Report No. 22032-BR Brazil Progressive Low-Income Housing: Alternatives for the Poor December 21, 2002 Brazil Country Management Unit Finance, Private Sector and Infrastructure Management Unit Latin America and the Caribbean Region Document of the World Bank Currency And Equivalents Currency Unit: The Real (R$)l R$1.75 = US$1 2000 R$2.05 = US$1 2001 R$3.35 = US$1 2002 (October) Fiscal Year: January I - December 31 Managers and Staff ResponsibRe for this Report Vice President David de Ferranti Country Director Vinod Thomas Sector Director Danny Leipziger Lead Economist Joachim von Amsberg Sector/Task Manager Mila Freire ACKNOWLEDGEMENTS The report is based on the findings of several missions to Brazil in November 2000, and March and October 2001 and discussions during 2002. The report was prepared by a team comprising Mila Freire (task manager and principal author), Dean Cira, Vitor Serra, David Sislen, and Elizabeth Mehta under the general supervision of Gobind T. Nankani, Vinod Thomas, Joachim von Amsberg and Danny Leipziger. The work draws on and continues the analysis initiated in 1999 by V. Serra, B. Alberdi and A. Dubel. Peer reviewers for this report were Robert Buckley and Tova Solo. Additional inputs were received from Barjor Mehta and Ivo Imparato. Much of the analysis was carried out in collaboration with IPEA and SEDU. Diana Motta and Maria de Piedade Morais provided important information and insights into issues dealing with land and housing policy and low-income housing. Ricardo Lima, coordinator of Urban Policy in IPEA, provided invaluable support and collaboration. The mission is grateful for the support of Dr. Flavio Candelot, Maria Lucia Brizzi (SEDU), Fldvia Brasil (FJP), and the Municipality of Belo Horizonte for valuable comments and for making it possible valuable meetings with low- income residents and local officials. Earlier drafts of this report were discussed informally with government officials in Brazil and the information and concepts included in the document were key as background to the ongoing dialogue between the Bank and the Brazilian authorities on the National Housing Policy for Brazil. Both this report and a policy paper - Housing Finance in Brazil and Social Housing - were discussed with the Brazilian authorities in August 2002. The present report reflects those discussions. Abbreviations And Acronyms BID Interamerican Development Bank (Banco Interamericano de Desarollo) BNB Bank of the Northeast (Banco do Nordeste)l BNDES National Development Bank (Banco Nacional de Desenvolvimento Econ6mico e Social) BNH National Housing Bank (Banco Nacional de Habita,do) CAS Country Assistance Strategy CDBG Community Development Block Grant CEDAE State Water and Sanitation Company (Companhia Estadual de Aguas e Esgotos) CEF Federal Loan and Savings Bank (Caixa Economica Federal) CHF Cooperative Housing Foundation CIDA Canadian International Development Assistance COHAB Cooperative Housing Agency, Public Housing Company COHAB-CT Housing Company (Companhia de Habita,do de Curitiba) COIVIESS Community Water Cooperative COMPESA The state water company of Pernambuco COMUL Committes for Urbanization and Land Regularization (Comissoes de UrbanizaCdo e Legaliza,do da Terra) EMHAP State Housing Agency (Empresa de Habita9do de Pernambuco) FGTS Workers Severance Fund (Fundo de Garantia do Tempo de Servico) FHA Federal Housing Administration FIDEM Metropolitan Agency for Development (Funda,co Desenvolvimento Regido Metropolitana do Recife FJP Joao Pinheiro Foundation (Funda,do Jodo Pinheiro, Belo Horizonte) FMR Fair Market Rent FONAVIPO National Fund for Social Housing (Fondo Nacional de Vivienda Popular) FONHAPO National Fund for Social Shelter (Fondo Nacional de Habitacion Popular) FVP Fundacion de la Vivienda Popular (Venezuela) GDP Gross Domestic Product GSS Global Strategy for Housing GTZ German Aid Agency (Gesellschaftffur Technische Zusammenarbeit) IBGE Brazilian Institute of Statistics and Geography (Instituto Brasileiro de Geografia e Estatistica IBRD International Bank for Reconstruction and Development ID Institutional Development IDB Inter-American Development Bank ILPLANRIO Municipal planning institute of Rio de Janeiro IMIP Municipal Institute for Research and Planning (Instituto Municipal de Investigaciones y Planeaci6n) IPEA Institute for Applied Economic Research (Instituto de Pesquiza Econ6mica Aplicada) IPTU Urban Property Tax (Imposto Predial Territorial Urbano) MAC Metropolitan Area of Caracas MAP Mutual Help Plan (Plan de Ayuda Mutua) MCHF Micro-Credit to Housing Finance NGO Non-Governmental Organization OECD Organization for Economic Cooperation and Development OGU Federal Government Budget (Or,amento Geral da Unido) BRAZIL LOW-INCOME HOUSING: ALTERNATIVES FOR THE URBAN POOR Contents Page Executive Summary 1. Introduction .......................................................1 A. Purpose of the Report, Major Issues ......................................................1l B. Content and Methodology .......................................................3 C. Analytical Framework ........................................................3 D. Public Housing: Policy Options .......................................................5 2. Low-Income Housing - Deficit and Needs ...................................................... 11 A. Context and Background ...................................................... II B. Housing Deficit ...................................................... 13 C. Demand for New Housing ....................................................... 17 3. Federal Low-Income Housing Programs ...................................................... 20 A. Introduction ...................................................... 20 B. Federal Housing/Land Programs ....................................................... 20 C. Federal Spending on Social Housing, 1995-1999 ...................................................... 23 D. International Collaborations ....................................................... 33 E. Conclusion ...................................................... 33 4. Best Practices in Brazil ....................................................... 35 A. Introduction and Criteria for Assessment ...................................................... 35 B. Brazil's Low-Income Housing Experiences ...................................................... 36 Rio: Programa de Assentamentos Populares- A Citywide Approach ............................. 36 Curitiba: Innovations in Partnerships ...................................................... 40 Sao Paulo - A Tale of Two Approaches ...................................................... 42 Belo Horizonte: Participatory Planning and Global Development Plans .......................... 46 Recife - PREZEIS and Metropolitan Cooperation ...................................................... 50 Ceara ...................................................... 53 C. Summary ....................................................... 54 5. Building the Main Blocks of a Strategy ...................................................... 56 A. Understanding the Way the Informal Sector Works ...................................................... 56 B. Converting the Need for Housing to Demand ...................................................... 57 Resident Housing Savings Program ...................................................... 57 Microcredit for Housing ...................................................... 58 Linking Informal Demand to National Housing Finance ................................................. 63 C. Improving the Stock ...................................................... 64 Land Management and Increasing the Supply of Land ..................................................... 65 Land Occupation and Regularization ...................................................... 67 D. Institutional Arrangements ...................................................... 68 E. Conclusions and Financial Framework ...................................................... 70 R eferences .......................................................... 9 1 Annex A: Urban Low-Income Housing: A Review of the Literature .................................I Urban Growth and the Issue if Housing the Urban Poor .....................................................I Evolution of Low Income Housing: The Main Approaches .............................I................. n Self-Help Housing Services and Sites and Upgrading ...................................................... III New Partnerships: The Enabling Approach .......................................................... IV Critical Issues .......................................................... VII Political .......................................................... VIII Administrative Institutional ..........................................................X Technical .......................................................... XI The Issue of Land .......................................................... XI Issues Concerning Standards .......................................................... XII Directions for the Future .......................................................... XIV Annex B: Statistical Appendix .......................................................... XXIX Tables in the Text Table 1. Housing and Urban Expenditures, 1997 .2 Table 2. Location of Urban Growth, 1950-90 .11 Table 3. Housing Stock by Tenure, 1970-98 .12 Table 4. Housing Indicators (%) .13 Table 5. Housing Deficit, Brazil, 1991 .13 Table 6. Housing Deficit by Income Level (%) .14 Table 7. Housing Deficit by Regions and Spatial Groupings, 1991 .15 Table 8. Informal Housing Units by Metropolitan Region, 1998 .15 Table 9. Informal Housing Units, 1991-98 .16 Table 10. Brazil: Housing Affordability Estimates, Brazil .................................................. 18 Table 11. Brazil: Expenditures in Housing and Urban Development, 1996-97 ................... 20 Table 12. Public Sector Housing Programs in Brazil ........................................................... 23 Table 13. Summary of Main Programs in Housing, 1995-2000 .......................................... 24 Table 14. Housing Programs by Region ........................................................... 25 Table 15. Structure of Housing Deficit and Central Gov't Investment in Housing ............. 25 Table 16. Regional Distribution of Habitar Brazil Funds, 1995-99 ..................................... 26 Table 17. Regional Distribution of Pr6 Moradia Funds ....................................................... 26 Table 18. Interest Rates ........................................................... 27 Table 19. Carta de Credito: Financing Limits ........................................................... 27 Table 20. Regional Distribution of Carta de Credito Program, 1995-1999 ........................ 28 Table 21. Housing Deficit in 1991 Corresponding to Geographic Regions ........................ 28 Table 22. Investments per Year in Selected Federal Housing Programs ............................. 30 Table 23. Housing Program Investments, 2000 ........................................................... 31 Table 24. Regional Distribution of Program Resources, 2000 ............................................. 31 Table 25. Financing Infrastructure Goods ........................................................... 36 Table 26. Income Distribution in Favelas and Loteamentos in Rio de Janeiro ................... 39 Table 27. Curitiba Households on Waiting List ........................................................... 40 Table 28. Housing Type, Metropolitan Area of Sao Paulo, 1999 ........................................ 42 Table 29. Distribution of Household Heads by Income Level ............................................. 44 Table 30. Belo Horizonte - % People Living in Favelas ..................................................... 46 Table 31. Ceara-Housing Deficit .......................................................... 53 Table 32. Difference Between Formal and Informal Housing ............................................. 56 Table 33. Comparing Housing Microfinance Programs ...................................................... 59 Table 34. Institutions Terms for Housing Microfinance ...................................................... 61 Table 35. Institutions' Terms for Microfinance in Brazil (excluding Housing) .................. 63 Table 36. Cost Breakdown of Urban Upgrading, Recife 1998 ............................................ 71 Table 37. Basic Infrastructure and Housing Needs of the Urban Poor in Brazil ................. 73 Table 38. Expenditures in Urban and Housing .......................................................... 73 Table 39. Scenario for Financing a 10 Year Low-Income Housing Program .74 Table 40. FGTS - Average Credit Per Family by Region, 1995-2000 .76 Boxes Box 1. The USA Experience with Low-Income Housing .8 Box 2. Overcrowding and Violence among the Poor .14 Box 3. Informal/Progressive Housing .19 Box 4. Estatuto da Cidade .22 Box 5. 2000-2003 Housing Programs .32 Box 6. Morros Dos Cabritos .38 Box 7. Cingapura .46 Box 8. Projeto Integrado da Vila Senhor Dos Passos .48 Box 9. Residents Saving Program .58 Box 10. Housing Microcredit Programs in Fortaleza, Brazil .60 Box 11. Improved Banking, Palmeira, Brazil .61 Box 12. Microcredit in Brazil .............. 62 Box 13. Cali .67 Figures Figure 1: Households, Services .12 Figure 2: Brazil - Informal Housing Units in Metropolitan Areas, 1998 .16 Figure 3: Brazil - New Entrants in Housing Market .19 Figure 4: Low-Income Housing Finance Program, 1995-2000 .24 Figure 5: Federal Funds Invested by Program, 1995-99 .29 Figure 6: Cufitiba, Households Looking for Low-Income Housing .40 Figure 7: Participatory Planning Process for Finished Housing .49 Annex A: Review of Literature Box 1. The World Bank-assisted Lusaka Project . xXV Box 2. The Madras Urban Development Project in India. XVII Box 3. The Kampung Improvement Program in Surabaya, Indonesia .XIX Box 4. The Urban Community Development Project in Hyderabad, India XX Box 5. The Million Houses Program in Sri Lanka .XXI Box 6. The Community Mortgage Program in the Philippines. XXII Box 7. The Urban Community Development Office (UCDO) in Thailand XXIII Box 8. Khuda Ki Basti Incremental Housing Scheme, Hyderabad, Pakistan . XXIV Box 9. The National Fund for Popular Housing (FONHAPO) in Mexico .XXV Box 10. El Mezquital .XXVI Box 11. Direct Housing Subsidy and Housing Financing .XXVI Box 12. The ARGOZ Case in El Salvador ............................................... XXVI Annex B: Statistical Appendix Table 1: Population Trends and Demographic Densities in Brazil, 1940-2000 ............ XXIX Table 2: Annual Growth of Population (in percentage) ............................................... XXX Table 3: Population/Households, 1988 ............................................... XXXI Table 4: Brazil - Housing Deficit in 1991 ............................................... XXXI Table 5: Housing Deficit by Regions and Spatial Groupings, 1991 ........................... XXXII Table 6: Type of Housing Occupancy by Region, Poor vs. Non-poor ...................... XXXIm Table 7: Selected Elements of Low-income Housing, 1993/1998 ............................ XXXIV Table 8: Brazil: Average Monthly Urban Household Expend. on Rent/Mortgage ................................................ XXXIV Table 9: Latin American Expenditures on Housing Improvement ......................... XXXIV Table 10: Cost Breakdown of Urban Upgrading, Recife 1998 ..... XXXV Table 11: Poor Households in Metropolitan Areas, 1993/1998 ..... XXXV Table 12: Total Funds that could be mobilized by the low-income population ..... XXXVI Table 13: Hypothetical Micro Credit for Housing Loans ..... XXXVI Table 14: Summary of Main Programs in Housing, 1995-2000 ......XXXVII Table 15: Investment in Housing during 1995-2000 ..... XXXVII Table 16: FGTS Carta de Cr6dito Programs in Housing by Region, 1995-2000 ..... XXXVI Table 17. FGTS - Average Credit Per Family by Region, 1995-2000 ..... XXXVIII Table 18: Access to Urban Services (1992/1997) ..... XXXIX Figure 1: Income Distribution in Favelas .....................XXXIX Table 19: Federal Expenditures with Social Sectors .XLII Table 20: Brazil: Per Capita Expenditures with Housing and Urban Development, By State R$ constant prices of 1995 .. XLI Table 21: Expenditures in Housing and Urban Development, By State in constant prices of 1995 .. XLII Table 22: Brazil: Municipal Spending in Social Sectors, 1997 .. XLHI Table 23: Rate of Urbanization and Human Development Index by State, Brazil .. XLIV Brazil: Progressive Low Income Housing Page i Executive Summary A. Introduction 1. Brazil is facing a pressing challenge to help poor households improve their housing conditions and find affordable shelter as part of a coherent poverty reduction strategy. This report provides a framework to discuss the issues facing the low-income housing sector in Brazil and to analyze policy alternatives'. Its message is twofold: First, there is a role for the government to play in the low-income segment of the housing market; indeed, market failures and sheer poverty call for a coordinated and revamped intervention by the public sector. Second, while the needs are staggering, improving the situation for poor households is feasible - but only if there is a combined strategy to boost demand through resource mobilization and resident savings, and to ease supply constraints at the same time. Underlying these conclusions is the ongoing need to continue capacity-building efforts at all levels of municipal and urban management. 2. In February 2001, Brazil's Congress passed a constitutional amendment guaranteeing the right of all Brazilians to adequate housing. The question now is how the Brazilian public sector can make this a reality while maintaining fiscal discipline. Because housing is a long-lived investment (houses last about 70 years) the sector is still suffering from problems created decades ago - the debacle of the national housing financing system in the 1960s and long-lasting macroeconomic imbalances, such as hyperinflation until recently, and continuing high interest rates. Today Brazil faces a triple challenge: to address the accumulated housing deficit, to provide for the improvement of the housing stock, and to accommodate an annual population growth of about 3.2 million people. While some of the growth in demand will be satisfied by the market, public intervention is required to address the specific needs of the very poor. 3. According to official estimates, 5.4 million households live in less than adequate housing conditions in Brazil2- that is, they are homeless, they live in houses made out of substandard materials or in overcrowded conditions, or they lack access to basic infrastructure, naniely potable water and sewerage.3 Sixty-five percent of these families earn less than three minimum wages,4 and 46 percent live in the Northeast. 4. Every year, an additional I million households enter the market for affordable shelter. One-fifth of them (those earning more than 10 minimum wages) will purchase their own homes with savings and through mortgage markets; another 20 percent will qualify for some kind of subsidized credit;5 and the remaining 60 percent will not be able to afford the down payment and the monthly installments to purchase a finished house in Brazil. Many of these households turn to the informal housing sector. 5. The informal housing sector has particular characteristics very different from the formal sector. In the formal sector, the production of housing units involves four steps: acquisition of a plot of land, planning, construction, and sale. In the informal sector, the process starts with Issues related to housing finance, notably SFI and SFH are not dealt with in this report. The contents of the report relate to policies and events up to June 2002 and do not include the policies of the new admnistration. 2Based on the 1995 estimates of Fundacao Joao Pinheiro; recent estimates produced in 2001 indicate a deficit of 6.3 million units but the concept of deficit is not comparable with the 1995 definition. In this report we will be using the data produced in 1995. 3Households may fall under more than one category. 4One minimum wage is equivalent to R$200 (April 2002). 3These values are consistent with the fact that in the last 10 years housing production has averaged about 900,000 units a year. The official housing delivery system has financed about 300,000. Page ii Brazil: Progressive Low Income Housing occupation of a piece of land, then moves to transition and consolidation. Occupation is the act of physically taking over a piece of land. It can start with one person building a shack or a tent and "living" there, although it usually involves some kind of middleman. When occupation is not opposed by the authorities or the landowners, the transition phase begins. During transition, other occupants join the first family, construct rudimentary houses, and begin demanding basic services and utilities (mainly water, electricity, and paved roads) from state and local governments. Consolidation occurs when the families obtain all basic infrastructure and urban services and receive titles to the land (Siembela and Moreno, 1997). 6. The informal sector has helped respond to the demand for low-income housing in Brazil. In spite of zoning and severe restrictions on land use which served to further reduce access to land by the poor, informal developers have managed to circumvent the law and supply illegal land, and the poor, with remarkable ingenuity and resourcefulness, have been able to build their own shelters, even in conditions that could be considered beneath their dignity and actual needs. This situation is unsustainable and misguided from a policy point of view, however. In fact, while the informal sector manages to provide housing solutions where the public sector and market appears to fail, the result is clearly second best. The informal sector cannot properly undertake the collective action role of the public sector, ensuring that, for example, settlements have adequately defined rights of way, properly titled property, access to urban services such as water supply and sanitation. Failure to properly plan for these rights within the property rights bundle, means greater capital outlays in the future when government action is almost always required to rectify the negative externalities generated by informal settlements. This often requires such actions as resettling families in high risk areas (such as flood plains), re-platting public rights-of-way so that emergency vehicles and collective transport can access these communities and untangling legal claims on property that can take years or even decades to resolve. An effective public policy should be based on a good understanding of this informal sector process and at which juncture public intervention is needed. 7. There are four main reasons for the considerable backlog of investment in low-income housing: (a) low-income levels; (b) high supply costs; (c) market failures; and (d) distortions in public policy. a) Poverty. The main obstacle to having access to "decent" housing by the poor is their lack of purchasing power. Assuming that households cannot allocate more than 30 percent of their income to housing, fewer than half of Brazilian families earn enough to afford housing on the open market. As a result, many Brazilians live in substandard structures, in shanty towns, loteamentos clandestinos (informal subdivisions), corticos (tenements), or in overcrowded conditions. b) High supply costs. In the formal housing market, a house of about 35 square meters rarely costs less than R$18,000; the average price is around R$20,000 (about the same price as in most Latin American countries). Were the R$18,000 to be financed by the mortgage market, monthly principal and interest would be about R$200, a level that would be accessible to households earning at least four minimum wages. This group makes up less than half of the Brazilian population. c) Market failures. Present regulations on land use and housing standards, together with monopolies in land ownership and the construction industry, prevent housing prices from falling. Special provisions need to be instituted to allow for lower and more flexible standards of housing construction for the poor as well as for less constraints to land Brazil: Progressive Low Income Housing Page iii subdivision.6 Experience in Brazil and elsewhere has shown that with flexible standards, a house could be built for about R$7,200, or half the price of a formal housing unit. *d) Distortions in public policy. Since the collapse of the National Housing Bank (BNH), the Brazilian public sector has made limited housing policy interventions and only recently public discussion has been initiated on the need to prepare a coherent low-income housing strategy. This neglect has had several consequences, notably: * Lack of coordination of housing policies across government levels. * Limited role of the federal government in dealing with policy issues, notably those related with land use and standards. * Limited role of the federal government in financing targeted low-income housing programs. Public expenditures for housing and urban development in the last five years have averaged R$12 billion a year. Of this, 80 percent has come from the municipalities, 12 percent from the states, and 8 percent from the federal government. Spending on housing represents less than 1 percent of federal social expenditures 7 which is much lower than other countries at similar level of development. * Public expenditures on housing mostly benefit middle-incomne families. Between 1995 and 2000, housing programs at the central level totaled R$12 billion. Ten percent of this amount was financed by the central budget; the remaining 90 percent came from the workers' severance fund (FGTS), which benefited households earning more than five minimum wages. * Very few programs are oriented towards the very poor, and the ones that exist cannot be scaled up to reach a significant proportion of these households who are the most in need of housing. * There is a scarcity of serviced land. Brazil has neglected the urban expansion of its cities. Overcentralized urban regulations have focused on the enforcement of zoning and construction permits but have failed to plan for or to accommodate the increase in population. The natural consequences are the informality and illegality of new settlements that have sprung up. B. Analytical Framework 8. Housing markets have unique characteristics. First, housing is the most basic form of asset accumulation for the poor - in countries like the USA and Mexico it may represent about 50 percent of the assets of the households. It also provides the base for income improvement and social inclusion. Second, housing services represent about 25 percent of household consumption both in developed and developing countries. Third, housing markets are segmented - low- income and high-income markets have distinct characteristics -'middle-income households will "bid" for houses provided by the public sector for the poor if the market does not supply enough middle-income housing. Finally, housing is a bundled good and can be unbundled for purposes of analysis and policymnaking. Location, external and internal finishing, area, and infrastructure all have their separate and identifiable demand and supply functions and characteristics. These features are important in the design of cost recovery policies and progressive loans. A common way of unbundling housing services is by separating the services that are private goods from public goods. 6For years, attempts at lowering the cost of housing have focused on lowering construction cost through technology. This has had very limited impact as the problem is linked more to the public goods provision than the private goods side of the housing issue: 7While not directly comparable it is interesting to recall that education and health account for 6 percent and 12 percent, respectively; social security, 40 percent of federal expenditures . Page iv Brazil: Progressive Low Income Housing 9. Public sector intervention in low-income housing is justified on several grounds: (a) housing generates positive and negative externalities in neighborhoods; (b) housing (basic shelter for the very poor) is a merit good, something that society views as more important than what would be consumed if left totally up to individual choices;8 (c) housing has a strong impact on human productivity and welfare; and (d) minimum standards for housing units raise the cost of housing above what poor households can reasonably afford. Housing subsidies are a method of compensating low-income households for being forced to allocate too much of their incomes to housing. 10. How to intervene. Public intervention in low-income housing can take place in the supply side of the market - e.g., by directly providing housing or by intervening in the market to ease such constraints as monopoly control of housing construction or land availability - and / or on the demand side through direct subsidies to offset the effect of housing regulations that raise the cost of housing units, and to boost the purchasing power or to facilitate the organization of savings schemes of families in need of shelter. 11. Public sector intervention in low-income housing began in the 1960s when massive migration to urban areas led to the formation of many irregular settlements of very low-quality housing. To control the proliferation of these settlements governments used a two-pronged approach: slum clearance (and the consequent displacement of the population), and construction of low-income housing by the public sector. Both approaches failed to produce sustainable results. On the one hand, the supply of public houses was clearly insufficient to meet ever growing demand. On the other hand, the new settlements were generally situated in distant (and cheaper) locations, making it costly and inconvenient for the resettled poor to commute to and from their jobs. As a result the poor would sell the new "public" houses and move into new settlements near town. Gradually, projects focused on slum upgrading and, in the 1980s, self- help construction was introduced. These yielded more success, but government agencies soon faced the same problem, that is, lack of public funds. Costs were still too high and the lack of a cost recovery policy hindered the capacity of the public sector to expand the programs. 12. The enabling approach. In the 1990s, the emphasis was shifted to the "enabling" paradigm. Here, the public sector is seen as the main guarantor of well-functioning markets, through adequate housing and land legislations, rather than as a provider of housing. Public policies are adopted to support demand (through property rights, developing mortgage financing, rationalizing demand subsidies), to help organize supply (providing infrastructure for residential land development, regulating land and housing development, organizing the building industry), and to promote institutional development. This approach also encourages partnerships that combine the resources of national government agencies, municipal authorities, the private sector, NGOs, international funding agencies, voluntary or community organizations, and households and individuals. 13. Focusing on the poor. Within the enabling approach, Simbieda and Moreno (1997) propose to "convert the informal sector housing needs into effective demand for housing and increase the supply of land, urban infrastructure, and housing." On the demand side, the objective is to ease the financial constraints of the poor. This could be achieved by means of small savings schemes, pooling group savings into community trusts and linking the trusts to national housing finance programs, access to staged and progressive loans, technical assistance, community-based 8A merit good is a good that creates benefits for the community as a whole. Housing, health, recreation, and education fall under this category. Brazil: Progressive Low Income Housing Page v production and employment, and small targeted subsidies in the form of up-front capital grants for the very poor who cannot afford a down payment. 14. On the supply side the emphasis is on lowering costs, increasing the supply of serviced land, and engaging residents in the provision of housing. This calls for administrative reforms to: * make the land tenure regularization and land transactions easier; * make building and land subdivision regulations more flexible; * allow for and encourage progressive infrastructure, whereby public investment focuses on major trunk infrastructure and leaves to the residents (in partnership with developers) the process of installing connections to their plots; * develop innovations in materials and reduce system installation costs and cost of materials; and * provide financing in a staged manner linked to local flows of project financing. C. The Brazilian Case 15. Three levels of government share responsibility for housing policies in Brazil. The central government, through the Secretaria de Desenvolvimento Urbano (SEDU), the Ministry of Planning and the Central Bank, is responsible for the overall design of urban and housing policies - including land use and construction codes, tax instruments, and expropriation of land - and for the overall investment program and financial regulations. State governments have their own programs but no specific responsibilities. Municipalities have the largest responsibility and authority over housing policies, through the Planos Diretores and zoning, and have the capacity to make land available for housing for the poor. 16. Public investment. The Brazilian public sector spends on average R$13.7 billion per year on housing and urban development: R$11 billion are financed by the federal, state, and local governments; and R$2.2 billion are credits financed through the Caixa Economica Federal (CEF), using the workers' severance fund (the Fundo de Garantia do Tempo de ServiVo or FGTS), for middle-income families. Local governments are the most important agents for urban and housing investment, contributing on average R$8.5 billion per year, or 80 percent of the total investment; the states' and the central government's respective shares are about 12 percent and 8 percent of the total. 17. Federal programs include matching grants and subsidized credits to states and municipalities, and credits to individuals and communities. The matching grant program is targeted at households earning less than three minimum wages. * Habitar-Brasil, financed by the central government budget, transfers resources to subnational governments as matching grants to improve housing infrastructure- states and municipalities finance 10 percent as counterpart funds. In the last five years, 574,000 families have benefited from the program, each receiving R$2,500 on average. The bulk of the program's resources have gone to the Southeast region. * Pr6-Moradia targets families earning less than three minimum wages. It finances states and municipalities (at subsidized rates) to develop specific projects, including sites and services and slum upgrading. Counterpart (municipal) funds are supposed to finance at least 10 percent of project investments; in practice municipalities contribute 30 percent on average. Average investment per household is R$3,500. The Pane vi Brazil: Progressive Low Income Housing program benefited 337,000 inhabitants during the 1995-2000 period, or almost 50,000 per year. Most of the resources were allocated to the Northeast. Carta de Credito (Individual and Coletivo), which is financed by FGTS, is the most important housing program in Brazil, accounting for 80 percent of federal expenditures in social housing. In the last five years it had invested R$11 billion, benefiting 150,000 families a year with projects averaging R$13,000. The program is targeted at families earning less than six minimum wages. but in fact only 35 percent of the beneficiaries are in this category. The eligibility income limit has been raised to 12 and 20 minimum wages (for individual and collective credit, respectively). Resources are absorbed mainly by the Southeast - 55 percent of the Carta de Credito Individual and 77 percent of the Cr6dito Associativo have gone to this region. Over time, the programs financed by FGTS had gained importance in terms of total investment, while Habitar-Brasil and Pro Moradia declined from 37 percent in 1995 to 5 percent in 2000. 18. Federal investments in housing for the lowest income groups (earning less than five minimum wages) have declined markedly since 1995. At the same time, investment in housing for those earning 12 to 20 minimum wages has increased under the PAR and Carta de Credito programs. Moreover, these programs are concentrated in the Southeast - 56.2 percent of total resources - while the Northeast, where the housing deficit is largest, receives only 15 percent. D. Selected Cases: International and Brazilian 19. Several experiences in Brazil and outside highlight the importance of sustainability, cost recovery, and community participation in low-income housing programs. Some of the more successful experiences in Latin America are: * The housing program in Guatemala's El Mezquital which is relatively low cost - an investment of US$1,200 per household (R$2,600). Cost recovery is the responsibility of the community, which organizes itself to collect payments and to ensure that every household pays. . In Caracas, social assessment is used as a major tool to design the project and to understand the community's capacity and willingness to pay, especially for water and wastewater services. * The case of El Salvador shows how the private sector can play an important role in developing affordable housing for the poor, mainly through improving land supply. * The direct up-front housing subsidy scheme in Chile is a successful example of a program designed and implemented by the central government with clear and transparent rules. The program is based on a flat subsidy that is used in combination with residents' savings and affordable mortgage schemes. 