Report No. 20475-BR Brazil Attacking Brazil's Poverty A Poverty Report with a Focus on Urban Poverty Reduction Policies (In Two Volumes) Volume 1: Summary Report March 31, 2001 Brazil Country Management Unit PREM Sector Management Unit Latin America and the Caribbean Region Document of the World Bank CURRENCY EQUIVALENTS Currency Unit - Real (R$) December 1996: R$1.04/US$ December 1998: R$1.21/US$ December 2000: R$1.95/US$ WEIGHTS AND MEASURES The Metric System is used throughout the report. FISCAL YEAR January 1 to December 31 ABBREVIATIONS BPC Disability and Elderly Pension (Beneficio de Presta do Continuada) FAT Worker Support Fund (Fundo de Amparo ao Trabalhador) FGTS Severance Fund (Fundo de Garantia por Tempo de Servi o) GDP Gross Domestic Product IBGE Brazilian Institute of Statistics and Geography (Instituto Brasileiro de Geografia e Estatistica) IPEA Institute for Applied Economic Research (Instituto de Pesquisa Economica Aplicada) LAC Latin America and the Caribbean Region LSMS Living Standard Measurement Survey NGO Non-Governmental Organization PAT Workers Food Program (Programa de Alimentaqdo do Trabalhador) PIS/PASEP Labor Funds for Private and Government Sector Workers (Programas de Integraq o Social e de Forma4do do Patrim6nio do Servidor Puiblico) PLANFOR Labor Ministry Training Programs (Programa Nacional de Qualificacdo do Trabalhador) PME Monthly Employment Survey (Pesquisa Mensal de Emprego) PNAD National Household Survey (Pesquisa Nacionalpor Amostra de Doniicilios) PPV Living Standard Measurement Survey (Pesquisa sobre Padroes da Vida) PREM Poverty Reduction and Economic Management Network SUS Unified Health System (Sistema Unico de Sazude) VT Employer Provided Transportation Pass (Vale Transporte) Regional Vicepresident: David de Ferranti Country Director: Gobind T. Nankani Sector Director: Ernesto May Sector Manager: Norman Hicks Task Manager: Joachim von Amsberg Attacking Brazil's Poverty A Poverty Report with a Focus on Urban Poverty Reduction Policies Summary Report FOREWORD 1. INTRODUCTION ..............................................................1 2. POVERTY TRENDS AND PROFILES ..............................................................2 3. POVERTY REDUCTION SCENARIOS ..............................................................7 4. POVERTY REDUCTION POLICIES ..............................................................9 OPPORTUNITY THROUGH GROWTH .............................................................. 10 CAPACITY BY STRENGTHENING THE ASSETS OF THE POOR ............................................................. 13 SECURITY THROUGH SOCIAL PROTECTION .................. ........................................... 17 EMPOWERMENT THROUGH EFFECTIVE PARTICIPATION ............................................................. 19 ANALYSIS, EVALUATION, AND COORDINATION OF SOCIAL SPENDING AND POLICIES ..................... 20 POLICY REFORM SUMMARY ............................................................. 25 5. LIST OF BACKGROUND PAPERS ............................................................. 28 TABLES TABLE 1: DISTRIBUTION OF POOR BY REGION AND AREA 3 TABLE 2: POVERTY REDUCTION SCENARIOS 2015 7 TABLE 3: SUMMARY OF CONSOLIDATED SOCIAL SPENDING, 1995 WITH UPDATES 21 TABLE 4: POLICY REFORM SUMMARY 26 FIGURES FIGURE 1: IMPROVEMENTS IN INFANT MORTALITY 2 FIGURE 2: IMPROVEMENTS IN PRIMARY SCHOOL ENROLLMENT 2 FIGURE 3: POVERTY RATES 1981-1999 3 FIGURE 4: POVERTY AND LEVEL OF EDUCATION 4 FIGURE 5: EXTREME POVERTY RATE BY REGION AND TYPE OF AREA 4 FIGURE 6: POVERTY AND EMPLOYMENT STATUS 5 FIGURE 7: POVERTY AND DEPENDENCY RATIO 6 FIGURE 8: THEORETICAL AMOUNT NEEDED TO ERADICATE INCOME POVERTY 6 FIGURE 9: TOWARD THE 2015 POVERTY REDUCTION GOAL - ALTERNATIVE SCENARIOS 8 FIGURE 10: GROWTH AND POVERTY REDUCTION OVER TIME 11 FIGURE I 1: IMPROVEMENTS IN SCHOOL ENROLLMENT 1980- 1999 1 3 FIGURE 12: INCIDENCE BY EDUCATION LEVEL (1996/7) 14 FIGURE 13: INCIDENCE OF HEALTH FACILITY USAGE (1996/7) 15 FIGURE 14: STRUCTURE AND TARGETING OF FEDERAL SOCIAL SPENDING, 1997 22 FIGURE 15: TARGETING AND REACH OF SOCIAL PROGRAMS 24 FOREWORD This report was produced under the supervision of Gobind T. Nankani, Director, Suman Bery, Lead Economist, and Norman Hicks, Sector Manager Poverty, by a World Bank team led by Joachim von Amsberg, and comprised of Francisco Ferreira, Peter Lanjouw, Kimberly Nead, Marcelo Neri, Judy Baker, Ricardo Paes de Barros, and Zhi Liu. Additional background papers were prepared by: Sonia Rocha, Martin Ravallion, Maria da Piedade Morais, Claudio Hamilton, Diana Meirelles da Motta, Silvio Caccia Bava, Paulo Sergio Pinheiro, Rubem Cesar Fernandez, Ana Clara Torres Ribeiro, Grazia di Grazia, and Rosane Mendon,a. Julia Conter, Leo Feler, and Carolina Sanchez assisted with the production of the report. The peer reviewers for this task were Jeffrey Hammer, Jesko Hentschel and Josef Leitman. This report was produced in close collaboration with vai ious institutions and individuals in Bra- zil. In particular, the extensive collaboration with the Instituto de Pesquisa Econ6mica Aplicada (IPEA) is gratefully acknowledged. The working papers prepared for this report were presented and discussed at various seminars. Many useful comments were received and are appreciated. The team wishes to thank all individuals in the Government, the World Bank, and elsewhere, who have contributed with their time, thoughts, and comments to this report. While the report has been discussed with many institutions and individuals of the Brazilian Government, the views expressed in the report are exclusively those of the World Bank and not necessarily those of the Brazilian Government. Most of this report is based on work undertaken in 1998, based on data from the PPV 1996/7 and the PNAD 1996. It reflects information available at that time. Only selective updates have been made in a few cases to reflect more recent data or policy developments. The Summary Report (Volume I) and the Full Report (Volume II) are based on several back- ground papers prepared for this task, which are listed at the end of the report. In the text, no spe- cific references are made to these documents. The Summary Report contains the most important findings and policy recommendations without providing the analytical support that is contained in the Full Report and the background papers. The background papers have been prepared under the responsibility of their respective authors and are not official World Bank documents. They are available from the following website: http:llwww.worldbank.ory/br. Attacking Brazil's Poverty Page 1 Attacking Brazil's Poverty A Poverty Report with a Focus on Urban Poverty Reduction Policies well as voicelessness and powerlessness. 1. Introduction For the quantitative analysis, however, this report uses a much narrower definition of 1. Thefirst central message of this report poverty as insufficient income or insuffi- is that Brazil has over the last years cient consumption. While the relationship achieved great progress in its socialpolicies between income levels and other household and indicators, especially in health and edu- characteristics is analyzed, a more compre- cation. Brazil has achieved considerable hensive approach for quantitative analysis of progress in the reduction of income poverty. other forms of deprivation or social exclu- Overall, recent and ongoing policy reforss sion is not attempted. Also, issues pertain- are on the right track, and the perspectives ing more broadly to the inequality of in- for Brazil's poor for a better life in the fu- come, assets, and opportunity are not treated ture have considerably improved. in this report.' 2. The second central message is that 5 Recognizing the rich literature on pov- poverty remains unacceptable high for a erty in Brazil, this report attempts a few country with Brazil's average income lev- specific contributions. First, it presents a els. The worst remaining income poverty is new 1996 poverty profile with a breakdown mostly concentrated in the Northeast re- by city size, incorporation of the imputed gion, and in the smaller urban and rural value of owned housing, and regional price areas. At the same time, the poor in large deflators. Second, it provides analysis of the urban and metropolitan areas suffer from incidence of selected public social spending additional deprivation related to crowding, based on the 1996/7 Pesquisa sobre Padroes poor health, violence and crime and require da Vida (PPV). Third, it selectively synthe- continued attention. sizes other work on poverty in Brazil in or- der to present an overview of cross-sectoral 3. The third central message is that, with comparisons of the effectiveness of policy decisive action, Brazil can achieve ambi- interventions. Fourth, it provides sugges- tious targets for further improvements in tions and selective inputs for the develop- social indicators, including the objective of ment of a national poverty reduction strat- reducing the rate of extreme income pov- egy. The report updates a previous World erty by 50% by the year 2015. This report Bank poverty assessment and builds on ear- proposes a strategy for reaching this goal lier work toward an urban poverty strategy that includes renewed broad-based and in- for Brazil.2 The report is directed at policy clusive growth as well as further reformns of makers at all levels (federal, state, and mu- social spending and social policies. This nicipal) as well as policy analysts. strategy requires continued actions and commitment at all levels of Government. The recent Voices of the Poor study for Brazil contains a fuller account of the non-income dimensions of poverty in 4. Poverty is a complex and multidimen- Brazil. Some of the findings are summarized in Volume 11 of sional phenomenon. A broad definition of this report; see World Bank. Consultations with the Poor: Brazil - National Synthesis Report. Poverty Reduction and poverty as unacceptable deprivation in hu- Economic Management Network. Washington, D.C.: The man well-being underlies this report. This World Bank, 1999. See also definition of poverty includes insufficient http://www.worldbank.org/poverty/voices/. income and consumption, unsatisfied basic 2 World Bank. A Poverty Assessment. Report no. 14323- needs such as basic education, health, nutri- BR. Washington, D.C.: The World Bank, 1995a; World tion, and housing, insecurity and risk, as Bank. Brazil: Toward an Urban Poverty Strategy. Washing- ton, D.C.: The World Bank, 1995b. (unpublished). Page 2 Attacking Brazil's Poverty and has to be considered in Figure 1: Improvements in Infant Mortality the design of poverty reduc- Each line in this figure represents one country showing average income tion policies. and the rate of infant mortality at three points in time (1980, 1990, 1998). The international regression line shows the level of infant mor- tality expected at the respective average income level. Brazil has not 2. Poverty Trends only greatly improved this indicator but also approached the interna- tional regression line. and Profiles 21S1 \ 7. Over time, Brazil has .~ |made great advances in im- proving its social indicators. sMo- l \ Twenty years ago, many of Brazil's health and education _.20 j, iA A indicators were far worse _l20O 1 -AS > - . than one would expect at 0 I Brazil's average income 304 i - - . -: -BRAZIL level. These indicators have I &S>> not only improved continu- 40 "-. .3a _ -ously and substantially, they 21 0.