%, compared with 36% from the formal sector. When adjusted by the size of the initial sector, the largest contributors per worker are the informal salaried followed by the self-employed and then the formal salaried. TABLE 2.8: SELF-EMPLOYMENT AND TURNOVER RATES Argentina Brazil Mexico LAC OECD % Workforce in Informal Self- 28.5 23.2 26.5 31.5 12.9 Employment Average Tenure 8.9 6.25 5.8* 7.6 10.5 years/employee, Manufactures) *Generated from firm level survey so not strictly comparable to the other figures based on household survey. Part of this phenomenon is due to the higher rates of turnover in these sectors. Small enterprises everywhere have very high mortality rates and hence both entrepreneurs and their employees may find themselves without work more often than formal employees. In addition, the informal salaried, as elsewhere in the region, are among the youngest, frequently enter from school, perhaps serve as apprentices, and then move on to other jobs quickly. Though turnover of these workers is lower than in Mexico or Brazil, the basic demographics of the sector very similar. (see Table 2.8) Nonetheless, the trend is toward more unemployed from the informal sector. While unemployment for the formal sector rises (4 to 8.8%) among the informal salaried it triples to (6.1-18.3%) and only jumps slightly less from self-employment (7.4 to 18%). What seem clear is that the informal sector is not primarily serving or able to serve as a safety net. Wage gains and losses are extreme and widely differ by sector, education and experience. (Table 2.9) 1. The big gainers over the last decade have been the informal salaried with a rise in 43.7% compared to 25% for the formal salaried and 14.5% for the self - employed. The gains have been particularly concentrated among those with primary education and with under 5 years of experience (over 50%). 85 2. That said, Table 2.9 show that the decline in wages observed in figure 2.3 is virtually entirely borne by the informal sector while formal sector wages remain fairly rigid downward. 3. Despite the fact that, in most classes of experience, the informal salaried outperformed their formal salaried counterparts, there are two notable exceptions: Table 2.9: Median Real Wage Changes of Male Workers, Buenos Aires By Sector of Employment, 1988-97 ( in %) Formal Informal Salaried Informal Self- Employed 93/88 97/93 97/88 93/88 97/93 97/88 93/88 97/93 97/88 Primal 39.3 -5.7 31.4 69.4 -11.0 50.8 31.5 -18.0 7.9 Prima2 37.5 -3.4 32.8 72.0 -9.8 55.1 43.1 -14.5 22.4 Prima3 17.5 4.0 22.2 57.8 -36.1 0.8 62.0 -20.8 28.3 Secunl 32.3 -16.5 10.4 41.2 -10.5 26.3 28.9 -40.9 -23.8 Secun2 15.6 4.0 20.2 48.3 -8.5 35.7 43.5 7.9 54.9 Secun3 10.2 -3.5 6.3 32.3 -28.9 -6.0 12.7 1.4 14.3 Univl 40.2 -19.7 12.6 29.3 4.6 35.2 -14.1 5.3 -9.5 Univ2 31.4 -11.0 16.9 -19.2 20.6 -2.6 -3.6 -1.7 -5.3 Univ3 -5.3 -4.2 -9.3 - - - 55.8 -9.8 40.6 Overall 24.2 0.63 25.0 64.4 -12.6 43.7 34.7 -15.0 14.5 Prima= some or completed primary; Secun = some or complete secondary education; Univ= some or complete college; I= less than I year of exp., 2= 1-5 years of exp., 3=more than 5 years of exp. Source: Based on EPH, May-Oct waves for each year, Buenos Aires only. Insufficient number of observations Since 1993, long tenured informal salaried workers with primary or secondary education have had serious falls in income (36.1%, 28.9%), along with those with secondary education just starting a business (40.9%). All three of these groups felt a decline to or below 1988 levels and would be likely candidates for driving the poverty numbers. 4. Low wage, young workers show the greatest gains across the period, particularly among those with lower levels of human capital. This may be because they are joining expanding industries and are not trapped by their age and depreciated human capital. 5. Table 2.9 suggests that among formal sector workers there is a sharp narrowing of the skills premium either measured from 1988 to 1997 or from 1993. A probable scenario The relative rigidity of the formal sector wage suggests that downward rigidity in nominal wages combined with the now near zero inflation rates of the 1995-1997 period prohibited any formal sector adjustment through wages during the Tequila crisis. This, combined with further privatizations and formal sector downsizing of labor force the 86 informal sectors to play the safety net for particularly older displaced formal workers, and young workers unable to access formal employment. This offers a reasonable explanation for the biggest wage losers-the self-employed with 1 year of tenure. A closer look reveals that this group is in fact largely middle aged. Workers over 45, laid off from formal sector firms are often unable to find jobs in the formal sector33 and are relegated to starting casual businesses. Though not the majority of the sector, the addition of this involuntary element to the stock of middle aged informal businessmen makes the sector appear less profitable. However, the high rates of unemployment emerging from the informal sector suggests that the sector was experiencing extremely negative shocks at the same time. * Interest rates were forced to rise in the wake of the Tequila crisis caused a brief recession in the construction sector. * Price stabilization may have reduced the advantage small firms had in the flexibility of setting wages and prices and the accompanying "rationalization" of prices may have revealed the lack of profitability of many enterprises over the long term (Pessino 1993). * The apertura may have forced a restructuring of the micro-firm sector as well. Anecdotal evidence suggests that shops that specialized in repairing locally made products confronted with the more sophisticated, varied and more reliable imports may have lost a large market for their services. Foreign marketing chains and a move towards super-markets offering higher quality and more standardized service, cheaper and more quickly may also have competed with traditional providers of these services. It may also be that larger firms who used to rely on micro-firms for intermediate inputs, partially to avoid the labor costs of producing in-house, now import superior quality substitutes. Each possibility constitutes an adverse demand shock to the sector and, more fundamentally, requires a restructuring of the sector. These kinds of shocks could explain the other big wage losers- older primary and secondary informal salaried workers who may be those whose human capital has become obsolete and who may not be able to find a new job. Unemployment and Poverty Clearly both employment and wages are important for poverty reduction. But how much would poverty fall if there were no unemployment? To answer this, we first estimated a Mincerian earnings equation for employed workers of the form: In (earnings from the primary job) = f (years of schooling, experience [=age-s-63,experience squared, female dummy, regional dummies) Arias finds the longest spells of unemployment among older workers. 87 Separate equations were estimated for each year, 1990-98. Using these equations, we then predict earnings for the unemployed workers. This "full employment" income is then combined with other sources of household income to recompute average income per adult equivalent, and the numbers below the poverty line, as well as the poverty gap and squatted poverty gap are then reestimated.34 The results are given the in the tables below. Table 2.10: Simulation of Full Employment Impact on Poverty Headcount, Poverty Gap, and Poverty Depth, 1990-1998 Part A: From Beg An Indgent From Being a Moderately Poor Non-poor Unemploy Year -ment rate Staying Becoming Escaping Staying Escaping Non-poor Indigent Moderately poverty Moderately poverty poor altogether Poor 1990 7.2 11.3 8.7 2.1 0.5 30.1 27.4 2.8 58.6 1991 6.4 5.9 4.7 0.9 0.4 24.5 22.0 2.5 69.6 1992 6.8 4.4 3.3 0.7 0.4 19.7 16.9 2.9 75.9 1993 9.3 4.3 2.9 0.8 0.6 17.5 14.1 3.4 78.2 1994 10.9 3.7 2.3 0.8 0.6 17.9 13.8 4.1 78.5 1995 17.0 6.1 3.1 1.9 1.1 21.1 14.7 6.4 72.8 1996 16.6 7.3 3.9 2.1 1.3 22.7 16.2 6.4 70.0 1997 14.7 6.8 3.8 2.1 1.0 22.6 17.0 5.8 70.6 1998 12.6 7.1 4.3 1.9 0.9 22.3 17.2 5.1 70.6 Sources: Calculated from the EPH surveys, average of May and October. Part B: Poverty H4eadcount Extreme Poverty Poverty Gap Poverty Depth Unemploy Year -ment rate Actual Full Actual Full Actual Full Actual Full employm employm employm employme ent ent ent nt 1990 7.2 41.4 38.1 11.3 8.7 16.4 14.1 8.8 7.2 1991 6.4 30.4 27.6 5.9 4.7 11.2 9.6 6.0 5.0 1992 6.8 24.1 20.9 4.4 3.3 7.8 6.4 3.7 2.9 1993 9.3 21.8 17.8 4.3 2.9 7.4 5.6 3.6 2.6 1994 10.9 21.6 16.9 3.7 2.3 7.2 5.1 3.4 2.3 1995 17.0 27.2 19.6 6.1 3.1 9.9 6.2 5.1 2.9 1996 16.6 30.0 22.2 7.3 3.9 11.3 7.3 6.0 3.5 1997 14.7 29.4 22.8 6.8 3.8 11.1 7.6 5.8 3.6 1998 12.6 29.4 23.3 7.1 4.3 11.2 8.1 5.9 4.0 Sources: Calculated from the EPH surveys, average of May and October. Part A of Table 2.10 shows how "full employment" affects both the moderately poor and the indigent, while Part B gives the adjusted poverty headcount, poverty gap and poverty depth (squared poverty gap) estimates. For 1998, the overall poverty rate of 29.4% would become 23.3% under full employment, with the extreme poverty falling from 7.1% to 4.3%. The fall in extreme poverty is made up of 1.9% of the labor force 34 For the estimated equations, see H. Lee, Background Paper No. 1. 88 moving from extreme to moderate poverty, and .9% moving from extreme poverty to the non-poor (Table 2.10, part A). These estimates make the extreme assumption of zero unemployment; however, even with a 50% drop in unemployment, poverty would drop by 3 percentage points, more than a 10% reduction in poverty. These estimates also assume that more unemployment is attained at current real wages. If full employment is attained through greater labor market flexibility, instead of growth in demand, real wages would fall, reducing the overall effect. Not only would some of the recently employed not escape poverty, but it is possible that some of the non-poor would become poor if real wages fell. While we can see that full employment, particularly if generated by rising labor demand, would have major effect in reducing the overall poverty rate, it is also true that full employment by itself is not enough. Given the characteristics of the poor, in terms of their education and skills, even fully employed they would note earn enough to raise themselves over the poverty line. Labor Market Reform and Poverty Though the adverse movements' action in wages and unemployment seems concentrated in the informal sector, it is also true that the adjustment in all sectors could be facilitated by faster job growth in the formal sector. Several factors suggest that current labor legislation may impede the absorption of labor. * Centralized or sectoral collective bargaining agreements limit firms' abilities to adapt to competitive conditions. The Ergo omnes clause extends these agreement to all workers and firms in the sectors, even if not signed. 90% of present contracts are "Ultra-active" remaining in force even after expiration and lock in work arrangements that may no longer be appropriate to market conditions. High non- wage benefits drive labor costs up relative to the costs of capital. Restrictions on firing discourage firms from hiring additional workers. Recent experience with temporary reforms in the 1 990s supports the view that these contracts are binding. Partial reform lowered payroll taxes somewhat, improved the pension system, and provided for temporary contracts that were exempt from payroll tax and severance payments. Employment under these modalidades promovidas reached nearly 5% of total employed workers by 1998, and accounted for 40% of the new jobs created between 1996-98. However, the modalidades promovidas were abolished in 1998, which may further add to unemployment in 1999. The labor law changes in 1998 also shortened probation periods to one month, and further centralized collective bargaining arrangements (although draft legislation is now pending that could reverse these changes). Since the modalidades promovidas were targeted toward the young who are less likely to be heading households, the direct impact on poverty may by minor. However, to the degree that they decrease the pressure on the informal salaried wage, the indirect impact may be more important. 3 Guasch (1999) Labor Market Reform and Job Creation: the Unfinished Agenda in Latin America and Caribbean Countries. World Bank, Directions in Development Series, 1999 p 49. 89 Figure 2.4 A,justed Mean Tenure .1 Ven F Figure 1~~~~~~~~~~~~~r 15 ~~~~~~~~~~~~~~~~~~~0 CA~~~~~~~~~~~r a ~~~ ~ ~~~~~FeC Bel E S W UK ueat t Gel Par economic D and prdutiit mea.n Austher cop.ud thFrbem sareutoe (0~~~~~~~~~~~~~~~~~~~~ F l ex b e .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . - R ig i Figure I o r CS0 The severane payments dsmistem in Argentina are quitehostly candrided with lxegalv unertantear tha srincpuamotof'reasevotssgiicantlys any coexsswthe aopnain ueplomete insurancescheme, bgai alltogretherend us any tociat e igabor costs. Mreovesoftelbr, abodintrreaiona ofo disnmicscause witou icrause" which includes 121pfanulcmenainfr year of service plus athet o "preaviso",plsayohrcmeatn pacsted backwards in the reform process. In addition the process is highly contentious, adding additional legal costs which do not benefit the worker or the firm. *Argentina, as most Latin Countries, shows high shares of the work force in self- employment, often thought to imply formal sector rigidities, but lower tenure in the manufacturing sector, sometimes thought to imply flexibility (see Table 2.8). However, such raw measures are misleading. In countries with lower formal sector labor productivity, the opportunity cost of being one's own boss is lower, and the share of the work force in self-employment therefore greater. Sirnilarly, lower level of education, and a younger work force are also associated with higher rates of turnover. After adjusting for these variables, Argentina in 1995 looks rigid on both counts. Figure 2.4 plots self-employment share and tunover adjusted for demographic and economic variables for a variety of OECD and Latin countries. Argentina is found in the northeast cordler with an unusually large self-employed sector and low labor tuc over along with Italy, Spain and Greece, countries with very rigid labor laws. 36 p6 For details, see W. Maloney, Background Paper No. 2 (Vol. rI). 90 * That said, the reduction in wages or labor costs may have ambiguous effects on poverty. As Guasch notes, the wage elasticity of labor in Argentina is .5 meaning that a 1% fall in labor costs will lead to only a .5% increase in employment. Though the long run social benefits may be positive, this conceivably would imply a lower income. However, if the reduction in costs comes from greater efficiency in the provision of non-wage benefits or elimination of costs that do not benefit workers, employment rises with no fall in worker welfare or adverse distributional consequences. Conclusions and Policy Recommendations * The tightness of the labor market is perhaps the most important determinant of poverty indicators. Sharp fluctuations in aggregate wages and unemployment are highly correlated with head count measures. * The rise and fall of wages and unemployment that seem to drive the poverty figures after 1993 appear to be concentrated in the informal sector. The pattern of who is most hurt appears suggests that this is mostly due to older, uneducated workers suffering as a result of the restructuring of the sector and not necessarily a long run or irreversible trend. * Labor market reforms that both induce investment and reduce the relative cost of labor in the formal sector are necessary. The existing Labor legislation is ill-suited to a new competitive environment makes adjustments to adverse shocks, for instance the tequila crisis, difficult and discourages firms from taking on additional workers. Proposed exemptions for firms of less than fifty employees (PYMES) that permit a special regime for collective bargaining and facilitating labor contracting are important to grant emerging firms the flexibility to adapt and grow. * Despite the recent adverse shocks to the sector, self-employment and small and medium firms are important sources ofjobs over the long run. They probably should be seen in the same way that small firms are viewed in the industrialized countries. As yet, there is limited evidence on barriers to future growth. Work by FONCAP suggests that there is a demand for credit not served by formal financial institutions, but it may also be the case that weak institutional structures-contracting mechanisms, judicial access- make expanding risky and hence credit is not a binding constraint. * Given that the bulk of the unemployment. in the 1990s emerges from the informal sector, it is probably unreasonable to consider the informal sector as a safety net for unemployment. Further work is needed to determined how the unemployed survive and how safety nets may be extended to cover these groups if necessary. * Existing labor market protections need to be made more efficient. Improving the quality and cost-effectiveness of labor benefits is a critical element in any program to 91 extend them to a broader range of the population Avoiding the inefficiencies in existing protections is rational for firms that can remain informal. If the implicit tax on wages is far higher than the actual benefits received, workers have every incentive to avoid formnality. The current high levels of non-wage costs and the inefficient provision of services make this a likely scenario. Improve the unemployment insurance system. One useful reform would be to eliminate the existing severance payment system and unemployment insurance tax, which together cost 8-10% of formal sector wages, and institute a system of unemployment insurance based on individual accounts that are portable, capitalized and fully funded. Such as system would reduce the disincentives to work found in most unemployment schemes, is easy to administer, reduces litigation over firings, and reduces labor costs to the employer. 92 Appendix A Female Worker Transitions Among Sectors of the Labor Market Across 5 Quarters Informal Defined as Unprotected Argentinean National Urban Employment Survey, 1993-1997 Probability of moving from initial to final sector, Pij, in percent Initial Sector OLF School Unemp Unpaid SE IS FS Other Total Out of Labor Force I 7 1 4 4 2 0 100 School 6 4 21 1 3 7 9 0 100 Unemployed 34 6 1 7 13 10 0 100 Unpaid 25 6 20 l _ 20 6 2 4 100 Self Employed 20 1 1I I 4 4 100 Informal Salaried 17 90 -1| C = O | 100 Formal Salaried 7 2 6 0 4 5 _ O 0 100 Other 20 0 0 0 40 lo o100 Total (P.) 47 6 1 1 10 l1 13 0 100 Notes: Sample aggregates 7 panels, each panel composed by individuals followed for four semesters beginning with a first panel in May, 1993. This generates roughly 3,700 females 16-65 Informal salaried defined as individuals in firms under 5 workers not covered by medical or social security benefits. Self- employed are those who declared so and individuals working in firms with less than 5 workers. Formal salaried are those with medical or social security. Boxed cells represent work, light shading=paid work, darker shading=probability of remaining in same sector. By compensating for final sector size, Pij/P.j gives a measure of fluidity among sectors. By compensating for rates of turnover in initial and final sectors, Vij gives a measure of disposition (logic) to move to a sector 93 Appendix B Male Worker Income Transitions Initial Quintile Final Quintile OLF Unemp lst Q. 2nd Q. 3th Q. 4th Q. 5th Q. Out of Labor Force . 9 6 3 O 0 1 Unemployed 11 21 20 8 7 4 1st Quintile 7 9 23 6 12 14 2nd Quintile 2 12 i1 3 14 11 4 3th Quintile 2 6 10 19 ; 33 4 4th Quintile 2 6 10 I I 9 19 5th Quintile 2 4 11 3 3 19 _ Final Distribution 10 9 15 18 10 20 18 Notes: Sample aggregates 7 panels, each panel composed by individuals followed for four semesters beginning with a first panel in May, 1993. This generates roughly 2500 males 16-65 household heads. Shading F probability of remaining in same sector. Boxed Cells Represent Positive Income. Figure 4b Female Worker Income Transitions Initial Quintile Final Quintile OLF Unemp 1st Q. 2nd Q. 3th Q. 4th Q. 5th Q. Out of Labor Force 7 5 2 2 1 1 Unemployed 50 l 1 7 4 4 2 Ist Quintile 30 7 8 8 6 10 2nd Quintile 16 9 1 yj17 7 4 3th Quintile 12 5 13 17 19 3 4th Quintile 9 6 9 8 1 16 Sth Quintile 6 4 6 2 4 16 Final Distribution 55 7 9 6 7 8 8 Notes: Sample aggregates 7 panels, each panel composed by individuals followed for four semesters beginning with a first panel in May, 1993. This generates roughly 2500 females 16-65 household heads. Shading = probability of remaining in same sector. Boxed Cells Represent Positive Income in the initial period. 