20. The projects analyzed in Brazil offer interesting lessons. Favela Bairro shows the "new way" of doing upgrading, taking into account the investments of poor people in their shelter. Guarapiranga illustrates how to deal with overwhelming slum pressures that threaten the environment. Minas Gerais illustrates the use of participatory budgeting and slum upgrading in accordance with master plans. Curitiba exemplifies novel practices capable of attracting investments by private developers in low-income housing by allowing denser development. Brazil: Progressive Low Income Housing Page vii 21. The above experiences are mostly municipal initiatives, as Brazil at present does not have a clear federal policy on low-income housing9. There are good and innovative examples, such as schemes providing incentives to promote private sector participation; the concept of collectively built self-help housing; slum upgrading; the use of low-cost technologies; and transparent budgeting processes like the Or,amento Participativo or participatory budgeting. Yet most programs,remain small and inefficient. Some of the issues encountered are as follows: a) Sustainability. Given the limited cost recovery, few programs are sustainable and have potentials for expansion. Cost recovery is hindered by the lack of political will and the difficult process of regularizing land tenure, which itself limits the use of taxes or betterment fees. b) Subsidies. Most of the subsidized housing and upgrading programs are targeted on a geographic basis. Up-front demand subsidies are not used. Housing credits are offered at below market levels. c) Affordability. Most programs fail to produce housing affordable to the poor. Progressive housing schemes, like those found in Ceara, have a unit cost of around US$2,500 (R$6,000, including all basic infrastructure, land, and housing) and could be progressively improved and extended. Adequate information on the true capital and operating costs of housing options and community participation in decision-making is essential to affordability. d) Sector objectives. There is a need to overcome local regulatory barriers to affordable housing. Most local land use regulations favor larger, more expensive housing and relegate low-cost housing to the urban fringe. Processes for land titling are long and costly, and tend to limit the number of property settlers who can obtain title, particularly in existing urban slums. Alternatives like usucapido - granting titles after five years of occupation - are clearly second-best tenure arrangements. They are also difficult to enforce owing to long and complex procedures for obtaining titles, indicating inefficiencies in the legal system governing property transactions and registration. A big step forward has been the recognition of slums through programs like ZEIS (Zonas Especiais de Interesse Social) and the designation of affordable housing districts. The Estatuto da Cidade is another step forward by defining the capacity of slum dwellers to demand full title of their land. Still, even innovative programs like the private-public partnership program of density bonuses found in Curitiba can be cumbersome and time- consuming to administer. More can be done to make the entire housing and real estate sector function more efficiently. E. Building a Strategy for Brazil 22. A low-income housing strategy for Brazil requires: (a) a good understanding of how the progressive informal sector works;10 (b) converting the need for housing to effective demand, and organizing demand through socio-economic and technological support; (c) easing the current supply constraints resulting from both regulations and market failures; and (d) adopting a clear and appropriate institutional framework. The manner by which informal housing settlements are 9The situation is changing. The federal govemment has launched a serious analysis of the National Housing Policy in Brazil and very interesting work has been produced as input to a generalized discussion and eventually proposal for an integrated and consistent policy. The Bank has been a partner in this ongoing work. Forthcoming policy notes will sharpen the main issues at stake in low income housing and housing finance. '° Not enough research has been done on this issue, particularly the process of informal land subdivision. Page viii Brazil: Progressive Low Income Housing created - from the invasion of public or private land through to transition and consolidation when infrastructure and regularization are obtained - necessitates gradual steps that can enhance the benefits of this progressive process. In terms of organizing demand, several alternatives are available, notably increased resident savings or mutual assistance programs for housing, microcredit schemes, and community-based production generating employment in construction or materials banking, for example. 1) Translating Informal Needs into Housing Demand 23. The most crucial aspect of the low-income strategy is that of organizing demand, a key element of demand-responsive and well-targeted programs. NGOs, private firms, and municipal agencies involved in social intermediation need to play important roles in mobilizing and promoting associations of poor households. 24. Residents savings programs for housing have been used successfully in Brazil and Mexico and can be expanded. Once these programs are established, savings can be tied to incentives programs sponsored by outside sources and to community development efforts. Benefits to participants can be in the form of construction materials, labor, or down payment for land purchase. Individual savings can be pooled into a legal trust, which provides a mechanism to attract additional financing, to protect against default, and to receive and manage subsidies. Building materials can be purchased locally. The government and bilateral organizations can provide technical assistance. There are cases when states provide land under loan agreements, by which residents earning less than one minimum wage pay 5 percent of their wages over six years. Residents build progressively and thus add value to their plots over time. The funds are used for land acquisition, and for housing and infrastructure construction. 25. Microcredit for housing is another vehicle for enhancing the financial capacity of poor households to improve their shelter condition. As demonstrated in many countries (see Chapter 4), microcredit is a promising financing mechanism that adapts well to the characteristics of the informal sector. Microloans are typically small (US$500 to US$5,000 or R$1,200 to R$12,000) and short-term (2-5 years). Moreover, they can fund steps in the progressive process (land acquisition, a sanitary core, etc.) at affordable terms because of the small size of the loan. Housing microfinance programs have followed two distinct approaches: (a) microcredit to finance housing, which started with microcredit provision for small entrepreneurs and expanded into housing credit to respond to the housing demand of clients; and (b) shelter advocacy for housing finance programs, which arose out of advocacy agenda that defend the right of the poor to land and shelter, as well as to adequate infrastructure and services. In some sense, the approaches are complementary, as microcredit to finance housing focuses on the purchase of housing, while shelter advocacy also involves obtaining access to land and infrastructure. 26. In Brazil, the few microcredit schemes for housing that exist have emerged from the advocacy role of NGOs which organize communities to pressure public authorities for tenure regularization and services. Examples are Casa Melhor and Programa de Auto Assiste^ncia na * Casa (PAAC) in Fortaleza. These two housing improvement loan programs promote the formation of community-based savings and loans associations, which qualify for matching funds from the government. 27. For purely housing finance, there is no housing microcredit institution in Brazil. Not only is the Brazilian microcredit industry undeveloped compared with other countries; existing microcredit institutions are legally prohibited from branching into housing. In 1999 the Central Bank created a new type of institution, a microenterprise credit society intended to provide NGOs with the formal means of organizing themselves with the Bank's oversight. These institutions, Brazil: Progressive Low Income Housing Page ix however, cannot offer consumer loans or housing finance. Given the large demand for housing improvement, it is important that this restriction be lifted and that microcredit for housing, land purchase, or infrastructure services be extended. In Bolivia, Chile, El Salvador, and other Latin American countries, examples of successful microfinance institutions for housing exist. 2) Linkinz Informal Demand to National Housing Finance 28. While individual savings can be pooled to form a collective savings schemes to be run by residents associations, maximum impact can be achieved if these savings schemes were linked with the national housing finance system. For families unable to access existing subsidized credit, direct demand subsidies need to be considered. For example, in Mexico, the program FONHAPO is targeted at families making less than two minimum wages. The program is funded by the public sector (federal and state governments) and finances the purchase of construction materials and land to enable poor families to build their own houses at affordable prices. Other schemes include leveraging community savings to attract additional private sector financing (along the line of the Curitiba experience) or revolving funds that allow state agencies to acquire land for resale to civic associations on the condition that these associations provide for land regularization and basic services. Such funds are generally mobilized from transaction fees or property taxes. 29. In Brazil, existing community savings programs should be expanded and linked with the ongoing official programs, especially those run by CEF. These funds can be used to finance pilot programs around the country. One such pilot could be to encourage private developers to organize low-income housing projects by allowing them to conform to less restrictive construction standards. Another would be to adopt the "SOFOLE" approach to transform residents savings associations into mortgage institutions. SOFOLEs are small and flexible private housing mortgage institutions in Mexico' that have been very successful in responding to local housing needs. Such projects can be adopted in the infornal sector but this would require a two- pronged approach: on the supply side public agencies can guarantee the regularization of land sites in the short term, while on the demand side a combination of instruments (e.g., a SOFOL- type institution) can promote a mutual assistance savings program to finance progressive or incremental housing projects. 3) Improving the Stock 30. NGOs. To improve the hundreds of thousands of self-constructed units in the informal sector, Brazil could establish small funds. The Cooperative Housing Foundation, an NGO in Ciudad Juarez which established a small fund of US$50,000 to finance small improvements, is a good example of how this could work. CHF manages the program and selects its participants. It grants loans that are interest-bearing and disburses direct payments for labor, materials, or improvements. Loans are made on the basis on a person's employment history. Participants are not required to hold legal title to their land or to live in a regularized subdivision. The default rate is very low because of the participants' desire to keep their creditworthiness and because of good program management. The use of NGOs makes housing loan operations simpler and allows funds to go directly to house improvement without passing through any official agency. Some industries are trying similar programs to help their workers. I For example, SuCasita, a SOFOLE in Mexico City, has obtained house construction financing for a group of 400 families who had collectively purchased a parcel of land payable over a four-year peiiod. With the land secured, SuCasita set up a special reserve fund for families who could not meet the minimum salary requirements, and whose income varies from month to month. In this manner, SuCasita could establish a credit history and qualify the borrowers. The reserve fund acts as insurance against default. Page x Brazil: Progressive Low Income Housing 31. Linking to the industrial sector. The industrial sector can be tapped to contribute to the provision or improvement of housing for workers. Employers assist their workers by means of: (a) rental units subsidized by the firm; (b) rent-to-own schemes, whereby workers pay rent during a fixed period of time, after which they are given the right to purchase the unit at a discounted price; (c) home improvement loans guaranteed by the companies themselves; and (d) subsidized down payment. 32. Materials banks. Materials banks play a key role in improving access to and affordability of housing for the poor. Construction materials can be locally purchased, and bulk purchases can be made directly from manufacturers and distributors. Materials banks provide interesting opportunities to trade work for materials and are also a vehicle for business and technical training. To start operation, materials banks require some financial subsidy and technical support. A typical financial mix is 50 percent from community savings, 30 percent from the government, and 20 percent from international funds. Fortaleza has successfully used materials banks as part of its program of mutually assisted housing construction.'2 In Pereira, Colombia, materials banks are part of the state-sponsored housing fund for low-income households through which materials are manufactured, built, and sold, and profits are plowed back into the community in the form of service buildings. 4) Land Management and Increasing the Supplv of Land 33. Land management is a key component of low-income housing policies. The price and supply of land influence the quality of housing, living conditions, affordability of housing services, and economic and infrastructure development. In Brazil the cost of poor land management is very high. The percentage of households and housing units in the informal sector is about 50 percent of the total units in some metropolitan areas such as Belem, Recife, and Salvador.'3 This is partly due to urban regulations that are not responsive to the socio-economic conditions of low-income households, and the lack of programs to improve access to land and housing by the poor. This has led to the development of urban settlements which do not conform to urban regulations and plans, and where households have no land title or ownership. 34. The case of Brasilia illustrates the negative impact that urban regulations can have in the supply of land. Since Brasilia was created in 1960, there have been six urban plans. Most of them had to be revised because of their rigidity and their attempt to control urban growth. While half of the land in the Distrito Federal is classified as environmentally protected areas no provisions are made to accommodate urban growth. As a result, land prices have increased and informal urban settlements accommodating mainly middle-income households have proliferated. By contrast, the urban plan of Curitiba of 1965 which is still in place is a good example of how to orient urban development with a flexible approach that responds positively to urban growth. Curitiba has adopted specific tools of urban management which allow the private sector to subdivide reserved areas into urbanized plots, leaving out environmentally sensitive areas with the objective of preserving them. 35. Land supply instruments. The supply of land in Brazil is largely regulated by two instruments: the Federal Law 6766/1979 and the municipal Planos Diretores. Federal Law 6766 12 In Brazil, some materials banks have been short-lived because of policy changes from one administration to the next (e.g., Recife and Santos in the mid-1990s). It would have been preferable if they are run by NGOs. CHF, however, prefers to reach agreement with local suppliers who agree to give discounts to CHF customers, who pay with CHF vouchers that are later redeemed by the stores. This does not undercut the local market. On the contrary, it enlists the local suppliers as allies in the home improvement scheme, which is a win-win situation. The role of local materials suppliers is key to the whole self-help construction process, since they often provide short-term financing and tend to operate for a long time. 13 IPEA 1998. Brazil: Progressive Low Income Housing Page xi was superceded in 1999 by Law 9785. The latter introduced changes to laws governing land expropriation and registration, and the subdivision of urban land. This law has significant implications on the availability of land for the poor. Some of its main elements are: * Land subdivision for urban purposes is authorized only in urban zones. No rural zone can be parceled for urban use. * Developers have to provide full infrastructure before selling any plot. * Construction on flood-prone areas or on steep slopes is prohibited. * Up to 35 percent of urban land must be reserved for urban amenities and infrastructure, including parking and other public spaces. * An urban lot must have a minimum area of 125 square meters, and a frontage of at least five meters. 36. The municipal Plano Diretor. According to Brazil's Constitution, the use of land - its planning, zoning, subdivision, and / or development - is the responsibility of the municipal government. Municipalities with more than 20,000 inhabitants need to have a Plano Diretor, which guides urban expansion and guarantees that urban property serves a "social function." The Plano Diretor is by definition the most important urban instrument in Brazil. It determines the forms of land subdivision and occupation, environmental zoning, annual and multiyear planning and budgeting, and participatory plan and program development and management. 37. Brazil's new Constitution. The role of the municipalities as main agent of urban policy and the use of the city for "social purposes" were two key concepts included in the new Constitution of 1988. The Estatuto da Cidade recently approved in Brazil (12 years after its inclusion in the Constitution) has several important provisions, notably: (a) the use of the usucapido, whereby informal settlers gain full ownership of the occupied land if such occupation is not contested; and (b) the use of a progressive urban property tax, which is intended to discourage the maintenance of idle land for speculation. 38. New tools of urban policy. The new Constitution has also renewed interest in the preparation of the municipal Planos Diretores, especially after 1988-89 when many newly elected mayors were interested in the urban reform and in the democratic evolution of their government. Since then, several municipalities have experimented with new urban tools -- new forms of zoning and serviced land, progressive land taxation, transfer of building rights without transferring ownership, environmental protection zones, and control of land parceling. In addition, the participation of the civil society in the planning and management of the city has clearly increased. The use of cadastres, and the procedures for land and deed registration need serious reforms as the present system is biased against the poor residents, given its high cost and excessive complexity. 5) Infrastructure and Public Services 39. The provision of infrastructure in slums and squatter settlements can be an effective approach to enabling the poor to house themselves. However, investment in infrastructure needs to be gradual and phased with the regularization of tenure to avoid displacement of the low- income population due to gentrification. As land values usually rise after upgrading (piped water connection can double the value of land), the sale of land by poor households is likely to take place (even if illegally). 40. Fiscal, legal, and administrative constraints prevent the conventional infrastructure supply system from servicing much of the irregular housing sector. These constraints render serviced plots more costly and time-consuming to produce. Emphasis should be given to opening Paae xii Brazil: Progressive Low Income Housing up land for residential development, with minimal core infrastructure at the outset. Infrastructure and services, like housing, can be built in a progressive manner, making land more affordable to the poor and providing some form of self-screening as only those who prefer partially serviced land would apply. National low-income programs in Brazil should therefore adopt the notion of progressive infrastructure and service provision. Constraints on land supply should also be removed. Such changes would relieve pressures from land speculation and encourage private sector involvement. 41. The progressive approach to infrastructure provision, however, requires a high level of technical assistance to households and community groups. The adoption of intermediate technology could lower costs and could increase the number of serviced plots entering the market. Some of these systems could be operated fully or partially by private companies under public franchise. Within a framework of progressive provision, both the central and local governments play a key role. The central government could finance major infrastructure investments, and also encourage arrangements and partnerships between municipal and state governments. Local assistance centers could be set up by the state or municipal governments to help address the concerns of informal households through the different phases. These centers could be responsible for matters involving production technology, fiscal management, as well as the establishment of savings programs. 6) Institutional Arranerements 42. There is a vacuum of responsibility for housing and urban policies at most government levels in Brazil. While the Estatuto da Cidade brought some rigor to the use of the urban policy instruments available to cities, there is very little intergovernmental collaboration and division of labor among the government levels. It is clear that there is an urgent need to re- institute housing and urban authority at the central level. The fact that the Estatuto da Cidade took 12 years to get through Congress shows the low priority accorded by the legislative and executive to urban and housing problems in Brazil. 43. There is an urgent need for the central government to formulate a sustainable and equitable low-income housing policy, to provide guidance to the subnational governments, and to mobilize resources to support the policy. Specifically the central government must: (a) evaluate the national needs for housing improvement and basic shelter; (b) select and adopt models that have proved successful in helping the poor to improve their housing conditions; (c) estimate the financial implications of such programs; (d) decide on the allocation mechanisms (through the states, directly to municipalities, through a national urban council); (e) identify financial mechanisms for mobilizing resources, including the use of grants from the central government, local taxes, residents savings, and microcredit; and (f) prepare targeting and self-screening procedures to ensure that the funds would have the best use possible. 44. A new institutional framework is clearly needed in Brazil to guide housing policy. The framework must: a) oversee the performance of the low-income housing sector as a whole and coordinate the activities of the major public agencies (banks, land registry and titling, public works, planning, finance, etc.); b) collect, analyze, and publish data on the performance of the housing sector, including such information as construction costs incurred, public contracts executed, outcomes of programs especially aimed at the poor, among others, that would constitute a low-income housing data bank; Brazil: Progressive Low Income Housing Page xiii c) provide a forum for the participation of the private sector, NGOs, community-based organizations, and the general public to discuss housing policy at both the national and the municipal levels; d) review the effects of housing regulations, and initiate regulatory reforms whenever necessary; e) establish a regulatory framework for the delivery of housing finance by the private sector, including microcredit for housing purposes; f) introduce and assess the outcome of targeted up-front housing subsidies for the poor within a comprehensive scheme of collective savings and affordable mortgages; g) establish property rights, especially through regularizing tenure in squatter settlements and informal land subdivisions; h) adopt a progressive approach to infrastructure provision in low-income settlements; and i) bring together infrastructure agencies to coordinate infrastructure provision and help create an adequate supply of serviced land. F. Conclusions 45. Improving and facilitating the functioning of progressive informal housing markets does require major policy changes, as outlined before. There are however some positive points. The markets markets (including the informal markets) are working and manage to provide a large number of housing units, albeit underserviced and minimally constructed. Most occupants pay something for land and for shelter. An important requirement of the informal sector therefore is a more direct partnership with the government to support a strategy that promotes a variety of savings methods, flexible and progressive infrastructure delivery systems, regulatory reform, employment generation, and community involvement. 46. The proposed strategy should be based on several main principles: * First, housing (e.g. basic shelter) should be recognized as a fundamental service for the Brazilian poor and as an effective instrument to improve their assets and social inclusion. Housing has a significant impact on the health, welfare, and self-esteem of poor households, and offers them economic opportunities as it is often used as a workplace. * Second, housing must be an integral part of a national urban policv formulated and approved by the government, with the involvement of communities. The recently approved Estatuto da Cidade provides an excellent framework for defining the roles of different levels of government and creating a comprehensive policy response to effectively address the housing needs of the poorest. Vertical coordination is required, for which a National Housing Policy Council can be constituted to oversee policy implementation and preside over a national system for low-income housing management. * Through a participatory process, local governments will need to identify the housing priorities of different income groups, as well as their repayment capacity and savings potential. Small and well-targeted up-front subsidies can be designed and adopted at the municipal level to leverage private savings. Such a mechanism can be more effective than below market mortgage loans to address the needs of lower income groups. * Clear financing for housing from all levels of government should be established, with larger contribution from the central level to be used in conjunction with locally available Page xiv Brazil: Progressive Low Income Housing resources (including private savings) and affordable loans. A system of transfers to poor municipalities- using transparent criteria for targeting - should be designed, adopted, and monitored to determine their effectiveness. * Technical assistance is crucial for municipalities and for the private sector, both formal and informal (private developers, NGOs and residents associations) to support the progressive and informal housing processes of the poor. Organizing the demand is key to ensuring the best use of scarce public resources for the low-income housing sector. * Finally, the collection and disclosure of accurate housing information, including data on housing costs, affordability, financing, production, quality and other performance indicators, would enable a wider audience of practitioners and public officials to learn from the successes as well as failures of housing policies and programs in Brazil. 47. A Financing Scenario. Brazil's housing strategy can be translated into a comprehensive low-income housing program. However, estimating the financial requirements of such a program is not easy given the lack of detailed data, particularly on the specific needs of each group of households. In building a scenario for financing a 10-year low-income housing program, it is estimated that about 3 million households would require assistance in improving their shelter conditions. Assuming the unit cost of producing a low-income housing unit to be R$7,200 - roughly the unit cost incurred in upgrading projects - the total financial program requirement would be more than R$21 billion. This estimate is twice what the amount the three levels of government spent on housing and urban development in 1997 and 1998, apart from the R$11 billion lent by FGTS every year as subsidized housing credit.'4 48. At present, Brazil's public sector expenditures on housing are very low. During the last five years, housing and urban development expenditures averaged R$12 billion per year, or 1 percent of GDP. Municipal governments financed the bulk of housing investments, i.e., about 70 percent; state governments contributed 20 percent; and the federal government, only 10 percent. In 1998 housing accounted for only 1 percent of federal social expenditures, down from 1.2 percent in 1996. 49. To make significant progress (in line with the above estimates), about R$2 billion a year would be mobilized for low-income housing over a 10-year period, forty percent of which would be spent on housing construction, 20 percent on land, and 40 percent on infrastructure. (Depending on the region, infrastructure and land may consume up to 80 percent of the total cost of the program). A simple scheme would call for municipalities to contribute the land, and states and the central government to finance the infrastructure on a shared basis; the 40 percent for construction would come from residential savings and microcredit leveraged by very small up- front capital subsidies, if needed. 14 Reorienting FGTS lending from middle-income families to the very poor would raise several difficulties. Lending to the very poor requires an administrative structure that the CEF does not have at this stage, and the risk is higher. CEF would be reluctant to shift into riskier business without adequate guarantee. Funding of FGTS comes from compulsory savings (8% of workers' wages) from the formal sector. It could be argued that those funds should be allocated to the contributing population (formal workers) rather than to low- and very low-income households who are not contributors to FGTS, the latter being mostly informal workers. In this context, one would argue that, on the basis of equity, aid to the very poor would need to come from the public sector, and that cross-subsidization through FGTS funds could not be justified in economic terms. Brazil: Progressive Low Income Housing Page xv Illustrative Scenario for the Financing of a 10-Year Low-Income Housing Program Annual Values (R$ Million) Source of Financing Purpose Total Annual Budget Annual Extra with Housing Amount Needed Federal Government Infrastructure/ Some Up-front 1,100 400 Capital Subsidies State Govemments Infrastructure Development 2,400 400 Municipalities Land Supply 8,600 400 Residents and Credit House Construction FGTS: 1,500 800 Or Microcredit: 8.2 Individual Contribution $2,500 per household 800 Total 2,000 Notes: Assuming that the total cost of housing is $7,200, and that the share of financing is 20% central government, 20% states, 20% municipalities, and 40% credit and resident savings. For a total of 3 million households, a 10-year program would attend 300,000 per year. The cost to be borne by each household would be $2,500, which seems compatible with the income level of the household. 50. A scenario as the one above would include the following items:: a) The federal government would increase its role in housing from policy making and regulatory view point but also providing incentives probably in the form of matching grants to encourage progressive housing and land supply policies to the poor. The outlined scenario would require a 40 percent addition to the housing national budget, bringing it to R$1.4 billion, or 1.4 percent of the government's social expenditures. Moreover, the federal government would assume an aggressive role to facilitate shelter improvements and infrastructure provision for the very poor. Given that the markets for different income levels are linked, the overall housing sector would be organized such that gentrification would not occur once informal settlements are improved. Moreover, a system of targeted up-front subsidies for the very poor should be designed to promote residents and collective savings, as well as progressive housing development. .b) The states' responsibility in housing would be one of coordination and intermediation. States have a key role to play in gathering and disseminating housing information, including best practices, to serve as a base for sound decision-making. They could provide municipal associations with technical support and kits for housing improvement, as well as facilitate new land transactions and registration. In financial terms, states would mobilize an additional 20 percent of what they are currently allocating to housing (R$400 million compared with the R$320 million they spent in 1997-98). c) Municipal govemments would need a comprehensive package of technical assistance. freedom to use the instruments created by the Constitution, incentives to promote popular participation, authority to use flexible systems of expropriation and usocapido to improve land titling and the liquidity of local land markets, and flexible standards for housing construction, land use, and infrastructure provision. In financial terms, municipal consolidated expenditure would increase about 5 percent. This seems to be realistic and possible, especially if progressive property taxation could be deployed. Additional resources from municipalities could come from: (a) the regularization of land titling that Page xvi Brazil: Progressive Low Income Housing would raise tax revenues; (b) sale of serviced land (part of the serviced land could be sold to the private sector for commercial development); (c) permit fees for densification, such as the one used in Parang; and (d) materials banks. d) Residents would contribute directly with savings (in kind or in cash) and with borrowed resources. Given the small size of the average loan (R$2,400), poor households would be able to afford down payment and loan repayment at commercial terms. e) Financial intermediation would be key element of this plan. The experiences of Caixa with the Carta de Credito Associativo and those by NGOs with microcredit provision can be guides for scaling up successful initiatives. 51. In terms of commitment, the above plan seems to be a tall order. Indeed, it will require some changes in law. In terms of financing, however, the amount is small in relation to Brazil's GDP as well as its overall social expenditures. The main challenge is faced at the level of policy reforms and institutional coordination as explained before. And for most people in Brazil, the need is real and pressing. Brazil: Progressive Low Income Housing Pave 1 1. Introduction A. Purpose of the Report and Major Issues - 1.1 Objective. This report aims to analyze key aspects of the low-income housing sector in Brazil. and to provide an analytical framework for reviewing alternatives to addressing the lack of adequate formal housing and urban services for the poor. It addresses four fundamental questions for policymakers in the housing sector in Brazil: First should the government be involved in policy interventions in the low-income segment of the housing market? Second, if policy action is appropriate, what roles should policymakers at the federal, state, and municipal levels play? Third, should the government's key interventions in the market be focused on the supply or demand side of the market? Finally, how can the government structure a comprehensive policy to deal with the failure of the housing market to provide adequate shelter for the poor? The report begins with a brief outline of the nature and extent of the low-income housing problem, describes the policies in place to address the lack of housing, suggests a methodological framework for assessing the paucity of low-income housing in light of international experiences, and identifies the main building blocks of a strategy for low-income housing in Brazil. 1.2 CAS Link. The report fits into the main objectives pursued by the World Bank in Brazil: poverty reduction, removing obstacles to sustainable growth and promoting more effective public governance. Housing and adequate shelter is not only a key part of the essential services needed by the poor but is also a source of asset accumulation and productive capital. About one-fifth of housing units occupied by the poor are used as places of employment. Housing investment accounts for two percent of GNP and the flow of housing services for an additional five percent; thirty percent of household income is spent on housing. The World Bank has financed several low-income housing initiatives in Brazil, most of them as part of wider programs. Examples are the environmental management projects (e.g., in Guarapiranga), municipal development programs (Somma, Ceara, Bahia), flood management (Rio de Janeiro), and several other initiatives. Lessons from these projects have been important in shaping the most recent programs being prepared in Brazil. 1.3 The Challenge. Official estimates indicate that in Brazil there are 5.4 million households living under less than adequate housing conditions. These conditions include homelessness, houses made out of substandard materials, more than two families living in the same house, and lack of access to water and sewerage. Of these households, 65 percent earn less than 3 minimum wages; 46 percent live in the Northeast. Every year, an additional one million households enter the market for housing looking for affordable shelter. It is estimated that 20 percent of these new entrants (earning more than 10 minimum wages) will purchase their own homes through savings and mortgage markets, another 20 percent will qualify for some subsidized credit scheme, and the remaining 60 percent will not be able to afford the down payment and monthly installments associated with the purchase of a finished house in Brazil. Many of them will turn to the informal housing sector. 