=_ have also approached or even - surpassed levels that would O 1 GO 2033 3.0%) AD 5(150 6000 73D0 83fiJ 30o0 103, Souw:W."d 5-k Wodd Dnt.PbI- ' GDP, PPP p.,-p[t ' ' ' ' be expected at Brazil's aver- age income levels today. 6. The present report excludes discussion Figure I and Figure 2 show this progress and analysis of policies specifically directed very clearly, using the examples of the in- at the reduction of rural poverty, such as fant mortality rate and the rate of non- rural development and land refornL Given enrollment in primary schools, respectively. their importance, specific rural poverty is- sues are discussed in a sepa- rate World Bank report cur- Figure 2: Improvements in Primary School Enrollment rently under preparation. Each line in this figure represents one country showing average income The present report excludes and the rate of non-enrollment in primary school at three points in time specific rural poverty policy (1980, 1990, 1998). The international regression line shows the level of issues simply to maintain the non-enrollment expected at respective average income level. Brazil has scope of the document man- not only greatly improved this indicator but also surpassed the interna- ageable. No judgement tional regression line. whatsoever on the relative 120 importance of urban versus 1 rural poverty policies is im- i plied by the exclusion of ru- l ral policy issues from this ' report. Moreover, it is not eol suggested that urban areas i . / are homogenous as a group X . vis-a-vis rural areas. In fact, , analysis in this report shows ,;40. 2. '' that small urban areas are B quite different from large UR,,20 cities and have similarities X with rural areas. Migration 0 . ;. links rural and urban poverty 0 .000 2.000 3.000 4,000 5,000 e,000 7,000 8.000 9,000 10.000 5*:0- ---n - WN.I 8 n-d GOP, PPP p., p-P0 Attacking Brazil's Poverty Page 3 Figure 3: Poverty Rates 1981-1999 core of indigent individuals. The rate of extreme poverty is highest in the North and Northeast and 9. To analyze the nature of in rural areas. The rate of extreme poverty is also very high in small the remaining poverty prob- and medium sized urban areas. However, PNAD data is likely to lem, the following poverty overestimate rural poverty. prof le defines a poor per- BO - . . . . _ ... _._._ son as one living in a 70- i ~ . A household with per capita 70.a.,.",... income less than the equiva- 60 . ,. ' - lent of R$65 per month at Sdo Paulo Metropolitan --- Area prices. Income is ad- justed for regional price dif- - \ /Y , A*~ ferences and for the imputed rent of owned or unpaid 20 . $6 , housing. Income is measured by the national household 10 survey undertaken in 1996 ______________________________________________ l (PNAD 1996). Comparative YV%1 1~1M1B 1%1897 MI 90192I . MI analysis is also based on in- SlPEA b.d m PNAD1BGE -A a R-1. _ Ub .,,,,,_ come and consumption data from the PNAD 1993, PNAD 8. Fluctuating with macroeconomic 1998, and the 1997 LSMS events, income poverty has also been de- household survey (PPV). The poverty line clining overall; however, a core of indigent of R$65 is determined by the cost of a basic individuals remain stuck in poverty. Ab- food basket (extreme poverty line). This stracting from the impact of an unsustain- report focuses on the extreme poverty line of able stabilization effort around 1986, Figure R$65 in order to produce a sharp profile of 3 shows a declining trend of income poverty the poor (the higher the poverty line used, since 1983, with a marked and lasting reduc- the more closely the profile of the poor re- tion in poverty levels achieved with stabili- sembles the profile of the overall popula- zation in 1994. Overall, economic stabiliza- tion). tion and growth have contributed signifi- cantly to the reduction of poverty in Brazil 10. At the chosen extreme poverty line, and will continue to do so. However, a sig- Brazil has a poverty headcount ratio (pov- nificant number of extremely poor with low erty rate) of 22.6%. This means 34.9 mil- education levels remain at the bottom of the lion Brazilians live in households with a per- income distribution, largely untouched by capita income below the poverty line. A economic development. These poor- little more than half of these poor (52.5%) maybe the bottom 10-15% of Brazil's in- live in urban areas. Of the urban poor, 39% come distribution-are often unable to bene- live in urban areas with a population less fit from economic opportunities, pointing to than 20,000; 23% live in urban areas with a the need for a basic social safety net for this population between 20,000 and 100,000; Table 1: Distribution of Poor by Region and Area Share of total poverty Northeast Centerwest North Southeast South Total Metropolitan Core 3.6% 0.2% 0.4% 1.3% 0.3% 5.8% Metropolitan Periphery . 4 ,,, 2.4% 0.5% 5.4% Large Urban 4.8% 0.7% 1.1/ 1.3% 0.6% 8.5% Medium Urban Y a7%.x -' 1.9% 1.3% 12.2% Small Urban . I4 , 1.Z 3.1% 1.2% 20.5% Rural 2 Z 1 . ,< 71%. . C .4,3% 47.5% Total 62.7% 5.0% 6.4/ 17.7% 8.1% 100.0% "Poor Areas" (defined as per capita earnings less than R$ 160 per month) are shaded in gray. Page 4 Attacking Brazil's Poverty Figure 5: Extreme Poverty Rate by Region and Type of Area The rate of extreme poverty is highest in the North and Northeast and in rural areas. The rate of extreme poverty is also very high in small and medium sized urban areas. However, PNAD data is likely to overestimate rural poverty. 70% 40% 00% 2 0% 10% L l Z fiNo t Northeast 0% ~~~~~~~~~~~~~~~~Centerwest Smlluralium Soteast Urban Urban Large Metropolit Meropolit Source: IBGE/PNAD 1996 Periphery an Core 16% live in urban areas with more than 100,000 inhabitants; and finally 21% live in the metropolitan areas. In terms of geographical distri- Figure 4: Poverty and Level of Education bution, 63% of the poor live The poverty rate decreases with increasing number of years of educa- in the Northeast, 18% in the tion. The highest share of the poor (45%) have had less than one year Southeast, 8% in the South, of education. There are virtually no individuals who are poor and have 6% in the North and 5% in more than 12 years of education. the Centerwest. Figure 5 shows the extreme poverty 45% . rate in different regions and 40% \ areas. Table 1 shows the distribution of Brazil's poor \5% between regions and areas. 11. Overall, there is a strong concentration of the 20%o _ * poor in the Northeast and 1,% _ _ -in the rural and the small _0% _ I\ and medium sized urban areas, both in terms of ab- solute numbers and in .0P._lY,0 _ ,t4Y~or. _%OY00ro l tol2Y00r. M0._ _ Y. terms of poverty rates. Both Ithon1 Yoor 1 S. 4 Y.. 4 b 8 Y- 8 t. 12 Y-r. Min a slt 12 Y( e .fsn Pq."- 3Sh-f P- =W l in absolute terms (number of Attacking Brazil's Poverty Page 5 safe water, infant mortality Figure 6: Poverty and Employment Status s es,sinfant withlity rises significantly with city The unpaid, unemployed, informally employed, and self-employed size, most likely related to experience the highest incidence of poverty. The largest shares of the crowding and pollution at the poor are either self-employed or infornally employed. local level. Also, crime and .. ..... . .. - violence are more severe in 45% .K the larger cities. In larger cities, crime and violence are 40% sometimes identified as the most significant barrier ob- structing the escape from 25% A"lll [ poverty. These observations 20% * / - fl i | suggest that an increased fo- cus on income poverty in smaller and medium-sized cities should not divert atten- 5% - L L - - * J 1E11 1 tion away from the persisting - ON _R] _ , _ _poverty problems of the lar- E1*plo P O c S-b F-W InKb SW-EV.y.d Irl- U-sed Un,d.. Er ESbY- ger cities, where a focus m ______________ olp- >apparticular on the non-income dimensions of poverty would poor) as well as in relative terms (poverty be justified. rate), poverty is the worst in rural areas; within urban areas, poverty is more severe in 13. Income poverty is closely associated small and medium sized urban areas; and with other forms of deprivation. The poor poverty is least severe in the center of the have less access to public services, such as metropolitan areas. From 1993 to 1998, water, sanitation, and garbage collection. poverty declined less in the Northeast, in the Even though country-wide electrification rural areas, and in the small and medium rates are high, a quarter of the poor still live sized urban areas. Consequently, poverty without access to electricity. 52% of the became more concentrated in these areas. poor have no canalized water, 68% have no This profile presents a stark contrast to the garbage collection, and 78% have neither dominant perception of poverty in the sewage connection nor septic tanks. Health favelas of the mega-cities of Sao Paulo and indicators, such as infant mortality, are sig- Rio de Janeiro, where poverty is most visi- nificantly worse for the poor. Finally, crime ble. In fact, the poor of the metropolitan and violence in urban areas typically affect areas (core and periphery) of Sao Paulo, Rio the poorest most severely. Interestingly, de Janeiro, and Belo Horizonte together residence in a favela is not significantly re- constitute only 3.7% of Brazil's poor or 7% lated to income poverty. of Brazil's urban poor. This finding sug- gests a stronger policy focus on the poor in 14. Most poor do not work in the formal small and medium sized cities. labor market. A large share of the poor work in the informal sector (22% are infor- 12. While not as severe in income terms, mal employees and 37% are self-employed) large-city poverty may be significantly or are inactive (15%). Only 15% of the poor worse due to other forms of deprivation are working in the formal sector (private or linked to crowding. There are indications public); only 5% are unemployed (Figure 6). that welfare outcomes (determined, for ex- This means that social policies tied to formal ample by health status indicators) in the big- employment or unemployment have only ger cities might be worse than income very limited reach among the poor. measures of poverty indicate. Controlling for poverty rates, education, and access to Page 6 Attacking Brazil's Poverty tion or discrimination, Figure 7: Poverty and Dependency Ratio explains most labor in- The poverty rate increases with increasing number of dependents. The come inequality in largest share of the poor has more than four dependents. The smallest Brazil. Low education share has two, one, or zero dependents. levels lead to low in- comes, which in turn, ! lead to low school at- tendance of children, perpetuating poverty. (c) Household size is 40% / strongly correlated with poverty (see Figure 7). Poverty 20% _^1 rates increase with the I - _ - dependency ratio 10% i _ /(share of income earn- ers over non-income 0% _ _ _ _earners in the house- d-< 2-&a 3-de4 d>4 hold). As observed all over the world, in- 15. The three factors most closely associ- creased income and improved health ated with poverty as likely underlying indicators would typically reduce de- causes are location in a poor area, low pendency ratios and thus further in- schooling, and large household size: crease