Ig995T Human and Social Development Group _ _ 77 LATIN AMERICA AND THE CARIBBEAN REGION -, ~ ^~ ~~HIDE OF THE POOR: - IN LATIN AMERICA AND THE CARIBBEAN ~~~~~ ~~~by -t Jacques van der Gaag Donald Winkler LASHC Paper Series No. 1 July 1996 Papers prepared in this series are not formal publications of the World Bank. They present preliminary and unpolished results of country analysis or research that is circulated to encourage discussion and comment; any citation and the use of this paper should take account of its provisional character. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, its affiliated organization, members of its Board of Executive Directors or the countries they represent. TABLE OF CONTENTS I. Poverty in Latin America and the Caribbean (LAC) II. Social Progress in LAC III. The Status of Poor Children IV. The Cost of Equal Access V. Targeting Basic Services on Poor Children VI. What the World Bank is Doing VII. Summary and Questions CHDREN OF THE POOR iN LATIN AMERICA AND THE CARIBBEAN Most of the poor are children and most children are poor (Children of the Americas, UNICEF, 1992). I. POVERTY IN LATIN AMERICA AND THE CARIBBEAN (LAC) According to official statistics, today more than 165 million people in the LAC Region live in poverty. The incidence of poverty has declined steadily, from 60 percent in 1950, to 37 Figure 1.1 Poverty in Latin America, 1950-1995 65 60 °.\ 17 °) 45\ 40 312 80 1 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 So JLLea0o. 199B Years Figure 1.2 Poverty in Latin America, 1950-1995 170 C 140_ / 1 r30_/ o120_ -/ 90 810l 70 7 1950 1955 1950 19S5 1970 1975 1980 1985 1990 1995 Years Source: J L L..doo 199Y CHILDREN OF THE POOR IN LATTN AMERICA AND THE CARIBBEAN percent in 1995 [Figure 1.1].' However, relatively rapid population growth has offset this decline: the number of poor continues to grow [Figure 1.2]. Though the incidence of poverty is highest in rural areas, the urbanization of the Region-- 73 percent of the population lives in urban areas--has had as a consequence that most of the poor now live in urban areas [Figure 1.3]. Figure 1.3 The Growth of Urban Poverty in Latin America, 1970-1995 180 160 140 o 120 EL 100 '- 0 co 80 .2 60) 40 20 0 1970 1975 1980 1985 1990 1995 arural a urban Source: J L. Lonromo 19s6 Years Of the 165 million poor in the Region, we estimate that about 58 million, or 35 percent, are under 15 years of age. Thus, the growing number of poor implies that the number of children living in poverty continues to grow. Indeed, by now almost 60 percent of all children are poor, while one-third of them belong to the poorest 20 percent of all households. In this paper we will address the social needs of poor children. How do they live? What access do they have to health services, education, safe drinking water and sanitation? What are their chances to escape from poverty and live a long and productive life? Londoilo, J. Luis, Poverty, Inequality, and Human Capital Development in Latin America 1950-2025, World Bank, 1996. 2 CHIM REN OF THE POOR IN L4TIN AMERICA AND THE CARIBBEAN First we will assess their living conditions relative to those of their parents when they were young. We will see that despite the growing number of people who live below an income poverty line, social indicators have increased steadily in the Region. This suggests that the living conditions of the poor, as measured by characteristics other than income, have also improved significantly, at least in parts of the Region. Next, we will access the conditions of poor children relative to their more fortunate counterparts. We will attempt to quantify the "social gap" between poor and non-poor children and assess the investments necessary to fill this gap. We will subsequently discuss policies and programs that need to be put in place to close the social gap for poor children and assess the World Bank's own efforts in this area. But first we will provide a brief overview of social progress in the Region in the past few decades. II. SOCIAL PROGRESS IN LAC The LAC Region has experienced rapid income growth during the seventies, but growth faltered in the early eighties. By 1990, per capita GDP had reached $1,850, a level already achieved in 1977 [Figure 2.1] Moreover, the most recent growth projections for the Region imply that for the foreseeable future, the number of poor will continue to grow. 3 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN Figure 2.1 Per Capita GDP in Latin America and the Carribean, 1970-1992 2400k 2. 2000 CD a2000 + 4 o 1600- C / EL/ CD 1400 1200 N . . E 0 : ; : Soute MMd Ba* Data Base Year Figure 2.2 Infant Mortality Rate and per Capita GDP in Latin America and the Carribean, 1992 100 I 9 Haib -°T Bolivia 870 'a \ Guatemala o 60T 1.50 - a) 40- - E 30- Z Colombia 20 - 0 :C~hile* Jamaica* * Costa Rica 1o - 0 200 1200 2200 3200 4200 5200 6200 7200 8200 9200 Soune Wold BarkDama Base Per Capita GDP in 1987 U.S. $ Against this erratic behavior of income growth, trends in social indicators look considerably more positive. Infant mortality decreased from over 80 in 1972 to about 45 in 1992. While three decades ago gross primary school enrollment was 75 percent, today 13 4 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN countries have achieved virtually universal enrollment, and the Region's average net enrollment is now about 88 percent. For secondary education, enrollment increased from 14 percent in 1960 to 55 percent in 1993. Immunization, virtually non-existent in the sixties has reached 85 percent in many countries. For the Region as a whole, the picture appears to be one of tremendous and virtually uninterrupted social progress against a background of uneven economic performance. At the same time, the Region falls far short as compared to the developed world. Infant Mortality, for instance is still six times higher, and secondary education is one third lower than in OECD countries. Moreover, individual country data reveal that much more needs to be done. Figure 2.2 shows Infant Mortality rates (IMR) for all LAC countries, ranked by GDP per capita. The IMR is highest in Haiti (93), followed by Bolivia (82) and Guatemala (62). Figure 2.3 Secondary School Enrollment Rate and Per Capita GDP in Latin America and the Carribean, 1992 90, 80s- 70 -hde = 60/ tolombia LU50 / 30 t Venezuela 20- HoncILrs 10 400 900 1400 1900 2400 2900 3400 3900 Per Capita GDP in 1987 U.S. $ Source: Wodd Bank Data Base At the other end of the spectrum, Jamaica, Costa Rica, Trinidad and Tobago, Barbados, and Puerto Rico, all have reached levels of Infant Mortality of 15 or lower. The Figure clearly 5 CHLDREN OF nTH POOR IN LATD AmERIcA AND THE CARIBBEAN shows (1) that there is a strong relationship between the IMR and GDP per capita, and (2) that policy matters. The latter is revealed by the large deviations from the trend line: countries can do much better or much worse than expected on the basis of the per capita income level. Figure 2.3 shows a similar picture for enrollment in secondary education. Clearly, a country's level of development matters, but some countries--e.g. Chile, Jamaica, Uruguay--do much better and others--e.g. Honduras, Venezuela--much worse than expected on the basis of their GDP per capita. In sum, the Region has seen tremendous but uneven social progress during the past few decades. This process implies that in terms of basic needs--safe water and sanitation, basic health services (including immunization), basic education--the children of the poor today are much better off than their parents several decades ago. Despite this progress, the social deficit looms large. It provides a major challenge to the governments of the Region, as well as to the development community, to be met within an uncertain environrnent for economic growth. This challenge is particularly relevant for the children of the poor, because they are the ones who face premature death, excess disease, and low probabilities to find and complete basic education. It is to these issues that we will turn in the next section. III. THE STATUS OF POOR CHILDREN Poverty is hereditary. Like a defective gene, it is passed on from generation to generation. The surest way to break this cycle and to provide children that grow up in poor 2 households with a better chance to escape poverty, is through education. Psacharapoulos et al. in a regional study for Latin America and the Caribbean concludes "Educational attainment has the greatest correlation with both income inequality and the probability of being poor". 2 Psacharopoulos, G.S Morley, A. Fiszbein, H. Lee, and B. Wood, Poverty and Income Distribution in Latin America, World Bank, 1993. 6 CHILDREN OF TEE POOR IN LATI AMERICA AND THE CARIBBEAN While underscoring the need for high quality schooling, recent research has stressed the importance of taking a more holistic approach to the development of the young child. Malnutrition at an early age (even as young as at birth), has been associated with stunted cognitive development, late school enrollment and poor school performance. Stemnberg and Grigorenko3 conclude from an extensive survey of the literature that "health deficits in children - whether caused by undernutrition. infection or intoxication - can have adverse consequences for cognitive development ..... delay in making suitable interventions can only result in millions of children being consigned to the development of cognitive skills that function at levels well below the children's potential". Moreover, Martorell4 has shown that the future consequences of ill health and poor nutrition are aggravated by poor socio-economic conditions. Indeed in general, growing up in poor households, with poor and badly educated parents, is a strong predictor for low school enrollment and bad school performance. Thus, the best way to break this vicious cycle of intergenerational transfer of poverty is to invest in the human capital of poor children, their health, their nutritional status, and their schooling. Numerous studies have addressed this issue. UNICEF, for instance, states: "Today's children will determine the future of Latin American and Caribbean communities . ..... Health, education and normal child development are links of a chain that is essential to human development. If any link in this chain is broken, be it through disease, malnutrition, failure in school, or exclusion from society, development is jeopardized and a regressive process sets in whereby poverty is passed from generation to generation"'. At the World Summit for Children, held in New York on September 30, 1990, more than 70 heads of states and an additional 88 official country representatives, agreed to a concerted 3 Stemnberg and Grigorenko, Effects of Children 's Ill Health on Cognitive Development, 1996. 4 Reynaldo Martorell, Undernutrition During Preganancy and Early Childhood and its Consequences for Cognitive and Behavioral Development, 1996. Children of the Americas, UNICEF, 1992. 7 CHILDREN OF THE POOR N ILATIN AMERICA AND TnH CARIBBEAN effort to enhance living conditions for children by the year 2000. And the Narifio accord, signed in April 1994, in Bogota, reaffirms the commitment of Latin American generations to investing in children as a means to sustainable development with equity. The Narifio accord also defines monitoring mechanisms to measure progress towards the regional goals. In this paper, we will complement these efforts with a profile of children who live in poor households. Data and Definitions. Over the past several years, the World Bank has helped Latin American and Caribbean countries carry out Living Standard Measurement Surveys (LSMS). The resulting data have been used by the Bank and others to construct poverty profiles for the Region. For this paper, we have selected four recent LSMS--surveys, for Ecuador, Jamaica, Nicaragua, and Peru--which have national samples and which contain detailed data on the status of children. In addition, we have used a recent household survey for Chile (CASEN, 1992), which includes similar data on children.6 What constitutes poverty can, of course, be defined differently for each country, or it can be defined as an absolute income or expenditure level per individual or household for all