94 CHAPTER III: POVERTY AND THE SOCIAL SECTORS Public social sector spending supports a variety of program;s that are usually considered important instruments for reducing or alleviating poverty. They range from short term palliatives such as emergency employment to long term investment in the education of poor children. This chapter reviews selected issues in social assistance, health and education. In particular, it examines the extent to which the poor are benefiting from social programs, identify coverage gaps, and explores ideas on how they can benefit more from government spending on social assistance, health and education. The setting: an overview of social expenditures Argentina has one of the highest levels of per person social spending in Latin America37.Argentina spends more per person than countries such as Chile, Costa Rica and Uruguay. Yet substantial poverty persists. In 1997 consolidated public social spending stood at 1,530 pesos per person (Table 3.1). This represented about 18 percent of GDP and 65 percent of total consolidated public expenditures. Of the consolidated public social expenditures in 1997, 54 % are accounted for by the national government. The balance comes largely from provincial governments, with a small but increasing share from municipal govemments. Since 1980, regional governments have become an increasingly more important source of social sector funding. Table 3.1: Consolidated Public Social Spending (GPS), 1980 - 1997 PERIOD GPS GPS PER CAPITA GPS/GDP GPS/GPC (bill 1997 Pesos) (000's 1997 Pesos) (%) ) 1980-1983 33.19 1.16 14.7 45.1 1984-1988 37.94 1.23 16.8 51.6 1989-1990 36.03 1.12 17.5 56.8 1991-1995 51.58 1.52 18.6 64.0 1996 54.53 1.54 18.2 65.5 1997 56.54 1.53 17.5 64.6 NOTES: GPS: Public Social Expenditures, GDP: Gross Domestic Product, GPC: Consolidated Public Spending Source: Govt. of Argentina, Ministry of Finance Caracterizaci6n y Evoluci6n del Gaso Puiblico Social en el Periodo 1980-1997, Buenos Aires, 1999. Government of Argentina,Direcci6n Nacional de programaci6n del Gasto Social - Secretaria de Programaci6n Econ6mica y Regional CaracterizacionyEvoluci6n del Gasto Putblico Social en elPeriodo 1980-1997, Buenos Aires, 1999 95 Table 1 shows that there was higher per capita social spending (GPS) in the 1990s than in the previous decade. This is due to rising national income and a slight increase in the share of GPS in GDP. Second, this trend has been accompanied by an increase in the share of GPS in public expenditures from 45 percent at the beginning of the 80s to 57 percent at the beginning of the 90s to 65 percent in 1997. These facts can be interpreted as an increase in government conmmitment to social issues, but it also reflects that there has been a downsizing of public spending relative to GDP. Social expenditures in Argentina may be divided into two broad categories (Table 3.2). One is social insurance spending; the other is social sector (or non-insurance) expenditures. The former consists of expenditures that are financed mainly by beneficiary contributions, and constitute about 38% of total social sector spending. The rest are financed by tax and other fiscal resources. Public spending on education is paid for 100 percent by fiscal resources; in contrast, 57 percent of public health spending is financed by mandatory fees and contributions. It is interesting to note that fiscal resources finance close to half of social security expenditures. Public social sector spending can be further sub-divided into two types of expenditures: universal and targeted programs (Table 3.2). Universal programs like support for education are intended to benefit the general public. In contrast, targeted expenditures are intended to benefit the poor and other disadvantaged groups. In 1997 the share of targeted social spending stood at 19.5% percent of total non-insurance social spending, and 2.0% of GDP. This is slightly less than the figures in the 80s, when targeted programs were as high as 25% of social spending and 2.5% of GDP. Table 3.2: Consolidated Social Sector Public Expenditures, 1997 (% of GDP) Total Social Social Non-Insurance Sector Insurance Year Total Universal* Targeted %Targeted 1980 16.1 9.1 7.0 7.8 2.6 24.7 1985 16.7 9.5 7.2 8.3 2.3 21.4 1990 17.6 10.8 6.8 8.1 1.9 19.1 1995 19.4 11.2 7.8 9.4 2.2 19.2 1997 17.0 9.9 7.6 8.3 2.0 19.5 * Covers education, culture and science and technical, health, water and sanitation, and other urban development ** Covers housing/urban development, social promotion, social assistance, and employment Source: Ministry of Economy, Secretary of Economic and Regional Programs, Caracterizationy Evolucion del Gasto Piblico Social, 1980-9 7, Buenos Aires, 1999. Overall Benefit and Tax Incidence The overall question that is important is "who benefits from social expenditures"? But any analysis of social spending also has to consider the impact of the tax system as well. Data for from studies by Gasparinil (see see Leonardo Gasparini, "Incidencia Tributaria del Gasto Publico Social y de la Politica Tributaria en Argentina" FIEL, April 1999 and "Incidencia Tributaria del Sistema Impositivo Argentino" FIEL, 1998. 96 Table 3.3 below) demonstrate that the poor receive a higher share of overall expenditures, and pay a lower share of taxes. Table 3.3 :Total Social Expenditures and Taxes by Quintiles, Urban Argentina 1996 (percent of total) Quintile: I II III IV V Total Expenditures: Social Sectors 29.8 18.8 21.7 16.8 13.0 100 Social Insurance 9.9 20.6 19.5 23.6 26.5 100 Total Social Expenditures 21.8 19.5 20.8 19.5 18.4 100 Tax Distribution 7.1 10.7 14.9 20.1 47.2 100 Income Shares 4.0 8.4 13.2 21.2 53.2 100 Source: L. Gasparini, FIEL, 1999 and 1998. hicome shares from EPH, average of May/October 1986. The lowest 20% in the income distribution received 22% of total social sector spending in 1996 while those in the upper quintile received 18% In effect overall social spending appears to be equally divided between quintiles. However, these results are distorted somewhat by social insurance spending which gives more benefit to the upper quintiles. Excluding social insurance, the poorest 20% received 30% of social spending, compared to 13% of the upper quintile. Gasparini also shows that the poor pay only about 7% of total taxes. Since the poor receive only 4.0% of total income, the tax system appears regressive. However, alternative estimates by Llach and Montoya38 seem to show that taxes are roughly proportional (see Table 3.4), although they confirm the progressivity of expenditures, particularly when expressed as a share of the income of the poor rather than as a share of total spending. Table 3.4: Tax and Social Spending Incidence in Buenos Aires, 1996 Quintile: I II III IV V Total Incidence of social spending: (% of income) 1991: 66.7 23.1 18.2 8.8 2.6 10.7 1996: 110.3 32.4 22.7 10.3 2.8 12.5 1998: 85.8 31.2 21.5 12.2 4.7 12.6 Tax Incidence (% of income,) 1991: 31.4 29.6 29.5 29.1 29.7 29.6 1996: 33.2 32.6 32.3 32.7 34.4 33.6 1998: 30.6 29.9 30.0 31.2 36.1 33.8 Net Subsidy(% of income) 1998: 55.2 1.4 -8.6 -19.0 -31.4 -21.1 Source: Llach and Montoya, En Pos de la Equidadc June, 1999 In terms of its impact on income, the progressity of social spending appears to have increased since 1991 as social programs have expanded. The Llach and Montoya 38 Juan Jose Llach and Silvia Montoya, En Pos de la Equidad, Buenos Aires, 1999. 97 estimates conclude that that social spending that benefited the lower 20% was the equivalent of 67% of their income in 1991, but this increased to 87% by 1998, resulting in a sizeable net subsidy to the poor (see Table 3.4, data includes social insurance). In terms of tax incidence, this study shows that the tax system is roughly proportional across income classes. Thus, the poorest pay about 31% of their income in taxes and receive benefits equal to about 86% of their income, for a net subsidy from the public sector of 55% of their income. Social Spending and the Business Cycle Targeted programs generally should expand during an economic crisis, as unemployment and poverty increase. At the same time, we know that declines in government revenues often produce reductions in budgets for social and other spending. What has been the case in Argentina? For the period 1980-97, Table 3.5 shows elasticities of government spending with respect to GDP, for total government spending, social spending, social security, universal and targeted programs. Table 3.5: Elasticities of Government Spending with Respect to Changes in GDP, By Type of Spending, 1980-9739 Spending Type Elasticity | (t-ratio) Total Government .954 (3.13) Total Social Sector 1.275 (2.42) Social Security .838 (1.47) Universal 1.874 (3.51) Targeted 1.855 (3.10) These elasticities reveal that all kinds of government spending are procyclical; they rise and fall with GDP changes. While the elasticity of total spending with respect to GDP is .95, for the social sectors is 1.3. Thus, social spending, seems to be more sensitive to the business cycle than overall spending. Within the social sectors, targeted and universal programs are very elastic with respect to the business cycle, while social security spending is more ambiguous. One would hope that targeted programs would be part of a social safety net that could expand during a crisis. Instead of being countercyclical, however, they seem to fluctuate by a factor greater than even total social sector spending. Thus, a five percent drop in GDP seems to be associated with a 9 percent drop in targeted social sector spending. 39 Estimated from equations of the form: dlnX = e(dlnGDP) + c where dlnX is the change in the log of the expenditure type, dInGDP is the change in the log of GDP, "e" is the estimated elasticity, and "c " is a constant term. All coefficients are significant at the 95% level, except for social security. An analysis of episodes when GDP has increased versus decreased shows that social security spending has a significant elasticity of 1.1 during increases in GDP, but no significant elasticity during downturns. 98 Targeted Social Assistance Programs Argentina's spending on targeted social assistance programs consists of a wide range of programs that seek to improve the quality of life of the poor, vulnerable, and Alternate Measures of Poverty: The Index of Unsatisfied Basic Needs (NBI) While income or consumption is a traditional measure of poverty, there are many dimensions of poverty that are not captured in such monetary measures. In many ways, social indicators capture some of these aspects. hi Argentina, as in many other countries in Latin America, selected social indicators are combined into an index of unsatisfied basic needs (NBI). By the NBI standard a person is considered poor if they live in a household having: . more than three persons per room (crowding); * living in a house made of irregular materials, or in rented quarters (housing); . not having an indoor flush toilet (sanitation); * having a child between 6 and 12 years that is not attending school (school attendance); . having four or more persons per person working and a household head with 2 or less years of primary school (subsistence capacity). The main problem with these kinds of indices is their purpose is mixed; it is not clear that this is a good proxy for low income, or is attempting to measure some other aspect of poverty. The present index is heavily biased toward housing quality (the first three indicators), and has only one indicator for education (school attendance), and none for health. Nevertheless, the NBI is the most common indicator used for targeting social programs, given the absence of nationwide poverty measures. The Government is presently considering a revision to the NBI to overcome some of its deficiencies. disadvantaged population. They include food, nutrition, health, employment, training, education, shelter, clothing, cash grants, and emergency programs. Most of them are targeted directly to poor families using the Index of Unsatisfied Basic Needs (NBI) to identify the poor (see box). Others are aimed at strengthening the ability of communities to address social and cultural issues (see Annex I for a summary matrix of these programs). Unfortunately, there are only a few evaluation studies that examine program impact and cost-effectiveness. Hence, the focus of the following discussion is on the benefit incidence and coverage of social assistance. Benefit incidence: an overview. Table 3.6 presents an overview of the distribution of social assistance, based on the 1997 Encuesta Desarrollo Social (EDS). Three main conclusions can be drawn: (i) As hoped, social welfare assistance appears to be strongly pro-poor on the whole. 47.2 percent of the social welfare benefits goes to the poorest quintile. In contrast, the highest quintile gets only 2.4 percent of the benefits.40 40 The 1997 EDS asked households whether for each of a series of items they received in kind or monetary assistance government agencies or other sources. The responses to these questions were used to measure benefit incidence, giving two points for regular receipt versus one for occasional. Source: 1997 EDS 99 (ii) Still, the scope for reducing leakage is considerable. Significant amounts of the social welfare benefits are going to the non-poor ( 35 percent). In fact, about a quarter have been enjoyed by the third, fourth and fifth quintiles. (iii) Surprisingly, about 70 percent of the poorest quintile are not getting public assistance and over half got neither public nor private help. This would indicate that a substantial percentage of the poor are outside Argentina's public and private social safety nets, although it remains unknown whether they actually need help. Table 3.6: Benefit Incidence of Government Social Assistance Income Distribution of Population % households quintiles benefits/a (%) distribution (%) receiving assistance Public All sources First 47.2 20 29.5 45.0 Second 29.0 20 18.8 31.4 Third 15.0 20 9.8 22.0 Fourth 6.4 20 4.9 16.1 Fifth 2.4 20 1.4 9.8 Total 100 100 12.8 24.8 Poor41 64.6 31.0 25.5 39.8 Non-poor 35.4 69.0 7.2 18.1 Total 100 100 12.8 24.8 (a) Weighted by provincial government spending on social welfare Food and nutrition programs for young children and senior citizens, are one of the most important thematic areas in the inventory of social assistance programs.42 Taken together, these programs are targeted to pregnant mothers, children 0- 14 years old and senior citizens (60 years old and over). As should be expected, results (Table 3.7) show that food and nutrition programs are strongly pro-poor. The poorest quintile is getting disproportionately more of the benefits than the upper income quintiles. But despite the progressiveness in the distribution of benefits, the number of vulnerable population that has not been reached remains substantial: 41 Poverty here is defined using this report's poverty line and the EDS data, which yields a slightly higher overall poverty estimate 42 Programs in this area include POSOCO (Politicas Sociales Comunitarias), PRANI (Programa Alimentario Nutricional Infantil), PROHUERTA (Program Huertas) and PROSONU (Progama Social Institutional). Feeding and nutrition related assistance are also included in other thematic areas. In health, programs like PMI (Programa Materno Infantil) and PROMIN (Programa Matemo Infantil y Nutricion) have feeding and nutrition components. In Tercer Edad, assistance to help senior citizens meet their basic needs is provided through ASOMA (Apoyo Solidario a Los Mayores). In addition to the above interventions, government is funding the PROHUERTA program which seeks to improve the nutrition of the urban and rural poor (NBI families) through the promotion of family and community gardens and farms. 100 Table 3.7: Benefit Incidence of Targeted Food, Nutrition and Housing Programs Total Household per capita income quintile Programs for 1 2 3 4 5 Food and Nutrition: Children 3-4 years old Coverage rate (a) 19.7 20.4 13.8 19.5 21.6 20.2 Distribution of benefits Nurseries, preschool & others 100 38.5 26.7 10.9 17.0 6.9 Infant feeding centers 100 56.4 25.1 14.8 3.7 0.0 Primary school students (b) Coverage rate (c) 50.1 69.7 49.8 33.0 30.5 13.2 Distribution of benefits 100 56.6 22.9 10.7 8.0 1.8 65 year olds and over Coverage rate 6.0 21.2 7.0 5.3 1.7 0.1 Distribution of benefits 100 48.4 25.8 19.1 6.3 0.4 Pregnant mothers and children 0-2 yrs. Coveragerate(d) 38.1 44.4 34.8 29.9 34.7 32.5 Distribution of benefits (e) 100 69.9 21.2 6.6 2.1 0.2 Housing: FONAVI /IPV/FM distribution (f) 100 21.8 23.9 16.9 17.0 20.5 (a) Includes beneficiaries from preschool, infant, and other feeding centers; (b) Comedores Escolares; (c) Over children aged 6-13 years; (d) Includes beneficiaries of milk delivered by Obra Social; (e) Excludes Obra Social milk beneficiaries; (f) Housing programs: FONAVI = Fondo Nacional de la Vivienda; IPV= Instituto Provincial de la Vivienda; FM=Financiaci6n Municipal. Source: 1997 EDS, except for housing, which is from 1996 Encuesta de Gastos, special tabulation. (i) The most vulnerable group to long-term negative effects of malnutrition are the children aged 0-2 in the lowest quintile. Of these, only 44% receive benefits from public nutrition programs, even though they receive 70% of the benefits. Thus, these programs appear well targeted, but insufficient in scope. (ii) The corresponding rate for preschool children aged 3-4 years in the same first quintile is even lower at 20%, but coverage of school-age children in the lowest quintile is very high at 70%, even though they receive only 57% of total program benefits. (iii) And, of the poorest 65-year-olds and over, only a fifth received food and nutrition assistance. While the effectiveness and impact of nutrition programs need to be studied carefully, this analysis suggests that some reallocations of expenditures might be in order. For example; government might want to re-allocate some resources now devoted to primary school age children to the neediest of pregnant mothers and children 0-4 years old, on the assumption that the unborn and these very young children are more vulnerable to the ravages of malnutrition. Housing. Table 3.7 shows data on the impact of three government housing programs: FONAVI, Instituto de la Vivienda, and Financiaci6n Municipal; all of which provide finance for the purchase, construction, repair or modification of owner occupied housing. 101 These programs are not well targeted to the poor, but neither are they biased against the poor. The lowest 20% in the income distribution receive about 22% of the benefits. Of all housing finance received by the lower 20%, 73% comes from these programs, and 31% of households in the lowest quintile have benefited from these programs (in 1996). These programs constitute the bulk of spending by the Government in housing and urban development, which totaled $1.2 billion in 1997, or about 2% of total social sector spending. Cost recovery is very low, and substantial savings could accrue to the Government if these programs were better targeted. 4 Employment and productivity development This thematic area consists of various training, employment, and labor productivity promotion initiatives and include workfare, employment subsidy to private firms, and economic assistance to unemployed workers (Annex 1). As with other interventions, little rigorous assessment of the value added and incidence of these programs have been systematically undertaken. It is worth noting here that while an evaluation of Programa Social Agropecuario (PSA), a program designed to help small rural producers through financial, training and technical assistance, had been undertaken, the study focused on organizational and administrative issues, and has little to say about impact, incidence and cost-effectiveness. The exception to the above state of knowledge is TRABAJAR, for which noteworthy evaluation studies have included impact assessment. TRABAJAR is a social safety net program based on workfare principles, financed partly by the World Bank. The program provides temporary financial assistance to poor workers, who are suffering from high levels of unemployment, and assists in improving social and economic infrastructure in poor communities. Municipal and provincial governments can propose from a menu of public projects for financing by the program. Such projects include minor construction, repair, expansion or remodeling of schools, health facilities, basic sanitation facilities, small roads and bridges, small dams and canals, community kitchens and centers, and tourist centers. The national government pays for one-third of the cost of the proposed projects; the rest is financed by local government. TRABAJAR II cost about $1.2 billion, of which $192 million are for workers' wages and $8 million for administrative expenses. Management information system data reveal a coverage rate of about 20 to 30 percent of the poor unemployed persons, depending on 43 It is possible that there is a great variability of incidence across housing programs. For instance, in the case of Instituto Provincial de la Vivienda (IPV) in the Province of Buenos Aires (an executing agencies for the funds provided by FONAVI), about 70% of the financing goes to programs for the three upper income quintiles. The value of the houses built under such programs ranges from $ 20,000 to $ 60,000, compared to very basic housing unit that could be obtained with $ 5-6,000. Anecdotal evidence suggests that municipal programs may be better targeted. 44 see Jyotsna Jalan and Martin Ravallion, "Income gains from Workfare and their Distribution: Estimates for Argentina's Trabajar Program" World Bank (processed), June 1998. 