perspectives for poverty reduc- tion. (a) Location (region and area) is the strongest explanatory variable (in a 16. Brazil's very higk income inequality statistical sense) for poverty. Even implies that the income gap of Brazil's households with the same characteris- poor is relatively small compared to macro- tics are more likely to be poor in poorer than in richer areas. A pov- Figure 8: Theoretical Amount Needed to Eradicate Income Poverty erty rate of nearly 50% The theoretical amount needed to eradicate income poverty is the sum in the Northeast sug- of the difference between the income of every Brazilian below a prede- gests a more explicit termined poverty line and that poverty line. Using the poverty line of approach toward re- R$65 per capita per month, this amount is about R$ 12 billion annually. gional equality and re- 200 _ gional targeting in so- ISO cial policies and fund- ing allocations. (b) Poverty is closely asso- 120/ ciated with low levels 2 100 of education (see / Figure 4): 73% of poor 0a households are headed 60/ by individuals with four 5 40 or less years of school- 20 ing. Educational ine- 0 quality, rather than Ia- 0 so 100 150 200 250 300 bor market segmenta- Poverty Une (RS per capita per month) Attacking Brazil's Poverty Page 7 economic aggregates. In other words, theo- poverty reduction scenarios are simulated. retically and in aggregate, Brazil has the The first scenario is based on the immediate resources necessary for solving its poverty eradication of poverty. The next two scenar- problems. The income gap of Brazil's poor ios are based on a more realistic reduction of was R$12 billion in 1996. Abstracting from poverty over time. Finally, the last scenario the problems of targeting, administration, assesses the impact of scenario of slow and incentives, this is the amount theoreti- growth and no progress on social policies cally needed to bring every poor Brazilian (see Figure 9 and Table 2): up to the extreme poverty line for one year. This income gap constitutes 1.6% of the (a) Given prevailing fiscal constraints, the 1996 GDP of R$750 billion or less than 5% Immediate Poverty Eradication Sce- of the income of the wealthiest 10% of nario is clearly not viable nor sustain- Brazilians. This amount is small in able. This strategy would require an- comparison with total social spending in nual transfers of R$27 billion to reach Brazil. For alternative poverty lines, Figure the poor. This is much more than the 8 shows, on the vertical axis, the annual aggregate income gap of R$12 billion amount of perfectly targeted transfers since uniform transfers to all poor are theoretically necessary to bring every assumed rather than unrealistic trans- Brazilian up to the poverty line displayed on fers equaling the individual income the horizontal axis. gap. Assuming that 40% of public spending actually reaches the poor, 3. Poverty Reduction this would imply fiscal costs of R$67.5 Scenaris .billion. Over time, growth and other Scenlarios elements of structural poverty reduc- tion would replace part of the transfer 17. The Government intends to reduce the payments. rate of extreme poverty by 50% by 2015, which provides a useful benchmark for (b) The Growth Scenario is based on an assessing the relative importance of differ- optimistic 6% annual growth rate per ent policies. Order-of-magnitude simula- year until 2015, complemented by tions show how this objective could be ob- educational improvements leading to tained by a combination of educational im- one additional year in the average provements, increased growth, and other education of the workforce, and measures (other structural social policies, additional transfers of R$1.5 billion such as land reform, rural and urban ser- reaching the poor (at a budgetary cost vices, and transfer programs). Based on dif- of R$3.8 billion annually). This ferent mixtures of growth, educational im- scenario would lead to a halving of the provements, and transfer programs, four rate of extreme poverty to 11% by Table 2: Poverty Reduction Scenarios 2015 Immediate Eradication 2015 / 50% Goal 2015 / 50% Goal through No Progress of Poverty through Growth Social Spending GDP Growth per Year N/A 6.0% 2.0% 1.0% Additional Average N/A 1 2 0 Years of Education Additional Annual R$27 billion R$1.5 billion R$5.4 billion R$0 Transfers Reaching the Poor Annual Budget Cost of R$67.5 billion RS3.8 billion R$13.5 billion R$O Transfers (40% Effectiveness) Poverty Rate Now 0% 22% 22% 22% PovertyRate2015 0% 11% 11% 23% Page 8 Attacking Brazil's Poverty Figure 9: Toward the 2015 Poverty Reduction Goal - Alternative Scenarios This figure shows four scenarios for the 2015 poverty reduction goal. The hypothetical immediate eradication scenario would bring the extreme poverty rate to zero through transfers and investments at an estimated cost of R$67 billion annually. The growth scenario would rely on 6% annual growth, one additional year of education, and transfers at an estimated cost of R$4 billion annually in order to cut the poverty rate to 11%. The social policy scenario would rely on 2% annual growth, two additional years of education, and tranfers at an estimated cost of R$14 billion annually in order to cut the poverty rate to 11% as well. The no-progress scenario with 1% annual growth would result in a poverty rate of 23% by 2015. 25% - 20%- 2 15% - 0 E 10%- 0 5% - 0% _ Eradication Scenario Growth Scenario Social Policy No-Progress Scenario Scenario * Remaining Poverty Rate in 2015 D Reduction Through Growth E Reduction Through Education 0 Reduction Through Transfers/investments Source: WB staff etsimates poverty to 11% by 2015 and demon- creased) social spending would be re- strates the large impact growth can quired to reach this objective with have on poverty reduction. slow growth. (c) The Social Policy Scenario is based on (d) In an undesirable No-Progress Sce- slow growth of 2% per year until 2015 nario, with growth at only 1% per year complemented by educational im- and social spending maintained at the provements leading to two additional current per capita level, poverty rates years in the average education for the would worsen slightly (to 23%) imply- workforce, and additional transfers of ing a significant increase in the poor R$5.4 billion reaching the poor (at a population (from 35 to 43 million). budgetary cost of R$13.5 billion annu- ally). This would imply a near dou- 18. A 50% reduction in the extreme poverty bling of transfers reaching the poor to- rate by 2015 is attainable with decisive day. This scenario would also reduce Government action. In a scenario of high poverty to 11% by 2015 but shows that 6% annual growth, continuing educational a major effort on improved (and in- improvements, and modest increases in tar- Attacking Brazil's Poverty Page 9 geted transfers, the rate of extreme poverty humility when it comes to proposing sim- rate would fall from 22% to 11% by 2015. plistic solutions to Brazil's poverty prob- Alternatively, the same target could be lems. At the same time, this report attempts achieved with low growth of 2%, large con- to show that there are many steps that fed- tinuing improvements in educational eral, state, and municipal governments can achievements, and a doubling of transfer take to further improve public policies and payments reaching the poor. increase their impact on poverty reduction without greatly expanding social spending 19. Taking together recent structural re- overall. forms to underpin improved macroeco- nomic performance, the social policy re- 21. The obstacles to further poverty reduc- forms undertaken in recent years and on- tion are multiple. Public policies to reduce going and the increased public and politi- poverty, thus, need to be composed of mul- cal attention to poverty problems, the pros- tiple strategies. Several obstacles to a more pects for Brazil's poor have clearly im- rapid and significant reduction of poverty proved Structural reforms of the economy remain. These obstacles include: (i) limited have laid the groundwork for economic sta- opportunities as the result of low growth in bility. Such economic stability will protect the past and the relatively small poverty- the poor from income fluctuations that, in reducing effect of growth in a situation of the past, were often related to failed macro- high income inequality; (ii) low educational economic stabilization attempts. The impact attainments; (iii) poor health; (iv) lack of of renewed growth on poverty reduction will physical assets and services for the poor; (v) be gradual but important. The major drive gaps in the social safety net and resulting for improved education will yield significant insecurity and vulnerability; (vi) problems poverty reduction results in the medium related to the quality and targeting of social term. The impact of improvements in other spending; and (vii) limits to the effective structural social policy areas, including voice and participation of the poor in the health, land reform, urban upgrading, pro- decision making processes of relevance to fessional training, and labor market is diffi- them. There is no single or simple solution cult to quantify but qualitatively important to poverty in Brazil, and the different obsta- and expected to continue. The poverty re- cles to poverty reduction must be addressed ducing impact of transfer programs has been simultaneously through policy reforms in large in recent years and is also expected to different areas. continue, especially if measures for better targeting are successfully implemented. 22. This report shows that Brazil can achieve the objective of reducing the rate of extreme income poverty by 50% by the year 4. Poverty Reduction 2015 and proposes a strategy for reaching this goal. Similar to the strategy proposed Policies in the World Bank 2000/1 World Develop- ment Report for poverty reduction at the 20. A large number of Government policies international level, the proposed strategy and programs at all levels are in place to includes: fight poverty and major efforts have re- cently been made to increase their effec- (a) Opportunity: Renewed and accelerated tiveness. Poverty in Brazil would be much economic growth. Economic growth worse without many of the ongoing policies based on economic and fiscal stability and programs. The wealth of experiences would create economic opportunities involving poverty reduction policies in all for the poor. sectors and at all levels provide great learn- ing opportunities. The existing experience (b) Capacity: Strengthened assets of the and the many ongoing efforts also suggest poor. The assets of the poor would be Page 10 Attacking Brazil's Poverty strengthened through better education, 24. Improved analysis, evaluation, and co- health, physical and urban services, ordination of overall social policy and and physical capital. Greater assets spending is essential to direct the decisive would give the poor the capacity to effort necessary to reach the ambitious make better use of economic opportu- poverty reduction targets of the Govern- nities. ment While this report provides indicative analysis of the allocation and targeting of (c) Security: Increased protection for the social spending, further and continuous ef- poor. Security for the poor would be forts for evaluation and