countries in the Region. The former definition takes account of each country's particular circumstances, while the latter permits cross-national comparisons. We have selected a compromise between these two alternatives and defined poverty as the bottom forty percent of the income distribution. This definition facilitates cross-country comparisons, while also permitting the poverty line (that level of income or expenditure below which a family is declared to be poor) to vary by a country's per capita income. Also, since children almost always live in households, we have elected to use households rather than individuals as the unit of analysis and to define the poor as the bottom forty percent of 6These countries were selected because these household surveys could be easily analyzed; future research will expand the number of countries to be more representative of the entire Region and to cover most of the children of the Region. 8 CHLDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN the income distribution of households (ranked by per capita income) define the very poor as the bottom twenty percent of households. Who are the poor children? Children are over-represented in poor and very poor households. In the five countries examined, over 30 percent of children are located in the bottom 20 percent of households and almost 60 percent of children are in the bottom 40 percent of households. Furthermore, as shown in Figure 3. 1, the concentration of children in poor and very poor households is much greater in rural areas than urban areas, in all five countries. Figure 3.1 The Percent of Very Poor Children in Urban and Rural Areas of Selected Countries 60 50 0 D~40- CD _ 30- *) 20-7 10 -- 0 Chile Ecuador Jamaica Nicaragua Peru *Urban o Rural Countries Source: CASEN-Qie (1992), LSMS-Eoaor (1994), Jamaica (1991), Ncaragua (1993), Peru (1991) That children are over-represented in poor households is not surprising7. Quite simply, poor families are larger than families that are not poor. The magnitude of this difference, however, varies widely across the five countries studied. Extremely poor families in Chile and Peru are only slightly larger than non-poor families, while in Ecuador, Jamaica, and Nicaragua, extremely poor families have almost two more family members than do non-poor families. See Annex Table 1 for detailed data on characteristics of poor and non-poor households. 9 CHILDREN OF THE POOR IN LATi AMERICA AND THE CARIBBEAN In addition to having larger families, poor households have higher ratios of dependent family members (those below age fifteen and above age sixty-five) to adults in the economically- active age group 15-64. Chile, which is further along in the region-wide process of demographic transition, has both significantly smaller families and lower dependency ratios for all income groups. Figure 3.2 Percent of Female Headed Households (of All Households with Children) in Selected Countries, By Poverty Level so 451 401 35 30 I 0- 0 Chile Ecuador Jamaica Nicaragua Peru * Very Poor a Poor a Non-Poor Countries Sur: CASEN-Chle (1992): LSMS-Ecuadr (1994). Jamaca (1991). Nlcargua (1993). Peru (1991) Two other family characteristics appear to be unrelated to poverty status. The age of the household head varies little across income groups, as does the percent of heads who are female. On the other hand, the percent of household heads who are female varies widely by country and. to a smaller extent, by urban-rural location. In Jamaica, 48 percent of al! households are headed by women and in Nicaragua 27 percent are headed by women, compared to about 15 percent in the other three countries. Also, when household characteristics are disaggregated by location of the household, female-headed households are somewhat more common in urban than rural areas, but, again, there is little difference across income groups. 10 CHILDREN OF THE POOR iN LATnN AmERICA AND THE CARIBBEAN Figure 3.3 Education of Mothers in Households with Children in Selected Countries, By Poverty Level 10 0 8- LU - T E 2- 0 0 Ecuador Jamaica Nicaragua Peru * Very Poor 0 Poor C Non-Poor countries Source: LSMS-Ea (1994). Jamaca (1991). N icaraga (1994), Pemu (1991) Finally, as shown in Figure 3.3, what may be the most important household characteristic--the education of the mother--differs greatly by income level. Moreover, the education level of the poor, especially, varies widely by country.8 In Ecuador, the average mother in non-poor households has 9.1 years of education, which is more than double the 4.0 years for mothers in very poor households. And the average 7.3 years of education of mothers in very poor households in Jamaica exceeds both the 5.7 years for mothers in non-poor households and the 1.2 years for those in very poor households in Nicaragua. Finally, disaggregating by urban-rural location shows that education levels are consistently higher in urban than rural areas for all countries and all income groups. In sum, from what is known in the literature on poverty, children born in poor households start with a disadvantage for success later in life: they have more siblings, poorer educated parents, and the very poor live in often remote rural areas--all strong correlates of future poverty. Research consistently shows that higher levels of parental education, and, especially, mother's education, contributes to school attendance and achievement, to better nutrition, to better health, and to lower birth rates. 11 CHI.DREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN Of course, public policy could compensate for this disadvantage in early life, by providing basic services (nutrition, health and education) that will equip the children of the poor to compete fairly as adults in the labor market. We have already seen that in this respect the poor children of today are much better off than their parents when they were young. We will now compare access to basic services between today's poor and non-poor children. Do poor children receive the same services as the non-poor? The 1990 World Summit for Children and the 1994 Narifio Accord specified a number of goals in health, nutrition, water and sanitation, and education. These goals provide a good basis for defining the basic services which all children should receive. In what follows, we examine the data available in the household surveys to assess the extent to which public policies have succeeded in reaching poor children. Figure 3.4 Vaccination Rates for Children less than 5 Years in Selected Countries 90 s0 701 cis~~ ~ 60er Por Oor7Nnpo 50- 40 - O 30- 20-. 