1.4 The Reasons. There are four main reasons why a considerable backlog in housing persists: (a) low-income levels; (b) high supply costs; (c) market failures; and (d) distortions in public policy. (a) Poverty: Sixty percent of the households entering the housing market every year in Brazil earn less than 10 minimum salaries. Assuming that households can allocate at Page 2 Brazil: Progressive Low Income Housing the most 30 percent of their net income for housing, one third of the newcomers will be able to afford a housing unit that costs around R$12,000. The remaining households would need to find much cheaper houses. (b) High supply costs: The costs of formal housing in Brazil are high. A single family unit of about 35 square meters costs rarely below R$ 18,000; the average cost is around R$20,000 (about the same price level found in most Latin American countries). Were the R$18,000-unit to be financed by the mortgage market, the monthly interest payment would be about R$200. This is accessible to households earning at least 4 minimum wages, the group that accounts for less than half of the Brazilian population. (c) Market failures: The present regulations governing land use, housing, and service provision, along with the monopolies in land ownership and construction industry, keep housing prices from falling to levels that could be accessible to poor families. The lowest price for a house constructed in a special zone for low-income housing is about R$7,000 or half the price of a housing unit in the formal market. (d) Public policies and expenditures: Since the collapse of the National Housing Bank (BNH), the Brazilian public sector has not made any significant housing policy intervention nor has it promoted a coherent low-income housing strategy. This apparent neglect has had several consequences: * There is a lack of coordination of housing policies across government levels. * The federal government has taken a very low profile in dealing with policy issues involving land use and building standards. * The federal contribution for low-income housing finance has been minimal. Housing represents less than 1 percent of the federal social expenditures. By comparison, expenditures on education and health have accounted for 15 percent and 21 percent, respectively, of federal social spending. * Public expenditures oriented to housing finance mostly benefit middle-income families. Between 1995 and 2000, the central government program for housing added R$12 billion to its budget: 10 percent of this amount was financed by the central budget, and the remaining 90 percent from the workers' severance fund (the Fundo de Garantia do Tempo de Servico or FGTS) in the form of subsidized credit at below market rates. The latter benefited households earning more than 5 minimum wages. * Finally, there are very few programs targeted at the very poor. Those that exist do not have the potential for expansion to be able to reach a significant proportion of these very poor households. Table 1. Housing and Urban Expenditures, 1997 (at 1999 constant prices, R$ million) Gov't Level Total Per Structure Share of Housing Expenditures Capita in % Expenditures 1997 (R$) In Total Budget Federal 1147 7.2 10 1.0 States 2043 16.1 21 2.3 Municipal 8345 52.4 69 20.7 Total 11536 75.7 100 -- Source: IBRD, IPEA, 2001. Brazil: Progressive Low Income Housing Page 3 1.5 In February 2000, Brazil's Congress approved a constitutional amendment guaranteeing the right of all Brazilians to decent housing. To accomplish this objective, the above issues need to take priority in the government's policy agenda. This report is intended to contribute to the discussion of these issues and to identify the necessary actions to resolve them. B. Content and Methodology 1.6 The report is organized in five chapters. Chapter I is an introduction to the housing situation in Brazil and defines the methodological approach for analyzing policy options for the sector. Chapter II describes the overall picture of low-income housing in Brazil, the existing and projected unmet demand, and the challenges the public sector faces in its attempt to find sustainable and equitable responses to the problem. It outlines an analytical framework for studying the nature of housing as a sector and as a service to the poor, examines alternatives for public sector intervention, and establishes some principles to guide policymaking. Chapter Im reviews federal public housing programs and policies, their merits and weaknesses. Chapter IV examines best practices in low-income housing at the subnational level, and tries to capture the elements underpinning their success that can be used in Brazil. Chapter V evaluates the main elements of a sustainable housing strategy for Brazil from the perspectives of both housing demand and supply, and concludes the analysis with recommendations for a coherent low-income housing public policy and program. The Annexes include a Review of Literature on Urban Low- Income Housing (Annex A) and a Statistical Appendix (Annex B). 1.7 The report draws from previous work undertaken in the Bank, notably from the background papers on Finished and Unfinished Housing,' Attacking Urban Poverty in Brazil,2 as well as from research and analytical work developed by IPEA and SEDU. The methodology used is a comparative analysis of programs used by developed and developing countries to address their low-income housing problems. Selected cases in Brazil and elsewhere are used as case studies. C. Analytical Framework 1.8 The present section reviews the analytical tools used in this study and outlines the methodological approach to analyzing housing markets. It also discusses why public intervention in low-income housing is necessary and what type of interventions is appropriate. 1.9 Characteristics. Housing is unique among consumer goods. First, housing is potentially very durable, with a useful life of around 70 years. Second, housing takes up a large share of any household's budget. In OECD countries, housing services, including imputed rent of owner-occupied households and utilities, account for 25 percent of personal consumption expenditures. In the USA and in other industrialized countries, affordable housing is generally defined as a housing cost at or below 30 percent of income.3 In less developed countries, this value can be much lower because food and clothing, by necessity, must absorb a larger portion of a low-income household's budget. In Mexico, expenditures for housing and maintenance are between 7 percent and 11 percent for the bottom half of the income-earning population; in Brazil, it is about 26 percent. IAlberdi and Serra 1999. 2 Joachim von Amsberg 2000. The ratio is a longer-term measure since a household's income can fluctuate; hence at some point a household may spend more than 30 percent of its income on housing. Page 4 Brazil: Progressive Low Income Housing 1.10 The social problem underlying housing affordability arises when low-income households, over the long run, must spend a high percentage of their income to obtain a decent housing. Local governments play a role in determining the ratio of housing costs to income for low-income households. The nature and extent of the affordability issue are defined by the standard for decent housing used. Most municipal governments have zoning and building codes that impose certain minimum requirements on the housing that can be offered for sale or rent. Often, however, the minimum housing standards as set by law are still high and housing inevitably becomes too costly for many low-income households. Although some prefers lower quality housing, such choices are often discouraged by municipal governments out of concern for public health and safety, and the overall quality of local neighborhoods. 1.11 Key Concepts. Housing markets are studied using two main concepts: filtering and hedonic demand models. Filtering relates to the segmentation of the housing markets. Low- and high-income housing markets behave separately but do interact. If there is an increase in the housing supply in the middle- and high-income markets, more housing will be available to low- income groups. Eventually, low-income units may not be used or needed, in which case developers may demolish them. In the absence of the desired filtering-down effect found in well- functioning housing markets, government subsidized housing schemes oriented for the poor are easily appropriated by middle-income groups (directly or indirectly through sales by the low- income beneficiaries) thereby creating a perverse filtering-up effect. The dynamics of housing markets and housing policies therefore needs to be considered in its entirety. Responses to the housing problems of low-income groups will not be sustainable unless the housing needs of the immediate income groups are also met. 1.12 The second concept is based on the nature of housing as a bundle of services. Unbundling housing into such characteristics as location, quality of construction, number of rooms, distance from church, etc., is a useful device to estimate income and demand elasticity, and even to determine the potential value of a house for tax purposes. Demand for housing can then be defined as the sum of demand for each of the characteristics (hedonic model). The concept of housing as a bundle of goods is also useful when tackling issues related to progressive informal housing. 1.13 Public Intervention in Housing Markets. The housing market is probably subject to more policy initiatives than any other market for consumer goods. In general the central government plays a major role, but local governments are also actively involved in administering federal housing programs as well as their own. Land and housing regulations are often the responsibility of local governments. 1.14 The rationale for public sector intervention rests on the fact that housing markets are subject to market failures and externalities: * The housing market generates both positive and negative extemalities in neighborhoods. Empty houses depress the market while well-maintained houses improve it. * In many countries, racial and wealth discrimination prevents minorities from owning a house regardless of whether or not they can afford it. * Housing is a merit good a good that society views as more important than what would be consumed if left totally up to individual choices.4 Providing adequate housing for the poor is an important policy for poverty alleviation in several ways: House ownership improves the opportunity of the poor to increase their assets, provides them with seed 4A merit good is a good that creates benefits for the community as a whole. Housing, health, recreation, and education are often mentioned as being in this category. Brazil: Progressive Low Income Housing Page 5 capital for future improvements, improves their health and living conditions, and contributes to enhancing their sense of social inclusion. * Housing has a strong impact on human productivity and welfare -- it stimulates saving habits, provides a base for economic activity, and improves one's sense of citizenship and participation. For most people, a housing unit will be the single most valuable asset they will own in their lifetime. This creates wealth that can be passed on to subsequent generations. * Housing and building codes that establish minimum standards for housing units cause housing prices and rents to exceed the amount that some households would choose to pay of their own accord. Housing subsidies (or another form of intervention) are often needed to compensate the low-income households for being forced to allocate a larger portion of their income for housing. 1.15 How to Intervene. Public intervention in low-income housing can take place in both the supply and demand sides of the housing market. Supply-side interventions can be in the form of direct provision of housing units and/or some intervention in the market to ease such constraints as monopoly control of housing construction or land availability. Demand-side interventions may involve some form of subsidy to compensate for housing regulations which often render housing unaffordable to low-income groups. In favoring subsidies to influence demand, most economists argue that housing provided directly or indirectly by the public sector is often more expensive than private sector-supplied housing as evidenced by several experiences worldwide. It is further argued that demand-side subsidies are more equitable (since the same amount of money when used more efficiently can serve more households) and allow households freedom to choose the type of housing and location they prefer. D. Public Housing - Policy Options 1.16 Public policy on housing is generally divided into two types: (a) policies designed to enable middle-class households to own homes, and (b) policies aimed at providing (or enabling the gradual ownership of) housing for lower income households. Middle-income housing policies are used all over the world, applying a variety of savings schemes and taking full advantage of the depth of financial markets. For example, in the USA, middle-class Americans are encouraged to own their homes through two long-standing policies: federal mortgage insurance and federal income tax deductions for mortgage interest payments and local property tax (Box 1). 1.17 Housing Policiesfor Lower Income Households have been developed in response to the constitutional mandate vigorously pursued in many countries "to provide a decent, safe, and sanitary living environment for everyone." This objective has gained strength as a result of UNHCS (Habitat) policies and has led to a multiplicity of national as well as and internationally supported programs across countries. Three types of low-income housing programs have evolved over time: public housing, sites and services, and "enabled" programs (see Annex A). 1.18 'Official' approach: Public housing. In the 1950s and early 1960s, most developing countries faced tremendous urban population growth caused by rural-urban migration. Lack of affordable housing in the formal market resulted in the proliferation of slums and squatter settlements that flouted existing regulation and building codes. These illegal settlements were often demolished and replaced with new, modern buildings as part of larger programs by governments to construct 'low-cost" public housing projects for direct sale or for rent. While intended to meet the overwhelming demand for housing, public housing had serious flaws. First, because of high standards of construction and service provision, they became far too expensive for their intended beneficiaries. Second, the lack of resources prompted governments to locate Page 6 Brazil: Progressive Low Income Housing public housing on less expensive land farther from the city center. Not only were beneficiaries not able to maintain the houses and pay for the added transport costs; they eventually sold their houses and moved into slums closer to places of work. 1.19 Self-help housing. By the 1970s, governments adopted a twofold approach to the low- income housing problem -- sites and services and slum upgrading. Sites and services involved the provision of tracts of vacant land for housing, where basic supporting services were set aside and low-income residents constructed their own dwellings or were provided "core houses." The rationale was to make housing sites affordable to low-income households, and allow the government-assisted self-help projects to fully recover costs and generate a surplus that could later be used for larger scale projects. Slum upgrading focused on improving existing housing sites of the urban poor rather than on developing undeveloped land. Since the late 1970s, most donors adopted the sites and services and slum upgrading approach. The main drawbacks were that the private sector, remained outside of projects and the possibilities for scaling up and expansion were very limited. 1.20 New partnerships: The enabling approach: In 1988, the General Assembly of the United Nations adopted the Global Strategy for Shelter (GSS) for the Year 2000 centered on an "enabling approach" for shelter provision. The GSS urges governments to adopt strategies to improve access to land, finance, infrastructure and services; remove restrictive regulations; introduce realistic building and land use regulations; and strengthen institutional frameworks to foster people's participation in the housing process. It encourages new partnerships that combine resources from national government agencies, municipal authorities, the private sector, NGOs, international funding agencies, voluntary or community sector resources, and individuals/households.5 1.21 Over the past decade a series of innovative, "enabled" approaches have been adopted, including: (a) community mortgage schemes to provide the poor with access to urban land; (b) community funds which offer housing finance with a collective guarantorship; and (c) community construction contracts that enable the poor to build their own infrastructure. Complementing these ideas is the new approach to "enable housing markets to work" which was formalized in the work of Mayo and Angel (1993). The central message is that governments should be the promoters of well-functioning markets rather than providers of housing. Public policies should be oriented to seven operational instruments: three focus on stimulating demand (develop property rights, develop mortgage financing, rationalize subsidies), three on supply (provide infrastructure for residential land development, rationalize regulation of land and housing development, organize the building industry), and one on institutional development (improve the institutional framework for managing the housing sector). 1.22 Interesting examples of this enabling approach include: * The Million Houses Program in Sri Lanka was based on 'minimal intervention but maximum support by the state and maximum involvement of the builder families.' The core element of the program were small housing loans to enable households to build or improve their houses. Several options and packages were available for upgrading or repairing an existing house, for developing a new house, for the provision of utilities, and for the registration of deeds. Between 1984 and 1995 about 2.5 million units were built. * FONHAPO, the National Fund for Popular Housing in Mexico, provided community- based finance and subsidized loans, specifically directed to low-income households. The system was designed according to the ability of the people to repay. The value of loan 5UNCHS 1996. Brazil: Progressive Low Income Housing Page 7 repayments could not exceed 25 percent of the minimum wage. The cost of the units were actually half of what they would have cost for an equivalent sized government-built house, attesting to the ability of private sector developers to provide affordable housing. The program, however, suffered from political interference, which caused it to lose a large part of its appeal and impact. * In the Philippines, the Community Mortgage Program (1988) helped low-income households acquire title to the land that they. occupied illegally while providing incremental investment to improve their homes and add value to their assets. To acquire the loan, residents were required to organize themselves into a community association which was responsible for collecting repayments and for ensuring that the loan continues to be serviced. The land, purchased on behalf of the members, remained under common ownership by the association. * The Urban Community Development Office (UCDO) in Thailand started in 1992 to strengthen the capacity of the urban poor and those living in illegal settlements to obtain higher and more secure incomes, appropriate housing with secure rights, an improved environment and better living conditions. To be eligible for a loan, each community must operate a savings program, with a clearly defined management structure, for at least three months. Loans could be granted for up to a maximum of ten times the amount saved. By the end of June 1995, loans totaling around US$13.6 million had been approved, benefiting some 140 communities with around 11,500 families. * The Incremental Housing Scheme in Hyderabad, Pakistan, sought to duplicate the success of the informal housing development process. In -the settlements being helped, the site was occupied first and infrastructure services were developed incrementally as the capacity of the inhabitants allowed. The long wait between advance payment and land possession, common to government programs, was eliminated and possession was immediate. Cost recovery on the plot investment, reviewed in 1989, was reasonably good.. The land title was granted only after payment was complete. The development remained entirely self-financed with only administrative overhead paid for out of the central government budget. 1.23 Assessment. Several factors hindered public housing, sites and services, and slum upgrading from reaching the low-income households: * Gentrification. In the early public housing and sites and services projects, units constructed for the poor were appropriated by lower to middle-income families who could better afford them and who also had unmet housing demands. Better infrastructure and services made neighborhoods attractive to higher income households who displaced the initial settlers. In a citywide upgrading program in Lusaka, for example, improvements were evenly spread over all slum areas of the city so that no one area was singled out as a potential middle-income neighborhood, thus avoiding gentrification. (Mathey, 1997). In contrast, in Mexico City and Tunis, one-third of the original beneficiaries were displaced by higher income households.6 6 Ferchiou 1982. Page 8 Brazil: Progressive Low Income Housing Box 1. The USA Experience with Low-Income Housing7 The first federal program started under Housing Act 1937, which created federally funded public housing. Under this program, local governments were authorized to establish public housing authorities that would receive federal funds to build public housing. The federal government set up the eligibility criteria for residents of public housing, who then paid a low rent to the local public housing authority. The program was expanded under the Housing Act of 1949. This program authorized the construction of 800,000 units of public housing. Local urban authorities received federal funds for the purchase of slum properties, demolished the slum building, assembled land into usable parcels, and then sold the land at below market prices to housing developers. Throughout thel960s, the housing policy branched out to encourage the construction of housing for low- income households by private developers. A small program of rent supplements for low-income households was created in 1965. Another program in 1965 permitted public housing authorities to lease existing private housing units and rent them on a subsidized basis to low-income households. In 1968, the Housing Act adopted the goal of constructing six million subsidized units. Another part of the act offered below-market interest rates to eligible home- buyers under the FHA mortgage insurance program.. In 1974, the Housing and Conumunity Development Act included a low-income rental assistance program and had separate elements for new construction, substantial rehabilitation, and existing housing. Under the new construction and rehabilitation, private and nonprofit developers obtained their own FHA- insured financing and the Department of Housing and Urban Development entered into a long-term contract of 20 to 30 years to subsidize the difference between fair market rent and 25 percent (30 percent after 1981) of the tenant's income. The program of project-based rental subsidies was eliminated in 1983 because of program cost. According to the Housing Department the annual subsidy was over $6,000. By the 1990s US housing programs evolved into three principal policy choices that prevail today, and which combine demand- and supply-side subsidies. At the core of the US policy are portable housing vouchers which are financed by the federal govemment, but administered by local housing authorities. Under these programs, eligible households receive vouchers which, in theory, allow them to search for rental housing anywhere in the local market. The household pays no more than 30 percent of household income in rent and associated housing costs, with the federal govemment subsidizing the rest up to an established "fair market rent" (FMR). In some cases, households rent housing units that exceed established FMR, but must pay the difference, increasing the household's out-of-pocket expenses to greater than 30 percent of income. Federal block grants play another important role in the Federal Housing Policy. The HOME Investment Partnerships program, modeled on the successful Community Development Block Grant (CDBG) program, transfers funds to state and local authorities which have relative freedom in the application of these housing resources that must be targeted at low- and moderate-income households, 15 percent of which must be earmarked for community-based nonprofit groups. These funds are often used to write down the cost of new housing construction, down payment assistance programs in the case of home ownership, and other locally designed programs. Also a part of the HOME program is HOPEVI, which provides grants to housing authorities for the rehabilitation, reconstruction, and possible privatization of public housing stock. On the supply-side, much new low-income rental housing is produced through the low-income housing tax credit program. In this program, individuals and companies that invest in low-income housing can take a tax credit (dollar-for-dollar offset against other taxes) equal to their investment in 10 annual installments. To qualify for tax credits, properties must rent at least 20 percent of the units to households eaming 50 percent of the area median income or less, or at least 40 percent of their units to households eaming less than 60 percent of median income. The rents charged may not exceed 30 percent of a household's income. From 1990-1998, about 500,000 affordable units have been produced under this program mostly by nonprofit housing development corporations and community-based nonprofits. This program has helped build the capacity of the nonprofit sector and has spawned the creation of specialized national, nonprofit financial syndicators of tax credits and other investment vehicles, most notably the Enterprise Foundation and the Local Initiatives Support Corporation. 7It is important to note that US federal housing programs targeted at the low- and moderate-income households are primarily focused on the rental market. However, other federal policies, notably the mortgage interest rate income tax deduction, provide strong incentives towards home ownership for those with incomes high enough to claim the deduction. In this instance, considerably more subsidies, through the tax system, flow to higher income households than do to low- and moderate-income households. Brazil: Progressive Low Income Housing Page 9 * Poor location. Budgetary limits prompted governments to locate projects on cheap land sites on the urban periphery. Since most jobs were found in city centers, the housing projects added transport cost to the already high cost of housing for the poor. As a result, many recipients either defaulted on payments or sold out to middle-income households who could afford to meet these extra hidden costs. * Public finance for housing and subsidies. Houses built by the government were more expensive than similar houses built by NGOs and private individuals.8 Such costly projects were made affordable only through heavy reliance on subsidies which imposed a heavy financial burden on governments. * The issue of land. Unavailability of land impeded government housing projects from reaching the poor as the choice of sites, often dictated by cost considerations, was ill- suited to their needs. Another difficulty was in acquiring land through expropriation and the long delays in doing so. In many cases governments were not willing to set aside land for housing the poor. * Standards. High standards of design and construction adopted in public housing led to high delivery costs. Sites and services projects attempted to lower these standards and supplied beneficiaries with unfinished "core houses" which they could complete over time. In many of these projects, however, it was the land standards which led to high costs because of overprovision of land. This is particularly serious in central city areas where pressure on land resources is greatest, resulting in the residents subdividing the plots by themselves and renting or selling them to others. The consequence is overcrowding or even slum creation. 1.24 Focusing on the Poor. The "enabling approach" viewed in terms of making the housing market work better and promoting broader participation in housing has become the more popular approach to helping the poorest obtain and/or improve their houses. It focuses on removing the barriers that impede the efforts of low-income households to secure land, housing, and urban services. These barriers exist in both the demand and supply sides of the housing market. According to Simbieda and Moreno (1997), any strategy to deal with low-income housing in a sustainable and efficient way calls for a double approach: (a) convert the informal sector housing needs into effective demand for housing; and (b) increase the supply of land, urban infrastructure, and housing. Community participation and self-help are the key ingredients of this process. The government retains the major role of a facilitator to ensure that financial support for housing, land, and service provision is provided. 1.25 On the demand side the approach calls for ways to ease the financial constraint of the poor in accessing housing services. Alternatives could include: (a) Small savings schemes catering to lower income families to improve the capital base of the families. Such savings could be pooled into community trusts, which in turn could be linked to national housing finance programs; (b) Access to phased loans to help households improve their housing in a progressive, affordable manner; (c) Community-based production and employment, promoting jobs in local construction- related businesses, and facilitating small home-based production enterprises; skejiofor 1999. Page 10 Brazil: Progressive Low Income Housing (d) Small targeted subsidies in the form of up-front capital grants for the very poor who will not be able to make down payments; and (e) Technical assistance in housing (design, construction, and infrastructure installation), finance, and employment activities (e.g., construction, materials banking), drawing on both government and community resources. 1.26 On the supply side the emphasis should be on lowering the costs of housing and serviced land. A common package would include: (a) administrative reforms to facilitate the regularization of land tenure and land transactions; (b) flexible building and land regulations; (c) progressive infrastructure provision whereby public investment focuses on major trunk and secondary infrastructure, leaving to the residents the process of installing connections to their plots; (d) innovations in materials and systems installation to lower costs; (e) incremental improvement of the existing housing stock; (f) negotiating growth management with residents associations in the informal sector; and (g) promoting innovations in financing home and community improvement through technical assistance and guarantees. 1.27 World Bank Experience in the Housing Sector. The Bank's approach to low-income housing has evolved since the first housing sector loans were made in the early 1970s. The first generation of projects - sites and services and slum upgrading - focused on affordable housing and infrastructure standards, land tenure regularization, and reform of the subsidy policies. By and large, these projects were successful and they increased access to shelter by the poor, primarily by reducing the cost of housing through reduced standards for services and housing units. However, for the most part, the projects were small in scale and not easily replicable mainly because of heavy reliance on public finance and inadequate financing from the central government. Subsidies continued to be used in the form of price writedowns for site preparation and housing construction, and the provision of government-owned land at below cost. 1.28 By the 1980s, the second generation of housing projects shifted from project-based lending to housing finance and housing sector development. Its objectives included transparent and targeted subsidies, increased resources for housing finance, and sound financial institutions. While the sector reform objectives of these projects may have been met, the poorest segments of the population were left out. The housing that could be purchased through the housing finance projects was beyond the lower income households' ability to pay. 1.29 The third and current generation of housing projects widened the policy agenda by introducing housing sector reform - securing property rights, further development of housing finance institutions, and better targeted subsidies for the poor. On the supply side, these projects focus on removing institutional constraints and helping reduce the cost of housing investments through infrastructure provision, and improvements in the regulatory environment and the efficiency of the building industry. Institutional reforms and closer link with macroeconomic policy are also part of the new approach. Brazil: Progressive Low Income Housing Page 11 2. Brazil Low-Income Housing - Deficit and Needs A. Context and Background 2.1 Context. Since 1940, Brazil has undergone a dramatic transformation from a largely rural to a predominantly urban country, dominated by two major metropolises and a number of large- and medium-sized cities. In 1940, 70 percent of Brazil's population lived in nrral areas, and the remaining 30 percent in urban areas. In 1970, the urban-rural distribution was about the same. By the turn of the century, Brazil was among the most urbanized countries in the world: 82 percent of the population was living in urban areas (Table 2). This dramatic shift in population from rural to urban areas, which accompanied the country's economic and social development, led to increased demand for housing in urban and peri-urban areas even more than the rapid national population growth would suggest. Table 2. Location of Urban Growth, 1940-90 Year Total Urban Urban North Northeast Southeast South Center- Population popula (%) (rnillion) (million) (million) (mnillion) West (adillion) tion (mnillion) Population 1940 41.2 12.9 31.2 0.4 3.4 7.2 1.6 0.3 1950 51.9 18.8 36.2 0.6 4.7 10.7 2.3 0.4 1960 70.1 31.3 44.7 1.0 7.5 17.5 4.4 1.0 1970 93.1 52.1 55.9 1.6 11.8 29.0 7.3 2.4 1980 119.0 80.4 67.6 3.0 17.6 42.8 11.9 5.1 1990 146.8 111.0 75.6 5.9 25.8 55.2 16.4 7.7 2000 169.5 137.7 81.2 9.0 32.9 65.4 20.3 10.1 Population Growth Rates (%) 1940-50 2.4 3.9 3.7 3.5 4.1 3.9 4.7 1950-60 3.0 5.2 5.1 4.7 5.0 6.5 9.1 1960-70 2.9 5.2 5.4 4.6 5.2 5.3 9.2 1970-80 2.5 4.4 6.4 4.1 4.0 5.0 7.7 1980-90 2.1 3.3 6.9. 3.9 2.6 3.3 4.1 1990-2000 1.4 2.2 4.3 2.5 1.7 2.1 2.8 Source: IPEA, IBGE 2001 2.2 Metropolitan areas have long been the predominant urban areas in Brazil. From 1950 to 1980, these areas grew more rapidly than both the nation as a whole and other urban areas. During the 1980s, however, population growth in a number of urban areas slowed down, particularly in the Sao Paulo and Rio de Janeiro metropolitan areas. This trend toward a more balanced distribution of population growth across a large number of cities may somewhat ease urban housing demand. When growth is concentrated in few areas, it is more difficult to provide sufficient land and urban infrastructure to accommodate such growth, often resulting in high land costs, poor housing conditions and an overburdened urban infrastructure system. 2.3 Demographics and housing demand. Household growth, and hence housing demand, rose sharply in Brazil during the 1990s as those born during the population boom of the 1960s and 1970s entered the peak years of household formation. Given continued rapid growth in the population over age 30, household growth will continue rising through the first decade of the next century, increasing to nearly 1.0 million per year. The most rapid population and household growth rates are in middle-sized cities of between 100,000 and 500, 000 in population. However, much of the country's household growth will continue to occur in the heavily populated states, e.g, Sao Paulo, Rio de Janeiro, and Recife. Page 12 Brazil: Progressive Low Income Housing 2.4 Housing Stock. In response to rising demand for housing, the housing stock had grown rapidly in the 1990s. Between 1961 and 1997, 23.7 million units were built in Brazil. Because of robust construction activities in the past few decades, most houses are fairly new, majority of which were built in the 1970s. The housing stock has more than doubled since then. In the 1990s a large majority of housing units were single-family homes, hence housing development was generally very low density. By region, the Northeast has household sizes averaging 4.2 people, while the South and Southeast have lower densities - about 3.6 people per household. Table 3. Housing Stock by Tenure, 1988-96 (%) Total Fortaleza Recife Rio de Janeiro Sao Paulo Year 1988 1996 1988 1996 1988 1996 1988 1996 1988 1996 Total Housing Units 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Owned 58.8 69.8 62.3 73.6 62.8 76.1 54.5 72.2 57.9 65.6 Not Owned 41.2 30.2 37.7 26.4 37.2 23.9 55.5 37.8 42.1 34.4 Rented 29.3 18.8 28.3 18.0 28.7 15.0 29.7 19.7 32.4 21.8 Other/Borrowed 10.4 8.6 9.2 7.8 8.2 7.0 11.2 7.8 9.3 10.7 Not Specified 1.5 1.8 0,1 0.6 .4 1.9 4.6 0.5 0.4 1.9 Source: Motta 1999. 2.5 Home ownership in Brazil has grown by nearly 10 percent in recent years -- from 59 percent in 1988 to 70 percent in 1996. During this period the most significant increase in home ownership occurred in Recife and in Rio de Janeiro, i.e., by 14 percent and 18 percent respectively. The increased level of home ownership compensated for the decline in the rental housing market. The latter's share of the total housing stock fell from 29 percent in 1988 to 19 percent in 1996 (Table 3). 2.6 Housing conditions.9 Housing conditions, including the size and quality of homes and the availability of urban services, is a critical indicator of a nation's well-being. With rapid Fig. 1: Households. Serviees . ~~~~~~~~growth in households, many of which have R.dI. ~~~~~~~very low incomes, Brazil continues to face M-.'hi.. the formidable challenge of providing decent T.I.phn ... N; EeE I| [_- housing for its citizens. Yet the last few 0 | !2! decades saw significant improvements in housing conditions and the availability of c ....". Ii. _ | " e ! urban services in the country. Between 1980 0 20 40 o I and 1990 the percentage of temporary ___________________________________ housing fell from 4.6 percent to 2.4 percent as wooden shacks in most irregular or informal settlements were replaced by houses made of brick. Density also declined to about 3.8 people per housing unit. 2.7 In addition significant improvements have been achieved in the provision of basic services. In 1970, less than half of Brazilian homes had potable water, electricity and sewerage connection. 9This analysis is based on the diagnostic work produced by Fundag5o Jo5o Pinneiro in 1995. The update produced in 2001 is not easily comparable with the earlier analysis and will not be used in this report. In termns of policy implications, the results are not influenced by the utilization of the 1995 results or the 2001 results. Annex C contains a brief explanation of the new methodology.. Brazil: Progressive Low Income Housing Page 13 Over time steady progress had been made in extending these services to more homes. By 1999 almost 90 percent of homes had access to these services (Table 4 and Figure 1). Table 4. Housing Indicators in 2.8 The housing situation, however, remains Urban Areas, % dismal for an estimated 1.3 million urban households (about 5 million individuals) who still live in Year 1992 1999 substandard housing. The vast majority of these Sewerage_______ 46___0__75___6 households are poor. Moreover, official indicators at Sewerage 46.0 75.6 times mask deficiencies in services provided. In some Potable Water 86.0 91.2 Waste Collection 81.5 93.6 urban areas, for example, water supply is intermittent, Phone connection 23.2 44.2 i.e., households get water only for several hours a day Rent to Income Ratio 15.3 27.4 and / or do not get water everyday. Households with 12.1 35.6 Rent/lncome ratio B. Housing Deficit 2.9 Housing Deficit is a difficult concept to define. It varies with what one considers the minimum housing standards that people need. In Brazil, the most widely accepted definition of housing deficit is that by the Funda,do Jodo Pinheiro (FJP). The concept is viewed as having two dimensions -- quantitative and qualitative. The quantitative deficit is measured by calculating the number of households living in: (a) improvised shelter, close to the notion of homelessness; (b) permanent housing with rudimentary construction materials; or (c) housing units sheltering more than two families. Qualitative deficit encompasses situations when: (a) the rental cost is above 30 percent of the household income; (b) there are too many people living in the same place (at least four people per room); and (c) basic infrastructure is lacking. (FJP has updated her methodology in 2001 which renders difficult comparison with earlier values. The analysis of this report is based on the 1995 analysis. See Annex C for more details). 