analysis by the Gov- increased through effective social in- emient itself are necessary and should feed surance mechanisms and a basic social into the ongoing decision making processes. safety net. More effective social in- surance mechanisms and a basic social Opportunity through Growth safety net would help the poor manage the risks of adverse shocks either at the 25. Both international and Brazil-specific economy-wide or individual level. studies confirm the strong relationship be- tween growth and poverty reduction. Even (d) Empowerment: More participation of though the effect of growth on poverty re- the poor. Greater empowerment in- duction in Brazil has been less than the in- volves participation and access to so- ternational average (due to high income ine- cial institutions and decision making quality and Brazil's particular pattems of processes. Moreover, greater empow- growth), Brazil's problem has rather been erment requires regulatory reformns to the absence of sustained high per-capita in- overcome the de-facto exclusion of the come growth. Accelerated and sustained poor from certain formal markets. growth that generates employment opportu- unities for the poor will have to be a key con- tributor to the sustained reduction of poverty recommended in this report are guided by in the future. five principles: 26. While growth may not be arithmetically (a) Market dlstortions need to be removed o necessary to end extreme poverty in Brasil, and regulations need to be reformed to growth is still critical for economic oppor- make mrtwke. tunities and sustainable poverty reduction perspectives. Given the small size of the (b) Social spending should be better tar- aggregate income gap of the poor, Brazil geted at the poor using different ap- may not seem to require growth to end pov- proaches that are appropriate for dif- erty. However, the cost of eliminating pov- erty through income transfers are much higher than theoretical calculations suggest (c) Soca pedg shoulde te to if the real life constraints of targeting and est benefits to the poor per amount of adverse incentives are taken into account. eublicstpenedits ttepreaonf Also, the political viability of large-scale income redistribution remains dubious. (d) Social programs should be oriented Growth will bring about a sustainable reduc- tion in poverty that does not depend on con- wrdith deximan choiesfor the prnd tinuing transfer payments. Also, growth permi mf. creates economic opportunities and pros- pects for the poor beyond surpassing the (e) The internal efficiency of programs poverty line. Moreover, growth generates shouldebe)increased. additional benefits for millions of Brazilians should be increased, who are above the poverty line but still ur- gently seek improvements in their living Attacking Brazil's Poverty Page 11 conditions. In contrast to income transfers, extreme poverty rate constant at 22%. In the growth makes the alleviation of poverty absence of any redistributional policies, much more politically palatable. growth of almost 8% per year would be re- quired to halve the poverty rate by 2015, and 27. Growth and directly targeted interven- growth of 2% and 6% would cut the poverty tions have to be seen as complementary rate in 2015 to 19% and 13% respectively. and not substituting policies. Growth will 2.5% annual growth is required to keep the take time to reduce poverty and will not absolute number of poor constant. Growth reach all the poor. However, growth can, of 2% and 6% would bring the absolute over time, reduce the need for transfer pay- number of poor from 35 million to 40 mil- ments, or allow them to be directed at rela- lion and 25 million, respectively, in 2015. tively less impoverished populations. Figure 10 shows the simulated evolution of the extreme poverty rate over time with dif- 28. Simulations show the potential for sig- ferent growth rates. nificant poverty reduction with high growth rates; they also show the increase in pov- 29. Growth policies should continue to in- erty that would occur with economic stag- clude efficient public investment and im- nation. The simulations are based on Bra- provements in the economic environment zil's actual income distribution. Distribu- for private investment. The impressive sta- tionally-neutral economic growth would bilization policy of the Government has laid have to equal the population growth rate of the foundation foT sustained growth, which 1.2% (constant per-capita GDP) to keep the until now remained elusive. The Govern- Figure 10: Growth and Poverty Reduction Over Time This figure shows the changes in the poverty rate based on growth of 1%, 2%, and 6% per year. Growth of 1% per year would not be sufficient to keep the poverty rate constant. Growth of 2% and 6% per year would cut the poverty rate to 19% and 13%, respectively, by the year 2015. 25% - 20% -- _ _ - _ - - - - - _ - _ - 2 15% - IL10% 5%- 0% l CD CD O CD CD rN CD C- OZ C.D G CD CJ) ON O C> O: C) 0O C4 04 O< ON C4 CD Year Source: WB staff estimates - Growth of 1% p.a. - Growth of 2% p.a. - - Growth of 6% p.a. Page 12 Attacking Brazil's Poverty ment should continue to promote sustainable legally highly regulated but de-facto not ex- and high-quality fiscal adjustment as a cessively rigid. Nevertheless, labor courts, means of ensuring macroeconomic stability unions, and an uncertain regulatory envi- and declining cost of capital. Macroeco- ronment impose costs on companies that are nomic stability and less expensive capital difficult to predict. This uncertainty regard- will encourage investment and sustainable ing employment costs causes capital substi- growth, ultimately leading to poverty reduc- tution in all firms. In addition, by virtue of tion. Given the linkages between fiscal ad- their size, larger companies are able to man- justment, growth, and poverty reduction, it age and predict these costs more success- is important to emphasize the ongoing fiscal fully than more employment-intensive, me- adjustment effort as an essential prerequisite dium-sized companies (or small firms, to the for more effective poverty reduction. Espe- extent that they are subject to effective regu- cially to the extent that high quality social lation). spending remains protected from fiscal re- trenchment, the possible negative short-term 32. Informality in the labor market is cho- effects of adjustment are likely to be out- sen when the costs offormality (the rigidity weighed by the long-term benefits of stabil- of mandated benefits, payroll taxes, and ity, growth, and subsequent poverty reduc- others) exceed the benefits of formality tion resulting from fiscal adjustment. (unemployment insurance, pensions, other social programs, and better career perspec- 30. Educational improvements, asset trans- tives), or when the employer is informal fers to the poor, financial market reforms, (typically related to high overall taxation or and labor market reforms are expected to rigid regulation). While informal employ- contribute to more broad-based growth and ees are clearly concentrated in low-quality possibly to the reduction of inequality. Re- jobs, there is no indication that informality cent Government initiatives to support and itself causes low-quality employment. More strengthen small and medium enterprises are likely, informal jobs are of lower quality also important for supporting employment- because the current structure of formal em- intensive and poverty-reducing growth as ployment contracts is less appropriate for the long as they are driven by efficiency criteria. requirements of low-end jobs. Informality Public investments and credit programs to fills the space between inactivity and formal foster labor-intensive growth should be sub- employment, thus creating a cushion in case ject to rigorous economic analysis. FAT of formal sector employment loss. Infor- credit funding allocations should be re- mality can also serve as a path for social viewed and reformed to ensure more effec- ascent. tive social targeting of programs and effi- cient investments. 33. While informality is a solution for many, it also reveals a problem (the high 31. The effective working of labor markets cost of formality) and has a cost (in par- is critical for translating growth into ticular social exclusion, less access to so- broad-based employment opportunities and cial insurance and government social pro- poverty reduction. The Government is ap- grams, less access to credit, and, for the propriately pursuing important reforms of self-employed, less access to other mar- labor regulation, whose importance is high- kets). A strategy to deal with informality lighted by the significant increase in open should be directed at addressing the underly- unemployment rates during recent years. ing problems by reducing the cost of and Issues under active consideration include thus increasing access to formality. At the further regulatory reforms, changes to the same time, policies for those who prefer to labor court system, and other policies to im- remain informal should be directed at reduc- prove the functioning of labor markets. ing the costs of informality. As a result, the These reforms occur in the context of high distinction between an expanding but less employment costs and a labor market that is regulated formal sector and a shrinking but Attacking Brazil's Poverty Page 13 less excluded informal sector would gradu- new financing mechanisms, including ally diminish. Policies to address informal- per-capita federal transfers and spend- ity would include the following: ing floors. (a) Tax reform and simplification aimed at (b) Urban Services and Housing. Urban reducing the number of informal busi- services and housing are important as- nesses. Informal businesses typically sets for the poor. Access to these ser- employ informally; thus, reducing the vices is important, however, they are number of informal businesses should not particularly suitable as vehicles for also reduce the number of informal resource transfers to the poor. Sub- employees. sidization has in the past often con- tributed to flawed targeting. (b) Less rigid regulation with regard to mandated benefits and the structure of the em- ployment relationship Figure 11: Improvements in School Enrollment 1980-1999 (structure of benefits, This figure shows the great recent advances in school enrollnent. Net payments at dismissal, enrollment at the primary level is almost complete, and secondary temporary contracts, school enrollment is increasing rapidly, as more students complete pri- etc.), permitting a lar- mary education. ger range of choices to .... the contracting parties. 9 _ - (c) A closer tie between - _ social insurance bene- fits (pensions and un- employment insurance) 2 and contributions. E I 40 (d) Selective access to so- cial insurance mecha- 20 nisms also for the in- 10 formal sector. 