10 0 . - II II I Ecuador Nicaragua Peru EVer Poor 0 Poor 0 Non-Poor Countries Source: LSMS-Ecuador (1994). Nicaragua (1993). Peru (1991) As it turns out, the answer to the above question depends very much on the particular service. Figure 3.4, for instance, demonstrates that there is a high level of access to vaccination for young children. The differences between countries and between income groups are relatively minor, although most countries still face the challenge of reaching universal vaccination. 12 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN Figure 3.5 Access to Piped Water by Very Poor Households with Children, By Urban and Rural 100 90 80 0 80_ 70 CO 60, 0 z soa o 0 401 0 303 O) 20 10 0 Chile Ecuador Nicaragua Peru i Urban r Rural Countries Soire: CASEN-Chle (1992): LSMS- Ecuaor (1994), Ncaragua (1993). Peru (1991) Figure 3.6 Access To Sewerage by Very Poor Households with Children, By Urban and Rural 90 0 80- * ~70 a, 60i 504 0 40~ 0 0. 30 0 20.. lo 10 Chile Ecuador Nicaragua Peru Urban I- Rural Countries Soubrce: CASEN-Chile (1992): LSMS-Ecuador (1994). Nicaragua (199). Peru (1991) 1 3 CEILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN However, water and sanitation conditions, which are important determinants of health risks to children, vary widely across countries, across income groups, and, especially, between rural and urban areas. Figures 3.5 and 3.6 show large differences in access to piped water and sewerage between urban and rural areas. Closer examination of the data shows relatively small differences in access between income groups within urban areas and quite large differences in access between income groups within rural areas. For example, in Peru 80 percent of poor and 94 percent of non-poor urban households have access to water, but in rural areas only 55 percent of poor households have access, compared to 86 percent of non-poor households. The differences in rural areas are even larger in access to sewerage. For example, 14 percent of poor households have access in rural Chile, compared to 38 percent of non-poor households. Figure 3.7 Access to Health Services by IlIl Children less than 15 Years in Selected Countries, By Poverty Level 80 CD 70 I 60 CD 50 :030 0) C 202 10 0 Ecuador Jamaica Nicaragua Peru * Very Poor 0 Poor c Non-Poor Counties Sorce: LSMS-Ecuador (19964). Jamaica (991). Ncaragua (1993). Peru (1991) Since the conditions that put children at health risk are so unequal, access to health care services for children who are ill is very important in determining their health status. Here we find the patterns vary significantly by country. Even though the average household's access to 14 CHM[REN OF THE POOR IN LATIN AmERICA AND THM CARIBBEAN health care is high in Ecuador and relatively low in Jamaica, in both countries access to health care by children is relatively equal across income groups9. In contrast, as shown in Figure 3.7, there are very large differences in access by income group in Nicaragua and Peru. As shown earlier, most children in Latin America and the Caribbean now have access to primary schooling. It is the quality of schooling that varies tremendously between income groups, not the access to basic education. Still, the enrollment rates for children in Nicaragua show that some countries face large challenges in simply providing access. In Nicaragua, only 52 percent of very poor children aged 6-11 are enrolled in school, compared to 88 percent of non-poor children. Figure 3.8 Preschool (3-5 Years) Enrollment Rate in Selected Countries, By Poverty Level 90 80 0 60. so w 40 30 201 10- Chile Ecuador Jarnaica Nicaragua Peru *Very Poor oPoor 0Non-Poor Countries Source:CASEN-Chile( 1992): LSMS-Eajudor (1994), Jamaca (1991), Nicargua (1993), Peru (1991) 9 Access by the non-poor appears to be lower in Ecuador and Jamaica. This is probably due to the self assessment of health status: parents in richer households are more likely to report minor health conditions for which no professional care is necessary. 15 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARmBEAN Inequities in access to schooling now lie mainly at the preschool and the secondary school levels. For the preschool age group (ages 3-5), Figure 3.8 shows the disparities that exist between children of poor and non-poor households in all countries. For the secondar) school age group (ages 12-17), enrollment rates vary both by poverty level, as shovwn in Figure 3.9, and by urban-rural location. Figure 3.9 School Enrollment Rate for the 12-17 Age Group in Selected Countries, By Poverty Level 100 90 - 80 - * 70- -6 60. 40 00 Chile Ecuador amca Nicaragua Peru * Very Poor o Poor Cj Non-Poor Countries Soure:CASEN-Chile (1992). LSMS-Ecuador(1994). Jamaica (1991). Ncaragua (1993), Peu (1991) Conclusion. Growing up in disadvantaged socio-economic circumstances increases the chances of poor children to become the next generation of poor adults. Public policy--i.e. providing basic social services to the poor--can compensate for this, at least in part. Whether such policies in the countries in our sample have been successful, depends on the specific services. For vaccination and primary education, most countries in the Region are approaching universal coverage, and thus--by definition--coverage of the poor. Some countries (e.g. Jamaica 16 CHILDREN OF T1E POOR IN LATIN AMERICA AND THE CARIBBEAN and Ecuador in our sample) have also succeeded in providing equal access to health services'0 but in other countries the poor receive much less care than the non-poor. In all cases we find that access to safe water and sanitation is highly income-dependent. For education, inequities are most severe