102 whether repeaters are excluded or included in the calculation. Studies also reveal that TRABAJAR II, which uses low wage rate to attract only those who are poor and with few prospects of employment, is well targeted. Experience thus far shows that self- targeting is effective: evaluation of the program indicates that on average 40 percent of its resources have reached those in the bottom 5 percent of the income distribution, while 75 percent have benefited the bottom 20 percent. Moreover, targeting of poor communities has also been effective. Box: Gender Bias in the Distribution of Social Programs Is there a gender bias in the distribution of public social programs? In a short overview paper, the questions of whether or not the household structure and/or the gender of the household head is instrumental in the distribution of resources obtained from transfers (in-kind or money) or training programs were examined. Using thel997 EDS, the study the sample into single-headed households and couple-headed households and disagregated by gender. The results suggest that single fathers are the least likely to receive benefits while single mothers are the more likely; if there is any gender bias it is against men. Impact was measured in terms of increased school attendance. Of the non-recipients, the children in households headed by single fathers are the least likely to attend school. Among recipients of at least one of the transfers, school attendance is equally as likely among all four types of households. Thus, the receipt of transfers by unmarried fathers seems to benefit their children more than children in any other type of household structure who received benefits, despite the fact that single fathers are less likely to receive benefits in comparison to single mothers and couple headed households. Finally, in couple-headed households, the gender of the head does not seem to significantly influence school attendance. (see Background Paper No. 9. Does the Gender Recipient of Social Programs Matter in Argentina? by Pia Peeters and Wendy Cunningham. In terms of impact, an econometric study using a sophisticated version of the matching technique reveals that the net income gain of workers employed by the program's projects is about half of the workers' wages. The other half is equivalent to the amount that the workers would have earned anyway without the project. Another study indicates that the quality of the projects and their valuation by beneficiaries are on par with those of Social Fund projects on the whole. Community development and social integration There are several interventions in this thematic area (see Annex 1). One of them is a new -program called FOPAR (Fondo Participativo de Inversion Social) which is being supported by the World Bank. FOPAR, which now has a budget of about $ 118 million, finances projects prepared and proposed by poor communities. Unlike TRABAJAR, it seeks to build institutional capacity among the poor, and simultaneously to provide them with means to meet specific needs. This means that the projects under this program are defined not only to satisfy specific needs of the poor, but also to be a training vehicle to build and strengthen their ability to identify, define and prioritize their needs, mobilize and manage resources, as well as design and implement their own projects. The program started in 1995, when the Govermnent established a pilot Social Investment Fund, which would operate in the northern provinces with future expansion in mind. Evaluation of this pilot program has been conducted, reviewing 119 projects and 103 interviewing a stratified sample of beneficiaries. These various assessments indicate that 75 percent of beneficiaries come from NBI families. In addition, four-fifths of the remaining 25 percent were from households with monthly income of $300 or less. This effective targeting of the poor comes from a combination of careful geographic targeting and the establishment of a menu of projects eligible for financing that are small and of such a nature as to attract only the poor. The ex-post technical assessment of the civil works considers the quality of the projects as good or better, while interviews with beneficiaries indicate that they were satisfied by the projects. They rated community and beneficiary participation as excellent. A high 96 percent of the beneficiaries felt that the projects effectively satisfied their needs or solved their problems as designed. Furthermore, 97 percent thought that FOPAR was indispensable in implementing their projects. Finally, a large majority of the beneficiaries (83 to 97 percent, depending on the nature of the project being evaluated) emphasized that the financial administration of FOPAR-provided resources were carried out in an honest and transparent manner.45 The incidence of public and private income transfers on poverty rate. In most countries, low income households are often helped not only by the government but also by relatives, friends and other members of society. This assistance can be in a form of in- kind or monetary transfers such as pensions, remittances from relatives and friends, and social assistance from government. In this section, we ask the question: What difference does it make to the poverty rate if these income transfers were included or excluded from the calculation of the poverty rate? Who would be most affected by it? In the EDS survey, there are two kinds of income: earnings from work and non- work income. Included in the latter are such items as scholarships, unemployment insurance, retirement income, pensions, subsidies, remittances from relatives and friends, and food assistance money, and "other income".46 In Table 3.8, it is clear that the effect of transfers on the incidence of poverty varies, depending on the demographic group and the definition of poverty. On the whole, the effect is quite modest in regard to indigent households. The difference in poverty rate with and without transfers is only 2.6 percentage points. It rises to 3.9 percent when moderate poverty is included. 45 Govt. of Argentina, SIEMPRO, "Programa FOPAR: Evaluacion Ex Post", 29 June, 1998 46We interpret "other income" as income from interest and rents, since the questionnaire does not have any other items asking for income from property. To measure income transfers from all sources, therefore, "other income" was added to work income to form an estimate of the family's total income from work and property, which we refer to in the discussion as its "core" income. 104 Table 3.8: Household Income Transfers and Their Effect on Poverty Rate Percent Poor Difference Percent Difference (%) Indigent (%) With w/o % With w/o % transfers transfers transfers Transfers All 31.0 34.9 3.9 10.5 13.1 2.6 Male Household Head 30.4 33.9 3.5 9.9 12.0 . 2.1 Female Household Head 33.2 40.7 7.5 13.0 18.9 5.9 Age Group of Household Head 15-24 32.7 35.7 3.0 13.5 14.6 1.1 25-59 33.3 34.7 1.4 11.4 12.5 1.1 60-64 26.1 30.5 4.4 8.5 12.1 3.6 65 and More 21.0 38.9 17.9 6.0 18.2 12.2 Source: 1997 EDS. The poverty rate in this table (31%) is obtained using the World Bank poverty line and the EDS data. It differs slightly from the results cited previously which used the EPA data. The impact, however, of income transfers is quite substantial for certain demographic groups. The most affected is the age group 60 and over, particularly those above 65 years old. Without transfers, the 65 and over age groups would have experienced an increase in poverty rate from 21 percent to 38.9 percent. The percentage of indigent persons in this age group would have risen by 12.2 percentage points. In terms of gender, the poverty rate of female-headed households (widows and single mothers) would have increased from 33.2 percent to 40.7 percent. In contrast, the prime age group of 25-59 years would have experienced an increase of less than 1.5 percentage points only. Table 3.9 further reveals that the better off in terms of core income (labor earnings and asset income) are heavily favored by non-work transfers. While the richest quintile gets 125 pesos of income transfers per capita, the poorest quintile gets only 8 pesos. This pattern raises equity issues. First, these transfers are mostly from public sources; relatively little come from relatives and friends. Second, close to half of the financing of pension payments are from general fiscal revenues. Consequently, further analysis is called for to determine whether greater reduction in poverty rates can be achieved by re-targeting of this public subsidy to senior citizens with low core income. Table 3.9: Average Monthly Income Transfers Per Capita by Source and Core Income Status (Pesos/month) Core income class All sources Public and private Relatives/friends (quintile) organizations First 10.2 8.2 2.0 Second 21.1 18.9 2.2 Third 36.7 32.1 4.6 Fourth 56.8 49.0 7.8 Fifth 125.1 108.8 16.3 Total 50.0 43.4 6.6 Source: 1997 EDS 105 Summary and Conclusions (i) The distribution of social assistance spending is pro-poor on the whole; reduction of leakage of benefits to the non-poor and less vulnerable is, however, still possible and desirable. (ii) Coverage rate remains an important issue. A substantial percentage of the poor appears to be outside existing social safety nets, particularly the poor of the vulnerable demographic groups. (iii) Hard choices need to be made regarding target beneficiaries, considering Argentina's financial situation. (iv) Income transfers do have considerable impact on poverty rate among senior citizens and female heads of households; this impact, however, can be enhanced by improved targeting of government income transfers such as pension subsidy. At present, these transfers favor well off individuals. However, because pension payments are based on past contributions (payroll deductions) it may not be possible to make this change, despite its desirability. 106 CHAPTER IV: POVERTY AND EDUCATION Education Sector Overview Argentina's education system is one of the most advanced in the region. However, as noted in Chapter 1, this may only be appropriate given the high level of development in the country and Argentina's performance in many ways equals that of the OECD countries (see Table 4.1). It can boast of a 96.8 % of literacy rate for the adult population, 9.7 mean years of education of the adult population and 14.4 years of school life expectancy. Participation in education among those aged 5 to 29 is also high at 60.8%, very close to the OECD average of 61.3%. Argentina is close to universal primary education enrollment (113 % gross enrollnent), and relatively good participation at secondary level (78% gross enrollment). Table 4.1: Basic Indicators: Argentina in Comparison with LAC and OECD Countries Indicator Argentina LAC OECD Adult illiteracy rates (%). 3.8 13.4 1.0 Years of school life expectancy 14.4 16.4 Gross enrollment for Primary Level (%) 113.3 113 105 Gross enrollment for Secondary Level (%) 77.8 54.5 112 Gross enrollment for Tertiary Level (%) 41.8 18.4 47 Participation in education among the 5-29 60.8 61.3 age group Source: OECD, UNESCO However, retention rate at secondary education level remains a major difficulty for Argentina in terms of the country's overall competitiveness and social cohesion. The secondary graduation rate in Argentina is 52%, compared to a mean rate for the OECD countries of 80%. In terms of transition rates, a cohort analysis show that of 100 students entering primary school, 84 will enter the seventh grade, 76 will enter the ninth grade, 40 will enter the last year of secondary school, 35 will enroll in university and only seven will graduate. The number of students repeating the same grade has a strong impact on the education system's performance. The repetition rates in 1997 were 5.3% at the primary level and 9.3% at the secondary level (1997), although they vary greatly across the country. Drop-out rates are remarkably greater amnong secondary school students than among primary school students (approximately 42% and 12%, respectively, in 1996). Both repetition rates and drop-out rates: (i) are very high for the first two years of primary and secondary level; (ii) varies widely across jurisdictions, with higher rates in provinces with higher poverty rates. 107 Fig. 4.1 Education Attainment Profile, Age 20-24,1997 120.0% 4, 100.0%- 0 80.0% - Quintile 1 fo 60.0% -- Quintile5 e 40.0%- 0 20.0%0/ 0.0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Highest Grade Attained Private school enrollment has historically comprised a sizable proportion of total enrollment at all levels in Argentina, and this figure increased slightly during the 1990s. For primary and secondary education, the percent of private school enrollment is 24.6% of total enrollment in 1996.47 At university level, private enrollment reach a 14.3% of the total enrollment. The percentage of students attending private institutions varies widely from jurisdiction to jurisdiction. In general, private education is more developed in urban and densely populated provinces such as Buenos Aires, Cordoba and Santa Fe. For example, in the City of Buenos Aires, approximately 28% of all primary students and 44% of all secondary students are enrolled in private schools, compared to Formosa, which has the lowest percentage of students in private institutions (6% in primary and 8% in secondary). Inequities in Access In Argentina, as in most countries of the region, youth from less advantaged homes have less access to schooling, higher dropout and repetition rates, lower completion rates and poorer achievement level than those from more advantaged homes. A recent analysis of school attendance patterns by income quintile48 indicates that most children are enrolled in primary school regardless of income level (see Table 4.2). However, this analysis also shows that, while almost all youth from households in the highest income quintile are enrolled in secondary school, only 71% among youth in the lowest income quintile are enrolled. Consequently, the average years of schooling varies widely among income quintiles. Among the population between 6 and 24 years old, the richest quintile has an average of 11.2 years of schooling, while the poorest quintile has only 9 years. 47 In OECD countries, 82.4% of students enrolled attend public schools and 16.3% attend private schools (among which 12% attend government-dependent private institutions and 4.3% attend independent private institutions). 48 Secretaria de Desarrollo Social. Encuesta de Desarrollo Social. 1998 108 Table 4.2: Educational Coverage by Income Level Coverage Total Households Income Quintiles per Capita Structural Poverty 1 2 3 4 5 Total NBI No NBI Pre-School 88.4 82.8 94.0 94.9 89.5 9§8.77 88.4 79.7 91.7 Primary 99.4 98.5 99.8 99.9 loo 100 99.4 97.7 loo Secondary 82.8 71.1 80.8 90.0 93.3 96 82.8 63.1 86.6 In addition, those children from low-income families are more likely to be older than other primary school children due to higher incidences of late entrance and repetition among poorer children. As Table 4.3 shows, 32 % of the children from the poorest income quintile are delayed in primary level and 50.2 % at secondary level, while these figures are 9.3% and 25 % respectively for the richest quintile. As a consequence of lower access and higher dropout rates, the youth from poor backgrounds are in a much higher proportion out of the school. This phenomenon is particularly serious at secondary level where the 36.7% of the poorest youth among 14 and 18 years old are out of the school (Table 4.3). In terms of completion, the differences among income quintiles also shows a very large gap: while 76% of the richest youth belonging to the 18-24 age group have finished secondary school, only the 24% of the poorest income quintile youth have (Table 4.3). These facts have several negative consequences which affect social cohesion: high level of unemployment among the risk youth, family problems, drugs addiction, crimne and violence. Table 4.3: Indicators of Education Perfonnance By households Income Quintiles Per Capita Population Households Income Quintile per Capita Total 1 2 3 4 5 Out of school (14-18 Age Group) 26.0 36.7 27.5 18.8 15.6 7.7 6 - 24 age group out of school: Average years of schooling 9.7 8.8 9.2 10.2 11.6 11.2 Average age to dropout 15.7 14.7 15.8 16.5 16.8 16.6 6-14 age group in school Delayed Entry to Primary School 20.1 32.5 18.6 11.8 8.3 9.3 Delayed Entry to Secondary School 40.9 50.2 45.9 42.8 30.0 24.9 Repeaters of at least one year 14.1 24.6 12.8 7.0 4.0 3.6 Secondary School Completion (age 18-24) 47.4 23.6 33.9 49.7 64.5 75.8 Inequities in access are currently greatest at university level. Recent statistics indicate that only 10% of youth from the poorest families attend university, compared to 50% of youth from the richest families49. Consequently, since the public university system is virtually free for most students, those who primarily benefits are the non-poor. 49 Enrollment in private universities is only 14% of the total, and so most students attend government- fimded public universities. 109 50 Repetition, Drop-Out and Student Achievement: Provincial Disparities Provinces with a higher proportion of population living with NBI (unsatisfied basic needs) are more likely to have higher repetition and dropout rates. In 1996, the average. repetition rate for provinces with high concentrations of NBI households was approximately 10%, compared to 4% in provinces with low concentrations of NBI households. As illustrated in Table 4.4, drop-out rates follow a similar pattern. The gaps in both repetition and drop-out rates between provinces with low and high concentrations of underprivileged population segments seem to be broader at the primary than at the secondary level, although the highest absolute rates occur at the secondary level. Concerning income distribution and repetition, in Table 4.3 it is possible to observe that among the 6- 14 age-group of the population, 24.6 % of the children from the first quintile have repeated at least one grade, while the figure is only 3.6 % for the children from the fifth quintile. Table 4.4 Comparison of Repetition Rates By Group of Provinces According to Percentage of Population Living in NBI Households 1994 and 1996 Primary Secondary % Population 1994 1996 Change In 1994 1996 Change In NBI Percentage Percentage Points Points In In Repetition Repetition Rates Rates Low (7-16%) 3.7% 4.1% 0.38% 6.9% 9.2% 2.2% Average (17- 4.1% 4.2% 0.12% 8.7% 10.3% 1.6% 25%) High (26% +) 10.2% 10.2% 0.03% 9.0% 9.5% 0.5% Country Mean 5.6% 5.7% 0.16% 8.5% 10.0% 1.4% Source: Programa de Estudio de Costos del Sistema Educativo, NMCE and Censo National de Docentes y Establecimientos Educativos 1994. Notwithstanding important accomplishment in the last years, the Operativo Nacional de Evaluaci6n 1996 indicated that poorer provinces are still experiencing lower rates of academic achievement than wealthier ones. Students in provinces with low concentrations of underprivileged households gave approximately 20% more correct answers on math and language exam questions than students in the poorer provinces. Here, too, the differences are more pronounced at the secondary level. Fifth-year secondary students living in provinces with lower percentages of poor inhabitants answered 26% more math questions correctly than those living in provinces with higher concentrations of poor. 50 We lack sufficient data to be able to relate educational performance and achievement and income distribution. Therefore, these paragraphs describe a comparison between differences in level of educational performance and the differences in economic development by province. Nonetheless, these provincial- level data seem to indicate that children and youth from low-income backgrounds are more likely to repeat grades and to dropout than youth from higher-income households. 110 Dropping out of school. Why many Box The Role of Gender in School Attendance children dropout of school early and why A background paper for this study looked at the gender aspects there is a huge gap m of school attendance. Specifically, the study explored the role played by the educational household structure and gender on the decision to send children to school attainment of poor and in Argentina. The detenninants of the education patterns of boys and of well off children? girls based was examined based on the gender of the individuals who make (or influence) the school attendance decision, and the economic and social household structure in which this decision is evaluated. The In the 1997 EDS, results suggest that while income support is particularly important for respondents who drop keeping girls in school, social and economic stability in the household is out of schools were the primnary factor for boys' school attendance. Furthermore, the asked their reasons for mother's education is more important than the father's in both her son's dropping out of and her daughter's school attendance. (see "Household Structure, Gender, and the Economic Determinants of School Attendance in Argentina" by Walter school. They were Sosa Escudero and Mariana Marchionni, Background Paper No. 7) asked in particular whether the following factors were key reasons: (a) school problems(schedule, school difficulties); (b) economic difficulty (cost of education, need to work); (c) personal motives (lack of interest, need to care for family); and (d) others. Respondents who answered positively were then asked to select from a list of causes one that they would consider the principal factor. It would appear from Table 5 that school problems are not considered a major cause of early school leaving. Only 11 percent of the respondents aged 15-24 say it is a principal reason. In contrast, about half considers "personal motives" a key factor, of which 77 percent say lack of interest is the main cause. It is difficult to interpret what the underlying reason behind this response. A common interpretation is that school is boring. Yet, as mentioned, only a few of the respondents blame schools for their decision to drop out. Moreover, an equally small percentage find education useless. A close second (43 percent) to "personal matters" is economic difficulty. This is a more revealing response. Seventy-three percent of those who answered positively to this item report that they left school because they needed to work. As might be expected, economic difficulty is a much more salient problem for the low income families. Analysis of youth earnings strongly confirms the validity of the above response. We calculate that the potential income that a poor, out-of-school youth would have to forego to be a full time student would be about $ 114 per month.51 This amount constitutes about 30 percent of the income of a poor family. This means that for the poor family sending 15-19 year olds to school would significantly reduce their ability to cope with economic difficulties. 51 For the source of these and other calculations, see V. Paqueo, Background Paper No. 4. These calculations assume that a school dropout could find employment. 111 Table 4.5: Reasons for Leaving School (Age 15-24): 1997 EDS (a) Leaving school due to (ii) Income quintile (a) 1 2 3 4 5 I.Schoolproblems 10.7 10.3 9.8 11.1 9.5 6 2. Economic difficulty 43.0 49.8 43.3 41.9 36.3 28.5 3. Personal motives 51.5 45.3 49.7 56.7 55.7 58.3 4. Others 8.3 7.4 9.0 7.8 6.1 9.2 An examination of actual school enrollment behavior of 15-19 year olds, using multivariate logic analysis (See Background Paper No. 4 for details), further reveals the following: * as usual parental education and income tend to have significant positive effects on schooling. * distance to public secondary school remains a significant cost (disutility) factor that discourages some children from schooling participation. In this regard, children in the first quintile generally live further from school than those in the fifth quintile. - provincial unemployment rate among household heads is significantly negatively related to enrollment. This suggests that in areas where the unemployment rate among household heads is high, children's human capital formation suffers. * the relationship between enrollment and the provincial average standardized test score in math and language (SCORE) appears to be surprisingly negative. This variable is arguably an indicator of the quality of the provincial learning environment. Therefore, a positive correlation between SCORE and enrollment decision should be expected. The result, however, turns out to be the opposite. One interpretation is that some provinces have highly selective educational systems that implicitly or explicitly discourage average and disadvantaged students from enrolling in post-basic education. If this interpretation is correct, it can be further argued that the discouragement effect of a highly selective process would probably tend to be greater among poor children. Regression results indicate that indeed this effect (in absolute value) increases with lower family income. Benefits from higher levels of education. The question regarding education benefits is important not just for determining the social desirability of asking children to stay in school longer, but also in figuring out whether the incentives for doing so are adequate. 