10 Capacity by Strength- 19 1991 1Y994 ening the Assets of the +PI -Pxi99ly Poor (c) Microcredit Programs. Microcredit 34. Efficient social investments create the programs and other initiatives to en- ladders of opportunities out of poverty and able access to productive physical allow a permanent escape from poverty. capital are promising areas for new High priority social investments include: policies. In the case of credit, access to capital is typically more important (a) Education and Basic Health. Educa- for the poor than price. Credit pro- tion and basic health help strengthen grams should be designed to be com- the human capital of the poor and in- mercially viable and financially sus- crease their income. Major progress tainable. has been achieved recently on health and education, in particular through Page 14 Attacking Brazil's Poverty Figure 12: Incidence by Education Level (1996/7) This figure shows the incidence of public spending by level of education. Public creche, kindergarten, and primary education are clearly progressive with more students enrolled from the poorer than from the wealthier population groups. Consistent with low completion rates of primary education by the poor, the incidence of public secondary education is lowest for the poorest quintile. Of all students enrolled in public secondary education, less than 10% come from the poorest quintile and less than 30% are from the poorest 40% of the population. The incidence of public higher education is concentrated in the fourth and fifth quintiles. The IBGE-PPV survey did not include one individual from the two poorest quintiles that was enrolled in public higher education. : 80% - = 70% . "I 60% o O 50% 0 40% w 2 30% .020 '0% 0% 1st (Poorest) 2nd 3rd 4th 5th (Wealthiest) Consumption Quintile + Creche -U- Kindergarten education -- Primary education -e-- Secondary education Source: IBGE-PPV, 1996!7 -K- Adult education -0- Higher education 35. Over the last few years, the Govern- pact on short-term educational indicators, ment has given major emphasis to im- such as enrollment rates and school attain- proved education and implemented major ments, is already visible (see Figure 11). reforms in this sector. As a result, Brazil More importantly, they will make an impor- has achieved very significant improvements tant contribution to poverty reduction with in public education. These improvements the lags inherent in any education sector re- include a massive expansion of basic, espe- form. cially primary, education and a reduction in the regional differences of education en- 36. Data from the 1996/7 PPV shows that rollment rates. Recent advances to improve Primary and pre-primary education is educational quality by supporting educa- highly progressive. Adult and secondary tional spending and reforms in the poorer education benefits mostly middle-income regions have been significant and are likely groups, and public higher education is al- to produce further positive results. These most exclusively used by higher-income sector reforms constitute great progress in groups. Figure 12 shows the incidence of the right direction. Their strong positive im- public education by level. Attacking Brazil's Poverty Page 15 37. Despite the drastic recent advances in sued: the education sector, low educational out- comes and attainment among the poor re- (a) New educational investments and re- main the single most important obstacle to forms concentrate on early childhood reducing poverty and inequality in Brazil. and primary education in poor areas The low quality of public schooling in poor and communities. These include ini- areas is the main obstacle to improving the tiatives that promote the participation educational performance of the poor. Low of parents in school governance, tech- education levels lead to low income, which nical assistance and training for in turn, perpetuates poverty. The inferior schools that focus on improving learn- quality of education in poor areas is related ing outcomes, and testing and report- to traditional spending inequalities and is ing programs that increase local ac- being addressed by the Government's re- countability. form program. Further improvements in educational outcomes, in line with the cur- (b) Professional training programs should rent educational reform strategies of the be rigorously evaluated and, if appro- Government, have to be at the core of any priate, expanded through targeted pro- strategy to increase the economic opportuni- grams such as PLANFOR. Adult ba- ties of the poor. Several policies that are sic education, such as the Alfabetiza- essential for improving education of the a4o Solidaria program, needs to be poor are ongoing and should be further pur- strengthened. Figure 13: Incidence of Health Facility Usage (1996/7) Approximately 25% of the patients in public hospitals and clinics are from the poorest quintile. Less than 10% of the patients of SUS-financed private clinics appear to be from the poorest 20% of the population. Averaging over all publicly-financed facilities, the poorest 20% of the population comprises more than 20% of the patients overall. 0.5 _ _ _ _ _ _ _ _ 0.45 a0.4 LI o.s 0.35 03 0.2 0.3 Sr .2 0.25 LA.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' '3 0.25 0.1 U S 0.05 0 1st (poorest) 2nd 3rd 4th 5th (wealthiest) Consumption Quintile Public Hospital -u Public Health Post Contracted Private Hospital (no insurance) -- Contracted Private Clinic (no insurance) -_-All Public-Funded Page 16 Attacking Brazil's Poverty (c) Funding for public universities should basic physical services for the poor can sig- be re-targeted by introducing cost re- nificantly improve living conditions and covery mechanisms, accompanied by contribute to income earning opportunities. the expansion of bursary systems for Improvements in basic services result in the poor. more time and better access to paid work and education. Improvements in housing 38. Major reforms of the public health care facilitate the use of the house and the system are ongoing to strengthen the uni- neighborhood for economic activities and versal delivery of basic health care to all, improve employment prospects. including the poor. Recent initiatives to focus more on basic, female, and family 41. In many areas, coverage of the poor health services are of great strategic impor- with urban services has improved. How- tance and promise positive results. Further ever, the rapid growth of many urban areas public health care system reforms directed at coupled with financially unsustainable the needs of the poor-toward more effi- subsidy policies have implied a persistent cient and equitable funding mechanisms and access gap for urban services for the poor. incentives for effective service provision at Water and sanitation services are typically lower government levels-are underway. the highest priorities. However, access to This effort is all the more important as the effective public transport is also important. poor rely almost exclusively on public Several policies can be implemented to im- health care and cannot typically afford to prove the provision of urban services for the pay for private health services. poor: 39. Public health facilities are used mostly (a) Urban upgrading projects should inte- by poor and middle-income families, yet, grate the relevant sectors and incorpo- some private but publicly-funded facilities rate genuine beneficiary participation appear to be used predominantly by the in the design and selection of priori- better-off Figure 13 shows the incidence of ties. Consideration should be given to use of different publicly-funded health fa- projects that transfer resources directly cilities in 1996/7-before some of the recent to community associations for imple- reforms have taken place. The apparently mentation of their own priority activi- regressive spending on private clinics con- ties. tracted by SUS requires further analysis. If the reported findings are confirmed, and (b) Phased with improvements in overall there are access barriers to the poor for using urban cost-recovery policies, subsidies such facilities, these should be removed, and for urban upgrading projects should be spending in this segment of the health sector phased out in favor of full cost recov- should be reassessed. The Government is ery. Only the most basic services (wa- moving in this direction through a pilot ini- ter standpipes, for example) in nar- tiative to charge private insurers for SUS rowly targeted programs could remain services provided to their clients. Control exempt form a general full cost recov- over who benefits from publicly financed ery policy. health care will also be facilitated by the pilot-introduction of a SUS identity card. (c) Tariffs for water and sanitation and other network services should be re- 40. Limited access to physical assets and formed. Subsidized social tariffs, urban services for the poor is another im- when necessary, should be tied to a portant obstacle for poverty reduction. means test rather than consumption Lack of affordable housing and gaps in ur- levels. Any necessary subsidies ban services limit the quality of life and re- should be provided for access rather strict the economic opportunities of the ur- than usage charges. Cross-subsidy ban poor. Improvements in housing and mechanisms need to be designed such Attacking Brazil's Poverty Page 17 that they offer the provider with strong poor and regularizing the present informal incentives to serve the poor. housing situation of many of the poor. (d) The privatization of infrastructure ser- 43. Past housing financing was very re- vices should be pursued to increase gressive with most financial contracts ex- sector investment and efficiency. tended to households above the poverty Clear incentives and requirements to line. Poorly targeted but subsidized housing serve the poor and effective regulatory loan programs (in particular Carta de oversight need to be an integral part of Credito) should be discontinued in their cur- every privatization. rent form. Instead, financial sector and regulatory reforms should be pursued to fa- (e) Access and affordability of public cilitate the production of unsubsidized hous- transport for the poor should be im- ing for the non-poor. More unsubsidized proved through: (i) fare and schedule housing options for the non-poor will reduce integration; (ii) improved all-weather the crowding-out of the poor in housing access to poor communities; (iii) fa- markets and free resources for programs cilitation of non-motorized access to targeted to tb, poor. public transport; (iv) regularization of the informal transport market where Security through Social Protec- appropriate and efficient; and (v) di- tion versification of the public transport vehicle fleet. The conversion of Vale 44. A basic, efficient and well-targeted so- Transporte into a regular cash compo- cial safety net can be important to insure nent of the salary should be evaluated. against the risk of poverty. A social safety (f) Fiscal reforms need to improve incen- net can also guarantee a basic level of con- tives for revenue collection and service sumption for those not able to benefit from provision to the poor at the municipal the opportunities created by growth and level. Viable municipalities need to social investments. Growth will continue to build capacity in order to support the exclude some individuals at the very bottom managemtof urban services. Re- of the income distribution who depend on a management opmen serviche- social safety net. Also, overall poverty re- gional area development approaches duction can proceed more rapidly if growth might be more appropriate in the area and social investments are complemented by of non-viable micro-municipalities, social