for pre-schooling and secondary education. Equally striking as the differences among income groups, are the large differences in access to basic services between urban and rural areas. The absolute number of poor is growing in urban areas, but rural areas have higher concentrations of children in abject poverty, and their access to basic services is even lower than that for poor urban children. Finally, while these disparities are alarming and discouraging, the data show one lack of disparity which is particularly encouraging. Differences in access to basic services appear to bear very little relationship to the sex of the child. IV. THE COST OF EQUAL ACCESS Methodology. The statistical exercise reported above allows us to fairly accurately determine how many children fail to receive basic services. In what follows, we estimate the cost, as a percent of GDP, for countries to bring the levels of coverage for poor children up to the average level of their more fortunate counterparts. This goal would go a long way to providing children living in poor households with a level playing field. By providing them with the basic services needed to augment their human capital, they would no longer be at a disadvantage vis-a- vis children born in richer households. We will focus on pre-school, primary and secondary education, as well as on a basic package of health services. The latter is defined in the 1993 World Development Report" I. The '° We have no direct data on the quality of care received, but other studies have shown that the poor generally receive lower quality care. See, for instance, Judy Baker and Jacques van der Gaag in E. van Doorslaer, et al., Equity in the Finance and Delivery of Health Care, Oxford, 1993. We did not include estimates for water and sanitation, since it is impossible to sort out which part of the necessary investmen is for children. 17 CHauREN OF THE POOR IN LAT AMRIcA AND Tm CARIBEAN estimates are calculated by simply multiplying unit costs by the number of children which must be covered to attain this goal. This methodology assumes constant unit costs and ignores the 12 question whether existing resources could be used more efficiently' . Also, as noted earlier, we have ignored the issue of differences in the quality of services received by poor and non-poor children and, thus, have not included in these estimates the cost of raising the quality of services received by poor children. Table 4.1 Cost as Percent of GDP for Bringing Poor Children up to the Level of Access to Basic Services Equal to the Average of Non-Poor Children Basic Needs Chile Ecuador Jamaica Nicaragua Peru Preschool 0.05 0.29 0.11 0.69 0.23 Primary Education 0.01 0.06 0.01 0.91 0.00 Secondary Education 0.02 0.24 0.24 1.28 0.02 Health 0.00 0.00 0.00 1.28 0.07 Total 0.08 0.59 0.36 3.01 0.32 Estimated Costs. Table 4.1 reports the costs of meeting the goal for the five countries for which we have data. The estimates range from 0.08 percent of GDP in Chile to 3.01 percent of GDP for Nicaragua. The estimates for Nicaragua are high, of course, because the existing level of access to basic services is so low. From the rough estimates, it is clear that--with the possible exception of Nicaragua--the cost associated with bridging the social gap between poor and non-poor children is well within the financial ability of most countries. Moreover, some of the resources could come from more efficient and better targeted use of existing budgets. In addition, private resources, as well as 12 Unit costs for health were obtained from WDR, 1993. Those for education are mainly from UNESCO- OREALC. 18 CHBLDREN OF THE POOR IN LATIN AMERICA AND T=E CARIBBEAN ODA, could be used to provide basic services for poor children. Clearly, if the political will is there, and if all parties (government, private sector, NGOs, donors and development agencies) work together, this simple goal--providing a level playing field for poor children--is financially achievable. V. TARGETING BASIC SERVICES ON POOR CHILDREN Policy Matters. Figures 2.1 and 2.2 showed a strong relationship between per capita GDP and infant mortality rates and secondary school enrollment rates. However, the fit between income and coverage of basic children's services isn't always so tight. Figures 5.1 and 5.2 show that countries with almost identical levels of per capita GDP have very different immunization rates and pre-school enrollment rates. The Dominican Republic's inumunization rate is 47 percent and Honduras' is 94 percent, yet their per capita GDP are not very different. Quite clearly, the spending priorities and public policies of countries matter. Figure 5.1 Immunization Rate (DPT) for Children Under 5 Years and per Capita GDP in Latin America and the Carribean, 1992 95 Honduras Chile 85- Colombia 75 E E 65- 3- 55- IL . Dominican Rep. 45 - 35 200 700 1200 1700 220 2700 3200 3700 Soume:Wo~ddBankDataBase Per Capita GDP in 1987 U.S $ 19 CHLDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN One way in which richer countries attain better coverage of basic services is by spending more than their poorer neighbors. Yet. Figure 5.3 shows an imperfect fit between per capita GDP and social expenditures as a percent of GDP. Countries like Chile and Costa Rica spend more than expected given their incomes, while Peru and Venezuela spend less. Furthermore, the allocation of social expenditures across services and its distribution by income quintile vary by country. 13 Figure 5.2 Preschool Enrollment Rate and Per Capita GDP in Latin America and the Carribean, 1991 45 Jamaica 40 - n 30 251 r- 20 Chile 5- 0 0 500 1000 1500 2000 2500 3Q00 3500 Per Capita GDP in 1987 U.S. $ Soure: Wcdd Bank Data Base and UNESCO-OREALC 13 See, for example, CEPAL, El Gasto Social en America Latina: Un Examen Cuantitativoy Cualitativo, 1994. 