112 Box: School-based Parent Organizations One factor affecting the overall quality of education is the reliance of many provinces on the cooperadoras-school-based parent organizations-for the provision of everyday supplies and school maintenance. Parents are expected to make regular financial contributions to the cooperadoras to cover these costs, and most cooperadoras establish a quota for such contributions. In general, financial contributions to the cooperadoras reflect the income level of the parents whose children are served by the school. Furthermore, there are indications that relatively few parents (roughly one-third) actually contribute to the cooperadoras. For example, a survey of 65 rural and urban schools serving communities of different socioeconomic backgrounds across the country indicated that the average cooperadora quota is US$8 per month during the school year, and that, on average, only 35% of parents pay this quota. Based on these findings, a school of 250 students can be expected to have a disposable income of only US$7,000. Source: "Challenges for the Net Phase of Education Reform in Argentina" World Bank, Oct. 30, 1998. Table 4.6 presents estimates (based on the 1997 EDS) of private marginal rate of return (RR) by level of education. They show RR of 2.5 percent, 10 percent and 29 percent for primary, secondary, and tertiary education, respectively. This means that for those who had already finished secondary education, it would be highly profitable for them to enroll in tertiary education. For secondary education, however, the marginal rate of return is just about equal to the usual rate of discount of 10 percent. If out-of-pocket costs such as transportation and other school expenses are taken into account, an additional year of secondary education would not be a very attractive investment. This is one possible explanation for the sharp reduction in the percentage of children completing higher levels of schooling after primary education as observed in Fig. 1. Table 4.6: The Marginal Rate of Return to Education for Various Levels: Age 20-65 Regression Results for Log of Wages on: Coefricient t-value Intercept 0.1833 7.63 Age 0.0116 23.92 Gender (male=1) 0.1249 10.94 Illness -0.0717 -5.87 Education completed (S) S = I to 6 0.0246 9.37 S = 7 to 13 0.1013 27.42 S = 14+ 0.2905 53.91 Adjusted R-square 0.201 Sample size 18,499 Database: 1997 EDS. Further analysis reveals that the rate of returns to education is probably lower for individuals with poor parents, as measured by the education of the head of the family. Rough estimates show that 20-29 year olds whose parents had completed tertiary education have a marginal income growth from a year of tertiary education that is about 40 percent higher than corresponding individuals with parents whose educational 113 attainment is only primary education. Likewise, students whose parents have only primary education have an incremental rate of return from a year of secondary education that is 60 percent lower that those whose parents have tertiary education.52 These estimates need to be interpreted with caution. since the sample is limited to children aged 20-29 and, hence, the estimated wage effects might not yet fully reflect the impact of education on income. Nevertheless, the results partly explain school leaving differences between poor and well off children. The latter have stronger economic incentives to stay in school longer. Those rate of return differences also indicate that the income benefits from public education spending tend to be greater for higher socioeconomic classes, weakening the income inequality reducing effects of education. Enrollment Changes, 1992-97. The fact that the poor have lower rates of return from the same education might explain recent changes in enrollment rates at the secondary and tertiary levels. At the secondary level, enrollment rates between 1992 and 1997 have decreased, while enrollment rates for other deciles have increased (see Table 4.7). For higher education, the shift is even more pronounced. Enrollment rates.for those in the lower 40% all declined, while those for the upper 60% increased. Hence, the richer groups are responding to higher returns to education, while the poorer groups are not, potentially perpetuating the gap between rich and poor. Table 4.7: Enrollment Rates by Decile, Secondary and Higher,1992-97 Secondary Higher Decile 1992 1997 1992 1997 1 70.5 62.1 22.7 8.7 2 74.9 68.9 20.5 10.7 3 80.3 83.7 17.7 13.3 4 80.0 85.4 22.3 16.1 5 82.5 103.6 20.8 22.5 6 76.1 95.9 21.9 28.0 7 92.2 105.3 27.4 37.8 8 96.4 102.4 38.0 53.8 9 95.2 117.1 54.4 75.5 10 106.2 115.9 80.8 82.0 Source: Maria Echart, " Educaci6n Y Distribuci6n Del Ingreso", FIEL, 1999 Scope for increasing education benefits to the poor. The question at this point is how much room is there for improving the benefits poor children can derive from public education spending. Level of per student spending. It is often argued that one way of raising returns to education is for government to invest more per student so that school children can be provided with better teachers, educational materials, equipment, and classroom environment. The economic issue is the income payoff from such policy. We explore this issue by examining whether provinces that spend more resources per student are 52 See Background Paper No. 4. 114 associated with higher rates of returns to education - assuming that current differences in expenditures per student reflect persistent patterns in the past. The results reveal that the rate of wage increase from secondary education among 20-29 year olds is significantly positively correlated with high public secondary education expenditure per student (PEXS). Moreover, its coefficient suggests that a 10 percent increase in PEXS is associated with about 3.5 percent increase in the wage rate of a 20-29 year old with complete secondary education. Assuming this estimate is about right, rough calculation indicates that a secondary education graduate would earn about 200 pesos more annually during his working life, if the government were to spend about 800 pesos more per graduate. At 10 percent rate of discount, the present value of the additional stream of earnings of 200 per year is about 2,000 pesos - an amount that is more than the increase in cost per graduate. Table 4.8: Distribution of Public Education Expenditure Benefits by Level of Education and Income Status (Percent), 1997 Quintiles Total Primary Secondary Tertiary Bene Pop Bene Pop Bene Pop Bene Pop Fits 6-24 fits 6-12 Fits 13-17 fits 18-24 First 25.5 24.7 32.6 29.3 20.9 26.3 10.4 18.4 Second 23.4 22.7 25.1 23.5 24.0 24.7 12.1 20.3 Third 20.8 19.8 19.1 18.8 23.4 19.5 17.4 21.2 Fourth 18.0 18.0 15.1 16.1 19.3 16.7 26.0 21.1 Fifth 12.3 14.8 8.0 12.3 12.3 12.9 34.1 19.0 Total 100 100 100 100 100 100 100 100 Poor 40.4 38.5 47.2 42.7 37.9 42.9 17.6 30.4 Non-poor 59.6 61.5 52.8 57.3 62.1 57.1 82.4 69.6 Source: 1997 EDS Benefit incidence of public education spending. With universal enrollment, the distribution of primary education benefits appears to be somewhat favorable to the poorest group, as higher income families tend to send their children to private schools (see Table 4.8). The first quintile, which accounts for 29.3 percent of primary school age children, takes 32.6 percent of primary education benefits. In contrast, the share of the richest quintile, which has a population share of 12.3 percent, accounts for only 8.0 percent. The distribution of public secondary education benefits in contrast appears to be pro-middle class and anti-poor. The share of the middle income (third and fourth) quintiles in secondary education benefits is greater than its population share. The favorable distribution to the middle class is clearly at the expense of the poorest income group. While the richest group breaks even, the first quintile, which has 26.3 percent of the relevant population, gets only 20.7 percent of the benefits.53 53 The extent to which the distribution of secondary education benefits is biased in favor of middle income classes is probably understated. The reason for this is that private education subsidy and differences in per student public spending within each province, which have not been taken into account in the above analysis, is likely to be favorable to these families. In addition, it is possible that inclusion of rural children, 115 As in most other countries, tertiary education clearly favors the upper income groups. It is only the two richest quintiles whose share of benefits exceeds their population share. The rest of the income groups get a disproportionately lower percentage of the benefit distribution. Only 17.6 percent of the tertiary education benefits go to the non-poor, which accounts for 30.4 percent of the relevant population. The above analysis probably underestimates the degree of inequality in the distribution at this level for the same reasons as in secondary education.In summary, the results indicate that there is considerable scope for improving the share of the poor in education benefits. Efforts to Improve Equity in the System: Compensatory Programs The existing fiscal distribution process (coparticipaci6n) establishes some criteria (i.e., NBI levels, etc.) in an effort to compensate for socioeconomic inequities across provinces and promote more balanced development within the country. Concerning education, the Ley Federal de Educaci6n established equitable access to educational services as a main principle underlying Argentina's education system, with the goal of providing quality education and ensuring high achievement levels for all children. So far, most of the efforts are federal and there is aimost no specific provincial programs to address inequity55. The federal government uses two main instruments to reduce inequities in the education system: the Pacto Federal Educativo and the Plan Social Educativo. A principal objective of the Pacto Federal Educativo is to help the provinces meet the physical demands of extending mandatory education from 7 to 10 years. The Pacto Federal Educativo has allocated approximately US$398 million for this purpose between 1995 and 1998, with about 83% for infrastructure and 18% for educational equipment. In 1996, Pacto Federal Educativo resources represented approximately 11% of total federal education expenditures in pre-university education.56 Because most children who are not enrolled in secondary education come from underprivileged households, this major effort to expand physical capacity is expected to benefit poor children and youth in particular. The Plan Social Educativo (PSE) was created in 1993 to directly address inequities in the education system by targeting -resources to schools serving which are excluded in the 1997 EDS survey could further worsen the distribution of secondary education benefits. 55 One of the most ambitious programs to address inequity is the scholarship program for the Polimodal that the Province of Buenos Aires has implemented this year, .100.000 polimodal students will receive $ 100 per month during the academic calendar. The student awarded are selected take into account the social- economic situation of their parents: income, parent as a unique family support, etc. 56 This percentages were calculated on the basis of: (i) Total national expenditures in education for 1996 MCE's expenditures minus national expenditures on university education (based on the figures provided by Secretaria de Politicas Universitaria, MCE); and (ii) Total Provincial Expenditure in education for 1996. It is important to note that, due to the fact that these percentages may be lower-bound estimates because some resources at both the national and provincial level are allocated neither to primary nor secondary education (Secretaria Tecnica, etc.). 116 underprivileged children. The PSE serves 11,820 Inicial, primary and secondary schools and approximately 6 million children, or 75% of total enrollnent at these levels.57 Since its creation, the PSE has disbursed a total of US$703 million. In 1997, the PSE represented approximately 20% of all federal education expenditures and approximately 2% of total expenditures in the sector.58 Most of this investment has been directed to Inicial and primary schools. The PSE comprises three kinds of programs; improvements in school infrastructure; quality improvements in schools including programs targeted at indigenous students, rural schools, adult and special education; and a national program of student scholarships to keep children from poor families enrolled in secondary school The Ministry of Culture and Education (MCE) has set three criteria for the allocation of PSE resources: (i) funded programs must target the poor (i.e., be used in provinces or regions with higher NBI concentrations); (ii) the programs must reflect federally defined priority needs (e.g., broadening access to mandatory schooling or improving basic material conditions at the school level); and (iii) recipients must work in cooperation with the provinces (i.e., participating schools are picked by the provinces according to federal guidelines). The adequacy and effectiveness of PSE have not yet been rigorously examined, although feedback received have been positive. Indeed, it appears that PSE is strongly pro-poor. The 1997 EDS reveals that 60 percent of the PSE beneficiaries come from the first quintile and 40 percent from NBI families (Table 4.9). Moreover, the two lowest income quintiles account for over 80 percent of the beneficiaries. Still, it can be seen that about half of the children in the poorest quintile have not received assistance from PSE. The question then is what would be the impact on these children if they were provided with PSE assistance. Given the importance of this question for resource allocation, it is highly desirable to undertake a rigorous and comprehensive evaluation of PSE.59 Table 4.9: Coverage Rate and Distribution of Beneficiaries by Income Class Total Per capita income quintile Poverty 1 2 3 4 5 NIBI No NBI Coverage rate (a) 42.0 51.8 36.6 33.9 23.8 25.3 55.7 36.2 Distribution of beneficiaries 100 59.9 21.3 11.3 5.5 2.0 39.5 60.5 C(a) percent of children enrolled in public school. Source: 1997 EDS. SIEMPRO Tabulation Susana Beatriz Decibe, "Mejoramiento de la Educaci6n prar la Calidad y la Inclusi6n Social: El caso de la transformaci6n educativa en Argentina," prepared for World Bank conference on "Construyendo el futuro de Arnerica Latina: Asociaci6n publico-privada para la educaci6n," Washington, DC, June 4-5, 1998. And NMCE, Programa Estudio de Costos del Sector Educativo. 5"See note 55 about the way these percentages were calculated. 59 Other areas for potential reform include restructuring incentives to encourage good teachers to work in poor areas, and making the curricula more "pro-poor". See A. Fiszbein, "Institutions, Service Delivery, and Social Exclusion: A Case Study of the Education Sector in Buenos Aires" World Bank, December, 1999. 117 Summary and Conclusions. This chapter suggests the following conclusions and suggestions: First, there is considerable room for poor children to benefit more from public education spending. Second, the private rate of return to tertiary education of about 30 percent is sufficiently high to make it attractive for parents to send their children to college. It should be feasible to reduce the tertiary education subsidy for the well off to finance additional scholarships for low income secondary education graduates to go to college and at the same time keep incentives attractive for well off families to continue investing in tertiary education. As shown above, a large proportion of drop outs leave school because they need to work. Third, in contrast to the above finding, the incentive for primary education graduates to get additional years of secondary education, appears weak, particularly if the likelihood of their going to college is very small. Two strategies for making completion of secondary education more attractive appear promising. One possibility is to increase spending per secondary education graduate to raise its rate of return. The other is to reduce the cost of going to secondary schools, such as providing transportation allowance and scholarships to those deserving social assistance. Fourth, returns to education differ significantly between low and high socioeconomic status children, as measured by parental education. This phenomenon and the pattern of distribution of education spending among income groups undermine the effectiveness of education policy as a tool for reducing income inequality. Plan Social Educativo coupled with above-mentioned proposals could attenuate this educational inequality problem. It is, however, important to evaluate its impact and address the coverage issue. Finally, high educational standards are desirable; provinces, however, need to pay careful attention to how this is being achieved. If improving standards is done through selection of bright students as a way to ration limited places in good schools, it could discourage overall school participation, particularly among the lower income groups. 118 CHAPTER V: POVERTY AND HEALTH Health Status and Poverty Are the poor at a disadvantage in terms of death, disease and disability? As elsewhere, the poor in Argentina carry a disproportionate part of the total burden of disease and are disproportionately affected by a number diseases. The overall health indicators of Argentina are good when compared with those of other countries in the Region, and have improved markedly over the past ten years. In 1995, life expectancy stood at 76 years, 12 more than in 1960, the crude death rate at 8 per 1,000 population, infant mortality at 22 per 1000 births in 1995, down 16 percent from 1990, and maternal mortality ratio at 14 per 10,000 live births in 1995, down by 14 percent from 1990.60 However, these rates are worse than those of other middle income countries, and lower than one would expect for a from the economic and education indicators. For example neighboring Chile, with lower income than Argentina, had in 1995 an infant mortality rate of 12 per 1,000 live births and maternal mortality ratio of 6.5 per 10,000 live births. The relatively good national health indicators also hide significant variations between different income groups, the poor having much worse health status than the rich and having a different pattern of death, disease and disability. Overall, about 23 percent of the poor and rich report having been ill in the past 30 days; and about 16 percent of those from the lowest household income quintile report some form of chronic ailment, while as many as 23.5 of those in the upper quintile report chronic suffering.61 Objectively, when one looks at the burden of disease in terms of years of life lost62 due to premature death, evidence shows a quite different picture: the poor die earlier than they should, and are disproportionately stricken by largely preventable Group A diseases, i.e. communicable, maternal, perinatal and nutritional conditions. In fact there is a strong negative association between per capita household income and the years of life lost per 100,000 population due to syphilis, diarrhea, tetanus, abortion, protein- energy malnutrition, and with infant, neonatal and maternal mortality (see Table 5.1). Other negative associations exist, albeit weaker, between poverty and tuberculosis, bacterial meningitis, Chagas' disease, hypertensive disorder of pregnancy, and congenital heart diseases. Group A diseases account for 19 percent of total years of life lost in lower income and for only for 5 percent in higher income provinces. 60 Bos, E, Hon, V., Maeda, A. et at: Health Nutrition and Population Indicators, a statistical handbook. (Washington, DC: The World Bank, 1998) 61 Source: SIEMPRO, Social Development Survey 1996/97. 62 The calculation of Years of Life Lost in this study are very approximate. There are no corrections for underreporting, incorrect coding and other factors. Since the study is comparative between provinces, the lack of correction for the mentioned factors is less important because there are no reasons to believe that their would be a bias by province. 119 Table 5.1: Association between Average Per Capita Household Income and Years of Life Lost per 1,000 Population Disease which affect Linear Diseases which affect Linear disproportionately the poor Correlation disproportionately the rich Correlation Tuberculosis -0.361 Colon and rectum cancer 0.347 Syphilis -0.415 Pancreas cancer 0.326 Diarrhea -0.455 Breast cancer 0.387 Tetanus -0.455 Ovary cancer 0.361 Bacterial meningitis -0.334 Endocrine disorders 0.347 Chagas' Disease -0.383 Rheumatic heart disease 0.316 Hypertensive disorder of -0.370 Inflammatory heart disease 0.472 pregnancy Abortion -0.435 Protein-energy malnutrition -0.450 Congenital heart diseases -0.319 Cervix Uteri Cancer -0.318 Infant mortality ratio -0.62 Neonatal mortality rate -0.56 Maternal mortality rate -0.48 Proportion of low birth weight -0.58 babies (<2,500 g) On the other hand, the rich seem more vulnerable Group B diseases, i.e. non- communicable diseases. In general per capita household income is positively associated with non communicable diseases and the association is strong for several cancers (i.e. colon and rectum, pancreas, breast, and ovary), endocrine disorders, and rheumatic and inflammatory heart disease. Rich and poor seem to be equally affected by injuries. Differences in health status are consistent with differences observed in the major determinants of health: environmental conditions (i.e. access to adequate water and sanitation), health related behaviors (i.e. diet, exercise, smoking habits, alcohol and other substance abuse, and sexual behavior), and health care utilization. The relatively high incidence of communicable diseases and of infant mortality among the poor is likely to be determined by the direct environment conditions in which the poor live, i.e. sanitary conditions are likely to be associated with the high incidence of diarrhea among the poor, while housing conditions, crowding and malnutrition are likely to be associated with Chagas' disease, tuberculosis, leprosy, meningitis. A high proportion of the poorest live in dwellings of very low quality and with inadequate sanitation: 10 percent of those in the lowest per capita family income group quintile live in "precarious" dwellings (i.e. casilla or rancho), 42 percent have inadequate water supply and 47 percent have inadequate sewerage systems, two or three times higher the national average. In addition, for poorest family income quintile, the average number of people per dwelling is 4.8, almost twice as many as for the upper income families. The relatively high incidence of abortions, hypertensive disorder of pregnancy, syphilis and cervix uteri cancer, low birth weight babies, perinatal and maternal mortality are likely to be associated with inadequate coverage of poor women with quality reproductive health services, which we will describe later under the relevant section. 