2.10 Using the above definition, the housing deficit in Brazil is estimated at 5.0 million housing units, that is 15 percent of the total housing stock (Table 5). Of the total deficit, 84 percent are related to households earning 5 minimum wages or less (R$350 a month) (Table 6), and 65 percent to households making 3 minimum wages or less. Table 5. Housing Deficit, Brazil, 1991 Type of Deficit Metropolitan Other Rural Total Metropolitan Other Rural Total regions urban areas regions urban areas areas areas (in thousands) (% of total) Improvised/homeless 33 59 55 147 0.7 1.2 1.1 2.9 Precarious housing 230 434 1063 1728 4.6 8.7 21.3 34.5 Two or more families 980 1621 528 3129 19.6 32.4 10.6 62.5 Total deficit 1244 2114 1646 5003 24.9 42.3 32.9 100.0 Improvised/homeless - situation when people do not have a pernanent residence. Precarious housing - made of precarious materials, such as card board, waste material, etc. Sources: Funda,ao Joao Pinheiro 1995 and M.A.Morais 2000. 2.11 The housing deficit is mainly an urban problem: 70 percent of the households without adequate housing are located in urban areas. Regionally the housing deficit is concentrated in the Page 14 Brazil: Progressive Low Income Housing Northeast (45 percent) and Southeast (33 percent) (Table 7). The problem affects mainly low- income people; in urban areas 85 percent of the housing deficit relate to households earning less than 5 minimum wages per month. Among households living in improvised shelter, 43 percent earn less than 2 minimum wages. Over half of households living in inadequate shelter make less than 2 minimum wages. The most significant indicator of housing deficit is when two or more families live in a small housing unit. Sixty-three percent of the population faced with this problem are in metropolitan areas. The strong link between overcrowding and domestic violence makes the problem all the more worrisome (Box 2). Table 6. Housing Deficit by Income Level (%) 2.12 Several types of housing are categorized as informal housing. These Type of deficit Up to 2 2 to 5 > 5 Total include improvised units (equivalent to MW MW MW homelessness), substandard units (built Precarious 12 7 2 21 with temporary materials and do not materials conform to the minimum legal standards, Two or more 42 21 13 76 and having little access to basic services), FaTtales 29 15 100 and illegally occupied units (constructed Source: Fundagao Joao Pinheiro 2001. on illegally occupied private or public land). The percentage of informal housing varies considerably across metropolitan regions. The most problematic region was Recife, which has the highest share of informal settlements (32 percent of households), followed by Belem and Fortaleza (Table 8 and Fig. 2). Favelasl0 make up about 60 percent of the informal settlements. The proportion of favela households again varies across regions. In Fortaleza, Recife, Belem, and Rio de Janeiro, favelas are the predominant type of informal settlement. In the South, like Porto Alegre and Curitiba, invasoes are the most typical form of informal housing. Box 2: Overcrowding and Violence Among the Poor In Moser and Mcllwayne's Urban Poor Perceptions of Violence and Exclusion in Colombia, primary school teachers in P6rtico, Medellin, specifically identified lack of internal and external space and overcrowding as causal features of intrafamily violence among the poor." In Moser and Holland's Urban Poverty and Violence in Jamaica, young men in the Zinc City slum who were interviewed as part of the study's participatory urban appraisal specifically identified residence in Zinc City as a primary cause of lack of employment. Worse, Moser notes that the link between violence, lack of infrastructure, and poverty is a vicious circle: "Violence prevented the installation or maintenance of infrastructure, which in turn exacerbated crime ... and eroded community level cohesion. Lack of infrastructure or its adequate maintenance as a direct consequence of violence could indirectly increase fear and mistrust, and reduce community space for association. In Maka Walk [a local slum], for example, focus group participants recounted how local gang members had knocked out street lights to keep the area dim so that it would be easier to rob people. A focus group in Campbell Town [another local slum] analyzed the effect of a community having "light" (i.e., electricity). The lack of electricity was perceived as contributing to growing fear ... and the breakdown of community cohesion. Asked to develop community recommendations for building social infrastructure to realize their "dream community," focus group participants identified housing as the single most important infrastructure investment in local slums. ° Favelas are slum-type settlements, often illegal or informal, with a high density of population. Invasoes are illegal occupation of land plots by families who intend to stay. Moser and Mcllwaine. Brazil: Progressive Low Income Housing Page 15 Table 7. Housing Deficit by Regions and Spatial Groupings, 1991 Regions Improvised/ Precarious Two or Total Metropolitan Other Rural Total homeless housing more areas urban areas (in thousands) families areas North 24 18 196 238 67 155 - 222 Northeast 32 1,268 929 2,229 313 767 1,149 2,229 Southeast 47 272 1,342 1,661 753 690 217 1,660 South 23 80 408 511 110 244 158 512 Center/West 21 91 253 365 257 107 364 Total Brazil 147 1,728 3,128 5,003 1,243 2,113 1,631 4,987 (%) North 0.5 0.4 3.9 4.8 1.3 3.1 0.0 4.4 Northeast 0.6 25.4 18.6 44.6 6.3 15.3 23.0 44.6 Southeast 0.9 5.4 26.8 33.2 15.1 13.8 4.3 33.2 South 0.5 1.6 8.2 10.2 2.2 4.9 3.2 10.2 Center/West 0.4 1.8 5.1 7.3 0.0 5.1 2.1 7.3 Total Brazil 2.9 34.5 62.5 100.0 24.8 42.2 32.6 99.7 Sources: Fundaq&o Joao Pinheiro 1995 and Morais 2000 Table 8. Informal Housing Units by Metropolitan Region, 1998 (in thousands) Invasion in Formal Areas Metropolitan Improvised/ Substandard Corticos Owners Others Informal Total Share of regions homeless- favela and without title housing households informal others units housing in total (%) Recife 0.2 207.5 3.0 40.1 1.8 252.7 791.03 31.9 Belem 0 48.7 1.7 12.3 0.2 62.7 219.94 28.5 Fortaleza 0.41 77.8 1.7 52.6 5.8 138.3 653.42 21.2 Porto Alegre 0 36.1 3.2 124.2 9.2 172.7 997.24 17.3 Rio de Janeiro 0.58 333.3 29.0 55.6 29.0 447.47 3,171.07 14.1 Belo Horizonte 0.25 80.4 8.5. .42.3 3.7 135.18 1,045.56 12.9 Salvador 0.84 25.7 7.8 56.4 1.5 92.2 719.4 12.8 Curitiba 0 23.3 1.8 46.3 5.5 76.9 713.95 10.8 Sao Paulo 2.6 227.3 19.1 159.6 30.4 438.9 4,670.16 9.4 Distrito Federal 1.68 6.9 14.0 17.2 2.3 42.1 508.4 8.3 Total 6.56 1066.9 89.8 606.7 89.3 1859.2 13,490.2 13.8 Metropolitan Regions Sources: IPEA/DIRUR based on PNAD/IBGE 1998 and Morais 2000 Corticos, dwellings with several families per room; Comodos" in PNADs. Favela, households on sub-normal settlements 2.13 Causes. Housing deficit is the result of unmet demand for housing units. According to Instituto Cidadania (2000), between 1995 and 1999, 4.4 million of housing units were built in Brazil. Of these only 700,000 (16 percent) were supplied by the formal market. The remaining 3.7 million were produced in the informal market.12 This situation persists as the supply of housing in the formal market is not enough to meet housing demand. Demand for housing is currently estimated at about one million housing units per year. In addition, only 40 percent of the urban population at the most will be able to afford finished units from the formal market. Half of these will be able to finance their units out of savings and mortgages; the other half may be eligible for some public housing scheme. The remaining 60 percent are too poor to afford a 12 Instituto Cidadania 2000. Page 16 Brazil: Progressive Low Income Housing finished housing unit or to access financing to be able to purchase one. These poor households are the most likely to turn to the informal sector for shelter. 2.14 The most important causes of the Table 9. Informal Housing Units, 1993-98 housing deficit are insufficient income to (in Thousands) purchase or rent "adequate housing," and the lack of supply of "cheap" houses in the Metropolitan 1991 1998 Increase (%) formal sector. This situation results from a regions combination of factors: Belem 75 63 -16.0 Fortaleza 87 138 58.4 Low income in relation to the cost of Recife 65 253 52.8 finished housing. Twenty-two percent of Belo Honzonte 93 35 7247 the Brazilian population are below the Rio de Janeiro 284 447 57.8 poverty line. Their disposable income is Sao Paulo 260 439 68.7 spent mainly on food items, with little or Curitiba 32 77 143.4 nothing left for housing. Over half of new Porto Alegre 55 173 216.9 entrants in the housing markets earn less Distrito Federal 42 than 5 minimum wages. This is not Total RM 1,103 1,859 68,6 thn5mmu ags hsi o sufficient to access financing or to purchase Total Brazil 1,575 a house valued rarely below US$9000 to Source: Morais 2000 based on PNAD/IBGE. US$12,000. The high cost of available land and housing relative to average income and insufficient financing explain why a large number of the new entrants will live in informal settlements on the outskirts of urban areas, illegal rundown tenements in city centers, or in the homes of relatives or friends. Fig. 2: Brazil Informal Housing Units in Metropolitan Areas, 1998, % Distrito Federal ?, v. 4g 9 .: i Be ao Pauloe t Fortal e z 4 . l , ; tli tX T ' .'I l 1.1 Belem , , 'i Recife !|l>Qe 0 5 10 15 20 25 30 35 * The sheer magnitude of the problem. To solve the housing deficit (assum-ing that the correct policies were in placec), i.e., the problem of 5.6 million famnilies living in inadequate housing, about US$40 billion or 5 percent of GDP would need to be invested in the housing market. Brazil: Progressive Low Income Housing Page 17 This value assumes that an average low-income housing could be produced at R$7,000 (1999 prices in Recife). Most of the housing units produced in Latin America are twice as expensive. To solve only the most serious cases (the 1.8 million in inadequate conditions) would require a R$12 billion annual investment. This compares with R$14 billion invested in housing programs by the federal government and FGTS (Fundo de Garantia do Tempo de Servigo or the workers' severance fund) during 1995-2000 and with R$7 billion planned for investment during 2000-05. * The lack of targeting of public actions. Between 1995 and 1999, housing financed with resources from the FGTS reached R$ 6 billion (about US$3 billion). Of this amount, only R$700 million or 12 percent was directed to the Pr6-Moradia program which channels financing to municipalities for low-income housing. The rest was used to finance middle- income housing. 2.15 The Urban Context. Housing deficit is a direct consequence of rapid urbanization. While population growth in Brazil is now 1.64 percent per year, urban population is growing even faster and the urbanization rate is increasing. As mentioned earlier, 82 percent of the Brazilian population now live in urban centers; this may reach 85 percent in another ten years.'3 Urban growth occurs mostly in medium-sized cities and on the periphery of the large urban agglomerations. Urban poverty is a major problem. The urban poor account for 74 percent of the total Brazilian poor. 2.16 The most extreme cases of poverty are found on the periphery of the large and medium- sized urban centers where infrastructure is often nonexistent and the population has limited access to sources of employment and to basic urban services. Access to land is difficult both for newcomers and for "old settlers" due to supply-side constraints. Rigid and excessive regulations governing land and housing development have fostered disorganized occupation of vacant urban lands, rendering the provision of basic infrastructure to these areas difficult and unnecessarily expensive. C. Demand for New Housing 2.17 The Sector. The housing market in Brazil (as in other countries) caters to three segments of the population: * the upper income population which can afford to pay market interest rates and private bank mortgages, or direct purchase of housing units; * the middle-income groups who benefit from government programs through subsidized credit or direct supply intervention; and * the low-income households who earn less than 5 minimum wages and are mostly excluded from the formal housing market as they cannot afford the "cheapest" housing unit that it offers. Most of those who reside in informal settlements and who resort to illegal occupation of land, hoping to get their tenure regularized at a later stage, belong to this group. 13 Brazilian cities form a heterogeneous network that presents a serious challenge to urban management. IPEA describes this "maze" as comprising Ill urban centers, led by 13 metropolises. These urban centers include 440 municipalities and represent 56 percent of the total population. These centers form 49 urban agglomerations: 12 are metropolitan agglomerations, and 37 are not. Page 18 Brazil: Progressive Low Income Housing 2.18 The estimated demand for new housing units in Brazil is about I million units per year.14 Depending on income levels, families purchase their housing units in the formal housing market or the informal market. Using the income distribution of the Brazilian population as a proxy for the income distribution of the new entrants in the housing market, Table 10 presents the estimates of the number of families who can afford housing from the formal sector and those who would have to turn to the informal sector for their housing solution. Table 10. Housing Affordability Estimates, Brazil Household income No. of (%) Demand for Max. Max. value of Potential bracket households housing units monthly house markets ($/month) (000) (OOO/Year) payment'5 a. 10% year (R$) b. 19% year (in R$ thousand) Larger than 3,020 3,351.4 8.6 86 >906 >84 Finished housing units Or >20mw >53.8 Progressive formal 1,510-3,020 4,988.2 12.8 128 453-906 42-84 Finished housing units 10-20mw (27 - 54) Progressive formal 755-1,510 8,378.6 21.5 215 226-453 21 -42 Finished housing units 5-lOmw (13 to 27) Progressive formal 302-755 11,652.0 29.9 300 91-226 8,5 to 21 Progressive informal 2-5MW (5 to 13) Up to 302 9,119.0 23.4 234 up to 91 Less than 8.5 Progressive informal <2MW (Less than 5.4) No income stat. 1,480.8 3.8 38 n.a. - Progressive informal Total 38,970.0 100.0 1,000 - - Source: Serra and Alberdi 2000. 2.19 Considering the lowest price range of a housing unit in the Brazilian market to be between R$15,000 and R$20,000 (US$ 8,500-11,500) and calculating the number of families who could afford mortgage payments (assuming that financing would not be above 30 percent of their earnings), it is estimated that out of the one million new household entrants to the housing market, 22 percent would be able to purchase a unit directly or through credit, and another 22 percent would be able to access the market only if the interest rates were lower, if their incomes rose, or if provided with higher subsidies. The remaining 56 percent of the new entrants (more than 500,000 households) would resort to cohabitation, to rental in tenement housing or in slums, to overcrowded shelters, or to other types of progressive informal housing solutions (Fig. 3). 2.20 Progressive Informal Housing. Informal housing is often viewed as a problem by social policymakers; to large groups of low-income households, however, it represents a solution to the failures of the market to produce housing that are accessible to them. Progressive informal housing is adapted to the way poor people build their houses and is responsive to the variability of their incomes (Box 3). On the negative side, unplanned progressive housing increases the cost of service provision and infrastructure at a later stage. Resettlement of population may need to take place when the settlements are located in environmentally sensitive or high-risk areas. Moreover, informality hinders real estate development as the property value is reduced in view of the risk associated with informal ownership and the residents' limited ability to enter the formal market. i This corresponds to one-tenth of the 20 to 30-year age cohort, which is the variable used in the literature to explain the formation of households; see Serra 2000. '5 30 percent of monthly income. Brazil: Progressive Low Income Housing Page 19 ,' of arel (s,ite) acquelstion and. stage0lAt~nJ documeiii~ti6i' poceuesadecordingo prinding '-'F'3t'E'", . Ak hie\semi4res gsmilvat;;to,SmXeW=us'trnsecgor3af^HN'econ_iq o ~prqriicsite)' for. 3 Bhe~rmas hrazfil- Newui EnVol~trant stinethe Hocupation traetan |consodtin ~~cturewi 14 t -ADirec 0 9e,' ½'k" "''d ' ~~ the transition strtage beisI i~As6 10- Pccu ants"'fiRAP uv nccpie a . o f 5 sagd?.'99?cai. Wove ment i ts~idan i p 1Aceau-'esty'anPd in,r c lu'' d'ng'.- iesrorqtsanh ~.V asgeernents.'tructjeyrse and tue-b-ii Informal4 - < - - 13 feesi" oA an'reuanaio)- - I 's~~ 'gition, eouhn ort shelequenre i o thobsidi ~~it i~~~~~ i ~ ~gnph- -'nt' sibs;ginjo A tsf,I,owtr ei-par e>ohii,,he pa asb ouc g orce widt ,Bymortgage aqfina t c6nshss' jnie: wner%e4ans ge nsledsi) 22% 5oaa6ini liedinsod~ened,rp l ptnor ~-V~ Figgo 3:mBrazil-sew Entrants in tefHousingrMarke us t ps ffW%~~~~Mortgage 10-og i (unmet shelter Subilime elftorM.s~idietod/i~,,as(l~,o ya 56% ~ ~ 9 Page 20 Brazil: Progressive Low Income Housing 3. Federal Low-Income Housing Programs A. Introduction 3.1 In February 2000, Brazil's Constitutional Amendment No. 26, declaring access to housing services as a social right of all Brazilians was approved. Making this promise a reality is a daunting challenge. It requires a strong commitment at all levels of government and the mobilization of substantial resources. Ensuring access to housing implies: (a) access to land tenure; (b) access to credit; and (c) access to basic services. This chapter discusses the main programs for low-income housing in Brazil financed by the public sector. Emphasis is placed on federal housing programs as data on programs financed by states and municipal governments are very scarce. 3.2 In Brazil the main provider of urban housing for the middle- and upper-income segments of the market is the private sector. For low-income groups, the public sector takes on this task in compensating for the failure of the market to do so and the lack of purchasing power among the poor. In general, the federal, state, and municipal govemments share the responsibility for urban and land policies and housing finance. The Secretaria de Desenvolvimento Urbano (SEDU) is responsible for establishing a national housing policy and for overseeing the implementation of projects and programs as determined by the Ministry of Planning. State governments set the policy and financial framework for their own programs. They can also intervene in housing and urban development across municipalities in a coordinating role although they seldom do so. Local govemments regulate housing construction, and land use, development, and titling. 3.3 Investment Program. In the last five years, public expenditures in housing and urban development averaged R$11.5 billion a year. The largest contribution came from municipal governments (75 percent of the total), followed by the states (17 percent) and the federal government (8 percent). In addition, a large amount of resources is made available for house financing under several programs. The largest programs are financed by the workers' severance fund (FGTS) administered by the Caixa Economica Federal. Table 11. Expenditures in Housing and Urban Development, Brazil, 1996-97 (R$ million, 1999 constant prices) 1996 % 1997 % Federal govemment 972 8.4 1147 9.9 States 1583 13.6 2043 17.7 Municipalities 9078 78.0 8345 72.3 Total 11633 100.0 11536 100.0 Sources: Mission estimates; IPEA 2000. B. Federal Housing/Land Programs 3.4 Brazil has a wide range of low-income housing programs (Table 12). Some of the most important elements of the programs are: a) Titling and Land Ownership. Housing ownership in Brazil is estimated at 68 percent, with poor households having a higher rate of ownership, i.e., around 71 Brazil: Progressive Low Income Housing Page 21 percent. This high percentage is misleading, however, since ownership as reflected in census figures is self-reported and does not distinguish between formal and informal types of ownership. The Brazilian government recognizes that titling and land ownership is important to promote better housing conditions as well as to improve the liquidity of the housing market, making it easier to develop transactions in this market. Efforts are being made to improve public registration and to facilitate the enforcement of federal laws governing land parceling, expropriation laws, and the usocapiao, the constitutional provision giving households illegally occupying a plot of land for more than 5 years the right of ownership to that land. Most of these provisions are included in the Estatuto da Cidade, the main law that defines the powers and responsibilities the municipal governments have in executing urban policy (Box 4). b) Right to Adequate Housing. In 1997, Brazil had 1.3 million substandard housing units: 80 percent were located in metropolitan areas, and most were of households earning less than 5 minimum wages. While no comprehensive public policy is in place to deal with the problem, municipal governments throughout Brazil have made efforts to improve the housing situation by encouraging and financing mutirao (self-help) housing improvements, self-construction, self- management, slum upgrading, or the construction of new housing units. Some of the better known examples of these programs are: (a) the Programa Brasileiro de Qualidade e Produtividade financed at the central level; (b) the Programa de Arrendamento Residencial or PAR; (c) Favela Bairro; (d) construction of housing through Mutirdo in Cuiaba; and (e) PREZEIS, a program that allows differential zoning norms to enable the poor to have access to land and urban services. c) Equality in Access to Land. The Constitution recognizes social housing as a social right of all Brazilians, and the new law (the Estatuto da Cidade, see Box 11) defines the tools that local governments can use to buy and expropriate privately-owned land for the purpose of social housing provision. The new law also establishes a progressive property tax to discourage land speculation, charging land vacant for more than a certain number of years higher property tax rates. d) Access to Housing Finance. After the abolition of the National Housing Bank (BNH) in 1986, the housing sector lost importance in the federal government agenda. In 1994, however, the new administration adopted a policy that emphasized the role of the formal housing market in providing housing for the non-poor, thereby allowing government resources to be directed to the housing needs of the poor. In 1997, a new law was passed to promote the real estate financial system (SFI), although high interest rates have so far limited its implementation. Historically, housing finance has played an important part in the Brazilian housing sector. During 1964-96 the housing financial system (SFH) financed 6.3 million housing units or 27 percent of the 23.7 million produced (i.e., 0.9 million per year). However, high subsidies coupled with a high rate of default (30 percent), limited capital, and constant political interference in credit decisions led to the collapse of the Brazilian housing system. Page 22 Brazil: Progressive Low Income Housing !Dox4.,E,t'tdCi - Brazil's 1988.Constittion-include'dtailed-and precise provision's..to re-gulate,urbanpolicy 'and. iVe'klopmenf;'Howe,ver, there was no follow-up legislati onlo imp M ,pisions.In November '2600; 12 years af.te,r tfi-Constitution was ,approved,.the Senate approvedthe Xtarui dadCid4e, which defines the po.wer and'responsibilities'of mUnicipalities in 'the execiutiontof te 'Con'stitutional provision. he Lower Cambe,r h,as 'rot yet approvedthe law. a '. Te E,statut&do aCidade establishes that.-:. -I,'-. s- Thcity andurban andshaveasocifunion- that.Brazilian urban policyaims-to guarantee,.- isustainable citibsov,i,th democratic-man,ageme,nt,cooprationobetwee.n gov,6inments,'privateinitiati-ve,X'.' '. . planning,in the developnient of cities, and provision of infrastructurese'rvices to ensure the Welfare of everyone ' . -,, re federal rovernmentis responsible,in cooperation with state -and municip`al'gov ei its, for,:'` ttw ,promong programsof housing constiictionr nd-improvement, and basi saitation;-dafing w g,gidelines -for urban developmefft, housing, anditransportation; preparing and implenentiog regionalY.- andnationalnplan 'or regional developm,ent (orde ao ob te'' tro) and,for,s 'ecoo 'c d.i.vevelopmnent '^" '. ' , Ur,ba'n PolicyInstruments include national, regional, and state plans for' ,ono 'i and social , .., 'development'a,nd 'for ordenaao doterimtrio d alsoe;iolitan ane'din.ipa'p1ans Te . I- Municipal Urban Plan (Pakno-Diretor2 is-by definition.the inst importaithurban instrument. It determnese tlei orms of land subdivision iand occUpaon,,j,ijen on i,m D ning tiyeqr p ann ^ budget guidelines and maniual cipator -epak ) prr-`s. er instruments includ3e: (a) property taxes and betterment taxes5,() expropriatrin laws'and-adnmisrative ,. ,; 3 luit%tbe3'creati'n of conservatin uni an ZS, conceis for , ' . s l us,or.h'o.'ng,- sudiovis3io n,' nd construction,purposes;, and (d)isu,, . the right'to ladh i-;nd.regulato, and pa,rtcipaoryinana geent in the preparation of plans. * -.- Pronressive vr3iootv taxes are levied to' avoidcases.,of vacant-or underused properyx p pyx ,c'are highler'.of,undernised land). -terfiveyeas landmay be,exopr dpaid wi debt tiat " "ithe. valdue used-for the calculation of the property-tax.` .r t 3 Usfcaviao I ecal'de im6vel urbano'can be -used bycwhih' anyone who occupies 250 sq a tr , for five yeatwitho.utopposition would become the owner of that land provided that the person does .nothave"ahy dther Irban or rural property.For landof oretan 250 sq meters, 'oll pec tive ,. ,, ownershiipis'ppossible,'althoughcorsvwilldividethe.-ownershipso'.tharte&h occupanttreives,an<;'eky'>- 7;3 ,requal part. . ' j. The&'PIano Director`(PD) ensures thattlie, land has.a social function, ensuring the needs ofthe citizeASns -.in'terrmis.of qa~lity forlife;social justice; aid econmi,developftient It isset by 'municipA,al law ad is"' the 'basic inst,rument of urban development and.expansion. It is an integral p.art of the municipalplan T,he'multiyear-plan-should include the objectives.and.guidelines included,in the PD. The plan,covers; . s the whole'tefrfitory of,the municipality. It should be revised at least every 10 years and the'planning process.needs to be open:to public discussion, e .Urban' plans are .compulsorv for cities7with 20,000 or motre i'nhabitants, and for' municipalities in metropolitan regions, in tourist areas, and/or areas of environmental interest, For cities with more,than 500 000 inhabitants,.there is a need foran, integrated urban transport'plan coinpatible with the:P.:no Direfor. tfie Plano Diretor must include the.establishmrent and definition f urban -areas where '. subdivision, construction, and compulsory use of land and utban regulations can be waived,- tconsidering the existence of infrastructure -and the demand for owZincome housing; * a Democratic mana ement occurs through urban policy,. forumspat national, state;"and municipal levels,'--'g > debates and'public consultation, conferences on issues of urban relevance and referendums. All ,' . ni.etropboiitan' agncie need to empWlo)some:form of popular'governance.'' ' - Brazil: Progressive Low Income Housing- Page 23 Table 12. Public Sector Housing Programs in Brazil Objective Program/instruments Examples Titling and land Public registration, federal laws of land parceling, Teresina, 155 slums; 11,600 with ownership expropriation HFS, FGTS, constitutional provision ownership and titling that regulates usocapiao Adequate Mutirao, self construction, basic basket of materials -Programa. Brasileiro de housing for construction, slum upgrading, or construction of QualidadelProdutividade new housing -Programa. Arrendamento Residencial -Favela Bairro Construction of housing with Mutirao,Cuiaba - PREZEIS -- especial Access to land City Status (Estatuto da Cidade) . Expropriation law Public registers Law on Parceling the Urban Land Access to credit Habitar Brazil (Housing Brazil) Promoradia (Pro-housing) Carta de Credito Individual Carta Credito Associativo Program of Support to Production PPA - -Cartao de Credito(Credit card) Plurianual Plan -Comunidade Activa (Active community) -Gestdo de Desenvolvimento Urbano (Urban Development Management) -Gestao Urbana (Urbam Management) -Minha Casa/-Morar Melhor( My house/live better) -Nosso Bairro (Water and Sanitation) -Restructuring of Metropolitan Regions -Saneamento e vida (Sanitation is life) -Transport I urban roads -Habitar Bid ($417 million, of which $250 million are BID). Social inclusion Comunidade Soliddria Projeto Alvorada Caixa Municipio Aldeia Urbana (Quilombos) Street children, bolsa escola, fight against hunger Source: Secretariat of Urban Development. C. Federal Spending on Social Housing, 1995-1999 3.5 Federal investment in social housing is channeled through three main programs: (a) programs financed by the federal budget (OGU); (b) programs financed by the workers' severance payment (FGTS); and (c) programs financed by external donors with national counterpart financing. Between 1995 and 2000, the overall program represented an investment of R$14.1 billion'6 (about US$7 billion), reached 1.4 million families (Table 13), and generated 500,000 jobs. 3.6 Programs financed by the federal budget. The programs financed by the federal budget are passed on as grants or subsidized credits to municipalities and are, in principle, targeted at households earning less than 3 minimum wages. There are two programs of this type - Habitar Brasil and Pro-Moradia - accounting for less than 20 percent of the total federal investment in housing programs. The programs financed by FGTS are the most important, accounting for 80 percent of the total provision for social housing in Brazil (Fig. 4). Among the 16 SEDU 2000. Page 24 Brazil: Progressive Low Income Housing FGTS-financed programs are Carta de Cr&dito Individual and Carta de Credito Associataivo, followed by the Cr&dito de apoio a produqdo. FTGS programs are available to families earning up to 12 to 20 minimum salaries (in the case of the crddito individual or associativo). In the last five years, R$9.4 billion was invested in the latter program (benefiting about 640,000 families). Given the fact that the Carta de Credito programs are financed at a subsidized interest rate, it is clear that federal housing subsidies do not go the poorest segments of the population (making less than 3 minimum salaries), but rather to those families earning between 6 and 12 minimum salaries. Table 13. Summary of Main Programs in Housing, 1995-2000 Program Target Source Investment Loan Counter- % Beneficiary Investment (R$ million) % (R$ part families per family million) ('000) (R$) Housing 3,731 22.5 2102 1629 43.7 926 4,029 Habitar Brazil <3 mw OGU 1392 8.4 1136 256 18.4 574 2425 Habitar Brazil/IDB <3 mw IDB 58.1 7.0 151 8.7 6,700 Pro-Moradia <3 mw 1174 7.1 815 359 30.6 337 3484 Housing FGTS 12,877 77.5 11,224 1,653 13 943 13,655 Carta de CrSdito <12 mw FGTS Individual 9264 55.8 8338 926 10.0 3.5 12,621 Asociativo 3377 20.3 2762 615 18.2 734 19,749 ApoioaProdu,Ao 141 0.8 55 86 61.0 171 40,286 Resolution 166 72 0.4 55 17 23.6 3.5 20,571 Resolution211 23 0.1 14 9 39.1 31 742 TOTAL 16,608 100 13,326 3,282 19.8 1443 11,509 Of which, OGU 3,731 22.5 4084 1629 43.7 458 8,146 FGTS 12,877 77.5 9242 1653 12.8 985 13,073 Sanitaton and Infrastructure 6287 3704 4767 1520 24.2 12966 485 PASS OGU 1352 21.5 1085 267 19.7 1,943 PRO-INFRA OGU 1041 16.6 854 187 18.0 696 901 Pro-Saneamento FGTS 3141 50.0 2239 902 28.7 1156 900 PROSANEAR FGTS 249 4.0 191 58 23.3 3491 899 Resolution 185 FGTS 347 5.5 246 101 29.1 277 899 PASS-BID BID/ 74 26 0.4 21 386 PAR FAAR 131 2.1 131 0 0.0 Total with Sanitation and Infrastruture 22,895 100.0 18093 4802 21.0 14409 1,589 Source: SEDU 2000. Fig. 4: Low Income Housing Finance Program, 1995-2000 Carta de Credito Apolo a Assoc,iatvo producao Habitar Brazil 19% 1%1% Pro Moradia 9% Carta Credito Individual 61% 3.7 In terms of regional distribution, resources have been concentrated in the Southeast (56 percent), followed by the Northeast (17 percent) and the South (16 percent), while North and Center-West receive only 3 percent and 6 percent respectively of the housing investments (Table Brazil: Progressive Low Income Housing Page 25 15). Considering the regional distribution of the housing deficit, the concentration of housing expenditures in the Southeast rather than in the Northeast reflects the orientation of the program and the financing capacity of municipal governments. Table 14. Housing Programs by Region, 1995-2000 (R$ million, current prices) North Northeast SE South Center- West Brazil Pr6 Moradia 51 348 467 219 82 1,166 Habitar 155 481 328 117 102 1,183 Caixa Credito * Individual 153 1,339 5,125 1,723 682 9,021 Associativo 37 242 2,552 228 90 3,149 Other 90 439 318 134 191 1,173 Total 486 2,849 8,789 2,421 1,147 15,692 Structure (%) Prd;Moradia 4.3 29.8 40.0 18.8 7.0 100 Habitar 13.1 40.7 27.7 9.9 8.6 100 Caixa Credito Individual 1.7 14.8 56.8 19.1 7.6 100 Associativo 1.2 7.7 81.0 7.2 2.9 100 Other 7.7 37.4 27.1 11.4 16.3 100 Total 3.1 18.2 56.0 15.4 7.3 100 Source: SEDU 2000, Morais 2002 3.8 Habitar-Brasil transfers federal resources to states and municipalities as matching grants. The criteria for distribution are somewhat unclear and are mostly based on technical and political considerations. States or municipalities are required to provide at least 10 percent of counterpart funds. There is no conditionality for cost recovery.'7 In the last five years, Habitar- Brasil has financed R$1.14 billion in housing investments. Resources come mainly from the federal budget (80 percent). The program forms the Table 15. Structure of Housing basis for the Habitar BrasilfBLD program.'8 The Deficit and Central Government average investment per household was about Investment in Housing R$2,500. In terms of regional distribution, 43 percent Housing Central gov't of the funds were allocated to the Northeast region, Regions deficit investment in and 24 percent to the Southeast region (Table 16), the North 4.8 housing two regions with the greatest concentration of poor North 4.8 ~~~~3.0 Northeast 44.6 17.2 households and with the largest housing deficits. Southeast 33.2 56.2 Habitar-Brasil program therefore is more poverty- South 10.2 16.2 focused and responsive to the housing deficit Center/West 7.3 6.0 problem. Total Brazil 100.0 100.0 Source: Mission estimates 17 DB is supporting Habitar-Brasil. A condition for cost recovery is expected for the components financed by the loan. 18 Habitar Brasil/BlD Program became effective in 1998 and has faced a slow disbursement pattem as of January 2001. The program is discussed in some detail in the section discussing the PPA. Page 26 Brazil: Progressive Low Income Housing Table 16. Regional Distribution of Habitar Brasil Funds, 1995-99 Region Total investment % Federal funds No. of families Ave. investment (R$ thousand) invested benefited per household R$ North 154,900 13.6 140,239 58,884 2,634 Northeast 470,093 41.3 398,965 193,271 2,435 Southeast 301,760 26.5 213,851 125,254 2,408 South 109,943 9.6 80,479 42,495 2,587 Center West 101,574 8.9 85,149 40,177 2,526 All 1,138,270 100 918,683 460,081 2,518 Source: SEDU 2000. 3.9 Pro-Moradia channels housing finance to states and municipalities (at subsidized rates) to develop specific projects, including sites and services, construction materials packages, construction projects, infrastructure provision in existing housing projects; and urbanization of favelas and informal settlements. Created in 1995, the program targets households earning below 3 minimum salaries. The limit is raised to five minimum salaries in the case of projects financing infrastructure in existing housing settlements. The borrowers must be considered creditworthy and must present a Carta Consulta to the Colegiada Estadual which ranks requests for investment. Counterpart funds are a minimum of 10 percent of the total project costs. However, most municipalities especially the smaller ones do not have sufficient debt capacity, hence they are not eligible for these funds. Pro-Moradia funds, therefore, tend to benefit mostly the larger and richer municipalities. Table 17. Regional Distribution of Pr6 Moradia Funds, 1995-2000 Region Total Structure of Total FGTS Participation/ No. of Average Investment Invest. (%) Loans Counterpart Families Inv/Family R$,000 R$,000 (% of total Benefited ('000) investment) North 50,606 4.58 40,993 19.00 17,480 2.35 Northeast 288,082 26.05 234,574 18.57 110,313 2.13 Southeast 466,708 42.20 263,314 43.58 131,687 2.00 South 218,753 19.78 161,561 26.14 44,440 3.64 Center West 81,865 7.40 54,577 33.33 27,811 1.96 All 1,106,014 100.00 755,019 31.74 331,731 2.28 Source: SEDU 2000. 3.10 From 1995 to 1999, approximately 331,000 families benefited from the Pr6 Moradia program with investments totaling R$1.11 billion, 70 percent of which were financed by FGTS and the remaining 30 percent by the individuals or municipalities. Average investment per household was R$3,334. Financing was most heavily distributed to the Southeast (42 percent) and Northeast (26 percent). While the Northeast has a higher level of both absolute poverty and housing deficit, the fact that the Southeast received more funds is indicative of the greater capacity of the municipalities in this region to mobilize counterpart funds. The Southeast contributed 44 percent of the program funds; the Northeast relied most heavily on the FGTS allocation, but still contributed 19 percent of its own funds to the program during the 1995-99 period (Table 17). 3.11 Carta de Crddito is the largest and therefore the most important housing program in Brazil. In the last five years it has spent R$9.5 billion in house financing, benefiting about 640,000 families. The program has two lines of credit: Carta de Credito Individual FGTS for individual families, and the Carta de Credito Associativo for groups. Both are partially targeted at families earning less than six minimum salaries; the eligibility income limits are set at between Brazil: Progressive Low Income Housing Page 27 12 and 20 minimum salaries. The eligible income range (4 tol2 minimum salaries) corresponds roughly to the 46h to 82nd income percentile. This means that only the top 18 percent are excluded from the program and its benefits are concentrated in the 6h to 8h income decile, making it evident that the program is not poverty targeted.'9 In addition, since all funds from FGTS have an interest rate ceiling of 12 percent, an implicit subsidy is provided to beneficiaries. Only about 35 percent of Carta de Credito beneficiaries earn less than 6 minimum salaries. 