20 CEHLDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN Figure 5.3 GDP Per Capita and Social Expendiure as Percent of GDP in Latin America 25 0 Costa Rica ~ 20- C) C t 15 Chile CD ) DE * .Venezuela 0.0 LU U) .Peru 0 500 1000 1500 2000 2500 3000 3500 4000 GDP Per Capita in 1987 U.S. $ (1992) So,zo: Wofld Bank Data Base and CEPAL (1996) Whether it is because they spend more as a percent of GDP or because they target their social expenditures on the poor, some countries appear to consistently do better in providing services and improving the welfare of children than would be expected given their income levels. Figure 5.4. for example, shows that Colombia, Jamaica, Costa Rica, and Chile, among others, have lower malnutrition rates than would be expected given their levels of per capita GDP. Graphs presented earlier show that Jamaica, Chile, and Costa Rica also consistently do better than expected in terms of immunization, school enrollment rates, and infant mortality. Consistent and Comprehensive Policies Work. What explains the performance of Chile, Costa Rica, and Jamaica in providing basic services to children? One explanation may be a political commitmnent to spend on social services. Figure 5.3 shows that Costa Rica lies further away from the regression line than any other country, meaning it spends far more on social programs than would be predicted given its income. However, according to that same figure, Chile spends just about what would be expected, so high levels of spending can't be the only explanation for its good performance. Its performance may instead be partly explained by its success in targeting social expenditures on the poor. In addition to high social expenditures and, 21 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN more recently, targeting of expenditures on the poor, all three countries have long histories of providing comprehensive health and education services. Figure 5.4 Malnutrition Rate for Children Under 5 and per Capita GDP in Latin America and the Carribean, 1992 40 35 w 30 2- 0 20 200 700 1200 1700 2200 2700 3200 3700 4200 Per Capita GOP in 1987 U.S. $ Source: Word Bank Data Base The experience of these and other countries suggests poverty reduction requires sustained political commitment, which in turn requires a consensus by civil society about the importance of providing basic social services to all. While the capacity to design and manage policies and programs is clearly important, the essential ingredient in the recipe for poverty reduction would appear to be sustained conmmitment to deliver a comprehensive set of social services to which the poor have access. Innovations in Programs. The comrnitment to provide services does not dictate any single approach to financing and delivering those services. Indeed, the increased emphasis on popular participation, the decentralization of many government activities, the gradual elimination of general subsidies and price controls--in favor of more targeted interventions--and the reduced 22 CBIuDREN OF THE POOR IN LATIN AM!ERICA AND THE CARIMBEAN size and scope of the public sector have all contributed to a large and increasing variety of innovative programs to deliver services to poor households and poor children. Social service innovations in Latin America and the Caribbean include partnerships between public and private sectors, NGOs, and community groups; efficient targeting; contracting service delivery to the private sector; and development of low-cost service options. The Province of Mendoza, Argentina, supplies basic sanitation and other services to low income households through partnerships between public agencies, private contractors, NGOs, and neighborhood organizations through a program called MENPROSIF. Under this program, the beneficiaries participate in designing, financing, and managing public works, thereby lowering costs and permitting provincial government to leverage its own limited financial resources and thereby increase access to basic sanitation and water supply. Chile gives us two examples of efficient targeting. Through a program called PAE, Chile has since 1964 provided food rations to needy 6 to 14 year old children enrolled in publicly- funded schools. School lunches are targeted on those schools which forecast models predict have high nutritional needs. The actual provision of food rations is carried out by private firms which have won public bids that minimize cost for a given level of quality. Almost 80 percent of this program's benefits are received by the poorest 40 percent of school children. Under another Chile program called The 900 Schools, which started in 1990, the Ministry of Education used measures of family socioeconomic status and student academic achievement (as measured by a national standardized examination) to identify the 900 neediest schools. These schools were provided with a basic set of learning materials (textbooks, workbooks, libraries) with accompanying teacher training and didactic materials. The result has been a significant increase for those children in average test scores in both language and mathematics, with performance in the 900 schools now exceeding the national average. In several countries in Latin America and the Caribbean, business groups are making important contributions to education. These include measures to strengthen schools in industrial comnrnunities, establish innovative private schools for the poor, reform school management, 23 CHILDREN OF TE POOR IN LATIN AMERICA AND THE CARIBBEAN develop innovative curricula in mathematics and technology, etc. 14 Examples of school-business partnerships include the Clemente Mariani Foundation program in support of municipal education in Brazil, the EDUCO/FUSADES program in El Salvador, the Fundacion Corona's education technology initiative in Colombia, and the Business Partners for Education program in Jamaica. Community participation can both reduce the costs and increase the effectiveness of social services, as shown by the provision of early childhood care in several countries. In Colombia, the community is responsible for organizing and managing home-based care (Hogares de Bienestar Infantil), including supplemental feeding of children in poor neighborhoods. Volunteer mothers receive a nominal stipend and home improvement loans to provide the necessary facilities, and they receive training in child nutrition and health. The program started