120 Demography of Poverty The poorest households in Argentina tend to be younger, larger and have more children. The average age of the lower income quintile household is just 25 years, while the mean age of the upper income quintile households is 41 years. The average size of the lower income quintile household is 5.1 people and only four percent are single person households. Sixty percent of these households have children aged 0-19 years and only I percent of these households have no children. By contrast, while the average size of upper income quintile households is just 3.1 persons and about 29 percent are single person households. Also, only 36 percent have children aged 0-19 years and 10 percent have no children at all. Table 5.2: Average family age, size and com osition by household income quintile Per capita Household Income Quinhiles 1 2 3 4 5 Average Age 25 31 35 37 41 Average family size 5.1 4.2 3.8 3.5 3.1 Percent of households with no children 1 1 3 3 10 Percent of households w/ children aged 0-19 66 59 53 48 36 Estimated Crude Birth Rate(births per 1000)63 45 32 23 18 14 Estimated Fertility Rate(births per 1000 women aged 4.2 2.9 2.0 1.7 1.3 1544)64 ________________________ Source: EDS. Social Development Household Survey, 1997 and staff calculations. Total fertility rate was 2.7 per 1,000 women aged 15-45 in 1995, which is about the average for the Region, but are high when compared with other middle income countries such as Chile and Uruguay and higher than one would expect from the country economic and educational situation. This average parameter hides significant variation in the country and a negative association between birth rates and income level. While birth rates are generally not available by income groups, they can be approximated from the Social Development Survey (see Table 5.2). These estimates show a crude birth rate that varies from 45 for the lowest quintile, to 14 for the upper quintile. Likewise, the fertility rate varies from 4.2 per thousand in the lowest to 1.3 per thousand women in the richest quintile. Unfortunately, there is no way to estimate death rates, and the rate of population growth by quintile. However, from these estimates we can see that there will be a natural tendency for poorer families to increase faster than richer families, perpetuating a cycle of poverty. Given these demographics, programs that provide maternal and child care, as well as reproductive services, will be naturally tend to benefit the poor. 63 This estimate is made by taking the percentage r of children in 1997 aged 0-2, dividing by 2, and multiplying by 10, to obtain a rough estimate of the crude birth rate per 1000 population. Since the average age of the group 0-2 can be presumed to be one year, the calculated rate is effectively net of infant mortality. 64 Calculated with the same method as the crude birth rate, except the denominator is the number of women aged 15-64. 121 Health Care Coverage and Utilization Are the poor appropriately covered by essential public health services and do they have access to adequate personal health care? Available health sector resources are more than adequate for providing a comprehensive package of public health and personal health services for the entire population. The fact that avoidable, Group A disease incidence and mortality are relatively high, in spite of plentiful health sector resources suggest that there may be problems with the existing organization and quality of health care. Personal health care in Argentina is delivered by both the private and the public sector. The private sector responds mostly to the needs of those with mandatory or private health insurance and the public sector is meant to cater for the needs of the poor and the uninsured. The system has been heavily biased towards specialized and hospital care, but at present there is a move towards strengthening general and ambulatory health care network to improve the efficiency and quality of the system. About 54 percent of hospital beds are in the public sector, but the proportion can be as high as 70 percent in provinces where the population is sparse. Most doctors work part time in both the private and public sectors. As one would expect there are many overlaps: insured patients who seek service in the public sector, mostly for complex or expensive interventions for which the insurance co-payment would be very high, and uninsured patients who choose to pay out of pocket for convenience, to avoid waiting lists or to get what they may perceive as more personal care. The delivery of public health services is largely the responsibility of provincial public health services, with support from some Federal programs. Health Care Supply Health care resources in Argentina are generally plentiful and relatively accessible in every part of the country. Although the poorest in Argentina have more difficulties in accessing health facilities such as hospitals and pharmacies, more than 70 percent live within walking distance from such facilities, and even the poorest provinces the available number of doctors and hospital beds should be enough to provide a comprehensive package of public health and personal health services to the entire population. The health system in Argentina is well provided and distributed: about the 70 to 77 percent of the population lives within 10 blocks from a health center or a hospital, respectively, and 87 percent of the population does not have to walk more than 15 minutes (1 km) to the closest pharmacy. However there is a very strong negative association between per capita household income and access to public hospitals (see Fig.5.1). 122 Figure 5.1: Household Income and Access to Public Hospitals -- 35- o o 30 0 15 lo o O A 5 5 0 0 200 400 600 800 1000 1200 1400 Per Capita Household Income With an average of 367 inhabitants per doctor and 209 inhabitants per hospital bed, Argentina has enough, if not too many, doctors and hospital beds. The average number of doctors per 1000 population hides significant variation by province, from a low of 1 per 1000 population in poor Formosa to almost 9 per 1000 population in the wealthier City of Buenos Aires, as well as a strong association between the number of doctors available and the average per capita family income (R= 0.50). This situation is not as favorable for the number of specialized nurses and nursing auxiliaries: there are over 1800 inhabitants per specialized nurse and a ratio of 5 doctors per nurse. Even if we add the number of nurse auxiliaries to the number of nurses, supply is still very low, with 473 inhabitants per nurse or nursing auxiliary and more doctors than nurses and auxiliaries together, suggesting that many doctors must be carrying out functions which would normally be performed by nurses or auxiliaries. As for the number of hospital beds, the range of variation is much smaller, from 3.3 beds per 1000 population in San Juan and 7.8 in the City of Buenos Aires, but there is no consistent association between the availability of public hospital beds per capita family income. About 54 percent of hospital beds are in the public sector, but the proportion can be as high as 70 percent. As public hospitals are supposed to meet the needs of the uninsured poor, one would expect that in poorer provinces there would be relatively more public hospital beds, but such association does not exist (R= 0.06). Coverage with essential public health programs Overall coverage of the population with essential public health programs is high in Argentina. If we take prenatal care as an example, only about three percent of women do not benefit from some form of prenatal care. However, the poor are less likely to benefit from such essential public health programs, and when they do the quality of 123 service is lower. As in other countries, the "inverse care law"65 applies -- those who need the most get the least. If we use prenatal care as a tracer for the quality essential public health programs, there is a direct negative association between average per capita income and the probability of having had less than five prenatal exams (see Fig. 5.2) The same applies for immunization of children, for whom there is a strong association between average per capita income and the probability of having been correctly immunized (0.39 Better (state of the art) equipment and infrastructure - compared to the dilapidated state of public hospitals > Reception and waiting rooms of the private facilities that are much more comfortable and relaxing (with TV sets), reflecting client-orientation and sensitivity to patients' dignity > Much shorter waiting time in terms of appointment and return of diagnostic findings Knowledgeable local observers in Tucuman expressed the opinion that private health care providers could produce better quality services at a cost that is probably less than that of public hospitals. For example, birth delivery at the Hospital de Maternidad is about $400, compared to $250-300 in private hospitals. The clients of the above-mentioned private hospitals are from all economic classes. They include poor patients whose Obra Social has entered into a service contract with these providers. The point here is that for the same amount of resources being spent now by the govermnent on public health facilities, the poor should be able to get better quality services and more timely attention, which in turn should lower patient cost of waiting and repeat visits. Source: Argentina: Improving Health and Education in Tucuman, World Bank, 1997, Report No. 17758 AR To address the above issue, the above-mentioned provinces decided to adopt a health sector reform strategy, with World Bank support, that would concentrate provincial government health expenditures on the uninsured poor and change the payment mechanism. The hope is to both improve health care protection of the poor and raise the quality of services rendered to them by creating "choice and voice" mechanisms.72 The key assumption of the reform is that aside from ensuring that health sector "public goods" are financed adequately, the fundamental role of provincial government is to ensure that all citizens have access to good quality "basic health services". Given Argentina's social health insurance system, this largely means making sure that the uninsured poor have access to those services. A corollary to this principle is that the government health subsidy should be used to pay for services actually rendered to target beneficiaries. 72 For a description of the design of the PRL II strategy and its rationale, see World Bank (1997), Argentina: Improving Health and Education (Salta, Tucuman and San Juan). 132 Under these reforms, it is envisioned that in the long-run provincial funds would eventually no longer be given to public hospitals directly and automatically. Instead, it would go to some purchasing agency or fund holder (FH), which would be responsible for negotiating with public and private service providers for delivery of a pre-specified package of health services to a pre-determined group of eligible population ( the uninsured poor) on the basis of some agreed price. This FH could be an autonomous body under the supervision of the Ministry of Health or could be an agency that would handle health service contracts for both the poor and the provincial public workers (members of Obra Social Provincial). Under this vision, public hospitals, which would be given management autonomy, earn income to pay for their staff and their operations. They would do this by charging the above-mentioned FH and other insurance agencies' capitation or specific fees, reflecting the full cost of providing good quality services to their members. The idea is to give to FH the financial means to motivate and enable public hospitals to provide poor patients improved health service. Ideally, public hospitals would gain a network of service providers (a mix of public, private and municipal providers) to ensure the most cost efficient delivery of services). However, hospital self management reforms may put the poor at a disadvantage. Because autonomous hospitals are now allowed to collect fees from insured patients, and can retain the respective income and use part of the additional revenue to staff as performance bonuses, they also have an obvious incentive to increase their share of paying patients by offering them preferred services, i.e. better scheduling, less waiting times, jumping over waiting lists, and better amenities. Ideally, recent hospital autonomy reforms should be accompanied by some form of health insurance for the poor and uninsured that would put the poor one a more equal footing even in public health facilities Conclusion As elsewhere, the poor in Argentina carry a disproportionate part of the total burden of disease and are disproportionately affected by a number diseases. Diseases which affect disproportionately the poor are mostly avoidable communicable disease, maternal, perinatal and nutritional conditions, which could be controlled with universal access to a comprehensive package of quality public health and personal health services costing no more than US$30 per capita, about eight percent of 1997 public health expenditure per capita. The poorest households in Argentina tend to be younger, larger and have more children. Reproductive health services need remain a priority in a poverty alleviation strategy. Available health sector resources are more than adequate for providing a comprehensive package of public health and personal health services for the entire population. The fact that avoidable, Group A disease incidence and mortality are relatively high, in spite of plentiful health sector resources suggest that there may be problems with the existing organization and quality of health care. 133 The entire population is covered against catastrophic financial loss caused by disease, either by mandatory or voluntary private insurance, or by public health services. The issue is one of equity, as those with mandatory or voluntary insurance benefit from choice of provider, and better quality of service. Recent hospital autonomy reforms, if not accompanied by some form of health insurance for the poor and uninsured could put the poor at further disadvantage, even in public health services. There may be a problem with health insurance for those who become unemployed or are in transition between two different jobs. Total public health expenditures are sufficient to finance a very gefierous package of public health and personal health benefits for the entire population. Public funds available to finance personal health care for the uninsured population are about one half of those available to finance care for the insured patients. The share of the public health expenditures which is managed at the Federal level maybe too low for promoting national health policies. CHAPTER VI: RURAL POVERTY For a country at its level of development and sophistication, surprisingly little is known about the extent of rural poverty in Argentina. In part, this reflects the fact that most poverty analysis is based on surveys that cover only urban areas (see Chapter 1). This chapter is based on an analysis of a 1996 rural household survey undertaken by the Secretariat of Agriculture, Livestock, Fisheries and Food, with World Bank support, in the provinces of Misiones (the Northeast) and Salta (the Northwest), in order to provide a deeper analysis of rural poverty missing from previous studies of Argentine poverty.73 Basic Head-count Analysis of Poverty. All available poverty indicators confirm the high degree and severity of rural poverty in the two provinces sampled, whether compared to poverty in urban areas or the more prosperous regions of Argentina. In these two provinces, the poverty line in mid-1996, as estimated from expenditure data, came to about US$ 950 equivalent per capita per annum, and the line of indigency or severe poverty about US$ 600. About 77 percent of rural inhabitants in Misiones and 73 percent in Salta fell below the poverty line, as compared to 29 percent of the population in all urban areas (see Chapter 1). Moreover, some 31 percent in Misiones and 38 percent in Salta fell below the indigent line, as compared to 7 percent nationally. The incidence and severity of poverty in terms of NBI for these rural areas of these two provinces were substantially higher than was found for the corresponding provincial cities in these provinces (see Table 6.1). 73 This chapter is based on "An Analysis of Rural Poverty in Argentina", November 10, 1997, a paper prepared by Tom Wiens, LCSES. 134 Table 6.1: Consumption and NBI-Based Poverty, Misiones and Salta, 1996 Location Percentage of Population Below Line of: Poverty Indiaency NBI (sample) NBI (urban)' Misiones 77 31 75 32 (1992) Salta-- 73 38 88 27 (1992) Urban areas 30 7 19 (1992) 1/ INDEC, Encuesta Perinanente de Hogares (base reducida), October 1992. National statistics suggest that agricultural producers are a distinct minority of the rural poor, and that high proportions of farm producers are also engaged part-time in non- agricultural work. In addition, rural families depend partly on remittances from members working in urban areas and partly on pensions of the retired. Table 6.2a: Per Capita Income of Farm Households, Misiones and Salta (Pesos per month) Indigent Poor Non-Poor Total Farm Income 168 471 980 629 Crops 135 396 807 520 Livestock 12 3 49 26 Subsid. Products 21 72 124 84 Agric. Wages 119 97 106 106 Non-agricultural: Wages 149 137 273 201 Self-employment 43 47 168 101 Transfers 56 90 96 85 Rent and Other 1 4 66 31 Total 536 845 1,689 1,153 Household size 7.39 5.51 3.67 5.13 Pct. of farm hh 24 32 44 100 These points are confirmed from the sample data for Misiones and Salta, even though nearly half of rural households there have some income from agricultural production (Table 6.2). Agricultural income contributes 11 percent of inflows to the poorest, 26 percent to the richest rural households. Households depending entirely on non-farm sources of income are slightly more likely to be poor than other households. (The strictly non-farm poor often reside in small community centers, where much of the overt rural unemployment can be found.) Less than three-quarters of household heads with farms consider farming their primary occupation, and less than half of their income comes from farming. Non-farm wage labor accounts for 20-30 percent of net income except among the rich. 74 Or $79 and $50 per month, respectively, versus the poverty line for 1996 in urban areas used in this report of $158 per month, and a indigent line of $68. 135 The rural poor typically have larger families--7-8 members for the bottom quintile--but earn proportionately less per member. Thus the ratio between average per capita incomes in the top and bottom quintiles is about 4:1 for non-farm households, but 6:1 for farm households. Table 6.2b. Per Capita Income of Non-farm Rural Families by Source Indigent Poor Non-Poor Total Agric. Wages 363 425 1,054 587 Non-Agric. Wages 219 289 449 310 Self-employment 43 57 211 97 Income Transfers 96 206 -60 86 Rent andOther 1 0 0 0 Total 722 978 1,654 1,081 Household size 6.61 4.77 3.20 5.00 Pct. of non-farm hh 38 33 30 100 Some socio-cultural characteristics, including indigenous status, predispose a household toward poverty, though the determining factors may be education and occupational opportunities. Gender (in the sense of female-headed households) is not one of them, although female-headed farm households are more likely to be poor and gender is significant in determining opportunities for off-farm earnings. Educational levels of the rural population are very low, the majority of household heads and their spouses having not completed primary school (Table 6.3). Educational achievement levels, at either extreme, are partially transmitted across generations, but marked improvement has occurred and continues with the current generation. Table 6.3. Educational Achievement of All Household Members, 18 Years or Above School Attainment Indigent Poor Non-Poor Male Female Male Female Male Female Illiterate2 13.5 20.4 11.2 21.0 5.3 10.6 Primary Incomplete 40.4 38.0 40.5 39.5 42.9 31.0 Primary Complete 34.8 27.6 37.1 26.3 32.7 32.9 Secondary Incomplete 7.4 9.2 5.4 6.2 8.6 8.8 Secondary Complete 2.8 3.6 3.5 4.9 5.3 6.9 Tertiary or University 0.4 0.8 1.9 2.1 4.9 9.7 Other 0.7 0.4 0.4 0.0 0.4 0.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 IIncludes children still in school. 