safety net mechanisms. While well- 42. The poor often occupy shelter without designed safety net programs can cheaply formal property title in informal neighbor- and quickly lift a large number of poor out hoods without access to many public ser- of poverty, they are more limited in creating vices. Excessive housing and land use regu- upside opportunities. Social safety net pro- lations artificially reduce the supply and in- grams require continuous budget outlays, crease the cost of formal housing. As a re- and, if not carefully designed, they create sult, the poor are often excluded from formal the danger of poverty traps and undesirable housing, limiting their opportunities to grow incentives. out of poverty. Except for a few states in 45. Brazil's social olicies include several the South, there is no formal low-income p housing policy backed by adequate financial programs that constitute elements of a so- resources. A systematic policy for the in- cial safety net. Other than the social insur- formal housing market should be developed. ance programs discussed before, these pro- This policy should include regulatory re- grams include minimum pension programs forms aimed at making formal housing solu- and social assistance programs directed at tions (owned and rental) available to the specific vulnerable population groups. Workfare programs are used in particular in the rural Northeast during drought periods. Page 18 Attacking Brazil's Poverty Governments at different levels have also system should be reevaluated and pos- been experimenting with transfer payments sibly replaced with cash transfer pro- linked to children's school attendance. The grams. Subsidies implicit in untar- Federal Government is now massively ex- geted food stamp programs (PAT) panding these Bolsa Escola programs. should be eliminated. Overall, social transfer programs are of sig- nificant scale and scope and include the (c) The gradual expansion of minimum spending of nearly R$20 billion a year ac- income programs, means-tested and cording to some estimates (including mostly linked to structural improvements, special pension regimes). An analysis of the such as school attendance of children sources of income of poor households in (Bolsa Escola), should continue. Brazil reveals that these transfer programs have made a very significant contribution to (d) Alternative approaches for improving poverty reduction over recent years, in par- the social protection of informal sector ticular in rural areas. Shortcomings include workers without creating incentives targeting problems of some cash transfer against formality should be evaluated programs. Gaps in the social safety net re- and tested. main for income and employment loss in the informal sector. Also, current social policies 47. Social security (Brazil's pension sys- fail to provide general income support for tem), unemployment insurance, and sever- the poorest who do not fall into specific ance fund (FGTS) constitute important covered categories. social insurance mechanisms; however, they are not targeted at the poor. 46. Brazil can afford a gradual expansion of its social safety net based on targeting (a) Since general pensions (in particular improvements for expenditures on existing public sector pensions) are not targeted programs. Given fiscal constraints and pos- at the poor, they should be self- sible adverse incentive effects, the Govern- financing and should not rely on a very ment is understandably reluctant to create large transfer from the budget. Re- broad new safety net programs, broader sources could then be reallocated to workfare programs, or generalized minimum programs targeted at the poor. Com- income support. Remaining gaps of the prehensive pension reforms are ongo- safety net, for example for workers outside ing and should be further pursued by of the formal sector, would therefore be the Government.3 filled gradually through refinements and improvements in existing programs. At the (b) Reflecting the fact that most formal same time, ongoing experiments would be employees (eligible for unemployment carefully evaluated to design well-targeted insurance) are not poor, unemploy- and efficiently implemented additional ele- ment insurance benefits are heavily ments of the social safety net. These re- concentrated in the population group finements to the social safety net would fur- just above the poverty line. Since un- ther increase income security and reduce the employment insurance is funded number of poor. Several measures can im- through a special contribution from prove the social safety net: employers (PIS/PASEP), its inferior social targeting is not problematic per- (a) For income transfer programs, a rigor- se. More problematic is the overlap of ous quasi-means testing mechanism unemployment insurance benefits with and a recipients register should be cre- benefits from the severance fund ated. (FGTS). The large combined benefits (b) Food and nutrition programs that are 3For an account of the complex issues of pension reforn see not anchored in the education or health World Bank, 2000, Brazil: Critical Issues in Social Security, Report. No. 19641-BR. Attacking Brazil's Poverty Page 19 in case of dismissal from a formal sec- 49. In Brazil, mechanisms for the exclu- tor job are likely to contribute to inef- sion of the poor are typically subtle and ficiently high worker turnover and implicit, and are not typically based on other labor market distortions. Re- open discrimination. Access to formal in- forms should thus aim at integrating stitutions, such as formal labor contracts, the unemployment insurance program formal land ownership, and formal urban with the severance fund, forming a services, is critical for improving the living single system funded from contribu- conditions of the poor. However, access to tions. these formal institutions is often not avail- able to most of the poor due to high regula- (c) Professional training programs should tory standards and costs. Improving access be carefully evaluated. The Govern- to formal institutions is a critical strategy for ment already pursues their better coor- fostering social inclusion. Developing an dinated with labor market intermedia- active and explicit policy toward informal tion and unemployment insurance to labor and housing markets is consequently increase their effectiveness. of high priority. (d) The Abono Salarial (an annual salary 50. The empowerment of the poor and complement paid for lower income their representation and participation in formal sector workers from FAT) fails decision making are important dimensions to serve an insurance function, and is of poverty reduction. First, participation not targeted toward the poor. Given can improve the quality of a given program the need to prioritize spending, this or project. Participation is instrumental to program should be replaced in favor of revealing the demand of the poor. Participa- more poverty-targeted social spending tory processes to select priorities are thus programs. fundamental for social policies that meet the demand of beneficiaries. Participation can Empowerment through Effective significantly lower the cost of policies in Participation different areas. In-kind contributions from beneficiaries, participation of NGOs, super- 48. Empowerment of the poor as a poverty vision of contractors, and procurement by reduction strategy plays an increasing role community associations can all have a sig- rn the international thanking on develop- nificant cost-saving effect. Second, genuine in th intrnatonalthinkng o devlop- participation can foster the development of ment. An in-depth analysis of the issues re- part icipation can enablr the lating to empowerment of the poor in the social capital, which in turn can enable the context of Brazil is left for future work. At poor themselves to improve their well-being the same time, it is important to recognize beyond the specific intervention. Brazil has the potential importance of the related is- experience with a broad range of models of sues. Empowerment of the poor as a strategy beneficiary participation ranging from lim- for poverty reduction arises from the analy- ited attempts at superficial consultation on sis of institutional, social and political proc- substantially pre-determined projects to full esses that bring about policies that are more implementation of projects by community or less favorable for the poor. Empowerment associations or other local groups such as orcludess favorabe forthepoors: (a mpu dmeni- school councils. The latter model has been includes, amnong others: (a) public admini- used in Bank-funded rural poverty reduction stration accountable, accessible and respon- projectsBan d rural povert rects. sive to all, including the poor; (b) access for projects and school improvement projects. the poor to formal market institutions; (c) These initiatives could provide a framework the breaking down of social exclusion and for urban projects and policies as well. barriers for specific disadvantaged groups; and (d) effective participation of the poor in the definition and implementation of public policies. Page 20 Attacking Brazil's Poverty Analysis, Evaluation, and Coor- ble set of broader social indicator tar- dination of Social Spending and gets. The national poverty reduction plan would combine technical analysis Policies with a participatory process for the de- termination of priorities and actions. 51. The recent Plano Pluriannual (PPA) and the Projeto Alvorada of the Federal (b) Rigorous analysis and evaluation is a Government represent major progress with critical condition for the better target- respect to the planning and implementation ing and effectiveness of social pro- of social policies. The PPA creates the grams. Analysis and evaluation pro- framework for systematic and objective- cedures need to be strengthened and oriented planning of Federal Government built into any social program. There spending. If well implemented, the emerg- should be a central office (at the fed- ing evaluation framework of the PPA could eral level, for example in the Ministry prove essential for the improvement of the of Budget and Planning) that defines quality of social spending in Brasil. The criteria for impact evaluation and Projecto Alvorada focuses government ef- compares the findings across different forts and spending on a few selected priority programs. social programs and their targeting on poor states, micro-regions and households. This (c) It would be useful if the Government program is a promising to attempt improve defined an official poverty line that the prioritization, poverty-targeting, and co- would permit more systematic tracking ordination of social policies. of poverty indicators, together with a small number of additional social indi- 52. National policy coordination, the de- cator targets. velopment of an evaluation culture, and improvements in data and analysis are re- (d) IBGE should review and revise its sys- quired to continuously improve the quality tem of household surveys to improve and targeting of social spending. Cur- their usefulness for the evaluation of rently, efforts to fight poverty are still dis- efficiency and targeting of social poli- persed between different sectoral agencies at cies. The