3.12 To be eligible for these programs, Table 18. Interest Rates beneficiaries must meet certain basic requirements: Household Interest Rate % p.a. they should be a client of CEF; be creditworthy; have income Nominal i Effective legal standing; not be a recipient of other financing, Up to 3 mw. 3.0000 3.0415 for example, from the Housing Finance System 3.1-5 mw 3.5000 3.5566 (SFH); and not be a homeowner (unless the applicant 7.1-9 mw. 4.3000 4.3857 is applying for funds for house improvement or 9.1-il mw 5.1000 5.2209 completion). The loans can may be used for the Above 12 minimum wages = 12% plus the acquisition of new or used housing; acquisition of an reference rate for administration urbanized lot; construction of housing on a lot owned Sources: Informe Sobre a politica de by the boffower; acquisition of a site and construction Financiamentos e Subsidio Habitacional and article from the March 12,2001 edition of of a housing unit; completion, expansion and/or ;poca, pp. 78-82. improvement of an existing housing unit; and / or acquisition of construction materials for housing. Under the Associativo program, the use of funds is limited to the construction of housing on a site already owned; the acquisition of land, house construction, and the production of urbanized lots. 3.13 Limits on the loans vary between Table 19. Carta de Credito: Financing Amount R$7,000 for the acquisition of construction Limits, 2000 R$ materials under the Individual program, and Individual about R$43,000 for the purchase of site and. Acquisition of new or used housing unit, 34,800 conuction fo a housinghuni ud the construction of unit on owned site; construct0on of a housig unit under the acquisition of site and construction Associativo program (Table 19) 20 Debt capacity Acquisition of urbanized site 8,000 is established as a percentage of the gross income Completion of, addition to, or 17,500 of the household that can be assigned to cover improvement of existing dwelling Acquisition of construction materials 7,000 housing expenditures. A sliding scale is used, Associativo establishing the limit of housing expenditures at Production of urbanized site 8,000 20 percent of income. Sliding loan to value ratios Construction on owned site and/or 43,400 are also applied, allowing lower income families acquisition of site and construction to finance up to 95 percent of the selling price. Higher income families finance typically up to 85 percent of the cost of their housing unit. 3.14 Interest rates are established by Resolution CCFGTS No. 289 passed on June 30, 1998. The nominal interest rate is capped at 6 percent per year, with an additional 2 percent yearly charge to cover the costs of loan administration. However, these rates can be subsidized by up to 50 percent for families earning less than 6 minimum salaries, with subsidies decreasing as income rises up to I 1 minimum salaries. Loan terms vary according to the type of financing used and the age of the borrower. Typically loan terms are from 96 months for the acquisition of construction materials to 300 months for higher value transactions, such as for new construction or the purchase of a new or used dwelling. 19 World Bank 2001. 20 Cherkezian 2000. Page 28 Brazil: Progressive Low Income Housing 3.15 As for regional distribution, the Southeast captures the bulk of the resources from both the Individual and Associativo programs, i.e., 56 percent and 78 percent respectively; the Northeast with 15 percent and 9 percent respectively (Table 20). It is clear from this regional allocation of the credit line that that these programs were not well-targeted geographically nor were they oriented towards lower income groups. The Carta de Cre'dito Associativo program is aimed primarily at households earning between 6 tol2 minimum salaries (up to as high as 20 minimum salaries), and is thus targeted at a small percentage of households affected by the housing deficit problem but nevertheless fall within the income group earning more than five minimum wages (15 percent) (see Table 6). Table 20. Regional Distribution of Carta de Crifdito Program, 1995-1999 Region Total Total Structure of No. of Unit/famnily % investment credit program farnilies financed (R$ R$ (%) benefited thousand) Individual North 119,683 107,715 1.75 6,852 17.47 10.00 Northeast 1,027,283 924,556 15.03 92,902 11.06 10.00 Southeast 3,835,821 3,452,236 56.11 242,822 15.80 10.00 South 1,318,468 1,186,621 19.29 98,398 13.40 10.00 Center West 535,347 481,812 7.82 48,324 11.08 10.00 Associativo North 28,077 24,100 1.37 1,181 23.77 14.16 Northeast 195,833 166,049 9.47 10,326 18.97 15.21 Southeast 1,665,632 1,361,590 77.62 82,179 20.27 18.25 South 172,989 139,395 7.94 8,372 20.66 19.42 Center West 78,380 63,148 3.6 3,764 20.82 19.43 Combined North 147,760 131,815 1.67 8,033 18.39 10.79 Northeast 1,223,116 1,090,605 13.79 103,228 11.85 10.83 Southeast 5,501,453 4,813,826 60.88 325,001 16.93 12.50 South 1491,457 1,326,016 16.77% 106,770 13.97 11.09 Center West 613,727 544,960 6.89% 52,089 11.78 11.20 Source: Cherkezian 2000. Table 21. Housing Deficit in 1991 Corresponding to Geographic Regions Total General Deficit Percentage of Region Metropolitan Other Urban Rural Total Total Deficit Areas Areas North 67,042 155,771 - 222,813 4.47% Northeast 313,432 767,400 1,148,991 2,229,823 44.7% Southeast 752,925 690,301 217,045 1,660,271 33.28% South 109,916 243,549 157,823 511,288 10.25% Center West - 257,247 106,929 364,176 7.3% Brazil 1,243,315 2,114,268 1,630,788 4,988,371 100% Sources: Fundagco Joao Pinheiro 1995 and Morais 2000. 3.16 Resolutions 166 and 211: Resolution 166 and 211 programs are rather minor programs, accounting for less than 1 percent of the total FGTS funds invested in housing during the 1995-99 period. In 1994 the programs financed housing construction in areas that were previously acquired with FGTS resources. In 1996, the use of program funds was authorized for infrastructure provision in existing housing projects constructed under other FGTS programs. Brazil: Progressive Low Income Housing Page 29 Resolution 166 and 211 programs are targeted at families earning five or less minimum salaries. During the period 1995-99, their housing investments totaled R$68 million, benefiting about 34,500 families. Both programs were incorporated into the Pro Moradia program in 1996. Fig. 5: Federal Funds Invested by Program, 1995-2000 (R$ million) 2.000- 5 |- 00o 0 0 0 00 U I .~ o - ZF _. 0)~~~~~~3 3.17 Apoio a ProduCdo and Programa Arrendamento Residencial or PAR These programs are relatively small. From 1995-99, Apoio a Produ,dojinvested a total of R$137.4 million (of which R$54.2 million was FGTS funds). For an average per unit investment of about R$45,000, the Apoio program benefited 3,000 families. Similarly, PAR made an investment of about R$131.4 million, benefiting another 6,900 families. With the 100 percent financing provided under the program, families could lease a house for a certain number of years, building equity over time. The average unit cost was approximately R$19,000. Both Apoio and PAR programs are directed to families earning up to 12 minimum salaries. The former has since been incorporated into the Carta de Credito program. PAR continues as a stand-alone program. 3.18 Trends: 1995-99. Funding in most federal housing programs steadily increased over the period 1995-97 -- from about R$482 million in 1995 to about R$2.6 billion in 1997 (Table 22). During that period, this rise in investment is reflected in programs funded from the federal budget (OGU) and from FGTS, including those directed to households earning less than five minimum salaries. The largest increases, however, can be seen in the Carta de Credito programs which target higher income populations. 3.19 Federal investment in housing programs through the FGTS and the OGU has averaged R$2.2 billion to 2.5 billion per year since 1996. In terms of relative importance, federal investments in housing through Pr6 Moradia and Habitar Brasil declined from 37 percent in 1996 to 5 percent in 1999, while federal allocations for the Carta de Crdito programs, Apoio a Producao, and PAR increased sixfold, i.e., from R$366 million in 1995 to R$2.2 billion in 1999. During this period there was a clear movement towards the financing (and subsidization) of housing for those earning between 6 tol2 minimum salaries, and a marked decline in federal financing of housing for the very low-income group or those earning less than five minimum salaries. Page 30 Brazil: Progressive Low Income Housing Table 22. Investments per Year in Selected Federal Housing Programs (R$ million) Program 1995 1996 1997 1998 1999 Total Carta de Crfdito (I) 306 1,470 1,473 1,661.8 1,241.8 6,152.9 Carta de Cridito (A) 0 195.3 361.8 413.6 783.4 1,754.3 Pro Moradia 60 283.3 394.5 17.3 0 755.0 ApoioA Prod 0 0 38 7.2 9.0 54.2 Res. 166 0 54.5 0 0 0 54.5 Res.211 0 13.7 0 0 0 13.7 PAR 0 0 0 0 131.4 131.4 Habitar8rasil 116 196 283.5 214.6 108.5 918.7 All 482 2,213 2,551.0 2,314.5 2,274.2 9,834.7 Source: SEDU 2000. 3.20 The same trend was evident in infrastructure investments under the PROSANEAR and PASS/BID programs, the two other programs financed from FGTS funds. Both programs aim to provide basic infrastructure for populations earning less than five minimum salaries. Investments in these programs totaled about R$2.44 billion during the period 1995-99, with over half of them realized in 1997 alone. They represented about 20 percent of federal funds invested in housing and sanitation infrastructure for the poor and relatively poor. In 1998, investments in these programs dropped from R$1.31 billion to about R$229; in 1999, no additional investments were made. As with the experience with the Pro Moradia program, the weak capacity of smaller states and municipalities to service debt made it difficult for them to be eligible for finance under the sanitation programs. This explains the decline in investments made under these programs in recent years. Plano Plurianual (PPA), 2000-2003 3.21 For the period 2000 to 2003, the multi-year plan, or PPA, has established various objectives within the urban sector. These include: (i) improving the quality of life in urban agglomerations in metropolitan regions; (ii) increasing the supply of housing and improving the existing housing stock; (iii) increasing access to basic sanitation services and improving the urban environment; and (iv) improving the quality of transport and urban transit. These objectives are supported through ten programs, of which six are under the responsibility of SEDU -- the Carta de Cre'ito; Nosso Bairro; Morar Melhor; Infraestrutura Urbana; Saneamento e Vida, and Minha Casa (Box 5). The rest include the Comunidade Ativa; Gestdo da Politica de Desenvolvimento Urbano, Gestdo Urbana, Nosso Bairro, Reestructuracdo de Regides Metropolitanas, Saneamento e Vida, and Transporte Rodovidrio Urbano. 3.22 Of the approximately R$4.5 billion invested in the sector in 2000, the Carta de Credito programs accounted for 71 percent of those funds, and the PAR program, another 14 percent. In contrast, the total investment under federal programs that were oriented towards the low-income groups, specifically the Habitar Brasil/BID program (targeted at the lowest income group, i.e, those earning less than 3 minimum wages), was only R$54 million (Table 23). 3.23 In terms of regional distribution, the Southeast region continued to capture the bulk of housing resources in 2000 with respect to the most important programs: the Carta de Credito (68 percent) and the Habitar BrasillBID program (58 percent). Despite having the highest 21 As an example, in 2000, only the municipality of Fortaleza was able to enter into a contract to borrow funds under the Pr6 Moradia program. The municipality has secured FGTS funds via the National Development Bank (BNDES) in the amount of R$67,512,000 for 10 projects that will benefit 4,830 families. Brazil: Progressive Low Income Housing Page 31 incidence of poverty and a larger housing deficit, the Northeast absorbed only 11 percent and 24 percent of the resources of these two programs respectively. Under the PAR program, the Northeast fared better, having captured 39 percent of its total investments Table 23. Housing Program Investments, 2000 Program Source of Total Total FGTS or Total families federal investment OGU funds benefited funds (R$ thousand) Carta de Credito Individual FGTS 2,427,412 2,184,671 224,623 Associativo FGTS 1,236,527 1,008,513 65,469 Apoio a Producdo FGTS 3,458 506 60 Nosso Bairro Pr6 Moradia FGTS 67,512 59,643 4,830 Prosanear FGTS 5,905 5,216 6,561 HabitarBrasilUBlD (ID) BID/OGU 4,190 4,190 -- Habitar BrasiL/BID (UAS) BID/OGU 53,940 40,706 8,688 Morar Melhor HabitaVdo OGU 254.166 220,484 114,386 Saneamento OGU 180,319 146,590 360,638 Infraestrutura Urbana Prolnfra OGU 203,511 174,657 226,123 Saneamento e Vida Pro Saneamento FGTS 15,338 11,440 17,042 Minha Casa ProgramArrendamento FAR 608,346 607,520 30,904 Residencial (PAR) Total -- 5,060,624 4,464,136 -- Source: SEDU 2000. Table 24. Regional Distribution of Program Resources, 2000 Program Amount invested by region with FGTS or OGU funds only (R$ thousand) North Northeast Southeast South Center West Total Carta de Credito Individual 33,300 311,703 1,288,750 404,801 146,195 2,184,749 Asociativo 9,343 45,990 886,210 54,863 12,107 1,008,513 Apoio a Producdo 0 506 0 0 0 506 Nosso Bairro Pr6 Moradia 0 59,643 0 0 0 59,643 Prosanear 0 5,216 0 0 0 5,216 HabitarBrasi UBID (ID) 245 815 2,150 605 375 4,190 Habitar BrasillBID 0 10,341 24,105 6,260 0 40,706 (UAS) Morar Melhor Habita,do 32,866 81,267 44,279 15,423 49,105 222,937 Saneamento 16,541 40,404 31,324 14,443 44,971 147,683 Infraestrutura Urbana Pr6 Infra 21,901 65,145 15,468 37,154 36,812 176,480 Saneamento e Vida Pr6Saneamento 0 11,440 0 0 0 11,440 Minha Casa Program Arrendamento 18,467 234,792 227,218 67,234 60,607 608,318 Residencial (PAR) Source: SEDU 2000. Page 32 Brazil: Progressive Low Income Housing Box 5. 2000-2003 Housing Programs The urban sector programs,withinthe,purview of the Secretaria Especial de Desenvolvimento U (SEDU) include six primary programs each with several subprogramns. In addition to the Carta de Credito programs which, are d&scribe4d above,-these programs comnpnse Nosso Bairro, Morar Methor, Infra- estrutura Urbana, Saneamento e Vida, and Minha 'Casa. Nosso Bairro includes the Pr6 Moradia, Pr,osanear, PASS/BID, iand Habitar BrasilUBID. In 2000, only-.. the municipality of Fortaleza was able to secure'loan.resources from BNDES under-the Pro Moradia program. PASS/BII) is essentially a water and sanitation program that provides funds-for basic sanitation-. in municipalities with po,pulations betweeno15,000 ain 75,0, inihabitants. The most relevant of these ' . programs, however,is the Habitar Brasil/BID. The Habitar BrasillB II) is a flexible program for, improving thelvinvmg conditions of the urban pqor earning less than three minimum salaries. It comprises -two subcomponents: `the institutional development (DI) - _71' -component and the U,rbananizadao de Assentamentos UAS) component;. TheoDIcomponent is designed to provide municipalities.with assistance in developing-strategic plpns to addre sIs theifrastrucdtureneed 6of' substandard human settlements such asfavelas a.nd, tes clandestinos. The UAS component fioances works and services based on the strategic plas elaborated by the muicipaliti es. The program gves - priority to metropolitan regions and urban agglomerations as dfined by IPEA and'state.capitals. The use of funds is flexible-for land tenure,regularization, indemization, 'resetement, preliminary plans and engineering, provision of infrastructure and urban services, 'ho.usn'g iprovements , andotheryactivities typically associated with urban upgnidingprograms, .XThe pfiis flnanced by thielnteAnerlan..a . Development Bank (IDB) and is expectedetooreachcnpleto .a .,eriof fryears. T hancin,f theprogram is US$417million, f which US$25,0,m,illo,n isa ,-loaot.m theDBda,ndanotherU$167 .- million in counterpart funding. Funds are grantedt:'tatesikcit ate sfor texecutioWofproT To date, approximately US$22 miillion has been disbursed. Morar Melhor is'.the third block of the urban- housing program and includest OGU funds for sanitatiotJ under the PASS program, additional funds for-Habitar Brasil, fsth c llection;of solid waste in-areas ,ere children -li've' in 'proximity' to solid waste dumps, and he, financig of individual lifrastrire' improvements under a program known as-Prc% Comunidade. The Saneamento 4 eda:program is, thejfpurt block of programs, focused on, providing funds to public and. priVate water r,ipanies- to eipand bai* sanitation services. Together these progranms accounted for-only 9% of investmnOts in 2 .000- The Infra-eetructura Urbana program provides, less direct housing-related ,benefits haj theoh programs. This program is aimed essentially at the revitalizaiion of (ity centers in medium and largedcities anBd at improving`and expanding trunk infrastructure. The final program that is of importance to the housing sector is the PAR program. This program, lnown locally as leasing, allows families earning up to 12'nminirnum salaries to accessfinished housbt . leasing program tihat converts to ownership over time. It represents a growing part of the housimg potffolio In .2000, investments in this program totaled about R$607 Million, compared with R$131 million in 1999. Brazil: Progressive Low Income Housing Page 33 D. International Collaboration 3.24 International Collaboration. Several bilateral and multilateral donors have been actively involved in the preparation and delivery of low-income housing programs in Brazil: * The Gesellschaft fur Technische Zusammenarbeit (GTZ) is implementing PRORENDA, a program that helps organize poor communities to improve their capacity for more active participation in political and civic life. Currently, PRORENDA supports three projects: (a) urban development in poor slums in Fortaleza, Ceara; (b) urban improvements in low-income settlements in the Metropolitan Region of Recife; and (c) the promotion of sustainable development in the State of Rio Grande do Sul. * CIDA (Canadian Aid) is supporting the program Gerenciamento Participativo nas Areas de Mananciais in Santo Andre (community-based watershed management in Santo Andre) with the objective of defining a benchmark for the sustainable development of the municipality of Santo Andre and implement environmental management activities in settlements located in high-risk areas. Program components include re-zoning of areas of mananciais, conflict resolution, public awareness campaign, and fostering partnerships within communities. * The Government of Italy is supporting the project ALVORADA to help the municipality of Belo Horizonte upgrade the favelas in the metropolitan region, promote economic and social development, and strengthen local associations. These projects received positive assessment and are being replicated in Bahia (Novos Alagados) and in the Cidade do Cabo (Africa do Sul). * The Government of Japan is active mostly in the area of waste management in Sao Paulo and Pemambuco. In Sao Paulo two projects were concluded: management of industrial solid waste and water resource management. * The United Nations Development Program (PNUD), one of the most important development partners of Brazil, is supporting the strengthening of SEDU, and the development of municipal-level performance indicators in the states of Para and Parana. E. Conclusion 3.25 Federal housing policy in Brazil has been mostly ad hoc. Investments in housing for the lowest income groups (i.e., those earning less than five mininium salaries) have sharply declined since 1995. Conversely, under the PAR and Carta de Creedito programs, housing investments for those earning up to 12 to as high as 20 minimum salaries have increased. The subsidies embedded in these two programs indicate that federal housing programs are poorly targeted. The Habitar BrasilfBID program may have held great promise to households earning three or fewer minimum salaries, but more investments and subsidies will continue to go to higher income groups under the PAR and Carta de Credito programs. While the Habitar BrasilfBID program represents an investment of approximately R$830 million over a period of four years, in 2000 alone, R$2.2 billion were invested under the Carta de Cre'dito Individual program. The latter benefits mostly households earning 6 to 12 minimum salaries. 3.26 Even in terms of regional investments, federal resources are poorly targeted. The bulk of federal funds flow to the Southeast despite the fact that the housing deficit and the incidence of poverty are most acute in the Northeast Region. Urban poverty may be higher in the Southeast in percentage and absolute terms, but federal resources, especially those intended for the lowest income groups, could be better directed to the Northeast region. Page 34 Brazil: Progressive Low Income Housing 3.27 A more concerted effort is therefore needed to mobilize more resources for states and municipalities to enable them to implement programs that will meet the housing needs of the poorest segments of their population and that will help overcome the housing deficit. Subsidies should be channeled away from higher income groups and towards the lower income groups whose housing needs are greater. Instead of such programs as the Carta de Credito which are regressive, distort the credit market, and stifle competition from private banks that could expand the supply of credit for the higher income segment of the market, Habitar BrasilIBID and similar programs should be expanded given their potentially greater impact on low-income households in particular and on the housing sector as a whole. Brazil: Progressive Low Income Housing Page 35 4. Best Practices in Brazil A. Introduction and Criteria for Assessment 4.1 Improving access by the poor to housing services is a challenge for policymakers. Supply-driven approaches have failed not only because of the lack of sustainability and poor financing capacity of central and local governments, but also from an inadequate response to the needs of the poor. This chapter reviews some successful experiences with low-income housing and upgrading with the purpose of extracting lessons that can be used in Brazil. It describes the objectives of the programs and provides an assessment of how these programs rank according to five criteria: (i) financial sustainability; (ii) targeting and transparency; (iii) affordability and effective demand; (iv) institutional appropriateness; and (v) sectoral objectives. * Sustainability. Financial sustainability is the ability of a program to generate sufficient resources to keep the flow of the services that it was created to provide. Sustainability requires adequate cost recovery. Table 25 presents different types of cost recovery instruments in line with the nature of the goods and services being provided. It explicitly suggests that investment in public goods should be recovered through direct contributions, taxes and betterment fees, user charges (for urban services), and repayable loans (for private services and housing). * Targeted Subsidies. Subsidies may be needed to address market failures and to supplement income for redistribution purposes. In principle, subsidies need to be transparent and targeted. While demand-side housing subsidies are generally preferred over supply-side ones, in most low-income housing programs, individual targeting may be difficult. This suggests the need to: (i) use up-front demand subsidies whenever possible, as opposed to indirect subsidized credit financing; (ii) achieve full cost recovery of investment and maintenance costs; (iii) ensure some form of self-screening; and (iv) design the services such that the choice of technology is appropriate for the poor. * Affordability and Effective Demand. Demand refers to the ability and willingness of beneficiaries to pay for a given project. To use this concept for policy purposes22 one needs to: (i) have data on demand for infrastructure services, from both public and private providers; (ii) determine how much households must contribute as "proof' of their. willingness to pay; and (iii) contact communities to assess their willingness and ability to pay for the housing services. Based on these data one can determine whether the quality and quantity of the services to be provided by a program are in line with the preferences and willingness of the community to pay. * Institutional Appropriateness. The roles and responsibilities of all stakeholders should be clearly defined. In principle, central governments should have the responsibility for building an enabling environment, including a regulatory framework, to allow land and housing markets to function efficiently and to contribute directly or indirectly to the design of progressive housing schemes, with a focus on&poverty alleviation. States share the responsibility for ensuring an enabling environment by playing important financing and coordinating roles. Municipalities are best suited to identify the target population, articulate their demands, and serve as the executing agency for national programs. All other agents - the private sector, nonprofit organizations, and the community itself - have key roles to play. 22 Kessides 1997. Page 36 Brazil: Progressive Low Income Housing Table 25. Financing Infrastructure Goods _, , as ., r c .. . . . .. o o .Me6ianism - Aeeovery Mithod .. - Public goods are goods that are consumed Typically it is the municipality that has a. Direct contribution from the project "collectively" and therefore have a low level of responsibility for the provision of these beneficiaries. e.g., initial payment from the excludability. As public goods, there is no goods, and financing would come beneficiary community to ensure obvious market mechanism to match supply with through the capital budget. Public community buy-in. . demand, and there is usually some high social or goods are generally financed from the b. Betterment levies or benefit taxes- political value associated with them and/or high revenue source of the community, that special property tax increases can be levied positive or negative externalities. In the housing is, from property taxes. to recover costs directly from context, these goods would include streets and beneficiaries. drainaSee public lighting sidewalks and parksj c. Property tax is the most common form and public sLaces. of cost recovery for these goods, as it taxes incremental increase in property values brought about by the investments. Toll goods straddle the line between public goods Financing for these investments would Investment costs, as well as current costs, and private goods. They have high extemalities come from a variety of sources, including debt service, would normally be and are highly valued socially and/or politically. including states, municipalities, and recovered directly through user fees and They have a higher degree of excludability than public and private service providers. tariffs. purely public goods. Examples are secondary water supply, energy supply and distribution, sewerage and septic collection, disposal, and treatment; and trash collection and disposal. Private goods have a high degree of excludability Financing for these goods comes Costs for these goods are recovered in that they can only be consumed individually. directdy from the consumer, but typically through user fees and trffs, Examples are dwelling units the land for sometimes through banks, the supplier, credit and loan pavments, and lent. housing household water connection and or other intermediaries. consumption electricity connection and ConsumDtion. etc. * Sector Objectives. Programs to improve the housing stock need to rationalize the functioning of the housing sector to enable housing and land markets to work more efficiently. To the benefit of the poor. These efforts involve the reduction of transaction costs, in terms of the money and time spent in registering titles and in appraising the value of real property; streamlining development regulations; lowering mninimum lot sizes; reducing the level of public services required to subdivide and develop property; facilitating land titling and property rights development; and offering low-cost alternatives to finished housing. Private sector involvement is also necessary as it tends to drive down the cost of real estate and housing development. To encourage this involvement, information should me made easily available through' public registries, planning agencies, and public works departments. The building industry should be encouraged to participate in this effort and to contribute with low-cost solutions. B. Brazil's Low-Income Housing Experiences Rio: Programa de Assentamentos Populares (PROAP-Rio)23-A Citywide Approach 4.2 Rio de Janeiro has a population of 5.8 million, of which 1.6 million live in substandard conditions: 700,000 live in favelas and 900,000 in loteamentos24 along the periphery of the 23 This case is based on a site visit in March 2001. It also includes information from two studies conducted of the program. The first is a report prepared for the IDB in September 1998, titled, Programa de Urbanizacdo de Assentamentos Populares do Rio de Janeiro: Relatorio de Avaliacdo de Progresso. The second is a report prepared in December 1999, titled, PROAP 11 Avaliacdo S6cio- Economica. 24 Loteamento is a term typically used to describe settlements that, unlikefavelas, are planned and usually well ordered. with clear delineation of rights-of-way as well as public and private spaces. Loteainentos irregulares are typically subdivisions that for one reason or another never received final authorization from the municipality and/or the Registry of Properties. Loteamentos clandestinos are also "planned" subdivisions, but they were developed without official approval procedures and usually do not conform to the Brazil: Progressive Low Income Housing Page 37 cities. While the population of Rio is growing at 1.5 percent a year, the population living in substandard conditions is growing at an annual rate of 8 percent. There are approximately 800 favelas in Rio. 4.3 There have been a number of programs designed to address the favela problem in Rio. In the 1960s, with the establishment of the Sistema Financeiro de Habitacao (SFH) and the Cooperative Housing Agencies (Cooperativa de Habitacao or COHAB), there was a national effort to reduce the national housing deficit through the commercialization of housing production, while trying to eradicate "slums" through massive slum clearance programs. In Rio, an effort was made to selectively resettle families from favelas located in the southern zone of the city to large housing developments along the urban periphery of the municipality. These programs were largely financed by the National Housing Bank (BNH) and the State's COHAB. The policies of slum clearance did little to curb the growth of favelas, which continued to grow rapidly through the 1980s due to high rates of urban growth. 4.4 Some efforts were also taken to improvefavelas in situ by providing basic water supply and sanitation services. These isolated experiences were highly participatory and were reintroduced in the 1980s by the Secretaria Municipal de Desenvolvimento Social in the municipality of Rio de Janeiro. With assistance from UNICEF, the Secretaria extended sanitation services, combined with a program of health education, to the favelas. By 1988 the program was extended to 263 favelas. Around the same time, the municipal planning institute (ILPLANRIO) concluded the first exercise of cataloging and preparing the cadastres of all the favelas in Rio de Janeiro and completed studies on the regularization of loteamentos. This was a major step in recognizingfavelas and loteamentos as viable housing solutions and set the stage for Rio's most important progressive housing program, PROAP, or Favela/Bairro. 4.5 Project Description. Favela/Bairro is among the best known urban upgrading programs in Latin America. It was started in 1995 by the municipality of Rio de Janeiro with the first phase of the IDB-financed program. The total project investment of US$600 million is broken down into two phases of US$300 million each -- the first phase covering 1995-2000, and the second, 2000-2005. 4.6 The objective of the program is to improve the living conditions of the low-income urban population of Rio de Janeiro by increasing the availability of urban and social services in favelas and irregular subdivisions of the city. Unlike other urban programs, Favela/Bairro does not focus on the largest or densest of the favelas, nor does it specifically target preselected geographic areas. Rather it focuses on medium-sized favelas and loteamentos made up of communities of 500 to 2,500 households. According to official municipal estimates, medium- sized favelas represent about a third of allfavelas and about 60 percent of the favela population. 4.7 Favela/Bairro is in many ways a prototype of the latest thinking in progressive housing for the urban poor. In sharp contrast with the slum clearance programs of the 1960s, Favela/Bairro recognizes the large personal investments made by the communities in these areas, and takes an explicit approach to improving the living conditions in these areas, regularizing land tenure, and resettling only those families living in high-risk areas or in places where space is needed for infrastructure. Minimizing social disruption, integrating these areas into the so-called development norms and standards of the municipalities. There are many reasons for the development of such settlements. One of the most commonly cited is the usually burdensome bureaucratic process of formal land subdivision, as well as a demand for smaller lot sizes with less strict development standards from lower income populations. These areas, like favelas. often do not have al the essential municipal services, and land tenure is also sometimes an issue. Regularizing these areas is an important component of the FaveklaBairro program, as it is in other municipalities throughout the country. Page 38 Brazil: Proaressive Low Income Housing formal city, and rminimizing capital costs are the guiding principles of Favela/Bairro and other similar programs. 4.8 The first phase of FavelaJBairro benefited 210,000 inhabitants in favelas and another 51,000 in loteamentos, nearly meeting its goal of reaching 300,000 people. The program benefits 28 percent of ft7vela dwellers but only 5 percent in loteamentos . Investments in the first phase ranged from as low as R$1,130 per household to R$15,020 per household in the case of thefavela Serrinha. Typical investments included hillside stabilization; construction of resettlement housing; construction of day care centers, sports fields, and plazas; and the usual investments in drainage, water and sanitation, electricity, and access to roads. 4.9 The second phase will operate under the same rules as Phase I, but will include assistance in job training and social services for families. The goals are to urbanize 60 additional faivelas and benefit 52,000 families, or roughly 234,000 people. In addition, it is anticipated that 17 loteamentos will be regularized to benefit 6,000 families. Under its social services component, some 28,500 children and adults are expected to be served under a variety of programs and 84,500 individuals will benefit from job training and capacity building activities. Box 6. Morros Dos Cabritos Morros dos Cabritos is located near the Bairro of 'sil.ii Copacabana and is home to approximately 1,500 families, or l. , . . I 7,180 people Its population represents roughly the average I 8W2g1 , medium-sized favelas targeted by the project. Situated in a $P8ving . i,258432f hilly area of difficult topography, dos Cabritos is roughly |Ret5onw tltk s35$1a 10.02 hectares in area. Work in this area began in November l dp g- "I ,25457I 2000 and are expected to conclude by May 2002. The total 0 vmste ion | 4651-i > project cost is R$7.447 million (mainly for construction), PubfloIJ"bting representing a per household investment of about R$5,000, Tech4a#C1 si~Sp ' ;3, or about US$2,500. The single largest work items are the lSigia* Urbanxu6fizWls 11y642 ,. construction of buildings and the paving of roads. Additional LtI . vg 447,76 costs include R$441,927 for design and engineering. Buildings include a day-care center, a solid waste collection site, and housing units for the resettled families. Public and leisure spaces include a plaza and a football court. 4.10 Sustainabilizty and Cost Recovery. From a cost recovery perspective, the Favela/Bairro program receives a mixed rating. Beneficiary households pay for services such as solid waste collection, water and sanitation, and electricity consumption, but there is no cost recovery system in place nor any requirement for families to share a part of the investment costs of upgrading as a condition for eligibility. Capital cost is about US$3,000 per household in favelas, and US$4,370 per household in loteamentos. Property taxation (IPTU) is considered the main mechanism for cost recovery, but its revenue potential is limited by the lack of adequate cadastre and titling. In fact, the 1995 real estate cadastre shows that there were large portions of land where ownership was in question. The long and complex process of identifying and providing clean titles to areas that have been invaded renders the use of IPTU ineffective. Brazil: Progressive Low Income Housing Page 39 Table 26. Income Distribution in Favelas and Loteamentos in Rio de Janeiro Income Favelas Loteamentos (no. of minimum wages) Up to 1 13% 4% 1-2 23% 13% 2-3 16% 17% 3-5 12% 30% >5 4% 26% NS/NR/SR 32% 10% Total 100% 100% Source: PROAP 11 1999. Margin of error is 5%. 4.11 The situation is different in loteamentos. By the end of 1998, 34 loteamentos were regularized and registered under the program, allowing for the collection of IPTU. Through the latter, the Municipal Treasury estimates that urbanization investments made in the loteamentos under the program could be recovered within five years. The program would benefit from: (a) including a specific program on the regularization of land tenure and titling; (b) assisting families in paying the R$300 required to survey and register a property; and (c) investigating alternatives to legitimizing land occupation in contestable areas, such as the institution of the usucapido coletivas, which would also allow for a more systematic program of cost recovery. 4.12 Targeting. As in most upgrading programs, individual targeting is not used since the bulk of investment is in community infrastructure. However, an examination of the income distribution of beneficiaries reveals that a large percentage is indeed poor (Table 26). Data collected from Databrasil and the Secretaria Municipal de Habita,do shows that 96 percent of households eam less than five minimum wages; in loteamentos 64 percent have the same level of income. In addition, considering that more households in loteamentos have relatively higher incomes than those in favelas, there is equity in the fact that there are higher rates of cost recovery in loteamentos than infavelas. Still, the program would benefit by developing a specific cost recovery strategy, including setting a maximum per household subsidy based on income. 4.13 Affordability and Effective Demand. Affordability and willingness to pay were overlooked in the first phase of the program. Since there was no clear cost recovery strategy there was little incentive on the part of communities to choose low-cost options, when applicable, that would be compatible with their capacity and willingness to pay. Some of the first project designs were based on sectoral traditions and norms that did not take into account local economic circumstances. Greater participation would have enhanced the affordability of projects. This lesson has been incorporated into Phase II of the program, but no clear framework has yet been put in place to recover, even partially, investments in public goods such as access roads, sidewalks, and drainage. In this phase, investments in favelas will not be allowed to exceed US$4,500 per household and are expected to average US$3,800 per household. In the case of loteamentos investments will be capped at US$3,500 per household and are expected to average US$2,600 per household. 