2 Icluding some with 1-2 years of primary school. Non-formal training is scarce (received by less than four percent). Technical assistance is more frequently received (26 percent received it from individuals or through group meetings). The geographical distribution is very uneven (with training common in Salta, technical assistance in Misiones) and the class distribution strongly skewed in 136 favor of higher quintiles (who may seek it more aggressively, or be joiners of organizations which supply it, as well as being able to pay). Cooperatives, farmer's associations, and labor exchange represent three types of social organization, all much more common in Misiones than Salta. Labor exchange is equally common among all strata of farmers, but also has no measurable impact on economic welfare. Cooperatives and farmer's associations tend to cater to the better-off, well-capitalized farmers (who join cooperatives) or better managers (who join farmer associations). Where these organizations are established, farmers are more likely to be members; and member farmers are also much more likely to receive technical assistance, though not necessarily to have received credit. But the poor are less likely to have access to such organizations, or join them when available, and this explains the apparent bias of technical assistance delivery in favor of the non-poor. About one in six households (one in ten among the indigent, one in five among the non-poor) reported some form of accumulation of savings (or assets), mostly in buildings or stocks of food and forage. Hardly any reported cash savings, and the annual value of all savings averaged only $32 (with about equal numbers of savers and dissavers). Only one in 20 of the villages or towns sampled in the community survey had a bank branch. Overall there was little evidence of opportunities to develop savings institutions. Even fewer households (11 percent) had ever sought cash loans, these tending to be better-off and better-educated. Those who had not tend to be poor, female-headed, and/or indigenous. They typically thought they wouldn't be approved, feared the process, or were discouraged by the excessive documentation; although about half probably would not seek loans even if these barriers were removed. Only two percent of all households borrowed cash in the past year, more than half of these tiny amounts from informal sources, and mainly for purchases of consumer durables. The banking system offered only nine percent of these loans, these being in the $10,000-18,000 range. Loans from government programs were generally highly subsidized. Only one-sixth of loans received required explicit collateral. Consumer credit was received by 13 percent of households, mainly offered by stores for purchase of their goods. These loans were often over $1,000, not particularly smaller than cash loans. Thus the survey data suggest that rural producers, agricultural or non-agricultural, in these provinces scarcely benefit from the financial system, and the outreach of government credit programs, with their substantial subsidies, is extremely limited. Within this context, the poorest in income are also poorest in liquid assets, and the least served by the financial system. Among 2,200 household members over age 10, the potential labor force may be defined as all who are working or seeking work or, more broadly, to include those who have given up the job search, believing no jobs are available. On the first definition, the unemployment rate is 19 percent, while it is 31 percent on the second, and the 137 unemployed break down 3:1 between males and females in either case (Table 6.4). As far less jobs are opening up than the number of active job seekers, those who have given up may be realists rather than slackers. Unemployment is a major problem among the indigent and poor,. youth in general, and young women especially. Of those who had obtained jobs in the last year, 86 percent were young males. Adults in their prime, male or female, seek permanent and non-agricultural employment, whereas those seeking seasonal or agricultural jobs tend to be the youngest and oldest. A higher proportion of females than males were either still seeking work or had concluded that no work was available. Table 6.4: Employment, Job Search, and Reasons by Poverty Class and Age Group (%) Poverty Line Classification Poverty Group Indigent Poor Non-Poor Age Group' 10 20 30 50 10 20 30 50 10 20 30 50 All TOTAL (%) 100 100 100 100 100 100 100 100 100 100 100 100 100 WORKING 12.1 39.9 36.8 31.5 13.6 39.3 49.5 39.3 15.1 50.5 68.2 41.6 33.3 NOT WORKING2 87.9 60.1 63.2 68.5 86.4 60.7 50.5 60.7 84.9 49.5 31.8 58.4 66.7 Searching 7.3 15.2 13.9 7.1 4.7 17.0 8.5 4.1 1.4 7.4 7.5 1.1 7.7 Not Searching2 80.6 44.9 49.3 61.4 81.7 43.8 42.0 56.6 83.5 42.1 24.3 57.3 59.0 Student 57.0 2.9 0.0 0.0 63.8 4.5 0.0 0.0 70.5 10.5 0.0 0.0 22.8 Housework 12.9 25.4 32.8 22.0 10.6 25.9 30.3 24.1 10.1 24.2 20.8 29.2 21.3 Retired 0.0 0.0 0.0 17.3 0.0 0.0 0.5 16.6 0.0 0.0 0.0 17.4 3.7 Old/Sick 0.5 2.2 2.5 9.4 0.9 0.0 2.7 7.6 0.0 3.2 0.6 6.2 2.6 Doesn't Exist 8.9 13.8 11.9 11.0 4.3 12.5 8.0 5.5 2.2 3.2 2.9 3.9 7.4 Other 1.3 0.7 2.0 1.6 2.1 0.9 0.5 2.8 0.7 1.1 0.0 0.6 1.2 Unemployment (1)3 37.6 27.6 27.4 18.4 25.7 30.2 14.7 9.4 8.5 12.8 9.9 2.6 18.8 Unemployment (2)4 57.2 51.0 41.2 36.5 39.8 42.9 25.0 19.6 19.3 17.3 13.2 10.7 31.2 36.e Age groups 10=10-20 years; 20=20-30 years; 30: 30-40 years; 50= 50 and older. 2 Sum of rows in box immediately below. 3 Excluding individuals who claim work does not exist 4Including individuals who claim work does not exist For those who would not remain part of the unpaid farm or household labor force, the options are migration, local wage labor, or non-farm self-employment. Migration annually absorbs about 22 percent, local jobs 15 percent, and self-employment about seven percent of current active job seekers. For the poor, non-farm self-employment in the rural areas is a prospect offering lower economic returns than seasonal farm labor. Only among the non-poor can it be considered a source of economic opportunity. Migration is mainly for work and overwhelmingly to urban areas, usually within the same province, as the gains (as evaluated from remittances) from migrating to Greater Buenos Aires instead of local towns are marginal. Most migrants, and especially those from poor families, sent nothing home, so the main gain to the household was fewer mouths to feed. The great majority of rural families, including most farmers and all income groups, spend more money on food than the income they receive from agriculture. Thus uniformly lower food prices (to retail buyers and producers) bring a net gain to rural inhabitants, with the exception of large farmers. Only the top one-third of rural 138 Table 6.6: Income by Quintile and Fann Size, Misiones and Salta Annual Household Income Farm Crop incom income Farn Area Ql Q2 Q3 Q4 Q5 Total e /ha /crop (hectares) ha 0 - 3 3,258 3,572 3,320 2,796 2,029 3,150 358 2,724 3 -15 3,167 3,465 3,891 6,276 3,592 4,254 255 1,989 15-25 3,068 4,028 2,614 3,963 7,578 4,743 126 1,934 25-80 3,245 5,735 6,081 2,785 9,163 6,557 87 2,828 80-500 2,614 8,845 5,733 4,099 12,158 7,212 17 2,001 > 500 2,249 -200 4,557 3,278 4,862 3,712 1 512 Total 3,077 4,168 4,141 4,187 7,343 4,739 177 2,285 Crop Ha's/capita 0.2 0.6 0.8 1.3 3.3 1.4 Land Ha's/capita 24.5 23.1 75.7 70.7 195.5 87.9 households have expenditures on food which meet or exceed the local cost of the minimum daily requirement, based on Buenos Aires standards. Thus, in a country where food is abundant and cheap, the rural poor nonetheless face a serious household food security problem. Potential for Agricultural Productivity Increase. The relationship between land holdings and economic well-being of households is complex: overall, total household income is often no higher on farms of more than 500 hectares than it is on farms of less than three hectares, as farm income per hectare drops precipitously from $358 for the latter to $1 for the former (see Table 6.6). In general, the poorest families typically are very large, with excess family labor to pour into both intensified farming and non-farm activities which supplement farm income, whereas the larger farms either lack sufficient labor to specialize in crop cultivation or possess land mainly suitable for pasture, hence concentrate on extensive forms of livestock raising. The smallest farms-thus obtain the most return from land as a natural resource, but also rely heavily on off-farm income. Some 20 percent of households with 3-25 hectares and a small family size tend to be both efficient in resource use and viable (in the top two quintiles) without so much reliance on off-farm income. The major form of land tenure in Argentina, in general and also among the poor (nearly two-thirds), is individual proprietorship with clear title, though not necessarily registered (see Table 6.7). This is true of "farmers" (those for whom livestock accounts for less than 50 percent of gross income) in Misiones and Salta as well. There is little tenancy or "squatting" among farmers, but, among the bottom 60 percent, the majority of land owned is untitled. For "ranchers" in these provinces, on the other hand, almost all land is "occupied", without contractual or property rights. Land titling is associated with higher investment in fixed capital per hectare, and ranching with extensive cultivation and low yields on cultivated land. Squatters, all below the poverty line, confine their activity to livestock and annual crops (but otherwise are more diversified and labor- intensive in crop cultivation, just as likely to use cash inputs, and just as productive as other farmers). 139 Table 6.7: Farm Area by Tenure Status and Quintile Mean Area of Farms (< 50% Livestock Gross Product) (hectares) Quintile Tenure Status Qi Q2 Q3 Q4 Q5 Total Owned w/ title 1.8 5.8 5.8 8.1 55.8 19.6 Owned w/o title 2.1 12.9 6.1 7.9 11.1 8.5 Rented 0.8 0.6 3.2 0.1 1.4 1.2 Occupied (gov't) 0.1 3.1 1.4 1.1 1.0 1.3 Occupied (private) 0.1 0.9 0.1 12.8 0.0 3.5 Other 0.5 1.1 1.4 1.0 0.1 0.8 Total 5.4 24.4 18.0 31.0 69.4 34.9 Mean Area of Ranches (> 50% Livestock Gross Product) (hectares) Quintile Tenure Status Q1 Q2 Q3 Q4 Q5 Total Owned w/ title 0.8 0.3 0.0 2.0 192.5 45.9 Owned w/o title 0.0 0.0 0.0 0.0 0.0 0.0 Rented 0.0 0.0 0.0 0.0 5.0 1.2 Occupied (gov't) 0.0 37.5 50.0 0.0 0.0 14.7 Occupied (private) 75.0 100.3 0.0 0.0 25.0 47.1 Other 0.0 0.0 0.0 166.7 0.0 29.4 Total 75.8 138.1 50.0 168.7 222.5 138.3 Beginning with improvements in the agricultural context, one of the most straightforward options for a farmer is to change the mix of crops. This typically consists now of two or three cash crops, differing by province and even by quintile, plus one or two subsistence-oriented crops (e.g., corn and beans). Only the poorest quintile stands out in the greater significance of subsistence crops, but even this group plants a high proportion of area (80 percent in Misiones, nearly 50 percent in Salta) to cash crops. And there is a strong tendency in Misiones for farms heavily focussed on tobacco, and entirely dependent on the tobacco companies for credit, inputs, and technical assistance, to be poorer than those which emphasize tea or mate as their cash crop. Is there a possibility of gain from a farm strategy of either diversification or specialization? There is in fact a strong positive correlation between the gross and net productivity of farms (in terms of income per hectare) and the extent of diversification, and an inverse correlation with degree of specialization. But diversification and intensified farming, to increase productivity, can be interpreted as part of the survival strategy of poor farmers, more than a means of escape from poverty. There are, clearly, differences in the value-added per hectare attainable from different crops, although crop choice per se accounts for at most 19 percent of variation in productivity. While in Salta, reduction in the area of low value-added crops bears little relation to productivity differences, in Misiones this, and also diversification into fruit, vegetables, and miscellaneous minor crops (though on a minor part of area) play a role in productivity increase. But in Misiones too the poorest farmers have already mostly achieved what can be expected from appropriate crop choice. 140 It is reassuring, however, that no major tradeoff exists between increased farm productivity and increased off-farm income. In fact, the two are positively correlated, from which one might deduce that off-farm income provides the working capital needed for intensification. One element of a strategy for assisting the rural poor in escaping poverty is simply to identify those among them with substantial potential for increasing farm productivity and provide them with access to the technical assistance and other resources required to realize that potential. The questions here are how to identify those with such potential and what resources they require. Answering these questions from information provided by a cross-sectional sample of farmers is a tricky process. "Potential" here refers to underlying managerial ability which hasn't been fully realized. The sample provides no direct measure of managerial ability, but rather some partial indicators or correlates, ranging from objective personal characteristics (such as the level of education completed), to revealing behavior (such as having joined an organization or borrowed money in the past), or attitudes (such as reasons for not borrowing). If the potential of managers who have sufficient resources to draw on has been realized and evidenced in higher productivity, then it would be possible to isolate and measure "management ability" as a latent variable through the correlations between its partial indicators and variations in outcomes. In making such measurements, other, simultaneous influences on outcomes--notably the other resources available to farm managers--need to be considered and controlled for (which in turn enables us to identify what other resources are required for managers to realize their potential). The statistical exercise is one of trying to identify a mathematical model of the behavior of the farm as a unit (assuming the household head is the manager) which conforms with economic logic and common sense as well as statistically accounts for observed choices and outcomes. The details may be found in the background paper on which this chapter is based, but results suggest that the following "story" (interpreting a statistical model) best accounts for the data. Households with large areas of land, especially in Salta, tend to be ranchers. Although ranchers usually also grow crops, techniques are less intensive and crop yields much lower than those of farmers (i.e., non-ranchers), possibly reflecting differences in land quality and environment (as well as tenure) which cause them to choose ranching in the first place. Secondary school graduates also tend to become ranchers, perhaps for the social prestige rather than income and, as most of these have lucrative non-farm jobs or enterprises to rely on, it is possible that crop cultivation isn't managed to maximize income in such cases. Management ability is apparently reflected in aggressive pursuit of advantage through access to technical assistance, participation in coops and farmers associations, and attendance at meetings of the same. Good farm managers tend to have completed a primary (but not secondary) education; to have borrowed in the past and have savings now. Good managers tend toward diversity in cropping rather than specialization, use higher levels of current inputs, and thereby obtain higher crop yields and value-added for any given level of inputs. Put more precisely, the above characteristics identify some 141 unmeasured factor associated with higher productivity which we identify as management ability, and which answer the question as to how management potential can be identified. Receipt of technical assistance is strongly associated with "good" farm management, whether it is because the managers seek the assistance or the extension agents seek out good managers. Receipt of technical assistance in turn positively influences the decision to use fertilizer and other inputs (along with completion of primary education and being a farmer rather than rancher). However, this by no means explains completely the decision to use fertilizer and other commercial inputs. Given this decision, the amount of inputs purchased is largely explained by varying availability of liquid assets (for every additional one percent of financial resources, input purchase goes up by 1.5 percent). Possession of land title appears to be strongly associated with greater fixed capital investment, but not current input use (due to the dysfunctional credit market). However, the level of fixed capital use does not have a significant positive influence on crop yields, so this relationship doesn't provide an economic rationale for titling. Nor are yields increased by greater intensity of household labor, which as used in crop cultivation, appears to be in surplus. This would be consistent with a rationing of off-farm jobs which keeps excessive labor working on the farm. The larger the scale of crop area, the lower are crop yields, everything else equal--small farms have a distinct land productivity advantage. Just why this is so needs further investigation, as it is true even when the amount and proportions of household and hired labor, the extent of diversification, liquid assets, etc. are the same. Of the major factors of production, then, only the level of use of purchased current inputs has a significant positive impact on yields. The answer to the second question, what resources are needed for good managers to realize their potential, therefore appears to be mainly "liquid assets" to finance input purchases such as fertilizer and seed. Overall, the study has suggested that an effective strategy to address rural poverty needs multiple instruments or components going well beyond the narrow focus on agricultural productivity characteristic of traditional rural programs in Latin America. The components discussed above would deal with all of the correlates of low-income status. However, most would also not be effective in isolation. There is a need. for balanced attention to agricultural and non-farm elements of livelihood, and to long-term solutions, such as improved education and family planning, as well as short-term measures, such as rural organization and delivery of training, technical assistance, and finance. Attention to quality and relevance, e.g., of education and technical assistance, may be at least as important as increased scope of service delivery. And attention to institutional development problems, in the areas of finance, land tenure, labor market, and regional development, may be as relevant as targeted projects. 142 CHAPTER VII: URBAN DEVELOPMENT AND INEQUALITY Introduction Argentina is largely an urban economy. The urban population has grown from 83% of the population in 1980, to 87% in 1991 (Census data), to an estimated 89% or 33 million people in 1999. Projections made by INDEC estimate that in the year 2020, the total population of the country will grow to 45 million people, of which 93% will be living in urban centers (see Table 7.1). Thus, the majority of the poor are found, and will continue to be found in urban centers. Providing basic urban infrastructure for this population, and specifically meeting the basic needs of the poor will be a continuing challenge. However, the provision of urban services can potentially reduce the gap between rich and poor, and reduce the worst aspects of poverty. However, it can equally perpetuate or exacerbate existing social and economic differences. Table 7.1: Population, Urban and Total, 1980-2020 (millions) Year Total population Urban Population % urban 1980 28.1 23.3 83.0 1991 33.0 28.7 87.1 1999(est.) 36.6 32.2 89.1 2010(proj.) 41.5 37.9 91.4 2020(proj.) 45.3 42.9 92.5 Source: INDEC In terms of physical assets, the poor of Argentina are relatively well off, in comparison especially with other countries. Their poverty shows up more in terms of inadequate access to public facilities, such as water, sanitation, health and education In 75 terms of consumer durables, for instance, of the urban poor > 78% own a refrigerator > 85% own a television (31% have cable); > 59% own a washing machine, and > 19% own a car. However, most live in difficult circumstances in poor neighborhoods that lack basic infrastructure services. If we focus not on the poor, but the lower 20%76, we find that crowding is high; over 50% live in houses with two or more persons per room; 17% live in houses with 3 or more persons per room, defined as "critical" crowding. While only 4% lack electricity, most of these connections are probably illegal. Access to other public services is more critical; 47% lack adequate sanitation, 42% lack running water in the house, and 45% live on land with insecure tenancy, that is without legal title or a formal 75 Data here is from the Social Development Survey of SIEMPRO (1997); "poor" is defined on the basis of structural poverty (NBI) 76 The focus shifts here to the lower 20%, because the poverty definition being used is the NBI, which itself is based on housing quality indicators. 143 arrangement. Many do not have formal titles to the land they occupy, and their houses are constructed informally and illegally, often in areas lacking adequate drainage or prone to flooding, or near garbage dumps or other marginal areas (see Table 7.2). The fact that 34% of the poor are located in areas prone to flooding makes them particularly vulnerable to the risk of major flood damage.(see box) This, combined with the insecure titles to land held, also reduces the incentives for the poor to invest in upgrading the quality of their housing. Exposure to Flood Risk By any standard, Argentina Is highly vulnerable to economic loss from flooding. Within the developing countries of the world, Argentina falls within the top seven of risk compared to GDP. In Latin America, Argentina has the highest exposure to direct flood loss, and is second in comparison of loss to GDP Since 1957, Argentina has, had 1 1 major floods. Of these 11 floods, three of them have caused direct damage in excess of $1 billion. According to statistics developed by Swiss Re, Argentina is among the top 18 countries worldwide with potential flood losses in excess of $3 billion. In measuring flood losses as a percentage of gross domestic product (GDP), Argentina is only one of fourteen countries whose potential losses to floods is greater than 1% of GDP. In Latin America, only Ecuador has a higher exposure to GDP from flood risk. The major risk to flood in Argentina is in the river region in the Northeastern and central parts of the country. Especially in the plains, large areas along the river courses may be flooded, including riverside areas near the large cities of Buenos Aires, Rosario, Santa Fe and La Plata. In fact, it is the widespread flooding in the cities (especially in the Buenos Aires/La Plata region) that could result in economic losses approaching USD 3 billion. Source: Paul K. Freeman "Recommendations for Risk Transfer for Argentine Flood Risk", August 10, 1999. Table 7.2: Housing and Infrastructure Indicators, 1997 (percent of households) Indicator: Lower 20% Upper 20% Poor (NBI) Non-poor Crowding: - more than 2 persons/room 51.0 2.8 71.8 15.8 - more than 3 persons/room 17.2 0.0 46.8 .1 Without access to safe water 8.3 .2 14.0 1.7 Without access to running water 34.2 .8 54.4 6.3 in the house Without adequate sanitation 47.0 5.1 68.0 15.6 Without electricity 4.4 .3 6.7 .5 With insecure tenancy of land 44.6 11.5 58.0 18.8 Located in flood area 28.5 11.9 33.6 16.9 Source: SIEMPRO, Social Development Survey,1997. Poverty defined as having one or more deficient basic needs. 