in 1987 and now reaches about one million children in the country. Under a center-based childhood care program called PRONEI in Peru, the government trains and pays a paraprofessional caregiver and provides food rations, but the community is required to provide the physical facility, for preparing food, for selecting the caregiver, and for managing the center. Finally, under a quite different approach called Child to Child in Jamaica, children aged 9 to 12 are educated about the health, nutrition, dental care, and psycho-social development of young cEhildren. Since many of these children have child care responsibilities in the home, the quality of care improved. In addition, the knowledge of parents and guardians about child care improved. This brief summary of innovations in service delivery has by no means been exhaustive. However, just the few programs reviewed here demonstrate that basic services can be effectively and efficiently delivered to poor children through a variety of policy and program options. There is no shortage of policies and programs that can be adapted to particular country circumstances. 4 Jeffrey M. Puryear (ed.), Socios para el Progreso. La Educacion y el Sector Privado en America Latina y el Caribe. Dialogo Interamericano, Washington, D.C., 1996. 24 CHEILDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN VI. WHAT THE WORLD BANK IS DOING In the 1990 World Development Report on Poverty, the World Bank reaffirmed that the reduction of poverty is its overarching objective. This goal is to be achieved through a two- pronged strategy: stimulating sustainable economic growth and investing in people. The latter part of the strategy refers to the need to equip the population with the human capital necessary to compete successfully in the wide-open world economy. The World Bank's contribution to this strategy is apparent in its lending pattern: over the past decade, new lending for health and education has increased manifold, to reach $3.5 billion in 1995. Of this, one-third, or more than one billion dollars directly benefited young children, through feeding programs, basic health services, early child development projects, pre-school and primary education services. Table 6.1 shows actual social sector lending for the period 1991-95 and lending projections for the period 1996-98, for the LAC Region. Table 6.1: Actual and projected lending for health and education, for LAC, in $ millions 1991-95 1996-98 (actual) (projected) Health 1,573.60 1,864.40 Education 1,574.66 1,712.20 Total 2,148.26 3,576.60 As the Table shows, it is anticipated that the Bank's lending volume for health and education projects in LAC will increase from about $400 million annually, to almost $1.2 billion. Moreover, during the 1991-95 period, 40 percent of the funding for education is aimed at directly benefiting children under 15 years of age. When properly targeted, these financial resources should go a long way in assisting the region to increase access for basic services to the children of the poor and provide them with a better chance to escape poverty and live a full and productive life. 25 CHILDREN OF THE POOR IN LATIN AMERICA AND THE CARrBBEAN VI. SUMMARY AND QUESTIONS Children are over-represented among the poor in Latin America and the Caribbean. Almost 60 percent of children are poor, and about one-third of all children are found in the poorest 20 percent of households. Poor children are increasingly located in urban areas, but rural areas have higher proportions of children in poverty. Despite the debt crisis and the budgetary restrictions that accompanied economic reforms, most social indicators have improved significantly over the past two decades. However, these indicators still lag far behind OECD countries, and several countries fall far short of the Regional averages. In short, Latin America and the Caribbean has made a lot of progress, but most countries still have a long way to go. Much of the variation in social indicators across countries is related to differences in per capita GDP, but several countries fail to fit the pattern. These outliners demonstrate that policy matters. The poor are different than the non-poor. Poor families with children have larger family size, higher dependency ratios, and significantly less education than non-poor families. However, there appear to be no systematic differences in age or sex of the household head. In general, poor children receive fewer services than the non-poor. Immunization rates, access to health care by sick children, and access to primary schooling do not differ strongly and systematically between the poor and non-poor. Strong, systematic differences between the poor and non-poor are found in rural areas regarding access to water and sanitation facilities, but access is relatively equal in urban areas. The largest differences between poor and non-poor children are in access to preschool and secondary school, and, as shown by other research, in the quality of primary school. The cost of equalizing coverage between poor and non-poor children could be very high for some countries, but it is the large infrastructure costs of water and sanitation that account for much of this. The cost of equalizing access to education and health services is under I percent of GDP, except for Nicaragua, which has unusually bad social indicators for the Region. 26 CHELDREN OF THE POOR IN LATIN AMERICA AND THE CARIBBEAN Some countries in the Region appear to consistently perform better than expected across all social indicators. The answer appears to be a long-term historical, broad-based commitment to providing basic services to the poor and the provision of comprehensive services that meet most the needs of poor children. Countries which lag behind the Region in their delivery of services to poor children can select from a wide array of policy and program options for improving their performance. Continued inaction is due to problems of commitment and political will and not ignorance as to how to do it. Questions. This paper has identified some large gaps in access to basic services by poor children. While it is easy to criticize poor social indicators, most of the Region's leaders today are commnitted to reducing poverty and to providing all children with basic health, education, water and sanitation, and nutrition services. However, they find budgets are constrained, and voter conmmitment to reducing poverty is not sufficiently strong to permit shiffing budget allocations. This context leads to several questions: 1. If the budget does not permit easy solutions, should the priority lie with increasing services to the rural poor, who are usually the worst off, or should it lie with increasing services to the urban poor, who are rapidly growing in number and who may be easier to reach? 