4.14 Institutional Arrangement. The Favela/Bairro program has a well designed institutional structure. The program is managed by the municipality's Housing Secretariat within a small project management unit comprising two Departments (Favela/Bairro and Loteamento Irregular) and two subdepartments (coordination of works and coordination of programming/urbanization). The bulk of project management and supervision, involving works, design, engineering, and Page 40 Brazil: Progressive Low Income Housing construction, are outsourced to the private sector. The management unit also coordinates the work of other municipal and state agencies that have important roles to play in the project, notably the Secretariats of Labor and Social Development which oversee the social and job training components of the program; CEDAE, the state water company; and Rio Luz, the private provider of electricity in Rio de Janeiro. 4.15 As for community participation, this could have been more intensive, especially after the design phase was completed. Work in other parts of Brazil, such as in Sao Paulo, suggests that community participation is as important during the construction and post-construction phases. This ensures greater and sustained commitment on the part of the community and will impact the future maintenance of works and the ability to recover costs.25 Curitiba: Innovations in Partnerships26 4.16 Curitiba is the capital of Parana and the seventh largest city in Brazil. The city has a population of 1.6 million; its metropolitan region ,2.8 million. Curitiba is considered a model of municipal management, notably in the areas of urban transport and urban planning. It also has some of the most interesting low-income housing programs as described below. 4.17 Curitiba has a proactive stance toward low-income housing. The Companhia de Habitaqdo de Curitiba (COHAB-CT) keeps registers of low-income families in need of shelter and plays a key role in the design of local housing programs. According to COHAB-CT, the number of households looking for low-cost housing skyrocketed between 1991 and 2000, from 1,022 to 13,400 (Table 27 and Fig. 6). The majority of these households reside with family members in favelas and loteamentos. The vast majority are poor: More than 87 percent eam less than five minimum wages, 63 percent earn three minimum salaries, and 40 percent earn two minimum salaries or less. They are young -- 60 percent of the household heads are 34 years of age or younger, 48 percent are single-headed households, and 51 percent are headed by women. The average household size is 2.6 people, indicative of the large number of households headed by single mothers. Table 27. Curitiba Households on Waiting List Year Curitiba Metro Total Accumulated FiRg. 6: Curlbba, Huseholds Looldng for LowI Region total Inceom Housing 1991 1,022 100 1,122 1,122 1,0 1992 866 74 940 2,062 14,000 1993 1,076 109 1,185 3,247 i i_ | 1994 2,542 283 2,825 6,072 1995 6,071e, 645 6,716 12,788 10,003 1996 8,423 781 9,204 21,992 a10_ 1997 7,707 474 8,181 30,173 1998 7,217 572 7,789 37,962 4,000 1999 7,453 1,178 8,631 46.593 aooo 2000 13,467 1658 15,125 61,718 O 2001en 2,221 356 2,577 64,295 1990 1992 1994 1996 1998 2000 2002 d/ Feb Source: COHAB-CT 2001. 25 In addition to Favela/Bairro, the Rio de Janeiro municipality has programs to restore tenement homes, and for the renewal of the city center, the regularization of loteamentos, and the use of CRF Carta de Credito programs for a variety of housing options for families generally earning between 5 to 12 minimum salaries. 26 This case study is based on a site visit in March 2000 and on information provided by COHAB-CT and IPUCC. Brazil: Progressive Low Income Housing Page 41 4.18 Financing for Curitiba's housing programs comes primarily from federal, government programs such as the PAR, Carta de Credito, Habitar-Brasil and Pr6 Moradia. In addition, the city capitalizes and operates a municipal housing fund known as the Fundo Municipal de Habita,cdo (FMH), which was created in 1991. The Fund is capitalized from private sector contributions generated though development agreements, i.e., contributions to the fund in exchange for development bonuses. Available funds fluctuate with the fund rarely generating more than R$1.0 million annually. Though quite innovative, the funds generated through the program are relatively small in comparison to the demand for housing, although the city also relies on federal funds for programs when available. 4.19 Parceria Iniciativa Privada was authorized in January 2000 and is one of the most important and innovative instruments available to COHAB-CT to develop housing options for low-income households. It has two important elements: It provides incentives for private sector provision of affordable housing for the urban poor, and it focuses on the provision of urbanized lots, rather than of finished housing. 4.20 The private sector incentive is in the form of density bonuses being offered to landowners or developers wishing to subdivide and urbanize land for housing. Since the standard minimum urban lot size is 360 square meters, which is too costly for low-income housing, the municipal law (No. 9.802) allows developers to apply for an exemption for a lower minimum lot size. Developers may also apply for an exemption from the regulation governing the height of new buildings to be constructed. In exchange for the bonus, developers are required to provide either land or cash to COHAB-CT. In either case their contribution is based on the market values of land and/or the increased floor space allowed to be constructed. 4.21 If the contribution is in the form of land, the municipality grants the developer the right to subdivide his property in lot sizes up to half the standard size. In exchange for the density bonus, the developer must donate 20 percent of the lots, plus 10 percent for public uses (i.e., parks, schools, etc.). A fee of R$8,000 is also charged for administering the subdivision process in cases where fewer than 200 lots are to be developed, or a fee of R$40 per lot where more than 200 lots are to be developed. The subdivision process takes 8 to 12 months. COHAB-CT urbanizes these sites and sells them to families on the waiting lists. The financing terms are the same as the Carta de Credito, i.e., 10 percent down payment with financing at 6 percent per annum over 20 years. Most beneficiary families earn 3 to 8 minimum wages and can afford houses with prices ranging from R$6,000 to R$8,000 (US$3,000 to US$4,000).27 To date, 15,000 urbanized lots have been developed under the Paceria Iniciativa Privada program, and COHAB- CT expects to develop 10,000 more under this program by the end of 2002. 4.22 PAR Program. The municipality of Curitiba is also pursuing the new federal Programa de Arrendamento Residencial (PAR) as another altemative to providing housing for the lower- income segments of the population. Currently there are three projects being developed in Curitiba for a total of 264 units to be completed later in 2001. With monthly payments fixed at R$140 for the first 12 months, and tied thereafter to the rates paid on deposits in the FGTS, the program is expected to benefit families earning 4 to 6 minimum wages. Under this program the families must rent the unit for a period of 15 years, after which they have the option to purchase it. The value of units (including land and services) is not to exceed R$20,000. These units are approximately 40 square meters and include two bedrooms, a living cum dining room, kitchen, and service area. The potential of the PAR program to help fill the housing gap is minimal. Using only the 27 Under this program, with down payments for urbanized sites ranging from about R$600 to R$800 and the remainder financed over a period of 20 years at a rate of about 6 percent per year., monthly payments would be R$40-R$52 per month. Page 42 Brazil: Progressive Low Income Housing COHAB-CT waiting list as the source of demand, there are approximately 2,000 families that would potentially qualify for this program as against only 264 units to be produced in 2001. Table 28. Housing Type Metropolitan Area of Sao Paulo, 1999 No. of Population Structure households (%) (%) (in thousand) Standard Housing 1,245 43.6 49 Irregular settlements (loteamentos) 1041 37.5 25.5 Favelas 379 13.3 19.4 Tenement housing 161 5.6 6.1 Total 100.0 100.0 Sources: PMSP, SEHAB, HAB/FIPE 1999 and PMSP, SEHAP, RESOLO 1994. 4.23 Nossa Vila. There are approximately 262 favelas and loteamentos in Curitiba, home to about 232,000 people. In general, these families are poor, with 60 percent earning three or less minimum wages. Nosso Vila is similar to other upgrading programs in other cities, with the specific objective of improving infrastructure and basic services in low-income settlements. In addition, the program includes a component to regularize tenure of individual parcels in these areas. To date, 5,000 lots have been regularized and another 10,000 are in the process of regularization. In Xapinhal neighborhood, a typical favela covered by the program, 2,300 households have benefited from improved services, mostly water supply and sanitation, electricity, drainage and access, as well as from land tenure regularization. The average per family investment under this program is US$1,915. The Curitiba experience is noteworthy for its efforts to provide housing solution through a variety of vehicles, including the upgrading of the existing housing stock and the provision of progressive housing through sites and services. Sdo Paulo28 -A Tale of Two Approaches 4.24 The metropolitan region of Sao Paulo (MTSP) comprises 41 municipalities with a total population of 16.6 million, of which 60 percent live in Sao Paulo itself. The city is wealthy by Brazilian standards, with a per capita income of nearly US$5,000. The metro region accounts for 16 percent of Brazil's GDP and 15 percent of its industrial output. Sao Paulo had served as a magnet for rural migrants, which explains the growth of squatter settlements in the city. 4.25 Between 1973 and 1999, favela residents increased from 72,000 to 1.9 million, that is, from 1 percent to 19.4 percent of the city population (Table 28). Another 26 percent lived in loteamentos, and 6.1 percent in corti,os or tenement-type housing found mostly in older areas within the city center. Indeed, only half of the population lived in standard housing. Sao Paulo's housing policy29 is therefore focused on how to improve this situation. Its objectives are to: (a) regularize loteamentos; (b) contain the growth of informal settlements; (c) renovate the city centers, including tenement housing; (d) provide adequate infrastructure and introduce environmental improvement in watershed areas; and (e) control the use of real estate property. 4.26 The main programs in Sao Paulo are: (a) Lote Legal which aims to regularize loteamentos and control the use of real property; (b) the Cingapura project which aims to contain 28 For lack of information, the report does not include the Social Housing Program run by the State of Sao Paulo. The program funded by a specific tax (added to ICM) has substantial volume of resources and is allocated to the construction of housing for the very poor. Allocation of housing units is made generally by lottery. 29 A new administration took power in 2001; hence this policy is subject to change. Brazil: Progressive Low Income Housing Page 43 the growth of informal settlements through a verticalization process; (c) the program to rehabilitate tenement housing (cortiVo) and degraded areas; and (d) the Guarapiranga program to protect watersheds and improve living conditions in informal settlements. 4.27 Guarapiranga. The Guarapiranga project is named after the reservoir that was built south of Sao Paulo in the early 1900s for water supply, power generation, and recreation. The reservoir currently meets about 20 percent of metro Sao Paulo's water needs. Its catchment area occupies about 639 square kilometers and is home to 622,000 people, 120,000 of whom live in more than 180 squatter settlements with densities sometimes surpassing 500 inhabitants per hectare. In addition, there are 119 loteamentos housing more than 100,000 people in this area. 4.28 The project's initial objective was to protect and improve the water quality of the Guarapiranga watershed. However, a 1975 law protecting water catchments in Sao Paulo severely limited the use of land surrounding the reservoir and drove down the cost of land in this area, providing an incentive for informal occupation of land through both invasion and illegal subdivision. Combined with weak enforcement of existing land use and subdivision laws, the population of the area exploded. The- area was settled by mostly poor migrants seeking industrial jobs in Sao Paulo. By the late 1980s, the environmental impacts of uncontrolled occupation, including direct discharge of untreated sewage into the reservoir, spurred the state of Sao Paulo to take corrective actions. The urbanization of the human settlements of the Guarapiranga catchment area is only one of several components of a large and complex environmental improvement project. It represents about half of the total investment of US$340 million. 4.29 The results of the project are similar to those of the Favela-Bairro project. Approximately 250,000 people or 57,000 families benefited from the upgrading program; 139 favelas were upgraded as were an additional 200 loteamentos.30 About 7 percent of households were relocated. The average investment per family was about US$4,860, which is at the higher end of upgrading costs. A recent study of several upgrading projects in Brazil indicates that per family investment cost ranges from a low of US$510 to a high of US$6,400, or an average cost of US$3,400. This covers investment as well as planning, management, and other related expenditures. 4.30 Assessment. The Guarapiranga program lacks a cost recovery strategy and is likely to suffer from the same problem of financial unsustainability as does the Favela/Bairro program. According to several analyses that have been conducted in post-construction areas of the project, project beneficiaries were willing to pay user fees, local taxes, and tariffs for water and electricity consumption, but the rules concerning installation of meters, billing, and collection are not well defined. The collection of tariffs for water, for example, varies from one locality to another, with some families paying a flat rate, others paying metered consumption, and still others paying nothing. Defining a clear cost recovery strategy should be a high priority for the subsequent phases of this program. 4.31 Cost recovery through property taxation or from bettermnent fees is hampered by the lack of clear titling or alternative tenure arrangements. In favelas, very few households have a title. The situation is further complicated by the 1975 law protecting the water catchment areas which made the regularization of land tenure in favelas and loteamentos impossible. Although the law was amended in 1997 to recognize the legitimacy of land occupation in these areas, the new law did not define the legalization rights but instead delegated this responsibility to the municipalities for resolution later. This complicates an already complex process. It also adds to the feeling of 30 These figures are only for the municipality of Sao Paulo and do not include those favelas and loteamentos upgraded in other municipalities with the assistance of CDHU. Page 44 Brazil: Progressive Low Income Housing insecurity among residents, with adverse impact on cost recovery, on community attitude towards participation, and on the effective maintenance of investments. 4.32 Affordability/Effective Demand. As mentioned earlier, the Guarapiranga project was aimed at the environmental improvement and protection of the Gurapiranga basin. While the improvement of the informal settlements was included in the program, the size of this component was not very high. Original per family investment was estimated at US$1,800, which included the costs of infrastructure, resettlement, and a program of environmental education. When the focus changed from a traditional water supply, sanitation and drainage project to an integrated water basin management program that includes urban upgrading, the per family cost increased to about US$4,500. Although this falls at the high end of the cost range, this is about the level to be expected in a project that requires significant investments in drainage, erosion control, risk mitigation, and resettlement of families in high-risk areas. Still, the lack of cost recovery or subsidy policy does not augur cost-effective investments. 4.33 Targeting. More than 70 percent of the families in the basin area earn five or fewer minimum wages (Table 29). That percentage varies from one settlement to another and is likely to be much higher when only the households in favelas and loteamentos are considered. The population in these settlements makes up about 60 percent of the basin's population. Table 29. Distribution of Household Heads by Income Level District within the Income bracket Guarapiranga Basin (by no. of minimum wages) Up to 2 2-5 5-10 10-20 Over20 Undeclared CidadeDutra 23.13% 28.8 21.04% 9.47% 3.52% 0.54% Jardim Angela 27.32% 45.3 15.32% 2.70% 0.57% 0.75% Socorro 16.75% 16.2% 24.31% 18.29% 9.99% 0.93% Jardim Slo Luis 23.14% 40.72% 2.01% 6.83% 1.83% 0.62% Parelheiros 33.78% 42.76% 13.39% 3.27% 0.79% 0.99% Marsilac 41.71% 34.91% 11.52% 3.40% 1.18% 0.97% Total Basin Population 31% 409%v 18% 7% 3% 1% Source: IBGE 1991 census. 4.34 Like most upgrading programs, the Guarapiranga project is geographically targeted. An issue that arose during implementation was the potentially large number of relatively higher income families benefiting from this program. This, in itself, is really not a problem because on the one hand the geographic targeting will improve negative externalities for the public good, and on the other hand cost could be recovered with a sliding subsidy that decreases as household income increases. Indeed, this program has developed and utilized good planning tools that could be used to adequately target subsidies at the poorest households. Under this program, a highly sophisticated participatory planning mechanism was developed, which includes the use of socio- economic household surveys and geographic informational mapping, that allows for the poorest households to be accurately identified, and other important socio-economic, legal (such as the land tenure situation), and physical information (such as the level of services) to be collected. 4.35 For example, a map developed for one loteamento shows that 100 families or about half of the population earn greater than four minimum wages. Similar maps of the area identified that only about 12 percent of all families had regular titles to their property. A legitimate criticism of the Guarapiranga project is its failure to use such information based on which it could have developed and implemented sound cost recovery and targeted subsidy policies. Brazil: Progressive Low Income Housing Page 45 4.36 Institutional Appropriateness: The Guarapiranga program has an excellent institutional structure. Although originally designed to be managed through the state water company, the. complexity of the project warranted a different and broader institutional arrangement. The result was a horizontal institutional structure that consisted of a small project management unit and four separate executing agencies, each responsible for its own components of the project: SABESP for larger water and sanitation issues, SMA for environmental matters, CDHU for the urbanization of areas outside of the municipality of Sao Paulo, and PMSP for the urbanization of areas within the municipalities. The former program manager for the municipality of Sao Paulo credits this flat structure for allowing different levels of government to continue working in harmony, even when there were occasional political or other problems among them. 4.37 The involvement of the private sector was also very important for the future sustainability of the program. A critical mass of urban design and engineering professionals has developed the capacity to provide assistance in these types of projects on a large scale. Private institutions have developed social assessment and participation methodologies used during program implementation. In addition, construction firms that had never worked on these types of projects have now developed new lines of business and specialization. 4.38 Conclusions. The Guarapiranga project, like Favela/Bairro, is considered a prototype of urbanization programs. First, it uses geographic information systems, social assessments, and social outreach, which help improve the existing system of regularizing land tenure, develop a cost recovery strategy, and target lowest income households. The data collected from these instruments can also be used to monitor the long-term impacts of the program. Second, the program is adapted to the low-income housing environment. The institutional structure has allowed for an integrated approach to catchment area planning, superseding the previous sectoral approach to which illegal invasion of the area could be partly attributed. The entire basin area will be developed by municipalities based on established criteria to determine best uses, allowing for the orderly densification and development of the area. In addition, there is movement toward facilitating the legalization of occupied lots which is critical for cost recovery and hence the long- term sustainability of the program. 4.39 One of the lessons of the program is the need to improve social outreach and to promote community participation, not only during the planning phase but also during execution. In addition, there were no provisions in the program to continue providing social outreach after the completion of works. Such post-construction participation is critical to the long-term maintenance of investments and to consolidating social capital in the areas. 4.40 Cingapura represents an even more ambitious program for addressing the housing needs of the poor in the City of Sao Paulo. In contrast to the Guarapiranga model, the Cingapura program promotes the clearance of slums and replacing them with medium-rise apartment buildings in situ. In this regard, the program differs significantly from earlier generation slum clearance programs by keeping the slum communities in the same location, rather than moving them to new sites along the urban periphery. The Housing Secretariat refers to this process as "verticalization." In addition, the Cingapura program includes a substantial component that involves legalizing informal settlements or irregular lotes. In recent years, the Cingapura program has also evolved to include a combination of the verticalization approach with the upgrading approach used in the Guarapiranga program. This change came about since many families have expressed their preference for self-built housing with improved services rather than new apartment buildings. The irony is that poorer families have become the recipients of higher cost apartment units, while the relatively wealthier families remain in the newly urbanized but less formal settlements with strong informal economies and home-based businesses (Box 7). Page 46 Brazil: Progressive Low Income Housing Box 7. Cingapura The Cingapura program is at the opposite end of the upgrading spectrum, and pays little attention to questions of affordability and cost minimization. Since the program's inception in the early 1990s, a total of 16,700 apartment units have been built. The vast majority of these units are two-bedroom apartments (15,442), followed by three-bedroom units (51 1). Unit costs are high, considering that most recipients earn fewer than three minimum wages and many are informal sector workers. The average cost of a one- bedroom apartment is about R$38,000 while a three-bedroom apartment costs R$47,000 on average. Costs of the units are partially recovered. Beneficiary families receive a direct subsidy of about 50% of the cost of the unit, and are expected to finance the remainder under favorable interest rates (4% p.a.) over a period of 25 years. In addition to the direct costs, indirect costs include an extensive social component that extends for at least one year after occupation of the unit by the family. According to those involved in the program, the extensive social work is required for several reasons. This includes, inter alia, working with households on rules governing living in a communal situation like a condominium, paying for urban services, conflict resolution, etc. In addition, families must live in temporary shelters for long periods of time while the permanent units are being constructed. The social costs can run as high 3% of the unit construction costs. The Cingapura project is a costly government program that is clearly not affordable to its intended beneficiaries and will likely have problems of sustainability in the long term. Table 30. Belo Horizonte - % People Living in Favelas Administrative Belo Horizonte Favela-Vila % in regions Total population population Favela Bariero 237,089 22,016 9.29 Centro Sul 255,271 55,064 21.57 Leste 243,120 41,509 17.07 Nordeste 250,149 27,618 11.04 Noroeste 335,827 38,715 11.53 Norte 175,653 20,590 11.72 Oeste 252,288 71,671 28.41 Pampulha 120,916 6,383 5.28 Venda Nova 217,864 28,701 13.17 2,088,177 312,267 14.95 Sources: Censo DemogrAfico de Minas Gerais 1992 and 1996 Contagem Populacional, IBGE 1996. Belo Horizonte-Participatory Planning and Global Development Plans 4.41 Belo Horizonte, the capital of Minas Gerais, has a population of 2.2 million, of whom 21 percent (469,000 people) live in Special Zones of Social Interest (Zonas Especial de Interesse Social or ZEIS). ZEIS includefavelas and irregular settlements (vilas), as well as special interest housing zones where low-cost housing is situated.31 The state of Minas Gerais has been going through difficult financial times, thus the responsibility for the development of housing policy and all upgrading programs rests with the municipality of Belo Horizonte, more specifically the housing company called Companhia Urbanizadora de Belo Horizonte or URBEL. According to URBEL, the housing deficit in Belo Horizonte is about 161,000 units-a quantitative deficit of 50,000 units, and a qualitative deficit of 110,920 units. About 312,000 people live infavelas or vilas (Table 30). 31 These are more commonly known as loteamento irregulares or clandestinos in other parts of the country. Brazil: Progressive Low Income Housing Page 47 4.42 The case of Belo Horizonte is interesting because of the mechanisms the municipality has put in place to provide incremental housing investments, or to provide new housing options to attend to existing demands. Of these, the most important are the participatory budget process and the policy for housing interventions in the ZEIS zones. 4.43 Since 1993, Belo Horizonte has invested about R$30 million in infrastructure in vilas. (Favela/Bairro by comparison invested more than US$300 million over the same time.) The municipality has developed an efficient system of investing limited resources in progressive housing and infrastructure in the ZEIS areas in response to short-, medium-, and long-term investment needs as spelt out in the global plans, Planos Globais. These plans feed into the participatory budgeting mechanism known as Or,amento Participativo (OP) that guides municipal investments in Belo Horizonte. 4.44 Short-term investments in vilas are small-scale, i.e., less than R$150,000. Prioritizing is done through an open forum known as the Fora de Manutencdo das Regionais, which brings together municipal and state agencies working with the communities. Generally, priority is given to the maintenance of existing infrastructure and improved delivery of services such as solid waste collection. These works are overseen by the Diretoria de Manutencdo, a division of URBEL responsible for managing short-term investments. These investments are derived from a broader exercise that involves a characterization of living conditions in the favelas in each region, and ranking them using such criteria as topographic and risk conditions, and availability and level of infrastructure (access roads, sewerage, solid waste collection) to identify which ones require most urgent investments. 4.45 Medium- and long-term investments greater than R$150,000 are decided through Or,amento Participativo. For investments in housing and infrastructure for low-income families, two modalities are followed: Or,amento Participativo Regional (OPR) and Or,amento Participativo Habitacdo (OPH). 4.46 For investments in vilas and favelas, the municipality prepares the Plano Global, the instrument used to prepare integrated development plans for all informal settlements in the city. The Plano Global provides a blueprint of priority investments in housing and infrastructure, which serves as the basis for urbanization works, social programs, land tenure regularization, and resettlement whenever necessary (Box 8). Rather than undertaking a complete upgrading approach all at once-as in Rio de Janeiro and Sao Paulo-the Plano Global outlines the actions needed to urbanize the vilas and favelas, and the program of investments needed to enable the progressive improvement of these areas. 4.47 There are currently 13 Planos Globais completed, and 39 more are under preparation. For the years 2001-2002, R$17,156,000 (US$8 million) has been allocated through the OPR for the urbanization of favelas and vilas, which also includes the preparation of Pianos Globais in areas where they do not yet exist. Of investments being made in infrastructure under the OPR process, investments in the urbanization of favelas or vilas represent the second largest share of the total (24 percent), followed by infrastructure investments (47 percent). All investments made under the OPR are oriented to areas that have the lowest values of Quality of Urban Life Index, which measures deficiencies in urban infrastructure. Page 48 Brazil: Progressive Low Income Housing , ' . , Box & Piojeto lIntegrado da Vila Senhor Dos Passos The Projeto Integrado da Vi/a Senhor Dos Passos is a typical integrated project following the Plano ;3lobal concept. Like all informal settlements undergoing the process of the Plano Global, the case of Vila -SehorDos Possos in,volves three areas of analysis and proposals: (a) the regularization of land tenure; (b), - social and organizatioial evelopment; and (c) urban and environmental recuperation of the area. ,The-' 'inclusion,of social actions in-the,urbanization process appears to be an emerging trend in most upgrading ', prograns-at the local level in Brazil.32 VildaSenhor Dos Passos is located'in the northeast region of the city. It has a moderate density of 307' inhabitantsper hectare. The total population is 3,800 persons making-up 1,060 households. M6st,of the_ l Iand,that is occupied is public land; nearly 30% of the occupied land is private. Occupation of the area,' bega,n nearly 80 years ago., The Parno Global began with a diagnostic phase. Based,on the results of the study, Vila,Senhor is, classified.as tle 130 most vulnerable ofthe 77 planning units' in the, various regions of the city. The: dia gnostic phase evaluated a number of variables, includin, the land tenure situation, environmental health . hand natural risk conditions,-adequacy of vehicular and pedestrian access, and social conditions, based on, which a pliysical impro'vement plan was developed for the area..I addition, information collected regarding socio-economic conditions was shared with other municipal agencies so that they cbuld properly, target social development programs at deserving areas. The total physical investments necessary to urbanize the area, were estimated to be $R5.776. niillion, or R$5,449-per,household (approaimately US$2,720/household). This inclpded investments in risk m'iitigation,o 'water and sanitation, drainage, access roads and other infrastructure, as well as new housing for 88families that would. require relocation du,e to risk. A three-phased approach was'proposed for completing th,e urbanization process. priopotizihg the most important investments for the first, second, and third phases, iaccordingly. pri t ms ih ' Co$t recovery to soineextent is-being considered. For examplp, those families tbat will b resettled are. required,to pay-for- die, purchaseof their 'now apartments. Subsidy will be:provided, but rates vary accordingv to household'income. Iniaddition, in cases where the municipality is the owner of the occupiedfland, the land will be sold to the household-occupants in the process of regularizing land te'nure. Still, investrnent' costsjin infrastructure improvement,will not be recovered. The use, of betterment levies are not being contemplated, and the use of the property tax (IPTtJ) is not a viable option given the exemptions granted byl law following the regularization process and those granted following ZEIS designationl. 4.48 Finished housing, mostly self-built, is another plank of the housing policy under the Orgamento Participativo Habitacao process. The most common method of housing construction under this process is that commonly referred to in Brazil as muttrio. In this process, families are expected to contribute "sweat equity" toward the purchase of their unit by volunteering labor during house construction, which brings the development costs down. The mutirdos are assisted by professional engineering firms in the construction of their units. Total development costs under this method range from R$l 1,000 to R$17,000. Financinig is arranged through the CEF on terms similar to that of the Carta de Credito program. 4.49 Since 1997, 567 units have been developed under this process, 192 additionial units are under construction, and 488 more units are in the pipeline for 2001-2002. Resources under the Or,amento Participativo Habitacao come from cooperatives or associations of families (nucleos) that are registered with the municipality as families without housing (sem casa). There are currently 179 such nucleos, comprising 32,260 families. 32 The second phase of Favela/Bairro, for example, Is also using the process of geographiic targeting for physical upgrading, including the diagnostic phase that is an integral part of this process, as an Instrumiient for targeting social programis Brazil: Progressive Low Income Housing Page 49 1' 4.50 The process of resource allocation and of the urbanization of favelas is transparent and participatory. Prior to the start of the budgeting process, the Municipal Housing Council defines the potential investments under the various housing programs for the given year. The Council includes representatives from associations of potential moradores, cooperatives (Movimiento Popular por Moradia), and labor unions, among others. Meetings are held throughout the municipal regions for the various nucleos on the agency's waiting lists. From there the process moves to the formation of the Municipal Housing Forum which deliberates over the priorities for the allocation of resources and the number of families to be benefited, eventually deciding on which nucleos will be financed for the current budget year. For the budget year 2000-2001, the Municipal Housing Forum included 488 delegates, representing 179 nucleos. 4.51 The Belo Horizonte model is interesting for several reasons. On the one hand it is relatively large, but given its limited resources, the most the municipality has been able to do is to ease the housing deficit. The model is efficient and transparent; targeting is done reasonably well, especially in the case of urbanization of favelas and vilas. In this case, priority is given to poorer households in which the risk of social exclusion is highest, identified on the basis of extensive social exclusion maps developed for the entire municipality, its regions, and subregions. Figure 7: Participatory Planning Process for Finished Housing OP Habitacdo Deliberacoes do Conslho Municipal de Habitacao Defines investment programs in finished housing for the given budget period (20 members) I [ Cataloguing of Associations and Cooperatives 1 (179 in 2001/2001) l Reuni&o Regional Preparao6ria On a regional basis, information is provided and directives defined and disseminated to the groups who then elect delegates to the Municipal Forum I Fdrum Munwipal de Habitacdo Reviews the priorities for housing investments and determines which of the cooperatives and associations will be attended to during the budget year. (488 Delegates in 2001/2002 representing 32,000 families) 4.52 The cost recovery and subsidy policy aspects of the programs are not completely developed, though there has been some effort to address this weakness in the program. With finished housing, cost recovery is done through the traditional mortgage programs of the CEF, and includes substantial subsidies in the financing. Subsidies in this case generally decline as incomes rise. The upgrading programs, like most others, are less clear on the cost recovery and Page 50 Brazil: Progressive Low Income Housing subsidy policy, but the mechanisms are in place to improve that situation, although this may require changes in policy regarding exemptions to the property tax in ZEIS-designated areas, or the institution, if legally feasible, of a program of betterment levies (contribuiqdo de melhorias). 4.53 The institutional mechanisms for the programs appear well designed. The participatory budgeting process, which requires the involvement of a number of state and municipal agencies, makes investment decision-making transparent. In addition, the use of novel procedures, such as the phasing of work over time according to priorities and based on clear investment criteria, ensure the efficient use of resources and responsiveness to actual demands by target beneficiaries. The process seeks to minimize investment costs in order to increase affordability. Likewise, innovative practices like the mutirdo hold down costs and are likely to enhance cost recovery of finished housing. Recife-PREZEIS and Metropolitan Cooperation 4.54 Among the major metropolitan regions, Recife has the highest proportion of its total population living in informal settlements. It is in Recife where ZEIS, as a land use planning tool, was first used. The metropolitan region of Recife (Regido Metropolitana do Recife or RMR) is situated in the state of Pernambuco in the northeast of Brazil. With a total population of around 3.0 million, Recife is the fourth most populous city in Brazil. Like many urban centers, the RMR is a study in contrasts. While 76 percent of the total income is produced in the RMR, it is also characterized by acute poverty. Over 40 percent of the regional population, for example, live in favelas or other informal settlements. There are more than 600 such settlements, mostly in the municipalities of Recife and Olinda. The vast majority of the region's poor live in these areas, including 30 percent of the state's extremely poor population. 