lessons and experiences various government levels, each pursuing from the PPV should be incorporated their own agendas. These agendas only in this review. partly coincide with a national poverty re- duction strategy. Several steps can be taken (e) Additional poverty research and analy- to coordinate national policy and establish a sis should focus on: (a) more work on culture of program evaluation and analysis: poverty dynamics; (b) more detailed analysis of the relationship between (a) The development and implementation the pattern of growth and poverty re- of a national poverty reduction plan duction; (c) analysis of the two-way could serve as a vehicle for coordinat- relationship between poverty- ing the activities of different entities reduction interventions and the forma- and orienting systematic and cost- tion of social capital; and (d) a rigor- effective action. With coordinated ac- ous process of classifying Brazil's tion and the implementation of a na- poor into groups with relatively similar tional poverty reduction plan, Brazil characteristics, permitting better tar- would be one step closer to the strate- geting of social programs to these gic goal of reducing the extreme pov- groups. erty rate by 50% by 2015 and a possi- Attacking Brazil's Poverty Page 21 53. The following paragraphs indicate the 54. The following indicative analysis shows type of comparative analysis of social that most social spending is not targeted to spending that should be undertaken on a the poorest. Table 3 summarizes the main continuous basis. This analysis is based on items of consolidated social spending in many simplifying assumption and should Brazil in 1995. The overview shows social not be misused for simplistic conclusions. spending of approximately R$132 billion, of However, further more detailed efforts along which R$68 billion refer to social security the same lines would greatly contribute to and public service benefits. Of total social informing the public debate on social poli- spending, R$18 billion or 14% accrue to the cies and improving the quality of social first quintile of the income distribution spending. (roughly corresponding to the poor as de- fined in this report). These benefits to the Table 3: Summary of Consolidated Social Spending, 1995 with Updates Total Budget Effective Targeting Total benefits to Budget Spending Spending (R$bn/a) to bottom 20%* poor (R$bn/a) per Total benefit to Poor ereral Investment Uniform Transfer Payment 20% 6.3 5rfechWe 1 1 4 0454. Kindergarten 42% 2.4 Basic Education 11.6 26% 3.00 3.8 Secondary Education 1.7 7% 0.13 13.5 University Education 5.1 0% 0.00 Adult Education/Training 0.4 5% 0.02 22.2 Universal ub ic Health Care 21 ater onnection Sewer Connection 4% 25.0 Urban Public Transport 2.6 9% 0.24 11.1 Housing (Carta de Credito) 7.2 2% 0.14 50.0 Favela Upgrading 34% 2.9 Microcredit Programs .0 Land Refom 2.0 70% 1.40 1.4 BPC (LOAS) 1.2 70% 0.76 1.6 Disabled Services 0.1 50% 0.02 2.5 Child Services 0.2 42% 0.07 3.0 Subnational Social Assistance Programs 1.9 70% 1.06 1.8 W a orPra"dicat"lon" NM Minimum Income Programs (subnational) 0.0 70% 0.01 1.8 F t School Lunches 0.7 25% 0.14 5.0 Matemal Nutrition (Milk Programs) 0.1 29% 0.03 4.3 nempomentnsurncee Severance Payments (FGTS)- 13% 9.6 Abono Salarial 0.6 13% 0.06 9.6 Drought Workfare (in drought years) 1.2 70_ _ _ _0.76 1.6 SummarylTotal 131.8 14% 18.00 * Targeting numbers in italics are staff estimates, not based on household surveys No budgetary spending Sources: IPEA, Femandes et.al., World Bank staff estimates, includes updates to reflect more recent program changes. Page 22 Attacking Brazil's Poverty Figure 14: Structure and Targeting of Federal Social Spending, 1997 This figure describes the structure and targeting offederal social spending. The box represents total federal social spending of R$ Ill billion in 1997. The horizontal axis (areas shaded in black) describes how much of this social spending is targeted to the poor. The vertical axis gives the specific amount spent on a social program. For example, the amount spent on pensions and public servant benefits is approximately R$68 billion of which 8% was received by the poorest quintile. All the shaded areas within the box show the share of public social spending that was targeted to the poor. The box in the top right right corner illustrates the size of the aggrgate income gap of the poor compared to social spending. Higher Education 100 80 Pensions and Public Servant Benefits c 60 0 , 40 Secondary Education Urban Transport- Labor _ ~~~~~~~~~~~~~~~Other Education_ 20 Health _ ~~~~~~~~~~~Housing and Urban Nutrition _~~~~~~~~~dcto :Land Reform | ~~~~~~~~~~~~~~~~~~Social Assistance' 0 i I 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Share of Spending Targeted to the First Quintile (Poorest) Shaded poor can be divided into cash (R$6.5 billion) 55. Excluding social security and related and in-kind benefits (R$11.5 billion). Bene- items, social spending amounts to R$64 fits include those accruing immediately billion, of which about RS13.5 billion (or (transfer programs) or over the lifetime of 21%) accrue to thefirst quintile. Figure 14 beneficiaries (education). shows the structure and targeting of federal social spending in 1997 (compared to con- Attacking Brazil's Poverty Page 23 solidated social spending in 1995 shown in reach among the poor in existing pro- Table 3). The chart excludes state and mu- grams. The more complete the reach to the nicipal spending, which is significant espe- poor, the more difficult it is to control leak- cially for education and health. The entire age. This is the challenge faced in up- box in the figure represents the size of fed- scaling small and well-targeted social devel- eral social spending (R$ 111 billion). Each opment programs. The challenge is to either horizontal slice represents a different cate- reallocate funds from programs with inade- gory of spending, ordered by the targeting quate reach and targeting to programs closer ratio (share of spending accruing to the bot- to the top right corner, or to redesign exist- tom quintile). For each slice, the shaded ing programs such that they move toward area shows the share of spending in that par- the top right corner, representing better tar- ticular category that accrues to the bottom geting and wider reach among the poor. quintile of the population. The total shaded area in the chart therefore represents the 58. Through sone socialprograms that are share of total federal social spending that not well targeted, it costs R$25 or more of accrues to the bottom quintile. budget spending to bring benefits of R$1 to the poor. It is instructive that several large 56. Most programs lack either broad reach programs, which are considered part of among the poor, good targeting to the poor, overall social spending, are less cost- or both. This report employs different effective in bringing benefits to the poor (as methodologies to compare the effectiveness cash income or in the form of other benefits) of social programs in different sectors. than a hypothetical program that would Figure 15 graphically compares programs transfer a uniform amount to every Brazil- along three dimensions: each bubble repre- ian. Programs that are not targeted or ad- sents one area of social spending; the size of versely targeted (for example in the area of each bubble is proportional to annual spend- urban investments, secondary, adult, and ing per beneficiary household (annualized in higher education, social security, and unem- the case of investment programs) showing ployment insurance) should not rely for their the relative importance of the program to justification on the objective of effective beneficiaries; the horizontal position of the poverty reduction. Of course, these pro- bubble shows the level of targeting of the grams may be worthwhile for other reasons program to the bottom quintile; the vertical not captured in the poverty analysis. position of the bubble shows the reach (cov- erage) of the program among the bottom 59. Further social spending reform is nec- quintile. Programs in the lower left corner essary to increase targeting, allocation be- are poorly targeted and do not reach many of tween programs, internal efficiency, and the poor. The largest of these are pensions, demand orientation. Given fiscal con- unemployment insurance, sewage provision, straints and the importance of fiscal equilib- and secondary education. Programs in the rium for the long-term prospects of the poor, bottom right corner are those well-targeted it would be difficult to increase the level of but only reaching a small share of the poor targeted social spending until social security (typically social assistance programs). Pro- and other fiscal reforms begin to free up re- grams near the top left corner are universal sources. In the meantime, improvements in programs, especially water and public the targeting, allocation, and efficiency of health. Public pre-primary and primary social spending are critical. education is better targeted but reach differs by level. For reference, the impact of dis- (a) Better targeting requires different ap- tributionally-neutral, annual growth of 4% is proaches for different programs. For shown in the top left corner. universal programs, such as education, health, and urban services, the target- 57. The analysis summarized in Figure 15 ing of public spending can be im- suggests a trade-off between targeting and proved without abandoning universal- Page 24 Attacking Brazil's Poverty Figure 15: Targeting and Reach of Social Programs Each bubble in the figure represents an area of social spending. The size of each bubble represents the estimated spending per beneficiary household and, thus, the relative impportance of the program for beneficiaries. The horizontal position of the bubble represents a program's targeting to the poor (share of spending to the bottom 20% of the population) while the vertical position represents reach among the poor (share of the bottom 20% of the popualtion who are program beneficiaries). For example, public kindergarten education is roughly at 0.4 along both the horizontal and vertical axis, indicating that the program is approximately 40% targeted to the poor and accessible to approximately 40% of the poor. In other words, the poor comprise 40% of all kindergarten students and 40% of children from the poorest quintile of the appropriate age are covered by public kindergarten education. 