144 Since the NBI is a weighted index of five housing/infrastructure indicators, it is a good overall index of these deficiencies.77 There are considerable differences between individual towns and varying degrees of poverty as defined by the NBI (Unsatisfied Basic Needs) as well as other indicators (see Table 7.3).7S While the population with unsatisfied basic needs is only 8.1% in the Federal Capital, the average for the 25 urban areas is 16.5 percent or double the level in the Federal Capital. Moreover, 14 of the 25 towns have NBI levels more than double this average. These differences are also reflected in major differences between towns in housing conditions and deficiencies in basic needs. While only 8% of the population of the Federal Capital had unmet basic needs in 1991, the level was as high 25% in Formosa and 22% in Corrientes. While the scores on individual indicators vary considerably, the overall picture is of a country with enormous differences within and between its urban population. Table 7.3: Poverty and Inequality Between 25 Cities and Towns, 1991 Urban Pop. NBI Pop. Illite % Not- Homes Low Area Pop. Without rate Comple Without Quality (%) Health Rate Ting Toilet Home Coverage (%) Sec School (%) (%) Metro Buenos Aires 7,924,424 18.9 38.1 14.5 77.1 71.5 8.8 Federal Capital 2,871,519 8.1 19.1 8.3 56.6 5.8 1.1 Gran C6rdoba 1,193,443 14.4 35.7 13.6 66.9 77.0 1.9 Gran Rosario 1,108,917 17.2 25.7 14.3 70.6 54.0 6.6 Gran Mendoza 766,454 13.3 35.5 15.0 72.3 40.6 4.1 Gran La Plata 627,927 13.4 25.2 12.1 65.7 43.1 5.5 Gran San Miguel deTucuman 616,598 20.7 30.8 16.4 70.1 48.8 10.8 Mar del Plata 505,402 11.7 28.9 11.8 71.8 31.6 4.2 Santa Fe 402,303 18.6 29.9 15.4 68.7 73.8 4.7 Gran Salta 366,994 22.9 39.1 17.0 70.7 27.0 6.1 Gran San Juan 350,478 15.5 35.5 15.7 73.9 80.6 7.1 Resistencia 289,618 23.2 37.9 18.9 73.2 71.4 7.0 Santiago del Estero-Banda 260,997 21.2 37.3 18.0 71.8 68.2 3.0 Bahia Blanca 256,900 12.1 26.0 13.0 71.8 28.1 1.3 Corrientes 255,869 22.0 38.3 18.7 67.0 46.5 6.5 Neuquen 241,530 16.1 36.0 18.5 72.4 59.1 2.4 Gran Parana 209,657 15.7 23.9 15.6 68.9 61.0 3.3 Gran Posadas 208,808 23.1 38.6 19.8 71.5 78.6 11.0 Gran San Salvador deJujuy 177,912 27.1 36.9 17.2 73.1 50.7 10.4 Formosa 147,002 25.7 36.7 19.7 73.6 56.7 10.6 Gran Rio Cuarto 137,325 11.2 30.5 14.5 73.7 54.1 0.8 G.S.F. del Valle de Catamarca 131,394 19.2 24.7 17.4 69.6 60.6 2.4 Comodoro Rivadavia 122,146 17.4 25.3 16.1 77.0 29.1 5.1 San Nicolas de los Arroyos 117,936 14.3 24.2 14.4 75.6 55.5 4.5 Concordia 115,699 24.7 36.6 19.7 75.7 56.7 10.5 Source: the data, provided by SIEMPRO is based on the 1991 Population and Housing Census made by INDEC (National Institute of Statistics and Census). 77 The NBI (Index of Unsatisfied Basic Needs) consists of a composite index which includes access to water, crowding, housing quality, sanitation, school attendance, and subsistence capacity (see Chapter 3). 7 Data obtained from the Ministry of Social Development, SIEMPRO. 145 As the major generator of economic product, it is not surprising that the population in these cities enjoy levels of income and welfare, in the aggregate, considerably above the national average. The issue, however, is not so much what the averages portray, but rather, within a rich country with many thriving sectors of the respective urban economies, what are the differences among groups within cities and towns. An initial breakdown of urban areas by level of inequality is presented in Table 7.4. This table presents inequality within cities using the degree of variation in NBI by presenting the percentage of population with NBI in top and bottom 3 deciles of districts. This table shows great differences in the proportion of NBI population at the top and the bottom deciles. In Santa Fe, the difference is 10 times, and other big cities outside of MBA and Gran Cordoba have great concentrations of poverty in the bottom. Surprisingly, some large cities are relatively homogeneous, especially greater Rosario and greater Buenos Aires. In the latter case, within the exception of the three wealthy districts of San Isidro, Vicente Lopez, and Hurlingham, and of course the Federal capital itself, most of the metropolitan area is homogeneously poor and lacking services Table 7.4: Percentage of Population with NBI in Top and Bottom 3 Deciles of Districts in 25 Cities, 1991. % NBI Urban Area Population Top 3 deciles Bottom 3 Relation deciles Santa Fe 3.2 34.5 10.8 G. Rosaro 3.2 31.3 9.8 M. del Plata 2.2 20.0 9.1 G. La Plata 2.9 25.1 8.7 G. Mendoza 3.1 24.4 7.9 B. Blanca 3.0 20.6 6.9 G. Salta 7.8 39.7 5.1 G. C6rdoba 5.1 24.7 4.8 Federal Capital 3.9 18.2 4.7 Resistencia 8.4 36.9 4.4 G. San Juan 8.9 36.9 4.1 G.S.M. de Tucuman 8.5 31.3 3.7 G. Paran6 7.4 27.2 3.7 Neuquen 7.6 26.1 3.4 Corrientes 11.6 39.6 3.4 G. Posadas 11.4 35.8 3.1 G.S.S. de Jujuy 18.6 57.3 3.1 Sgo. Del Estero 12.0 33.1 2.8 Metropolitan B A 10.8 25.3 2.3 Source: the data, provided by SIEMPRO is based on the 1991 Population and Housing Census made by INDEC. The Distribution of Investment: the Case of the Federal Capital: The previous sections examined the availability of infrastructure, education and health services in various urban areas. However, little is known about the rationality of the allocations of investment funds, particularly for infrastructure and social services. This section presents public investment data by individual sectors, for the Federal Capital 146 area with Buenos Aires. Studies undertaken by the Government of the City of Buenos Aires in 1997-1998 include data on infrastructure and education investment broken down by the 21 school districts of the city. Within the Federal Capital there are wide differences with unsatisfied basic needs. While 13 of the districts had less than 7.6 % of their respective populations lacking basic services in 1991, four districts had more than 2 or 3 times that level. Indeed, two districts in the south of the city had 20.3 and 26.3 % of their populations respectively with unsatisfied basic needs (see Table 7.5). Infrastructure:. Data on investment in public infrastructure such as streets, sidewalks, public buildings, parks, playgrounds, and traffic lights over the period 1991 to 1997, as presented in Table 7.5, show that some districts receive much more than others. Per capita expenditures for the 1991-97 period (in 1997 prices) vary from a high of $300 to a low of $10. In relation to population, a minority of population received a disproportionate share of this investment. For example, two districts with 6.2 % of the population received 30.5% of total investment. These remarkable disparities do not seem to be related to relative economic development. A simple correlation between NBI and average infrastructure spending has an correlation coefficient of only -.08. However, if one looks more carefully at the distribution (Figure 7.1), an interesting pattern emerges. Districts with low NBI (low poverty) show wide variations in the level of infrastructure investment. However, districts with high NBI populations generally have low levels of public investment. There seems to be a bias away from poorer areas toward richer areas, but the bias is not uniform or constant. However, there does not seem to be any clear bias in favor of poor areas. 147 Table 7.5: Buenos Aires, Federal Capital Public Investment (1991-97) in Infrastructure and Education, Population and NBI (1991), by school district Population Infrastructure Education District NBI 1991 (share) ($ per (share) ($ per l_______ . person) person) 1 7.1 3.1 2,141 19.8 492 2 6.9 3.3 434 4.3 360 3 12.9 3.7 272 2.9 246 4 26.3 7.9 362 8.4 117 5 20.3 3.9 553 6.4 272 6 9.9 3.4 529 5.4 419 7 4.6 3.1 389 3.5 380 8 4.1 5.4 127 2.0 191 9 5.1 2.0 1,105 6.4 631 10 3.7 3.1 1,175 10.7 428 11 5.3 9.4 256 7.1 87 12 4.9 8.9 191 5.0 102 13 8.7 3.5 89 0.9 222 14 5.8 5.7 251 4.2 110 15 3.7 3.0 235 2.1 247 16 3.2 4.8 150 2.1 297 17 2.6 3.6 69 0.7 261 18 3.2 8.9 109 2.8 113 19 16.2 5.0 69 1.0 128 20 10.0 3.9 260 3.0 181 21 17.9 4.4 96 1.2 175 Total 100 100 Mean 8.69 60.29 422.00 259.95 Standard 6.50 71.01 497.05 146.25 dev. Correlation -0.0798 -0.0798 -0.249 with NBI Source: staff estimates based on various sources, including 1991 Census, SIEMPRO, and the City of Buenos Aires and public sector budgets. See Background Paper Table 7.5 also presents investment in public education by school district within the Federal Capital in 1997. These data show that on a per capita basis, some districts receive more than five times than others. Wide differences in public spending for education are also accompanied by wide differences in the percentages of student population who finish primary and secondary school. In 14 of 21 districts, more than half of the students do not complete secondary school. Furthermore, spending seems to be negatively correlated with poverty as measured by the NBI; the overall correlation coefficient is -.25 between school district spending and the NBI of the school district. 148 Source: See Table 7.5 Figure 7.1. NBI and Investment 2,500 - a 2,000- - @ 1,500 - 1,000 0 ' 500 - +**- 0 10 20 30 NBI (1991) The negative correlation between NBI and education spending is somewhat stronger than infrastructure, but not robust. As poverty measured by the NBI goes up, there is a tendency for education investment to decline. Privatization and the Poor Argentina' reform program during the 1980s and 1990s included a major downsizing of the public sector and a privatization of major urban services, including telephone services, transport, banks and urban services. It is often alleged that such privatizations impact adversely on the poor, since they often result in higher prices for basic services which were previously subsidized. However, we know also that privatizations also potentially can improve the quality of services being provided. The impact of privatization for households in Buenos Aires can be roughly estimated on the basis of two household surveys undertaken in 1986 and 1996. Comparing the household surveys for 1986 and 1996 reveals a sharp increase in consumption expenditures for items related to public services; electricity, public transport, telephone, sanitation services. Table 7.6 gives information for total per capita household expenditures for the first and second quintiles of the population. 149 Figure 7.2. Prices of Public Services, Buenos Aires (19883100) 400000- 350000- 300000- 250000- 200000- 150000 100000 50000 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 |-CPI " EIecridty -*san servces, gas -pub transport - cornunications Table 7.6: Household Expenditures for Public Services, Buenos Aires (Percent of Total Expenditures) 1st Quintile 2nd Quintile 1986 1996 1986 1996 Electricity 2.1 3.7 1.8 3.0 San Services, Gas, other 2.4 3.7 2.2 2.9 Public transport 3.7 8.0 3.8 7.5 Telephone (and mail) 0.9 2.0 0.7 2.5 Total 9.1 17.4 8.5 15.9 Expend./Household (per month, 593.1 452.5 834.8 640.8 1996 pesos) % change, 1986-1996 -23.7% -23.2% As shown in Table 7.6, overall, the average consumption level of the lowest two quintiles has declined over the period. Average consumption spending for the lowest quintile drops by about 24% over the period, while that for the second quintile drops by about 23%. Despite this drop, the share of the household budget going to public services has increased sharply. For the lowest quintile, the share goes from 9% to about 17%; for the second quintile, it rises from 9% to 16%. The major increases appear to be for public transport, which more than doubles its share of the budget in the lowest quintile (3.7 to 8.0%). The share for electricity goes up 2.1 to 3.7%, an increase in the share of 76%. These increases, however, can be traced to two possible cause, either price increases or increased real spending (volume increase). With regard to prices, Figure 7.2 traces price increases for public services between 1988 and 1998, and compares these prices to an overall consumer price index which is weighted to reflect the consumption basket of the poor. In general, all public services lagged behind the overall CPI during the period up to 1996. After 1996, only communications (telephone) and public transport rose to be slightly above the CPI; electricity and water/fuel/sanitary services still lagged behind. Thus, any increase in consumption of public services must arise from increased utilization of services, perhaps brought about as a result of increased efficiencies resulting from privatization. 150 Estirnating increases of real consumption of services is difficult because of the wild swings in the deflators during the hyperinflation of the 1980s. Nevertheless, Table 7.7 gives an approximation using the expenditure and price data given above, covering the period 1986-96. In general, the estimated increases in consumption of public services by the poor are quite remarkable. Overall, the lowest quintile increases its consumption by 66%, while the second quintile by 94%. This is even more remarkable when one remembers that total household consumption was declining over the period (-24% for the first quintile). Table 7.7: Real Consumption of Urban Services, 1986-1996 (Buenos Aires, 1996 constant prices, household expenditures per month) 1st Quintile 2nd Quintile 1986 1996 %change 1986 1996 %change Electricity 9.73 16.74 72.1% 11.73 19.22 63.8% San Services, Gas, 14.17 16.74 18.2% 18.28 18.58 1.7% Public transport 30.07 36.20 20.4% 43.47 48.06 10.6% Telephone (and mail) 5.85 9.05 54.8% 6.40 16.02 150.3% Total 15.57 25.79 65.6% 18.13 35.24 94.4% The largest increases for the first quintile are in the areas of electricity (+72%) and communications(+55%). For the second quintile, the consumption increases for communications are even higher (150%), while the consumption increases in other areas (public transport, sanitary services) are somewhat lower. While these conclusions are somewhat tentative, give the quality of the data, it suggests that the poor have not suffered because of higher prices of public services during the period 1986-96; in fact, prices seem to have barely kept up with inflation. Clearly, the poor have increased their consumption of public services, despite falling household income during the period. One possibility is that privatization, by improving the efficiency of firms providing services, increased the supply of or access to services, and consumers responded by purchasing more. However, it is also possible that privatization improved the collection of bills and the reduction in illegal taps and connections, so that the apparent increase in consumption represents only the payment for services previously obtained free. In a similar study, Navajas looks at the distributive impact of changes in all relative prices, 1988-98. He finds that public sector prices did not worsen income distribution, and that privatization appears to have increased access of the lower income groups to these services.79 A study by Chisari, Estache and Romero, using a CGE model of the economy, investigated the general equilibrium ramifications of privatization on income distribution.80 They found that the operational gains from privatization of utilities reduced the average cost of services by 41%, or 0.9% of GDP, even assuming that ineffective 79 Fernando Navajas, "El Inpacto Distributivo de los Cambios en Precios Relativos en la Argentina Entre 1988-98 y los Efectos de las Privatizaciones y la Desregulacion Economia" in La Distribution del Ingreso en la Argentina, FIEL, Buenos Aires, 1999. 8O Omar Chisari, Antonio Estache and Carlos Romero, "Winners and Losers from the Privatization and Regulation of Utilities: Lessons from a General Equilibrium Model of Argentina" The World Bank Economic Review, 13(May 1999), 357-378. 151 regulation allowed the new owners to keep a large part of these gains as quasi-rents. With effective regulation, a further gain of 16% could be achieved. With ineffective regulation, the direct gains were significantly higher for the higher income classes. However, effective regulation tended to shift the gains to the poorest classes. Conclusions: The poor distribution of urban services both contributes to, and is explained by, the overall level of income distribution. Moreover, it seems clear that the provision of urban services can contribute to maintaining socio-economic patterns and differences, or changing them. In the past, the power of the national government and the absence of elected local institutions, has meant that local problem-solving, at the neighborhood level, has not been a central feature of political and institutional history. The arena for problem- solving has been "la niacion" and not "el barrio". Respondents in a 1998 survey list local government as third after the national and provincial governments as institutions which have been effective in alleviating poverty81. While an accurate reflection of the country's history, it would appear, based on experience in other countries, that Argentina has under-valued the potential of improving welfare of individuals and families through local action. As a result, much more attention needs to be devoted within individual cities to spatial differences in levels of economic welfare. The debate about welfare is not only about employment, but how place-based variables affect levels of real income and quality of life. The above suggests that effective social targeting of public expenditures also requires more attention to place-based programs. It also suggests, moreover, that the privatization public services has not reduced the welfare of the poor, and may have improved it. 81 see S. Cesilini and E. Zuleta, Background Paper No. 2 152 Annex 1 Inventory of Social Programs for the Poor and Vulnerable, 1998 Theniatic Area000000 Obctives S e Provided 7 Food and Nutrition * Community Social * Improve the quality of life of the NBI population * (i) Provision of free meals; and (ii) Other social services Policies (Prov. Gov./POSOCO) * (i) Improve living conditions and access to sufficient . (i) Daily dietary rations in infant feeding centers, infant * Infant Nutrition and food for nutritionally vulnerable children from 2 to 14 care centers and school cafeterias; (ii) Strengthening Feeding Program years old who live in unfavorable socioeconomic PRANI (monthly box containing 9 meals - 1200 (Prov. Gov.IPRANI) conditions, through free meals and support for basic calories/day); (iii) Technical Assistance and training for education; (ii) Evaluate and restructure the system for Provincial, Municipal and NGO officials and technicians; infant feeding centers and school cafeterias; and (iii) and (iv) Equipment and infrastructure support for feeding Gradually transform infant feeding centers into infant centers/cafeterias care centers . Community Farm a Improve nutrition of the NBI population with * (i) Community, family and school-attended farms Program insufficient nutrition supported by key investments, technical assistance and (Prov. Gov.IPROHUERTA) training; and (ii) Training sessions * Improve the nutritional status of the target population . Free meals * Social Nutrition Program (Prov. Gov./PROSON/ Trainina and Strengthening . National Center for . (i) Articulate community actions in a framework . (i) Coordination, training and various services for Community complimentary to public and private fields; community organizations; (ii) Distance training; (iii) Organizations (ii) Facilitate development of community Workshops for the National Volunteer Program; (iv) (CENOC) organizations through human resource training, Training for Indigenous People (Program for Training formation of networks and alliances at the regional, Indigenous Peoples) local and national levels; and (iii) Recognition and diffusion of the sector's activities through the community organizations considered as active players 153 Thematic Area Objectives Services Provided in the development of social policies . (i) Training ;(ii) Technical Assistance to social programs; Information, Monitoring * Providing Government organizations with the (iii) Tables, CD and documents with information on the and Evaluation System necessary resources to operate a social information social needs of the population; (iv) Social survey; (v) for Social Programs system, including monitoring, evaluation and training Studies on qualitative aspects of the poor social sectors; (SIEMPRO) in social administration (vi) Evaluations of social programs; (vii) Information on the provision of social programs (services on. demand); (viii) Reports on the status and analysis of the social situation; (ix) Maps or documents with indicators; (x) Technical plans for dissemination and equipment; (xi) Publications/bulletin; and (xii) Integrated monitoring system. Community DeveloDment and Social Integration . National Institute for * (i) Promote integrated development of indigenous * (i) Financial assistance; (ii) Technical assistance; and (iii) Indigenous Issues (INA! communities, with a focus on preserving and valuing Training their cultural heritage; (ii) Plan, design and supervise the designated resources for technical assistance and financing for the communities and public and private organizations in relation to their area of competence; and (iii) Organize a national registry of indigenous communities (RENACI), in accordance with Decree 155/89 * Attention to Priority . (i) Technical assistance and financing for institutions Groups . (i) Improve living conditions of the populations with specializing in care for the handicapped; (ii) Training for social risks (handicapped, adolescents, children, providers of home care; and (iii) "Tourist days" (semi- adults over 60, the chronically and terminally ill, etc.), complimentary and complimentary) by facilitating the satisfaction of basic needs; and (ii) Promote strategies of social support, favoring integration and participatory solidarity * (i) Financing of projects with participatory planning * Attention to Vulnerable . (i) Contribute to decreasing exposure to social risks processes (PPP); (ii) Financing projects for direct Groups (PAGV) and improving the quality of life for the most vulnerable assistance; (iv) Financing projects for institutional groups in urban centers and indigenous populations strengthening through implementation of a model of participatory ._______________________ administration at the local level and articulation with 154 Thematic Area ObJoctive S _&vices Prov7ded community initiatives based on the strengthening of local public organizations and civil society; (ii) Implement a method of administration which supports the reform process and social administration at the local level, promoting coordination of community initiatives to decrease risks; (iii) Support the