4.55 A 1996 study of 50 informal settlements in the RMR, housing 4,100 families, highlights their poor environmental, social, and economic conditions. As with most informal urban settlements, the population of these areas is predominantly young. In the RMR, about 61 percent of the total favela population fall in the age group 25 years or younger, with close to 20 percent aged 0-6 years. Of the school-age population (26 percent), a high percentage-nearly 19 percent, according to a 1996 survey-does not attend school. In addition, most families are poor, with over 40 percent earning less than two minimum wages (less than US$1 per day per person). In the case of the Beberibe River Basin which has the highest concentration of informal settlements in the RMR, over 50 percent of the families earn less than two minimum wages; this goes up to over 60 percent in some areas like the municipality of Olinda.33 4.56 The most urgent issue in the informal settlements is the poor level of urban services. While 80 percent of the population is connected to the city's water supply system, poor technical connections and a generally low level of service citywide means that in effect only about 24 percent receives piped water on a daily basis. The vast majority gets piped water only every other day. The sanitation conditions are even worse. Only 8 percent of households are connected to a sewage network. Thirty-three percent have septic tanks, but this is not an acceptable solution in high-density urban areas. Direct discharge into streets, canals, and other public areas is more common. Most trash goes uncollected, often due to a lack of adequate access for service vehicles. According to a 1996 survey, 70 percent of households either dump their trash directly into nearby empty areas or into open waterways. These conditions lead to high incidences 33 This is based on the social assessment conducted for the PROMETROPOLE project. The focus of the first phase of the project is Beberibe River Basin. Brazil: Progressive Low Income Housing Page 51 of waterbome diseases; destabilization of hillsides, and greater occurrence of flooding due to clogged waterways. 4.57 Land use and zoning laws are the principal instruments used by urban planning officials to guide and facilitate the efficient development of urban land. In reality, such laws often have a distorting effect on land markets and can actually contribute to a substantially less than optimal distribution of land for various uses. In Brazil, for example, much has been written about excessive urban land use regulations as the major contributing factors to the growth of favelas and other forms -of unauthorized housing development. Such regulations, combined with high rates of poverty, tend to limit the supply of developable land and increase development costs. 4.58 As in most counties in Latin America, Brazilian policies toward illegal settlements are described either as negligent (with the occasional introduction of services when politically expedient or necessitated by emergency), or as a program of systematic slum clearance replaced by large social housing projects. Yet illegal occupation of land continues. It was not until the 1980s that most municipalities began to include informal settlements on planning maps, at which time these areas were designated as "provisional settlements." 4.59 A further change was introduced with new municipal urban planning instruments such as ZEIS (Zonas Especiais de Interesse Social, Special Zones of Social Interest) or Areas de Especial Interesse Social (AEIS). These instruments enable municipalities to establish new urban parameters for designated areas of human settlements, most notably favelas and other unauthorized settlements typically occupied by lower income households. The ZEIS objectives allowed for: * formerly unrecognized, illegally occupied settlements to be officially recognized and legitimized; * the introduction of urban services and infrastructure; * a decrease in the disparity and inequity of the level of services and infrastructure supplied to the low-income, informal sector vis-a-vis the formal sector; * planning and investment to be introduced in these areas, requiring strong community participation; * the potential to increase municipal tax collection in areas going through the regularization process; and . an increase in the supply of urbanized land and land markets for the urban poor. 4.60 As mentioned above, the municipality of Recife was the first to introduce the ZEIS instrument into its arsenal of planning tools. In 1983 Lei de Uso e Ocupa,do do Solo da Cidade was passed, granting legal recognition to informal settlements as an integral part of the urban structure. However, it was the 1987 law creating the Plan de Regularizacdo das ZEIS (PREZEIS) which established the mechanisms for urbanizing these settlements, and for fostering community participation in municipal decision-making. 4.61 PREZEIS is notable for being the first program to institutionalize the process of integrating irregular settlements into the formal urban planning-process and setting the stage for similar processes and programs developed in other municipalities. PREZEIS created the Urbanization and Land Tenure Legalization Commission (Comiss6es de Urbaniza,cio e Legaliza,do da Terra or COMUL), which is credited' with identifying and addressing specific problems in each ZEIS. There are currently about 70 irregular settlements declared as ZEIS areas in Recife. Page 52 Brazil: Progressive Low Income Housing 4.62 The COMUL is guided through a consultative mechanism known as the F6rum PREZEIS, which is composed of representatives of the communities, NGOs, and the municipality. To facilitate the consultative process, each F6rum has three councils: Urbanization, Legalization (of land tenure) and Finance. In 1993, a law creating the FUNDO PREZEIS, which funds the PREZEIS program including its operating and investment costs, was approved. The law stipulates that 1.2 percent of municipal tax collections be earmarked for the program, although the actual budget of the PREZEIS program is much lower.34 Compared with municipalities such as Sao Paulo, Belo Horizonte, and Rio de Janeiro, investments in ZEIS areas have been minimal. Between 1994 and 1997, investments under the program was only about R$4.0 million.35 The lack of resources has hampered program expansion. Out of 200 settlements, only 70 have ZEIS designation and institutional structure in place. 4.63 The State of Pemambuco, however, is now working with the municipalities to secure additional resources to scale up the Recife PREZEIS into a metropolitanwide program. Called PROMETROPOLE (Programa de Infra-estrutura em Aneus de Baixa Renda du RMR), the program proposes a first large-scale, coordinated approach to addressing the shelter and service needs of the urban poor in the RMR. If implemented, PROMETROPOLE will include the following important new modalities that could serve as a model of regional cooperation: * Metropolitan outlook. The Fundagdo de Desenvolvimento da Regido Metropolitana do Recife (FIDEM) will be the state agency in charge of metropolitan project coordination and planning. The program of institutional capacity building will strengthen its ability to operate as a metropolitan planning agency. FIDEM will provide technical assistance to the municipalities in developing policies to deal with the problems of shelter and the poor. * Decentralizing planning and execution of local works to municipalities. The municipalities will be the executors and will share in project financing, while the state will discharge oversight, policy advice, financing, and monitoring functions. The program of institutional capacity building will ensure the municipalities' readiness to participate in the project. FIDEM will enable municipalities other than Recife, such as Olinda, to take on the responsibility for urban planning and coordination, rather than allowing them to continue to rely on individual, uncoordinated, sectoral investments, as has been the norm. * Ensuring true community participation. PROMETROPOLE will expand the PREZEIS approach to the metropolitan level, incorporating the innovative and effective instruments of community participation that were successfully employed by the PREZEIS program. * Real estate analysis and definition of appropriate shelter options. The scope of projects to be undertaken by the municipalities depends on the demand for housing, land, urban services, as well as on the financing and cost recovery schemes available. Because the needs of larger cities like Olinda and Recife are very different from those of smaller cities such as Itamaraca, Igarassu, and Ipojuca, a real estate analysis would help municipalities understand the level and pattern of demand and to develop appropriate policy interventions. In larger cities, for instance, the trend is toward the densification of existing slums, hence the need to promote their upgrading. In smaller cities which are experiencing faster growth rates, the response might include facilitating the provision of access to land and services in order to avoid slum formation. 34 Uma Politica Inovadora de Urbanizacao no Recife: 10 Anos do PREZEIS. 35 Rolink 2000. Brazil: Progressive Low Income Housing Page 53 * Coordinating sector investments. For this, some of the main instruments which FIDEM can use are: (i) the Intersector Intergovernmental Advisory Committee created by the project consisting of representatives from COMPESA (the state water company), EMHAP (the state housing agency), and the municipalities of Olinda and Recife; (ii) the preparation of the Beberibe river basin urban structural plan that would guide the implementation of PROMETROPOLE and future urban investments in the area; (iii) the development of policy matrices showing what levels of service, investment, subsidy, beneficiary contribution, and tariff should apply for different infrastructure interventions (water supply, sewerage, drainage) at the household level. * Mobilizing resources and seeking private partnerships. Both community organizations and the private profit and nonprofit sectors have much to contribute to: (i) a more accurate definition of people's needs and aspirations; (ii) better targeting of beneficiaries; (iii) a more efficient use of resources; and (iv) promoting change in behavior with respect to maintenance of infrastructure and public spaces, environmental protection, and cost recovery. The project will encourage community participation in identifying real needs of communities and private sector participation in areas complementary to those supported by the project. It will put together a public education and communication plan to inform the population of the RMR of the objectives of the project and the means by which communities can participate in its implementation. Ceard 4.64 The current administration in the state of Ceara has placed emphasis on developing a sustainable housing policy for its low-income households. According to official estimates, the state's housing deficit is estimated at 190,000 units, the vast majority of which relates to households earning less than 3 minimum wages (Table 31). Table 31. Ceara Housing Deficit by Income Level Quantity Type of Housing Less than three Three to five minimum Total minimum salaries salaries Substandard housing 53,357 1,513 54,870 conditions Doubling-up of families 122,977 12,108 135,085 Total 176,334 13,621 189,955 Source: IBCE 1998-99. 4.65 The state government is planning to launch a major housing study in support of the state Master Plan for the Housing Sector (Piano Diretor para a Area Habitacional), a policy document that will identify lines of action to increase access to housing by the urban poor in the state. In the interim, the state's housing policy for the years 2001-2002 has incorporated the following elements: a) A differentiation between market segments. The current policy distinguishes between market segments. Households earning above 5 minimum salaries are considered bankable, and therefore can obtain housing without state subsidies through regular market mechanisms. For families earning between 3 and 5 minimum wages, the households should bear a greater portion of the cost of investments in housing, land, and Page 54 Brazil: Progressive Low Income Housing infrastructure. Families earning less than 3 minimum wages should, at a minimum, be required to pay for the cost of the housing materials (private goods); the public sector would provide land and infrastructure. In practice, actual cost recovery has rarely been taken seriously, but growing recognition of the need to recover the costs of investment is progress in the right direction. b) Approximating demand. State assistance in housing is designed to be strategic. The interim housing policy focuses efforts in areas where demand for housing is greatest, i.e., in the metropolitan area of Fortaleza, and in cities and regions with higher growth rates and potential for economic development. The policy also promotes the location of housing in areas close to economic opportunities, and the integration of neighborhoods to avoid creating islands of poverty. c) Social Work and Participation. Like other programs, the state policy encourages strong community participation in its programs, and uses the concept of sweat-equity in its collective (mutirdo) housing and individual self-help housing programs. The policy recognizes the need to provide a continuum of social services in all phases of a project, including the pre- and post-construction phases. Employment generation for the lowest income groups is also a high priority of the policy. d) Progressiveness. The policy promotes a progressive approach to housing, making units more affordable to low-income families. It favors small housing units that are easily extendable, building materials that are basic, and construction processes that are easy enough to allow families to construct their own housing with minimal training. It rejects large, multistory, multifamily projects that are unaffordable to the urban poor and that often result in "vertical" slums. e) Sustainability. The policy also advocates for the complete titling of property. This is proposed for both new constructions and for the upgrading of slums, which is another important component of the state's housing policy. The state recognizes the importance of titling for an efficient housing market, allowing property to be more effectively traded, and for tariffs and other taxes such as the IPTU (local property tax) to be collected. C. Summary 4.66 Because federal intervention in low-income housing has remained weak, states and municipalities have filled in through schemes that promote private sector involvement, and that introduce the concepts of collectively built self-help housing, slum upgrading, low-cost technologies, and transparent budgeting processes like orcamento participativo. Yet most Brazilian programs aimed at providing housing opportunities for the urban poor remain too small to have a significant impact on the housing sector as a whole. 4.67 Sustainability. The sustainability of many programs is uncertain. Only a few of them have clear and well-targeted subsidy policies. Most programs emphasize provision of housing or upgrading of the housing stock, yet cost recovery is hardly considered. The latter is hindered by the lack of political will and the difficult process of regularizing land tenure, which in turn inhibits the ability to collect land taxes. 4.68 Targeted Subsidies. At present, most Brazilian states and large municipalities keep information on families requiring housing, allowing for certain products to be targeted at the Brazil: Progressive LoW Income Housing Page 55 appropriate income levels. Targeting appears least effective in slum upgrading in which, in the absence of a clear cost recovery strategy and policy, all families benefit equally without discriminating between income groups. Most subsidies in the Brazilian programs are supply-side subsidies, with emphasis on direct public sector intervention. For upgrading schemes, however, these may well be the most appropriate form of subsidy. The exception seems to be the housing program of the State of Sao Paulo which distributes subsidies in line with the family income. Despite many management problems, this program run by the State of Sao Paulo has been implemented in the last 10 years. 4.69 Affordability. Most programs, with some exceptions like the Cingapura program, are designed to be affordable to the low-income segment of the housing market. Among the programs reviewed, urban upgrading programs are the most affordable, as do progressive housing schemes like those in Ceara, with a unit cost of around US$2,500 including all basic infrastructure, land, and housing. The units are small but are designed to make progressive improvements and expansion possible. Community participation is a key component of all low- income housing programs by which issues of affordability are effectively resolved. Also important in mobilizing support is for beneficiary families to be informed of the true capital and operations and maintenance costs of the various housing options available to them. Such options appear to be well understood in most upgrading and progressive housing schemes reviewed. 4.70 Institutional Appropriateness. States and municipalities are responsible for developing programs targeted at the poorest segments of urban society. As discussed earlier, most federal expenditures (in the form of subsidized program) have benefited the lower middle- and middle- income groups. States and municipalities have limited financial capacity to meet this challenge on their own. The private sector is mostly absent in this segment of the housing market, with the exception of the construction industry which benefits from publicly let contracts. Greater private sector involvement in the lower middle and middle-income housing market in areas such as finance (i.e., microfinance) and development (i.e., of housing projects and serviced lots) would allow the public sector to better direct attention and subsidies to lower income families. 4.71 Sector Objectives. Some innovations have been introduced to overcome local regulatory barriers to affordable housing, but most local land use regulations favor larger and costlier housing and push low-cost housing development to the urban fringe. Land transactions and titling processes are also long and costly and tend to limit the number of informal occupants who can obtain title, particularly in existing urban slums. A big step forward is the legal recognition of slums through programs like ZEIS and the designation of affordable housing districts. Alternatives like usocapido are clearly second best tenure arrangements. Even innovative programs, like the private-public partnership program of density bonuses found in Curitiba, can be cumbersome and time-consuming to administer. More can be done to make the entire housing and real estate sector function more efficiently. Page 56 Brazil: Progressive Low Income Housing 5. Building the Main Blocks of a Strategy 5.1 A strategy to help the Brazilian poor meet their housing needs requires a comprehensive approach and a coordinated intervention of all stakeholders. This chapter focuses on four building blocks that are considered to be necessary to draw a consistent strategy for this sector: (a) understanding the way the low-income real estate market works; (b) organizing the demand for informal housing and converting informal housing needs into effective demand for housing and service; (c) easing supply constraints by increasing the supply of land, urban infrastructure, and housing; and (d) providing a clear and consistent institutional framework. 5.2 These challenges are naturally linked. Increased demand will put pressure on the supply side to loosen constraints through land regularization, construction of basic infrastructure, and provision of access to low-interest loan schemes. Many community groups have also started to achieve more for and by themselves with less direct aid and government action. While improvements in the enabling environment facilitate actions and solutions, housing as a social good still requires the intervention of the public sector. A. Understanding the Way the Informal Sector Works 5.3 As discussed earlier, the informal housing sector operates in ways that are quite opposite from the way the formal sector works (Table 32; also see Box 3 on p. 21). While housing from the for-mal market comes with related infrastructure and services, informal housing settlements obtain them by demanding government assistance directly (government-installed systems) or inairectly (resident-installed systems but with technical and/or financial support from the government). The government's response often depends on political factors, e.g., the level of pressure that it comes under and how close it is to elections. Residents of informal housing settlements and slum dwellers are "vote banks," and political parties understand the importance of delivering such services as water supply or paved roads, and even schools or public markets to them. In fact many illegal settlements have been regularized or provided with essential services by political operatives wishing to establish a constituency. Table 32. Difference Between Formal and Informal Housing Formal housing sector Informal housing sector Development * Acquisition and planning * Occupation process . Infrastructure and services * Transition: building houses themselves, * House construction requesting basic services * Sale and occupation * Consolidation: obtaining titles * Regularization Legal Each stage has its own documentation procedures and requirements recording of permits and agreements Housing finance * House offered as a complete product * Opportunity for progressive or self-built houses * Normally requires a mortgage that carries in stages payment over time * Pay-as-you go system, allowing the resident to * Buyer in the formal sector faces uncertain equity make adjustments according to the family's situation economic situation * The progressive house is all equity 5.4 The government (mainly municipal governments) is also the main financier at the consolidation phase. Although some states, such as Rio de Janeiro, Sao Paulo, and Recife, have large programs for slum upgrading, the responsibility for the regularization of illegal or informal settlements, along with all related expenditures, rest directly with municipalities. Brazil: Progressive Low Income Housing Page 57 B. Converting the Need for Housing to Demand 5.5 With 65 percent of Brazilian households earning less than three minimum wages any strategy to improve housing demand must look for ways to mobilize available domestic resources-household income, labor, and mutual assistance-and orient them toward increased production efforts. Improvement in housing demand can be accomplished in several ways. Residents need assistance in housing (i.e., for design, construction, and public services) and in finance (i.e., through collective and household savings schemes). Likewise technical assistance is an essential requirement in any community-based housing activity. Based on positive experiences in Latin America, notably in Mexico and Brazil, L6pez and Fausto (1966) advocated a four-part program of collaborative actions at the community level: a) increasing savings in mutual savings societies or savings groups and linking a portion of those funds to national housing finance programs as part of a microfinancing strategy; b) creating employment that is related to construction (e.g., materials banks); this is called community-based production; c) using mixed technologies for infrastructure and housing construction; and, d) helping with the administrative and financial management of community-based business. 5.6 Improvement on the supply side requires: (a) administrative reforms to ease regulations (concerning development norms and standards) and facilitate state-level regularization of land, either public or private; (b) innovations in materials and system installation and financing them in a phased manner as more resources become available over time; (c) progressive provision of infrastructure; (d) providing low-cost land; and (e) improving the existing housing stock. Residents' Savings Program 5.7 A residents' savings or mutual assistance program can be an effective instrument of low- income communities to find cost-effective solutions to their housing needs. The funds can be used for land acquisition, and new house construction or incremental improvement. The program requires that residents and participating groups, including private sector interests (both formal and informal), are given status in the process to enable them to negotiate with potential partners. The residents' savings can be pooled into a legal trust, which provides a mechanism to leverage and attract additional sources of financing, to protect against default, and to receive and manage subsidies. This trust can maximize the benefits of individual savings, provide the seed capital for community infrastructure, and bridge loans and end-user finance. Such trust can be set up by an NGO or a community association. 5.8 An example of a mutual assistance scheme is the Mexican credit association called Plan de Ayuda Mutua or MAP, a self-financing program for residents of informal settlements. It is a system of dedicated savings to be used for land, housing, and infrastructure, and is assisted by an NGO or a public agency. Participants contribute a set amount every week over a savings cycle, generally less than 52 weeks to maintain the interest of the participants. Funds are allotted every month (or every week) based on a lottery, in which one participant receives benefits-in the form of construction materials, labor, or down payments on land-equivalent to the sum of savings during the cycle. No cash payments are made to the winners. 5.9 The potential of a mutual assistance or residents' savings program is considerable and needs to be further explored. While it will not be enough to cover all the housing finance needs of low-income communities in Brazil, such programs have the advantage of engaging large groups of community residents (including the very poor), building the community's resource base and Page 58 Brazil: Progressive Low Income Housing actual ownership of the program, as well as stimulating savings and loan activities (Box 9). The approach of combining residents' savings with grants from the national government directed to the poorest households and financing from the states for infrastructure investment has been successfully piloted in Parana and could be scaled up. ~ Do A~ResIents Savig Programs , - In Fortaleza~~razilthNGQ ,e Co ahia. e_ a~i d eard operates ~a' mututal as'sistan'c"e ~pro,gra t is ~t-ias l:rubsed prijectsdb,tinclude-sites- for small businese fur urba servi isuc ~Bii'Id'ng a-hd e agovernment ,nd bi,,eral, alssit feng,les shi4of rium -wagj,.apay,, f the99r,wages for six years.Re-sid entsbuild progressil Microcreditsiftm.'Si for Alosi S.10 Micrfinancepo_ houin isap werfltooletofaciiaeacs ofiac yfmle and short-term4P) an NG -has b tyicl for s aye -6 iio2Sb'4 organied 600amih in`o systes o6f selffinance Ini oniecseth pb1~h~$~Iai parcels of and enough f r thir eds and'sell dff plts to finiance hsd improvement, althoughtheloanscn cases th f s were used to purhe c onstruction of amaterls atr oto n id fe fromnstandardn moertgages in hallcasest it paicipantsrayg thites Microcredit for Housing 5.10 Microfinance for housing is a powerful tool to facilitate access to finance by families that are otherwise excluded from the housing market. Housing mnirofinance is generally small and short-term loans (US$500 to US$5,000 payable over 2 to 5 years), typically for home improvement, although the loans can at times be used to finance the construction of a small core unit. Housing microfinance differs from standard mortgages in that it involves paralegal titles and informal sector incomes; mortgage finance requires a mortgage lien and formal income. 5.11 Housing microfinance is more adapted to the way most poor households build their shelter.s Unlike mortgage finance, which deals with complete units, housing microfinance fits their incremental building process, as short-term loans can fund each progressive step at affordable market terms owing to the small size of the loan. In addition, housing microfinance does not require a formal title, and the shorter loan repayment period suits the poor better given the irregular and unpredictable nature of their income, and hence their aversion to long-term debts. 5.12 Housing microfinance programs have followed two distinct approaches that differ in objectives and terms.3 The first approach, microcredit to housing finance (MlCHF), started with microcredit initiatives for -small and medium enterprises. The objective was to expand economic opportunities for socio-economically marginalized groups, particularly women. Progressively the initiatives expanded into housing credit to respond to client demand. Many microcredit clients operate home-based enterprises, and investments in housing improvements essentially constitute investments in their income-generation. The second approach, shelter advocacy for housing finance programs (SAHF), arose out of an advocacy agenda defending the right -of the poor to equitable access to resources, particularly land and shelter, as well as adequate infrastructure and 36 The bulk of households build their units incrementally over five to 15 years, the process starting with gaining access to land and building a small makeshift unit, and later improving and / or expanding this unit; see Ferguson, Schor and Buchenau 2000. 37 Goldberg 2000. Brazil: Progressive Low Income Housing Page 59 services. It is important to understand these two concepts and their modalities since, in the context of scale-up and policy recommendations, the two systems are not necessarily aligned (Table 33). 5.13 In Brazil, the few schemes of housing microcredit that exist have emerged from the advocacy role of NGOs, which have organized barrios or favelas and other squatter settlements to pressure public authorities for land regularization and services (Box 10). NGOs and civic groups have developed the capacity to offer credit services to their respective constituencies as an extension of their development or charitable activities. There are also community-based organizations that extend small credit for housing and basic services, but microcredit for housing finance is still a novel approach and is not well developed in Brazil. Table 33. Comparing Housing Microfinance Programs Microcredit for Housing Finance (MCHF) Shelter Advocacy to Housing Finance (SAHF) Origin Microcredit programs for small and micro- Advocacy groups for low-income households' enterprises right to access land, shelter, and services Core belief Microcredit is financially viable, and the poor are Shelter is a right, and the poor are entitled to a bankable more equitable redistribution of resources Vision Unconditional access to credit by the poor Equitable access to land and shelter by the poor Objective Facilitate access to credit by low-income Address inequitable resource distribution as it households to improve their living conditions relates to land, infrastructure, services, and shelter through linkage between home-based and the income-generating enterprises Focus Housing construction and home improvements Land and infrastructure Services Microcredit for construction and improvement; Community organization and mobilization for provided some technical assistance land, shelter, and infrastructure acquisition Eligibility * Individual or collective loans * Collective loans requirements * Participation in savings scheme to,develop * Participation in a savings scheme to develop and savings habit and create a reserve against savings habit; deposits are left to the conditions default individuals' ability to pay * Minimum periodic deposits are required for * Collective liability for group default 12-18 months * No land title is required * Co-signatures and collective liability for * Below market rate on subsidized funds: terms individual default are structured according to the terms of the * Legal land title or occupancy right required capital source * Market interest rate on own funds * No other requirements: flexible operation * Below market rate on subsidized fund * Concurrent SME operation * Successful completion of one or more SME loan cycles * Minimum length of residence in the community. Driving Performance-driven: empowering the poor by Process driven: empowering the poor by concern providing credit in a financially sustainable way addressing the structural causes of poverty Performance Financial sustainability criteria Human development criteria Indicators Constraints Access to credit is the constraint and not the cost of Inequity in access to resources is the constraint housing Clients The entrepreneurial poor in the informal sector, The poorest of the poor, with special focus on the with a special focus on women homeless Source: Goldberg 2000. Page 60 Brazil: Progressive Low Income Housing Box 1. H' sin o,Ho Pr n !_Ograms in Fort.alzav, zfl Casaeg& hor. and ? (Programa de'AutctAjuda e Assistencia na asa) are two housing ke-rov 'ent lJ8'aiJ` trg an asdvocy,Iagend J'i990, with the failure of a,nationaI4, 'I'elp' pousingpjrogr aaun'chd three'y e Iars earlie,r, huns9f grassroos organizations 'nd'hou'sing,activtsts m Fortalemanded. ,b,et,t,er access' taland infrastructure,;and construction loans. ,After an intemational. X -onZfSnc on altemattiye housing, the two rogZams -were created. Individual loans are awardedto ~ehb~eis~f eligiblesavings andloan6 as:sociations ,and 're guaranteed by a usufuct rightitote lan/d and oilletiiveliability Peer,group: pressute andIncentives regarding itur access to credit (up to thre s onsecve"oans1'ensure timely rlyep nf l two programsni iffer in their modalities, gradual'sel'f reliance`Casa Meihr requires members t,take on larger loans to make'up -for the phasingdout r6-f~i-ll subsidy compoienn after the firsttwio loans,ihers PAAC mandates anincreasersavins Onehie'achuevements of thesepigmsis to link-their low-income constituencies to formlainail r 'Aft dcisfu tepa enit'of a maximum of thee sequentiial croloans, lwcincome Xapolicantsware theirne cred tworthy''andScan c'iract;h-ir-for loan from formal, fnancial institutions`4i igFortakza uiiiae ket' interest rate t seqssive investments combine individual s vingsj- ink ic' ) :y and '', n-de byan initernedtiay '- GO- 5.14 In other Latin American countries, intermediary organizations have been successfully promoting microfinance for housing schemes that facilitate access to public programs by lower income groups: (a) El Alto in Bolivia channels funding for Mutual la Paz toward home improvement loans through a land regularization program created in 1993 to give people security for a loan. (b) FUSAI in El Salvador provides bridge financing for poor households to help them become eligible for a national subsidy administered by the Fund for Popular Housing. FUSAI assists these households in acquiring the plots they intend to build housing on, in order to be eligible for the FONAVIPO subsidy, which finances building construction. The largest use of FUSAI funds is for housing purposes (e.g., for building materials). (c) Cobijo, a microfinancing initiative in Chile, provides households unable to accumulate resources on their own with the minimum savings amount required to receive state subsidies under the Progressive Housing Program. These low income families participate in a revolving fund and borrow collective loans to meet the savings requirement. (d) In Venezuela, the Fundacion de la Vivienda Popular (FVP), a barrio improvement program, helps organize community groups and channel small amounts of government funding to them. These community groups provide small loans (from US$500 to US$2,000) to households payable over 2 to 5 years in the form of building materials receipts. Families pay according to their ability. Peer pressure for community members to access credit has contributed to the excellent performance of the program and the improved living conditions in the barrio. (e) In Bolivia, BancoSol is now a well renowned microcredit and housing finance institution. BancoSol was established in 1983 by the Promotion for the Development of Microenterprise (PRODEM). After nine years of successful expansion, PRODEM decided to convert its operation into a private commercial bank for microcredit to enable it to access larger capital resources from the national and intemational markets and offer full financial services to its borrowers. During its expansion phase, BancoSol limited its products and focused exclusively on microenterprise loans for the poorest segments of Brazil: Progressive Low Income Housing Page 61 the population. By 1995, BancoSol had a microlending portfolio of more than US$32 million and more than 75,000 borrowers. It has also started offering housing finance in response to the demands of its clients who are ready for larger loans and are gradually moving up the income ladder. Table 341. Institutions and Ter-nms for Housing Microfinance SEWS Bank CARD PAYArAS HPFIUTSHANI Gramneen Genesis (India) (Philippines) SCANEGERS FUND Bank (Guateznala) (Philippines) (South Africa) (Bangladesh Date 1974-76 1986 1993 1990 1976 1988 Institution Type Microfinance Bank Portfolio, % of total assets US$720,000 $414,000. $182K, 7% $5 8 million $2 million, 50% 6% 7% 18% % Housing loans to 100% 1005 985 100% 80% NA Women Avg Weekly client income $24 $13 23 130 Na Na % members w/ land tenure JOA Na 0 40% Na IHousehold per group Average Housing Loan 300 349 526 1459 - 120 Loan Matunty 3-5 years 1 year 6-12 months 0-15 years 0- IO years 1-4 years Interest Rate 17% 205 18% 12% 8% 21% Collateral Group Group Liability Group Liability Group Group Liability LUability Liability Saving amount required 12 months $39 over one $139 group Na Na Na and half year Repayment rate 98 Na 95 98/100 89/93 Source Goldberg 2000. ' . 