1 *General Investment/Growth of 4% p.a. Public Health Primary Education 0.8 * 0 0 0.6 Wae Merenda Escolar 0 E 0.4- 0 0. E Kindergarten Social Assistance 0.2 - Sewage 0.2 Pensions Unemployme Secondary Education LOAS/BPC Insurance dult Education Creche 0 0 0.2 0.4 0.6 0.8 1 Targeting to Poor Source: WB staff estimates ity. True universalization of services transfers targeted to the poor. Income from which mostly the poor are ex- transfers and related programs should cluded is, in fact, synonymous with be strictly tied to a means-testing pro- targeting the poor for service expan- cedure. In particular, public spending sion. First, public spending should fo- for income transfer programs should cus on spending items that dispropor- be focused on the very poorest. tionally benefit the poor (such as schools, health facilities, and water (b) There needs to be rigorous comparison supply in poor areas, or education ini- of the efficiency and effectiveness of tiatives targeted at reducing repetition social programs. Demand and will- rates in school). Second, some Gov- ingness to pay serve as the basic emmnent services could require cost re- guidelines to estimate program bene- covery for the non-poor. For insur- fits. Investments and current transfers ance programs, such as pensions and should be compared on the basis of unemployment insurance, contribu- cost-benefit and transfer effectiveness tions and benefits should be set such analyses. Redistributive objectives that public funds are only used for should not be used to justify bad in- Attacking Brazil's Poverty Page 25 vestments. Social investments should the capacity of states and municipalities to pass an efficiency test demonstrating deal with poverty-related issues. that they are more cost-effective than income transfer programs in bringing 61. Given the large intra-regional and in- monetary or non-monetary benefits to ter-regional inequalities, Brazil requires the poor. explicit federalist social policies. Nation- ally funded policies should use uniform tar- (c) The design of social programs should geting criteria (such as a national poverty correspond to the demand of the bene- line adjusted for regional price differences) ficiary population and include genuine that would imply a strong spending bias in beneficiary participation. In general, favor of poorer regions. Subnational gov- this implies that service provision emments in richer regions should be encour- should be at the level of willingness to aged to apply more ambitious criteria for pay. Cash and voucher programs are incremental spending from their own re- preferable to in-kind service provision, sources. unless the latter is justified by better targeting or externalities. In some 62. Social spending reforns are compli- cases, a switch from in-kind to cated by rigidities in the form of Constitu- voucher financing can be an effective tional and legal entitlements as well as means to increase transparency, con- earmarked funding sources. Vocal con- sumer choice, competition, and inter- stituencies are attached to most programs nal efficiency. Beneficiary participa- from which funds might be reallocated to- tion in program design, implementa- ward better targeted programs. Progress on tion, monitoring, and evaluation is es- social spending reform is therefore a gradual sential. process and should be assessed against real- istic benchmarks. Social spending reform (d) While not analyzed in detail in this should involve implicit public or quasi- report, the reviewed material suggests public spending, including credit programs a very substantial scope for further administered by public banks and subsidies improving the internal efficiency of implicit in public services and utilities. social spending. Reform of public sec- tor management as well as an in- Policy Reform Summary creased role of the private sector in providing social services could in- 63. The most important elements of the crease the physical output or social poverty reduction strategy proposed in this impact for given budget resources. report are sustained growth build on fiscal adjustment, continuation of the massive 60. An increasing share of social policy is ongoing reforms to improve quantity and implemented at the state and municipal quality of education, and a deeper effort to level, especially in education, health, and improve the quality and targeting of social social assistance. Thus, any social spending spending overall. Table 4 summarizes the reform must include reforms at the state and policy areas that compose the proposed ele- municipal level. Social spending reform at ments of a poverty reduction strategy (the the state and municipal level revolves first column), the types of policy and spend- around three objectives. First, reforms ing reforms to be pursued (the second row), should improve incentives for providing the principles for policy reform (the third service to the poor. This could include con- row) and the elements of a reform program ditional fiscal transfers. Second, the respon- (shaded cells). sibilities of the different levels of govern- ment should be clarified where they are not. Lastly, reforms must attempt to strengthen Page 26 Attacking Bazits Ptverty Table 4: Policy Reform Summary Functioning of Markets Cost Effectiveness and Targeting of Social Spending Regulatory Reform Social Spe ding Reform Role of Area Making markets work for the poor Targeting EfRicient Allocation Demand Orientation and Internal and allowing access to formality Between Programs Beneficiary Choice Efficiency Principles and Growth and regulatory reforms to The transfers (subsidies) Reallocate resources be- Provide social services ac- Improve internal effi- Messages improve opportunities and access of social programs should tween programs accord- cording to demand and en- ciency of program imple- for the poor to formal markets be targeted to those below ing to transfer effective- courage active beneficiary mentation (least cost for should be vigorously pursued. a chosen poverty line, ness. Social investment participation in design. physical output and im- which should be uniform programs should only be Transfers of cash are prefer- pact). for all programs (local undertaken if their bene- able over vouchers, which are governments in rich areas fits (including the saved preferable over in-kind trans- could adopt a higher administration costs of fers unless there are justifica- poverty line). Per-capita transfer programs) cx- tions in terms of better tar- transfers should always ceed their costs. geting, public goods, or well- be higher for those below grounded paternalistic mo- the poverty line. tives. Opportunity: Creates incom e ic g h in- ! Reviw a. k asnotst opportunities for the eldn ssanbl,hg4qaiyinfiintcei porm Growth chronically poor. fiscl d . f edYm ge stFuryther deeopanfra t:d1lcatiOdassets and capacity of the SC:.;0| j1i : Education poor to make uselabor maktflcoinldn ao tJrban Services . Human capital and 'tkdpaL6l\0 i i Y k k - X--. -0 $ ii PhsiaAsetsosbaicaanel Assets; mostenkefr basicemrfihrdca: urbandveldr seri- 2-j iQ S: 0 intesdeeny l030 (Credit) vices are suireguationlefofos resource*transfers; Remoe disedtucrton against efficient prices. iev ;::+3 bn no: 00 0000 00t 004i; ::0f= labor intensive growth.~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~it ,* Stk guicpddmdvy th lWtvdowu 4 t0$¢- grow; 0 nthbelastct of ov reduction in Brazil. ___________________ ~ ~ ~ ~ ~ eduatin, ealiL rba Canacitv:~ ~~~~~~ posses Stegte th * Deeo polic fo inora *$0 Fou edriscaltriosfXto an * Fute streng: :then . -5 Evlut conerio of * Imleen public0- : 00 0f: V000 :0 Health tenruiiaities. tenure ntrodu cost r:co: ar. : W e a r i Attacking Brazil's Poverty Page 27 Functioning of Markets Cost Effectiveness and Targeting of Social Spending Rergulatory Reform _ Social Sp nding Retorm Role of Area Making *arkets work foir the poor Targeting Efficient Allocation Demand Orientation and Internal and allowing access to formality Between Programs Reneticiary Choice Efficiency Security: * Social insurance * Encourage development of * Introduce quasi- c Phase out Abono * Replace food and nutri- * Implement public protects from transient private un-subsidized social insur- meaTs testing system for Salarial. tion programs (not anchored in sector management te- Social poverty (no subsidies). ance mechanistns for non-poor. transfer programs. * Reduce public find- the health or education system) fonss. Insurance * Social assistance * Create register of ing for insurance prograMs, wNith means-tested cash trans- Social protects from structural social assistance beneficiar- in particular public sector fers. Assistance poverty and can reduce ies. social security. chronic poverty. . r Reform unemploy- ment insurance atid FGTS jointly and provide basic protection for the informal sector. * Expand BoLr Fscola type programs. * Phase out worker food stamp progranm Empowerment: * Permits demand * Targeting through * Genuine participation in * Cost-savings through oriented services. local councils and conmmu- program design to determine beneficiary participation Participation * Reduces costs. oity associations in poor beneficiary demand. and community procure- * Helps overcome neighborhoods. ment. social exclusion and build social capital. Planning, * Develop National Poverty Reduction Plan, based on technical analysis and a participatory process Analysis, and * Define an office poverty line and social indicator targets irn order to improve tracking and accountability. Evaluation * Create evaluation culture for all social policies and central office for evaluation in the Budget Ministry. * Reform system of household surveys to improve capacity to evaluate public policies. ______________ Support priority poverty research and analysis in selected areas. Page 28 Attacking Brazil's Poverty 5. List of Background Papers The following background papers for this report are the responsibility of the respective authors and are not official World Bank documents; they are available on the following website: http://www.worldbank.org/br. BAKER, Judy. Social safety nets for the poor: international experience with implications for reform in Brazil. Washington, DC. 1998. BAVA, Silvio C. Politicas e programas de combate a pobreza urbana no Brasil: experiencias e impactos. Sdo Paulo, SP. 1998. BARROS, Ricardo P. de, FIRPO, Sergio. Uma analise dos determinantes dapobreza, escolaridade e mortalidade infantil com base em informaq6es municipais. Rio de Janeiro, RJ. 1999. FERNANDES, Rubem C. Urban violence and civic action - the experience of Viva Rio. Rio de Janeiro, RJ. 1998. FERREIRA, Francisco H.G., LANJOUW, Peter, NERI, Marcelo. The urban poor in Brazil in 1996: a new poverty profile using PPV, PNAD and Census Data. 1998. LIU, Zhi. Urban infrastructure and housing for the poor.: issues and options. Washington, DC. 1999. MORAIS, Maria da P. A atuaq ao governamental no setor de saneamento bdsico no Brasil: principais ac6es e desafios. Rio de Janeiro, RJ. NERI, Marcelo. A note on the growth sensitivity ofpoverty across Brazilian states. Rio de Janeiro, RJ NERI, Marcelo. Nature and dynamics of micro-entrepreneurial activities :employment generation and poverty alleviation policies. Rio de Janeiro, RJ. 1998. PINHEIRO, Paulo S. Urban crime and violence in Brazil with particular references to Sao Paulo. Center of the Study of Violence. University of Sao Paulo. Sao Paulo, SP. 1998. PRADHAN, Meno, RAVALLION, Martin. Demandfor public safety. Free University of Amsterdam and the World Bank. Washington, DC. 1998. RIBEIRO, Ana Clara T. Interveng6es urbanas, democracia e oportunidade : dois estudos de caso. Fase. Rio de Janeiro, RJ. 2000. ROCHA, Sonia. Targeting at the poor - issues concerning social programs in Brazil. 1998. SANTOS, Claudio Hamilton Matos dos, MOTTA, Diana Meirelles da, and MORAIS, Maria da Piedade. A atuaa'1o governamental sobre o urbano no Brasil atual e seu impacto sobre as condiqoes de pobreza da populaqdo: Uma discussdo introdut6ria. IPEA. 1998 VON AMSBERG, Joachim, LANJOUW, Peter, NEAD, Kimberly. The poverty targeting of social spending in Brazil. 1999.