2. The improved targeting of social services, of course, could lead to improved services to the poor and better social indicators without increasing budgets. However, the political support for social programs could decrease if targeting leads to many voters who no longer receive benefits. How can the political leadership of a country determine what is the optimal mix of targeted and non-targeted programs? 3. The international agencies are themselves facing increasing constraints on their capacity to lend or provide technical assistance. Given this situation, what should be their strategies for ensuring poor children receive the basic services? 27 CHILDREN OF THE POOR IN LATIN ANMRICA AND THE CARIBBEAN Should they concentrate their activities on the very poorest countries, where the need is greatest, or should they concentrate on those countries (often the richer ones) where the comnmitment and ability to implement policies and achieve results is greatest? 28 APPENDIX Household Characteristics for Households with Children less Than 15 Years in Selected Countries Very Poor Poor Non-Poor Total URBAN Chile Family Size 5.0 4.9 4.6 4.7 Dependency Ratio 0.9 0.8 0.6 0.7 Age of Head 40.7 41.0 42.7 41.9 Female Heads (%) 18.6 18.1 14.4 16.1 Education of Head (Years) 7.7 8.0 10.7 9.4 Education of Head's Wife (Years) 7.7 7.9 10.6 9.4 Ecuador Family Size 6.4 6.0 4.8 5.2 Dependency Ratio 1.3 1.2 0.8 1.0 Age of Head 41.8 41.1 41.0 41.0 Female Heads (%) 19.0 15.0 17.0 16.0 Education of Head (Years) 5.2 6.2 10.0 8.6 Education of Head's Wife (Years) 5.0 6.0 9.7 8.4 Jamaica Family Size 6.9 6.2 4.4 5.0 Dependency Ratio 1.2 1.2 0.9 1.0 Age of Head 47.0 46.3 42.2 43.6 Female Heads (%) 52.0 52.0 50.0 51.0 Education of Head (Years) 7.0 7.6 8.4 8.1 Education of Head's Wife (Years) 7.4 7.6 8.8 8.3 Nicaragua Family Size 8.3 7.2 5.5 6.0 Dependency Ratio 1.6 1.5 1.1 1.2 Age of Head 45.2 44.0 41.5 42.1 Female Heads (%) 36.0 36.0 33.0 34.0 Education of Head (Years) 2.2 3.1 6.2 5.5 Education of Head's Wife (Years) 1.9 3.1 6.6 5.7 Peru Family Size 6.7 6.4 5.3 5.9 Dependency Ratio 1.3 1.2 0.9 1.0 Age of Head 44.5 43.8 44.6 44.2 Female Heads (%) 13.0 13.0 16.0 15.0 Education of Head (Years) 7.2 7.9 9.9 9.5 Education of Head's Wife (Years) 6.3 7.1 9.1 8.7 RURAL Chile Family Size 5.1 5.0 4.7 4.9 Dependency Ratio 0.9 0.8 0.5 0.7 Age of Head 42.3 42.9 45.4 43.6 Female Heads (%) 12.6 12.0 10.4 11.6 Education of Head (Years) 5.2 5.2 7.2 5.7 Education of Head's Wife (Years) 5.6 5.7 7.4 6.2 Ecuador Family Size 6.9 6.4 4.8 6.0 Dependency Ratio 1.4 1.3 1.0 1.2 Age of Head 43.8 42.6 39.9 41.8 Female Heads (%) 10.0 9.0 10.0 10.0 Education of Head (Years) 3.5 3.9 6.7 4.7 Education of Head's Wife (Years) 3.3 3.7 6.3 4.4 Jamaica Family Size 6.3 6.0 4.8 5.6 Dependency Ratio 1.3 1.3 1.1 1.2 Age of Head 52.5 51.9 45.4 49.8 Female Heads (%) 46.0 47.0 44.0 46.0 Education of Head (Years) 6.7 6.7 7.3 6.9 Education of Head's Wife (Years) 7.2 7.2 7.8 7.4 Nicaragua Family Size 7.3 6.8 5.6 6.4 Dependency Ratio 1.7 1.6 1.1 1.4 Age of Head 43.3 42.4 41.8 42.2 Female Heads (%) 19.0 18.0 20.0 19.0 Education of Head (Years) 1.0 1.2 3.2 1.8 Education of Head's Wife (Years) 1.1 1.4 3.4 2.0 Peru Family Size 5.9 6.1 5.6 5.7 Dependency Ratio 1.3 1.2 0.8 1.0 Age of Head 41.9 43.0 46.0 44.1 Female Heads (%) 10.0 12.0 15.0 15.0 Education of Head (Years) 5.9 6.3 8.5 7.8 Education of Head's Wife (Years) 4.7 5.1 7.8 7.0 TOTAL Chile Family Size 5.0 4.9 4.6 4.8 Dependency Ratio 0.9 0.8 0.6 0.7 Age of Head 41.2 41.5 43.0 42.3 Female Heads (%) 16.8 16.4 14.0 15.3 Education of Head (Years) 6.9 7.2 10.3 8.7 Education of Head's Wife (Years) 7.0 7.3 10.2 8.8 Ecuador Family Size 6.7 6.2 4.8 5.5 Dependency Ratio 1.4 1.2 0.9 1.0 Age of Head 43.0 41.8 40.8 41.3 Female Heads (%) 13.0 12.0 15.0 14.0 Education of Head (Years) 4.2 5.0 9.4 7.3 Education of Head's Wife (Years) 4.0 4.8 9.1 7.0 Jamaica Family Size 6.4 6.0 4.6 5.3 Dependency Ratio 1.3 1.3 1.0 1.1 Age of Head 51.3 50.0 43.3 46.8 Female Heads (%) 47.0 49.0 48.0 48.0 Education of Head (Years) 6.8 7.0 8.0 7.4 Education of Head's Wife (Years) 7.3 7.3 8.4 7.8 Nicaragua Family Size 7.5 6.9 5.5 6.2 Dependency Ratio 1.7 1.5 1.1 1.3 Age of Head 43.7 43.0 41.6 42.0 Female Heads (%) 22.0 24.0 30.0 27.0 Education of Head (Years) 1.2 1.8 5.5 3.8 Education of Head's Wife (Years) 1.2 1.9 5.7 3.9 Peru Family Size 6.1 6.1 5.4 5.7 Dependency Ratio 1.3 1.2 0.9 1.0 Age of Head 42.7 43.3 45.4 44.1 Female Heads (%) 11.0 13.0 15.0 15.0 Education of Head (Years) 6.3 6.9 9.1 8.5 Education of Head's Wife (Years) 5.2 5.9 8.4 7.7 Source: CASEN-Chile (1992); LSMS: Ecuador (1994), Jamaica (1991), Nicaragua (1993), Peru (1991) Other Reports in the LASHC Series No. 1: Children of the Poor in Latin America and the Caribbean by J. van der Gaag and D. Winkler, July 1996 (available in Spanish translation) No. 2: Targeting At-Risk Youth: Rationales, Approaches to Service Delivery and Monitoring and Evaluation Issues by M Schneidman, July 1996 (available in Spanish translation) No. 3: Evaluating Programs for Vulnerable Children and Youth by A. Harrell, July 1996 (available in Spanish translation) No. 4: Chile's Learning Network by M Potashnik. July 1996 No. 5: Review and Analysis of International Experience with Programs Targeted on At-Risk Youth by G. Barker and M Fontes, July 1996 (available in Spanish translation) No. 6: Measuring Public Hospital Costs: Empirical Evidence From The Dominican Republic by M A. Lewis, May 1995 I