development and integration of vulnerable groups through actions which facilitate their access to social services; and (iv) Maximize the use of public resources at the local level . Participatory Fund for . (i) Financial and organizational assistance for community Social Investment . (i) Financing projects for social investment in locations projects; (ii) 4-hour training modules for project design; and (FOPAR) and municipalities with extreme poverty which respond (iii) Training for institutional strengthening to the most pertinent needs of the community; and (ii) Develop local administrative capacity through participatory experiences which are concrete in terms of structure, management and execution of projects designed to improve the socioeconomic conditions of groups and communities in poverty situations * (i) Financial assistance for projects supporting community . Plan for Strengthening * (i) Strengthening the operational and administrative initiatives, and those focusing on parents; (ii) Financial Civil Society capacity of the community organizations through the assistance for projects for training (equipment for NGOs, training and technical support for the initiatives they initial products, training, preparation of training material); develop; and (ii) Guide the links between the State and (iii) Technical assistance for projects; (iv) Training for social Civil Society for the development of social programs administration in municipalities; (v) Training community leaders (initial, specific and more advanced); (vi) Basic and complimentary training;; (vii) Training for promoters; (viii) Training internships; (ix) Financing projects for social administration in municipalities; (x) Administration of projects (workshops); and (xi) Financing projects • (i) Technical assistance, training and economic support . Plan for Strengthening . (i) Strengthen the capacity of youths in the design and for the design and execution of solidarity projects; and (ii) Youth Development execution of projects through the guidance of solidarity Technical assistance, training and economic support for the (PFDJ) groups of youths which work for improvements in the design and execution of innovative projects quality of life for the poor, or socially vulnerable; (ii) __Support for governmental and non-governmental 155 Thematic Areq 1Oc,tives,? ,, rVi.. Provided organizations which provide initiatives and innovations linked to youth issues; (iii) Strengthen the autonomy and consolidation of youth groups through agreements on the construction of a just and integrated society with solidarity; and (iv) Strengthen the democratic forms of participation, favoring the model of decentralization from * (i) Financial support for the design and execution of small the Federal Government construction works and teaching equipment; and (ii) Technical assistance and training for NGO personnel . Action Program for * Improve the social conditions and amplify the Integral Support for the possibilities for social and labor integration of children Socialization of and adolescents at risk through technical support and Marginalized Minors in the financing for projects in execution, or to be executed, by . Financing of projects for shelters, meal rations, medical GBA NGOs attention, production initiatives, promotion of rights, school (PROAMBA) support, recreation, and labor market training * (i) Contribute to improving the living conditions of . Program for Attention to children and adolescents in especially difficult Minors in Especially circumstances, providing opportunities for them to Difficult Circumstances realize their physical, psychological and social potential, (PROAME) oriented toward the exercising of their rights through projects led by NGOs and through strengthening provincial governments and/or municipalities with obligations to youth-related issues; and (ii) Promote . (i) Legal assistance to minors; (ii) integral attention to adequate methodologies for intervention which promote mothers in situations of risk or abandonment; (iii) temporary agreement on, and development of, community capacity housing for children and the handicapped at risk or and initiatives abandoned; (iv) equipment, repair and expansion of institutes; (v) training of specialized personnel; (vi) . (i) Protect youths and families at great social risk as treatment and assistance for handicapped children and . Advisory Programs for consequence of abandonment, abuse, mistreatment, elderly people; (vii) research; and (viii) re-establishing Youths and their Families addictions, law-breaking, etc.; and (ii) Coordinate the families for exploited street children participation of institutions, organizations and/or publicity entities for the programming, execution and diffusion of local and regional activities to orient and promote family integration, using technical assessment, specific . (i) Expansion of schools; (ii) Grants for secondary material, studies, investigations, congresses and training students; (iii) Teaching materials in Wichi; (iv) Works for water provision; (v) Provision of ambulances; (vi) Provision * Improve the living conditions of the Wichi population, of medicine and milk; (vii) Improved and/or new wells; and 156 • Integrated Project for using their cultural principles as a basis, valuing, and (viii) Living quarters, possible with latrines and baths Aboriginal Communities in spreading these principles. The Project would promote the Ram6n Lista Dept. education, nutrition, health, capacity for self- . Subsidies to institutions which present proposals for (Formosa) (DIRLI) administration, production and organization. interventions in the areas of extreme need and/or poverty . Improve the quality of the NBi population at social risk, through the participation of governmental and non- . Institutional Subsidies governmental organizations, generating a process of self-sustaining development Education * Social Education Plan - . (i) Improve the quality of education through . (i) Specific equipment for special education schools; (ii) Program 1: Better pedagogical support to poor schools; (ii) Promote Equip schools (laboratory, sciences, etc.); (iii) Cabinet for Education for All participation of the education community; and (iii) Focus information in the school; (iv) Textbooks and manuals (MCYE/PSE-PI) available resources toward the poorest schools of the (primary and secondary); (v) Institutional projects; and (vi) country Services . Social Education Plan - . Make possible a Federal Law for Education, improving . (i) Administrative areas; (ii) Teaching areas; (iii) Program 2: Improvements and extending the physical infrastructure of poor schools Construction of service areas; (iv) Classroom construction; in School Infrastructure in the whole country (v) Constructions of polymodal rooms; (vi) School (MCYE/PSE-P2) establishments; and (vii) School Reconstruction . National Program of * (i) Increase years of school attended by poor youth; * Grants Grants for School (ii) Strengthen the school institutions to improve quality Retention (MCYE/Becas) and equity of education; and (iii) Contribute to improved employability of poor youth Emergencies and Others * Actions from the Office . Plan, coordinate and execute actions for the . (i) Food, coats, mattresses, etc.; and of Social Emergencies prediction, prevention and resolution of states of social (ii) Training municipal and/or provincial agents in dealing with (DIES) need, caused by emergency situations which emergencies and catastrophes compromise the survival of communities Employment and Productive Development . Labor Training for . Increase the competence and productivity of the . (i) Grants for living expenses and maintenance for Sectors or Types of workforce in a specific sector or type of activity at the unemployed people who receive no unemployment Activity (MTSS) regional, provincial or local level insurance or other social security services; (ii) Free labor 157 Thematic Area Objectives ServicS. Provided training; and (iii) Civil responsibility insurance . Re-establish the workforce in sectors and/or regions * (i) Free training; (ii) Civil responsibility insurance; and (iii) * Training for Employee in crisis, or in the process of restructuring, improving the Grants for living expenses and maintenance for Support employability of workers unemployed people who receive no unemployment insurance or other social security services . (i) Establish a program of transitory employment to . Non-remunerative economic aid, civil responsibility carry out economic and social infrastructure projects, or insurance and health coverage . Local Employee to provide services which contribute to community Development (II) development and improvement in the employability of workers; and (ii) Present the legal framework with which executing organizations can provide non-remunerative economic aid a (i) Credits for investment and operation; and (ii) training * Facilitate the productive capacity of human and for producers natural resources of small producers and indigenous * Rural Development for peoples from NEA through sustainable increase in Northeast Argentina investment and training in self-administration (SAPYA/PRODERNEA) a Civil risk insurance ($5/month) and variable grants . Increase employment and decrease training costs of firms * EMPRENDER * Monthly non-remunerative economic aid ($200), civil (MTSS/EMPRENDER) responsibility insurance and medical assistance coverage * Private employee promotion to stimulate transitory * FORESTAR occupation for unemployed workers in rural areas to (MTSS/FORESTAR) carry out work linked to forest activities . (i) Subsidies to private employees with base salaries not more than $1000 (people between 38 and 44: $100/18 * Generation of private employment for undetermined months; persons 45 and over: $150/18 months); (ii) . PROEMPLEO periods of time, targeting the unemployed with difficulties Training to support the employees, and grants for living (MTSS/PROEMPLEO) integrating in the labor market expenses and maintenance; and (iii) Coverage of moving expenses . (i) Courses on the search for employment (5/6 days) . (i) Increase the probability of maintaining or finding ("Proyecto Imagen"); (ii) Training for developing . Program for Support of productive employment; (ii) Improve conditions of the occupational skills (including living expenses, accident and 158 Thematic Area Obiftetives Services Prov7de7 Productive Reconversion beneficiaries through an orientation on how to look for medical insurance, childcare, etc.) ("Proyecto Joven"); and (MTSS/PARP) employment ("Proyecto Imagen"); (iii) Facilitate insertion (iii) Courses on business administration and technical in the labor market, through specific training ("Proyecto assistance ("Proyecto Microempresas") Joven"); and (iv) Collaborate in the development of micro-enterprises and other forms of self-employment ("Proyecto Microempresas") . (i) Subsidies for protected production workshops; and (ii) * (i) Promote labor market adaptation for people with Subsidies for acquisition of goods for training and other . Program for Support for handicaps (without discrimination based on type and expenses Protected Production grade) of working age, for integration into the protected Workshops (MTSS) and non-protected labor market; and (ii) Promote new workshops and support existing ones, in order to improve the achievement of the economic and social goals in order to generate in each area a plan for business development . Monthly subsidy ($100) for converted workers * Promote labor contracting in the private sector, as well . Program to Assist as the reconversion of the workforce in firms which Reconversion of Shearers contract shearers (MTSS) * (i) Productive projects with technical and financial * Strengthen the institutional capacity of the assistance; (ii) Training; (iii) Monitoring and evaluation; (iv) municipalities to induce and manage economically Institutional development; and (v) Strategic planning . Local Development sustainable development, oriented to social Promotion Program investments through the creation and consolidation of productive entities which generate employment . (i) Strengthening self-consumption projects; (ii) * Improve the productive capacity of small rural Technicians and promoters trained; (iii) Innovative producers by increasing regular income productive activities; (iv) Traditional productive activities; (v) * Agricultural Social Courses for specific demand; (vi) Training of rural agents; Program (vii) Financial assistance to productive projects; (viii) (MEYOSP/PS) Technical assistance for small producers; and (ix) Training for producers * Implement, through agreements, actions to promote . Monthly aid (between $100 and $200) and civil private employment, responding to regional demands responsibility insurance or productive sectors in emergency or crisis which 159 Thematic Area ObjecttVes .evtces Provided Special Employment contribute to improving the situation of the unemployed Programs and most vulnerable (MTSS) * (i) Labor training to increase the employability of . (i) Grants from $80 - $160/month and civil responsibility workers linked to certain sectors or branches of insurance; and (ii) Payment for training activity; and (ii) Stimulate the association among actors who need to develop training . Special Training Programs (MTSS) . (i) Establish transitory employment through social * Non-remunerative economic aid ($160/month), civil service projects for the community, oriented to responsibility insurance and medical assistance coverage improving the quality of life of the population; and (ii) Reduce the impact of the loss of earnings due to loss of employment . Community Services . Unemployment insurance ($150-$300/month), medical * Establish income and medical coverage among those services and regular corresponding payments who lost their jobs and are newly employed * (i) Subsidies (up to $300,000) for training workshops for . Integrated System of * Develop quality labor training through the creation or civil society; and (ii) Technical assistance and training for Services for the adaptation of institutions designed to meet the needs workers in micro-, small and medium- enterprises Unemployed of occupational training for human resources directly (SEGURO DES) related to the demands of the local productive markets, and those in the zone of influence . Occupational . Monthly non-remunerative economic aid (up to $200), Workshops * Develop transitory employment for unemployed accident insurance and health coverage (MTSS) workers in conditions of poverty or social vulnerability, reducing the impact of the loss of earnings in poor areas due to loss of employment, and improving the employability of these workers through works considered socially relevant * TRABAJAR II (TRABAJAR 2) Health * Immunization Program * Prevention of immuno-preventable diseases . Doses distributed (AS; BCG; D.AD; DPT; SABIN; TT) (MSYAS/PAO 160 The0atic Aea0 Ojcti S c • Maternal and Child * (i) Reduce infant mortality from mother-child . (i) Technical assistance for provincial maternal and child Health Program malnutrition; (ii) Health promotion for children and programs; (ii) Elaboration of norms for prevention, health (MSYA S/PM) adolescents in the entire country within the framework of protections and attention to mothers, children and health policies already established; and (iii) Promote and adolescents; (iii) Powdered milk; (iv) Medicine; (v) Training develop the strategy for primary care for maternal and workshops for health teams child health * (i) Extension, revision and equipping hospitals, health . Maternal and Child . (i) Reduce infant morbidity/mortality, improving the centers, infant feeding centers, nurseries and daycare Health and Nutrition focus, design, extension and coordination of the services centers; (ii) Training; (iii) Common and modified milk; (iv) Program and programs for health, nutrition, free feeding and Provision of modified milk and boxes of free meals (MSYAS/PROMIN) infant development; (ii) Promote the psycho-social development of children from 2 to 5 years of age; and (iii) Reduce the prevalence of maternal-child malnutrition * (i) Insecticides; (ii) Training workshops; (iii) Control of . In five years, eliminate the occurrence of new cases of blood sampling; (iv) Studies of those infected ; (v) . National Program for Chagas, developing actions for control of the parasite in household surveillance; (vi) Serology in those from 0 to 14 Chagas Control the entire country years of age (MSYAS/CHAGAS) a (i) Reagents for studies on the viral "load" of HIVIAIDS * (i) Prevent sexual, pre-natal, blood, hemophiliac and patients; (ii) Anti-retroviral medications specifically designed transplant, and drug use-related transmission; (ii) for HIV/AIDS and other opportunistic infections; (iii) . National Program for Reduce the individual, family, and socio-economic Reagents for blood banks to test for HIV and STDs; and (iv) the Fight Against Human impacts; and (iii) stimulate the production of information Medication for STDs Retroviruses (AIDS and which allows evaluation of the tendencies and STDs)_(MSYAS) projections of the epidemic * (i) Financing for projects for prevention, selected from * (i) Reduce incidence of infection of HIV and STDs; (ii) those presented by OCSs; (ii) Training of health care Inform the population and stimulate the adoption of personnel; (iii) Prevention campaigns for HIV/AIDS, and the . Project for the Fight behavior to prevent HIV and STDs; and (iii) Improve the production of informational materials; (iv) Formation of a Against AIDS and administrative capacity of the programs for the national network of education disseminators; (v) Teacher training for Sexually Transmitted and provincial programs against AIDS and STDs, as well AIDS prevention in schools; (vi) Institutional strengthening Diseases as giving integral attention to the population living with for hospitals and health centers (MSYAS/LUSIDA) HIV/AIDS and/or STDs at the hospital and ambulatory care levels 161 Thematic Area Obj_ctives Services Provided Income Subsidies . Attention to Non- * Administration of the system for non-contributing * (i) Others; (ii) Pensions for assistance to invalids, soldiers Contributing Pensioners pensions, and medical coverage with the limits in the Malvinas War, families of the Disappeared, mothers determined by Decrees 292/95 and 492/95 of more than 7 children, and the aged population. * Social Aid * Attention to emergency situations for the extremely * (i) Housing or food provided by treatment centers in poor, without medical coverage, and at high risk Buenos Aires; and (ii) Subsidies for health care Care for the Aged . Solidarity Support for . (i) Improve the living conditions of seniors (60+ years * (i) Free medication; (ii) Food ration; and (iii) Tours/visits Seniors of age) in situations of social risk, facilitating the (ASOMA) satisfaction of basic needs; and (ii) Promote the integration of seniors with other age groups to promote solidarity and share experiences and values * Provide boarding in Geriatric Institutes for the elderly * Boarding rooms in Geriatric Institutes * PAM I-Geriatrics (PAMI) poor • (i) Rent assistance; (ii) Complimentary clothing; and (iii) . PAMI- Economic * Improve living conditions of retirees and pensioners in Medication Subsidies (PAMI) states of emergency . Food packages * PROBIENESTAR * Improve conditions of the poor, and reduce social (PROBIENESTAR) isolation of retirees and pensioners in PAMI, and in situations of social risk Housing and/or Social Infrastructure * (i) Legalize the properties of land settlements; and (ii) . (i) Construction and improvements in housing; and (ii) . ARRAIGO Construction and/or reconstruction of habitats Legalize properties of land settlements (CTFN/ARRAIGO) * (i) Support to communities and municipalities with * (i) Credit support to productive sectors; (ii) Credits for actions for improving the quality of life of the population; equipment; (iii) Credits for institutional development; and . Financing to (ii) Improve financial administration at the provincial and (iv) Credits for works Municipalities municipal levels; and (iii) Reinforce the institutional (PRODISM) capacity of local governments * Reduce the housing deficit and improve housing and . (i) Housing solutions; and (ii) Housing constructed sanitary conditions, as well as the quality of life of the (constructed through contracted firms, intermediate entities, 162 target population municipalities or communities) . National Fund for Housing . Contribute to the development and improvdment. of . (i) Provision of basic equipment for communities; (ii) (Prov. Gov./FONAVI) conditions of rooms, housing, basic infrastructure and Infrastructure works (rooms for multiple uses, rooms for access to land in NBI homes and for vulnerable groups first-aid, etc.); (iii) Regularization and urban planning in in situations of risk, emergency and marginalization, as irregular settings and poverty-stricken areas;(iv) Housing * Improvements in well as those affected by environmental emergencies solutions (improvement, complimentary . housing, credit Housing and Basic Social funds); and (v) Houses Infrastructure (PROG17 SSXI) * Improve the sanitary conditions of small localities . Construction and/or expansion and improvement of small works for drinking water and basic health • Program of Social . Contribute to improving the quality of life of the NBI . (i) Financing of project for improved drinking water; and Assistance for Provision population in border areas in the La Plata Valley (ii) Financing for project for basic health of Healthy Drinking Water, and Good Health (MEYOSP/PASPAYS) * (i) Improve the living conditions of the NBI urban * (i) Public infrastructure with inter-sectoral connections, population, situated in neighborhoods with infrastructure included the health unit; (ii) social support, technical * Program for Social needs, environmental problems and/or legalization of assistance and training; (iii) guarantees of land entitlements Development in Border land settlements through the execution of a group of to neighborhood populations; and (iv) technical assistance Areas of NOA and NEA works and services, as well as the strengthening of and training for institutional strengthening with NBI populations community organization; and (ii) Facilitate access to (PROSOFA) basic services, with a social focus to guarantee the sustainability and replicability of the program * Neighborhood Improvement Program I P:\1copr\poverty\Part2-gray.doc