3~~~~~Iox 111. ~I[mproved l8nnkigi, lr,eiir2a, l8ranil Palmeira is a slum in Cear. with 30,000 inhabitants. In 1997, 90% of the households made less than two minlimum wages (US$148); 80% of the active population was unemployed; and 1,200 children were out of school. Small producers were selling their houses as business was shrinking, and there was no credit available. With the assistance of Ceard Periferia, an NGO, and GTZ-financed PRORENDA the residents association pushed forward the idea of a bank that could provide credit to the producers who could not get credit from any other source. The Palina Bank (authorized by the Central Bank in the form of a credit union) guarantees microcredit for families (up to six-month maturity), charges low interest rates (2-3%), and does not require registration, proof of income, or guarantor. Credits are used to support small businesses, wholesale, production, and housing materials bought in the neighborhood. All manpower is voluntary. The interest rates cover the administrative costs. The bank is managed with the participation of the community and has developed new products, such as a credit card (PalmaCard) to be used in businesses in the neighborhood, guaranteeing that families consume local products and services; a weekly fair to market products from the bank's clients; a job bank and solidarity exchanges of products and services; and Palrnafashion brand created by a group of seamstresses who are clients of the bank. The Palmas Bank began with 10 clients and a R$2,000 loan from Ceard Periferia and small donations from PROPENDA. Technical capacity (all voluntary) comes from Cearl Periferia as well as local producers and business people, and members of the residents association. Another indirect source of resources is intermediation. The bank serves as a guarantor so that its clients can access other systems of credit. Default rate is very low-between 1% and 3%. With two years of experience the bank has 870 clients, and local business sales have increased by 30%, generating 80 new jobs. In addition, sales of commercial real estate increased by 40%; 100 new jobs have been generated; and 400 families now have a PaimaCard, which secures a volume of more than R$8,000 circulating among the small business of the neighborhood. Many families can now get piped water and electricity. Recent research on the Palmas Bank, , shows that 80% of the clients increased self--esteem. Page 62 Brazil: Progressive Low Income Housing 5.15 Challenges Facing Housing Microfinance in Brazil. At present, there is no housing inicrofinance industry in Brazil (Box 12). In 1999, the Central Bank established a microenterprise credit society intended to provide NGOs with a formnal means of organizing themselves (Table 35). Under the broad oversight of the Central Bank, these institutions are not allowed to offer consumer loans or housing finance. Even participating NGOs already involved in addressing housing-related issues (e.g., seeking security of tenure for the poor through land regularization) as pait of their humanitarian mission cannot engage in activities related to housing microfinance. To maximtze the impact of these NGOs, the current restrictions on their administration of consumer and housing loans should be lifted. Box 12. Microcredit in Brazil Brazil lags behind most other countries in Latin America in the development of its microfinance industry despite haiving a large and dynamic microenterprise sector. This is surprising since Recife saw one of the first microfinance programs in Latin America, the Project Uno funded in 1973. Some of the constraints to the development of Brazil's microcredit's systems are: (i) the countuy's macroeconomic instability until fairly recently; (ii) the crowding-out effect of subsidized government credit lines; (iii) fairly successful and widespread consumer credit; and (iv) legal and regulatory impediments faced by nonbank institutions. Brazil has an estimated 9.5 million microenterprises, mostly concentrated in retail sales, etc. Of these less than 20% have access to credit, of which only 2% have access to banking credit Brazil's financial sector is well developed, with more than 1,900 institutions and 223 banks, and a density of bank/population around 10,000, which is about average for Latin America. The Northeast is the least- , served region, with the BNB accounting for 86% of all bank financing in the region. Public banks tend to be specialized: BNDES is a second-tier bank that provides credit at preferential rates in line with , (levelopment goals: CEF is the housing and savings bank; and Banco do Brasil concentrated in agricultural' credit. Credit to private sector is about 31% of GDP, which is the average for Latin America. Public banks account for 46% of the bank branches and 55% of financial assets; 26% of the private credit goes into' housing despite tthe size of the CEF. Microcredit programs have been operating since 1997. The most important are Viva Credit (RJ), PortoSoI (Porto Alegi e), CrediAmnigo (Banco do Northeast), and FENAPE (UNICEF, IDB). Of these, only CrediAnrgo (sponsored by the World Bank) is a regulated financial institution with capacity to receive deposits and savings. BNDES has a technical assistance program to support the development of mnicrocredit by means of strategic project, institutional strengthening, governance; regulation, 'establishment of microcredit institutions, development of new products, and investment in market infrastructure. Further development of microcredit would need several changes in regulations to facilitate the operations of microcredit institutions, notably in terms of minimal capital requirements, specific supervisory guidelines for microfinance portfolios, loan provision requirements, capital adequacy requirements, non-, traditional collateral loan documentation, and operational restrictions. In 1999 the Central Bank issued ' Central Bank Resolution No 002627, which established a new type of institution recognized by the Central, Bank. That institution is a microenterprise credit society intended to provide NGOs with a more formnal means of organizing tbemselves within the broad oversight of the Central Bank. Some of the rules are: (i) they must focus onily in microfinance; (ii) they must have a minimum capital of R$100,000; (iii) they cannot call themselves banks or lend more than R$10,000 to a client; (iv) they cannot offer consumer loans, make or receive interbank deposits, or capture savings from the market and accept saving deposits; and most important for the purposes of this discussion, (v) they cannot finance housing. 5.16 The second challenge relates to the need to develop a framework that is in keeping with the principle of helping the poor to help themselves and within which housing-related microcredit provision should operate Priority areas to which microcredit support could be provided are land Brazil: Progressive Low Income Housing Page 63 acquisition and infrastructure provision. The biggest challenge'remains how to reach the largest number of the poorest of the urban poor, including squatters on remote peripheral or marginal lands and /or those renters in overcrowded inner-city slum tenements. Table 35. Institutions' Terms for Microfinance in Brazil (excluding Housing) CrediAriigo VivaCredit Porto Sol FENAPE BancoSol (PRODEM) former NGO Date 1997 1997 1996 1987 1976 Institution type Bank Division NGO Civil Association NGO Sources of funds Transfers from Equity, IDB, Grants, GTZ Equity, IDB, Equity, USAID, BNB and IBRD Govt. grants Govt., UNICEF Govt. Commercial loans Estimated market 2.5 nillion 15,000 19,000 6.8 nillion I million No branches 51 3 1 45 42 Portfolio, % of total assets US$720,000.50% $414,000; 6% $ 182K, 7% $5.8 million. 7% Outstanding portfolios $8.2 million $631,000 $1.2 million $7.3 million $67.3 million No loans outstanding 34,000 950 1860 23,300 76,700 Avg. outstanding loan $241 $654 660 $314 $878 % members w/ land tenure 10A Na- 0 40% Na Interest rate (monthly) 5% 4.9% 3.5% 5.5% 4.0% Collateral Group liability Co-signed, proof Co-signed, proof Cosigner or Group liability of salary of salary solidarity Group collateral collateral Repayment rate 96 95 96% 96 95 Source: Schonberg 1998. Linking Informal Demand to National Housing Finance 5.17 While the mobilization of savings is possible through collective association, these savings will not be adequate to address the housing needs unless they are linked with the national housing finance system. In most countries, people living in inadequate housing conditions cannot afford commercial or even publicly funded programs (see Chapter 3 for further discussion of this issue). Government-supported or managed programs like FONHAPO in Mexico do not yet exist in Brazil38 but such initiatives present realistic and affordable financing alternatives. Funds for such programs typically come from federal and state budgets. However, given limited public resources, additional funds from the private sector can be leveraged, be it from developers, promoters, or investors. Another approach is to constitute a revolving fund that will allow state agencies to acquire land for resale to civic associations, with the purchase subject to regularization and basic services provision. This fund could be derived from transaction fees, property tax, or from an international loan with long-term repayment provisions. 5.18 Another innovation is a system that could link small community funds with the programs run by CEF or the federal government. Such combined funds can be used to finance pilot programs around the country. For example, Brazil's largest housing mortgage fund run by CEF under the carta de credito is capitalized through employees contributing 5 percent of their wages. It provides mortgage loans for housing provided by the private construction industry but approved by the fund administrators. Majority of the loans, however, are granted to employees earning more than five minimum wages. In addition, only about 300,000 units can be financed per year, far below the national demand and the funds'own production targets. A similar system 38 FONHAPO is a program aimed at providing assistance to families making two minimum wages or less. It finances the costs of construction and materials only. The program was successful in the beginning, but later came under political pressure and lost its independence in allocating funds to the poorest households. Page 64 Brazil: Progressive Low Income Housing of fund mobilization linked with a federal housing program can be promoted at the community level and later scaled up as more capacity is developed. 5.19 In Mexico, a relatively new type of private housing mortgage system has been established. Called SOFOLEs, these small and flexible institutions that finance below market rate (social interest) mortgages have been very successful in responding to local housing needs. SuCasita, a SOFOLE in Mexico City, obtained house construction financing for a group of 400 families who had collectively purchased a parcel of land payable over a four-year period. With the land secured, SuCasita set up a special reserve fund for families who did not meet the minimum salary requirements and whose income varied by the month. In this manner, SuCasita could establish a credit history and qualify the borrowers. The reserve fund also served as an insurance against default. Su Casita made loans on serviced lots, holding legal title to its project sites in urban zones. 5.20 With some adaptation, similar projects can be developed in the informal sector with federal or state support. Private institutions like the SOFOLEs and state housing agencies or similar bodies can be effective partners in addressing the housing shortage of the poor. On the supply side, state agencies can guarantee regularization of land sites in the short term; on the demand side, a SOFOLE-type institution can manage a mutual assistance savings program to finance progressive or incremental housing projects. C. Improving the Stock 5.21 NGOs. Several NGO experiences have been successful in improving the hundreds of thousands of self-constructed units in the informal sector. In Ciudad Juarez, Mexico, for instance, the Cooperative Housing Foundations (CHF) established a small fund (US$50,000) to finance small improvements. It works with NGOs to manage the program and select participants. The loans are interest-bearing and are used to make direct payments for labor, materials, or improvements outsourced to local contractors. Loans are given based on a borrower's employment history. Participants are not required to hold a legal title to their land or to live in a regularized subdivision. The default rate is very low because of the participants' desire to remain creditworthy and because of good program management. The use of NGOs makes the operation simpler and allows the funds to be disbursed faster and to go directly to house improvement without passing through any official agency. 5.22 Linking to the Industrial Sector. Businesses are increasingly willing to contribute to the improvement of their workers' housing. Several industries have attempted CHF-type programs, set up rental units sometimes with subsidized rents, or established rent-to-own schemes, whereby workers pay rent for a fixed period after which they are given the right to purchase the unit at an attractive price. Financing is arranged through a housing management company. This scheme is particularly attractive to companies trying to keep employees, since workers who leave the job before acquiring purchase rights lose their equity. Other companies offer home improvement loans guaranteed by the companies themselves, or subsidize down payments. 5.23 Materials Banks. So-called "materials banks" can improve access to and affordability of housing, and can lead to greater community involvement in home and infrastructure construction. The idea is to pool community resources to purchase construction materials. Some of the materials can be locally produced; the rest can be purchased in bulk directly from manufacturers at a discount. The materials banks also offer opportunities for individuals to trade work for materials, and they also can be the vehicle for business and technical training. To start operation, materials banks require financial subsidy and technical support. A typical financial mix is 50 Brazil: Progressive Low Income Housing Page 65 percent from community savings, 30 percent from the government, and 20 percent from international funds. There are many examples of successful materials banks, including one in Fortaleza where they have been integrated into a program of mutually assisted housing construction and building for microindustry, and in Pereira, Colombia, where they are part of the state-sponsored housing fund for low-income household. In Pereira, materials are manufactured, built, and sold, and profits are plowed back into the community in the form of service buildings. Land Management and Increasing the Supply of Land 5.24 Land management is a key component of low-income housing policies. The price and supply of land influence the quality of housing, living conditions, affordability of housing services, and economic and infrastructure development. Unlike many other markets, land markets are closely controlled by local and national government policies to preserve the social function of the land and the environment. Well-functioning land markets promote easy entry and facilitate transactions, depending on the availability of adequate land information, secure tenure arrangements, and appropriate registration and recording mechanisms. Nonperforming markets are generally associated with: (a) overcentralization; (b) inflexible and excessive regulations; (c) lack of resources and will to tackle the problem of the housing sector; and (d) administrative systems that lack efficiency, equity, accountability, and probity. These nonperforming markets make it difficult, if not impossible, for the urban poor to participate. Land use laws have strong redistributive effects. This is because such regulation not only affects the price of urban land and therefore its accessibility, but also because it affects the cost of investment in infrastructure.39 Land is normally regulated by laws regarding the use and development of land, ownership, title and transaction, environmental protection, patterns of construction and infrastructure, provision of public services, and contracts. 5.25 The cost of poor land management is very high-illegal settlements, lack of productive investment, and poor infrastructure. In the case of Brazil, it is estimated that in the last 20 years 80 percent of the housing units had been built outside the existing regulatory framework. Part of the reason are rigid urban land regulations geared to enforce zoning and subdivision laws that neglect the spontaneous growth of urban settlements and the needs of the poorest. As a result, informal areas have sprung up, where the poor live without land title, infrastructure, or services. Unclear and often contradictory norms that do not reflect actual housing conditions have led to a shortage of serviced and serviceable land. Moreover, speculation has pushed up the prices of vacant serviced plots, put them out of reach of the poor, and led to speculation even on nonserviced plots. 5.26 Legal City vs. Real City. The case of Brasilia illustrates the negative impact that urban regulations can have on the supply of land. Brasilia has had six urban plans since it was created in 1960. All have been ineffective because of rigid urban regulations aimed at controlling growth. Fifty percent of the land in the Distrito Federal is classified as environmentally protected and not usable for urban development. There were no provisions to accommodate the low- and middle- income households arriving in the capital between 1978 and 1992. The result was sky-high land prices within Brasilia and sprawling informal urban settlements produced by informal developers in a parallel real estate market. The reality of urban expansion meant a judicially based "legal" city juxtaposed with a socially constructed "real" city. It does not have to be this way. The urban plan of Curitiba, for instance, is a good example of how to orient urban development with a flexible approach that responds positively and effectively to urban growth. 39 Mona 1998. Page 66 Brazil: Progressive Low Income Housing 5.27 Land Supply Instruments. In Brazil, two legal instruments largely control the supply of land: Federal Law 6766 of 1979, which regulates land subdivisions, and the municipal Planos Diretores. Federal Law 6766 has been superceded by Law 9785 of 1999. It regulates informal settlements and changes expropriation laws and laws concerning the registration and parceling of the urban land. Many of its provisions have significant implications on the availability of land for the poor, viz.: * Land subdivision for urban purposes is authorized only in urban zones. No rural zone can be parceled for urban use. * Developers should provide full infrastructure before selling any plot. * Construction on flood-prone areas or on steep slopes is prohibited. * Up to 35 percent of urban land must be reserved for urban amenities and infrastructure, including parking and other public spaces. * Lots cannot be smaller than 125 square meters and should have a frontage of at least five meters. 5.28 The Municipal Plano Diretor. Brazil's Constitution mandates the use of urban land - its planning, zoning, subdivision, and development - as the responsibility of the municipality. The municipal urban plan (Plano Diretor) is the main instrument used to guide urban development and expansion, and to ensure that urban property serves a social function. The plan defines the urban development objectives and strategy of the city, the services the city must provide to its population, and the instruments that the city will use to develop urban lands, and bringing vacant lands to productive use for the common good. It is envisaged to provide for the construction or improvement of essential urban infrastructure, including water, sewerage, schools (a primary school within a 15 kilometer radius), roads and transport, and public lighting. 5.29 Brazil's new Constitution. During the process of reforming Brazil's constitution in 1988, urban reformers fought for the inclusion of tools to guarantee the social function of the city. This means that the design of the city should be in line with the needs of its residents, not only for infrastructure and services (transport, schools, etc.) but for the proper use of urban property, as well as the democratization of the city.40 For the first time in Brazil, the constitution included a section on urban policy, or the Estatuto da Cidade (see Box 4). This section has two key provisions: It contains sanctions over the urban property that would not abide to this "social function," that is, the city could expropriate lands that are not used for housing but for speculation, and the usocapido which entitles occupants legal access to the land they have been occupying for five years if no one complains or tries to get the land back. 5.30 New tools of urban policy. There has been renewed interest in the preparation of Planos Diretores especially after 1988 and 1989 when many newly elected mayors expressed keen interest in urban reform and in the democratization of their cities. Since then, several municipalities have experimented with new instruments of urban policy - new forms of zoning and serviced land, the use of a progressive land tax to discourage speculation and to provide additional resources, transferring the right to build on (rather than to own) land, creating environmental protection zones, and controlling land parceling. In addition, there was a clear commitment to broaden the participation of the civil society in the planning and management of the city. 5.31 Some of the most interesting approaches used by Brazilian cities to solve the housing problems of the poor are: 40 Pols 1996. Brazil: Progressive Low Income Housing Page 67 o Subsidized Housing Production. These programs are available for people with incomes below five minimum wages. Program components include serviced land plots, housing units of 25 to 30 square meters which can be extended at a later stage, progressive housing (autoconstru!do), and community self-help housing (mutirdo) managed by the public sector or co-managed with residents associations. The rate of implementation of these programs, however, is generally slow and their production levels too small as to create a dent on the overall housing deficit. There are also financing problems. One way to increase government resources for financing low-income housing is through property tax which can capture part of the increase in the value of land and housing as a result of infrastructure investments. O Regularization of informal settlements and urban upgrading are the most common approaches that municipalities use to improve the living conditions of the low-income population. When adopting these approaches, two main policy issues need to resolved: (a) whether to permit lower standards for land development and/or house construction, and what to do when these standards are in clear conflict with national and local urban legislations; and (b) how to legalize the ownership of invaded land. To deal with the latter, usocapido is often invoked. The bigger problem is when the invasoes are on public and / or environmentally sensitive areas. This gets even more complicated when a legislation prohibiting the conversion of green zones to housing or residential zones exists (e.g. in Sao Paulo). Such regulation has made it impossible to legalize the titling of land in many favelas. o Establishment of ZEIS (Zonas Especiais de Habitagdo de Interesse Social). The creation of ZEIS or especially zoned areas for low-income housing is today a well established practice in many Brazilian municipalities. They present an alternative solution to the housing needs of the poor - with its minimum house size and quality standards -- while adhering to urban laws. Being special zones of social interest, ZEIS are defined in the city urban plans. Housing and families living in these areas are granted waivers of flexible urban and construction codes. it 1993, the mayorof Call announced a project to provide 2&,000 lots of.60 square kilometers each with basic services to be developed by self-construction; and, in a second stage; another 22,500 lots to be developed as multifarnily housing. In January 1993 when the mayor presented the project to the city. c6uncil, the market price of the land was 365 pesos per squam meter, six months later the prices had risn- i6 1,580 pesos per square meter, and one year later tie price was 3,600 pesos per squtxe meter. The prwces rose so fast because of regulations that were too ambitious for the poor- 15% of the towt area for community centers and open spce; 20% for roids; and 15%-20 bfor sidewalks and public utilities. Id i addition, the procedures to buy municipal land were too cumbersome. Eventually the prices of the houses, weme not compatible with families' purchasing power, and the houses were sold to middle-income families. Land Occupation and Regularization 5.32 Infrastructure and Public Services. The provision of infrastructure in slums and squatter settlements has been an effective approach to enabling the poor to house themselves. Some large- scale programs, such as the Kampung Improvement Program in Indonesia, have successfully and efficiently improved roads, water supply, sewerage, and drainage for 3.3 million people over 10 years for an average cost of US$1600 per person. This is remarkably low compared with the average per capita investment of US$3,500 in Favela/Bairo and Baixa Fulminense in Brazil. Page 68 Brazil: Progessive Low Income Housing Infrastructure provision needs to be phased and progressively executed to avoid resettlement or displacement of the low-income population due to gentrification. It also has to be complemented with regularized tenure to encourage further investment in housing improvements. 5.33 Fiscal, legal, and administrative constraints prevent the conventional infrastructure supply, system from servicing much of the irregular housing sector. These constraints render plots more costly and time-consuming to produce. However, it is more important to focus on opening up land for residential development at the outset, rather than on providing infrastructure. Infrastructure and services, like housing, can be built in a progressive manner, making land more affordable to the poor and providing some form of self-screening as only those who prefer partially serviced land would apply. 5.34 This incremental approach to infrastructure provision requires a high level of technical assistance to families and community groups. The adoption of intermediate technology could lower costs and increase the number of serviced plots entering the low-income housing market. Some systems could also be operated fully or partially by private companies under public franchise. Within a framework of progressive infrastructure provision, both the central and local governments could play a key role. The central government could obtain financing for major infrastructure investments required at the municipal and metropolitan levels, and promote more flexible arrangements through political or legislative means. Municipal and state governments could play an active role in promoting area-based approaches involving self-help construction and /or small concessions with local contractors. Local assistance centers could be set up to address the concerns of households at different phases of infrastructure development. These centers could deal with matters ranging from technology options to fiscal management. They could help establish savings programs, coordinate support from other organizations and different levels of government, and expand into other activities as necessary. D. Institutional Arrangements 5.35 Housing has important features that make it a unique and highly sensitive area of urban policy, both economically and politically. It represents a large proportion of national expenditure and credit (about one-third in OECD countries). Housing problems are at the center of many urban development and planning issues, and housing policy requires effective coordination between and among major parties (developers, private households, government, financial institutions). In most OECD countries, housing receives substantial public funds, especially from the central governments which decide on the volume, form, and terms of government assistance for housing. Often several ministries or other central government authorities are involved. In many federal countries, financial aid and housing programs are the responsibility of federal housing councils, which include representatives of central and subnational governments, but implementation is delegated to the local level. Allocation of aid involves a coordination of programming between the upper and the lower tiers of government, which works in slightly different ways in different countries. In Sweden, for example, municipalities must draw up housing programs covering several years, which serve as the basis for the allocation of aid from the national level to the municipal level. In several other countries, metropolitan housing authorities have been created to allocate aid to municipalities. In Germany and Australia, the federal and state governments jointly decide on the allocation to the states; the states are then free to decide how this housing aid should be used. In addition to vertical coordination, OECD countries ensure close coordination between national level authorities (finance, banking, and housing authorities) and the private sector. Brazil: Progressive Low Income Housing Page 69 5.36 In carrying out urban planning and development responsibilities, the OECD experience demonstrates some common pattern. First, while urban planning is essentially local, national authorities intervene in several important ways: they maintain the right to approve land development and use, and they exercise control over the planning instruments and procedures local authorities use, including mandating citizens' participation in planning. Second, central authorities ensure that decentralized urban planning does not lead to a dissipation of effort and excessive competition among local authorities, which may further lead to an overall reduction in national efficiency. Third. the nature and content of urban planning have changed dramatically, moving from its major concern to control land use (defining what could and could not be done) to regional development that takes into account the particular context of a country and / or a region. 5.37 In Brazil, one finds a vacuum of responsibility for housing and urban policies at most government levels. While municipalities have the main responsibility for basic urban infrastructure provision as well as regulating the use of urban land, they also have the direct responsibility for addressing the needs of the poor and dealing with the associated problems of land invasion, informal development, regularization, etc. Municipal officials face considerable political pressure that comes from direct contact with the public and those who have elected them. Many voters live in slums and informal housing settlements, and political parties are aware of the political importance of delivering services and infrastructure improvements in these areas. In fact many illegal settlements have been regularized or provided with essential services either by political operatives attempting to establish a constituency or by the intense mobilization and political pressure informal residents were able to exert. 5.38 There is clearly an urgent need to return some housing and urban responsibilities to the central level, giving it the authority to formulate a sustainable and equitable low-income housing policy, to provide guidance to the subnational governments for defining its overall objectives and targets, and to mobilize resources to support it. Specifically, the government must: (a) evaluate the national needs for housing and basic shelter; (b) select and adopt models that have proved successful in helping the poor improve their housing conditions; (c) estimate the financial implication of such programs; (d) identify mechanisms to mobilize resources, including grants from the central government, local taxes, residents' savings, and microcredit; (e) decide on the modalities of resource allocation (either through the states, directly to municipalities, or through a national urban council, etc.); and (f) prepare self-screening procedures to ensure that the funds are allocated to the best use possible. 5.39 A new institutional framework is required for the government, with its limited resources, to manage the housing sector in a manner that provides adequate and affordable housing over a period of time and in a flexible fashion. This institutional mechanism is needed to: * oversee the functioning of the sector as a whole and coordinate the activities of major public agencies (financing, land registry and titling, public works, planning, etc.) that influence housing sector performance; * collect, analyze, and publish data on the progress of housing sector activities, (including number of housing projects built, public contracts executed, construction costs incurred, etc.), and outcomes of programs especially aimed at the poor, all of which are key to improving the process of housing policy formulation and implementation; * provide a link between housing sector planning and macroeconomic planning; * provide a forum for the participation of the private sector, NGOs, community-based organizations,. and the general public in housing policy formulation at both the national and the municipal levels; and Page 70 Brazil: Progressive Low Income Housing * review the effects of housing regulations and initiate regulatory reforms whenever necessary. 5.40 In addition, other institutional responsibilities need to be addressed: * establishing a regulatory framework for the delivery of housing finance by the private sector (both formal and informal), including microcredit for housing purposes; * administering housing subsidies for the needy; * establishing property rights, especially through tenure regularization in squatter settlements; * providing basic infrastructure in informal settlements; and * bringing together infrastructure agencies to coordinate infrastructure provision and help create an adequate supply of serviced land. E. Conclusions and Financial Framework 5.41 Improving progressive, informal housing markets does require policy changes in the sense of facilitating their functioning. Brazil has several positive features to start with. These markets already work in the sense that they already provide a large number of housing, albeit underserviced and minimally constructed. Most of the people in them pay something for land and for shelter. Moreover, because the informal housing sector is not well-articulated in the national economy, it suffers less from recession than the formal housing market. It therefore serves as a form of a reserve to which low-income groups turn to meet their shelter needs. At this stage of the housing development in Brazil, the main requirement of the informal sector is a more direct govemment support to a strategv that promotes a variety of savings methods, flexible and progressive infrastructure delivery systems, regulatory reform, and community involvement. 5.42 Basic Elements. Ensuring access to housing and services to low-income Brazilians requires a multifaceted program involving the active participation of as many major stakeholders as possible -the government, private sector, parliament, NGOs, universities, and professional organizations. The program must: a) Set the National Housing Agenda. The government should formulate a policy that recognizes housing as a fundamental service for the Brazilian poor and as an effective instrument to improve their assets and social inclusion. This housing policy should be an integral part of a national urban policy. The Estatuto da Cidade (the urban law under consideration) is an excellent framework that defines with clarity the roles of the different levels of government, the policy instruments available to them, and how these instruments can be applied. The central government should take the lead in executing this policy within the framework of the Estatuto da Cidade to ensure an effective response to the housing needs of the poor. b) Define the institutional roles. The role of the public sector in market regulation and the promotion of social policies should be heightened. A National Housing Policy Council should be constituted to oversee policy implementation and preside over a national system for low-income housing management. Among its functions should be to define the specific roles of the central government, states, and municipalities in low-income housing policies; and to guarantee the establishment and operations of housing funds at the three levels of government, including a transparent system of granting credit. Through participatory budgeting, the Housing Council would engage the participation of Brazil: Progressive Low Income Housing Page 71 as many stakeholders' groups as possible in setting priorities for the allocation of resources. At the local level, municipal governments would provide up-front capital subsidies and would identify the priority beneficiary communities, taking into account income levels, savings potential, and ability and willingness to pay. Other sources of funding would also be tapped, including the FGTS, NGOs and community-level associations whose savings can be leveraged, and state agencies that could provide serviced land. Despite the small scale on which many local NGOs and housing associations operate, the long-term impact is potentially huge. In the US, for example, the development of affordable housing by nonprofit organizations took off in the 1990s, producing over 500,000 low-cost housing units from 1990 to 1998. Most of these initiatives began at a very small, unsophisticated scale. * _______________________________ c) Establish clear financing and Table 36. Cos.t Breakdown of Urban subsidy policies: Subsidies Upgrading, Recife, 1998 should have a place in a national Item R$ per % of housing program, provided they household total are transparent, up-front, and Infrastructure well targeted. They can be used Access road 450 6.3 in conjunction with private Drainage 350 4.9 Piped water 300 4.2 savings and affordable loans. Sewerage 750 10.4 Transfers to poor municipalities bectricity 140 129 for infrastructure improvement Subtotal 1,990 27.7 should also be made clear and Land, Structure and Other transparent, and their House in , 500 7.0 performance assessed using clear Land acquisition 1,300 18.1 indicators, notably the Quality of Miscellaneous 595 8.3 Life Index. Regular monitoring Subtotal 5,195 72.3 should identify the main Total 7,185 100.0 beneficiaries of housing policies, I/ Brick, 32 sq, meters,including material and and should continually assess the construction. effectiveness of targeted housing Source: COHAB estimates based on Habitat-Brazil programs. projects in Recife Metropolitan Region. d) Ensure the accessibility of accurate housing information: Understanding the pattern of housing demand and housing improvement among the lowest income groups, and the effectiveness of related policies and programs, are necessary to better design poverty interventions in the housing sector. Data on housing services consumed by households at different income levels need to be analyzed as a basis for policy evaluation and design. Other inforrnation, such as on housing affordability, quality, finance, and production capability could assist in the development of a typology of shelter options, identifying key elements that are more responsive to the needs of low-income groups. These information, when collected and made accessible, would also be valuable inputs to the preparation of an integrated housing plan at the federal and state levels as a prerequisite to allocating housing finance. In addition, they would be useful in planning for housing in larger urban areas (with more than 20.000 population) which should be a precondition to housing policy intervention. All these would lead to the development of a housing data bank on the basis of which the performance of the housing sector in general, and progress in low-income housing delivery in particular, could be objectively assessed. The findings of such assessments are vital to the process of improving housing policy development and implementation. Page 72 Brazil: Progressive Low Income Housing 5.43 The Financing Scenario. In Brazil, as in many other countries, the main constraint for well functioning markets is an adequate enabling policy. For the sake of completeness we developed a scenario to estimate (even if very roughly) how much would it cost to shelter those people who are living under poor conditions. We use the cost of producing a house in a Recife ZEIS (i.e., around R$7,200; see Table 36) and the size of low-income households requiring assistance (about 3 million), the total housing budget needed would exceed R$21 billion, a figure close to the estimate made in a recent report by Brazil on urban poverty (Table 37). This is twice the total amount spent by the three levels of government on housing and urban development in 1997 and 1998, not including the R$11 billion lent by FGTS every year as subsidized housing credit that mainly benefited middle-income families. Reorienting the credit system to the very poor raises several issues: First, lending to the very poor requires an administrative structure that the CEF does not have, the risk is higher, and CEF would be reluctant to get into riskier business without adequate guarantees; and second, FGTS funds come from compulsory savings from the formal sector (5 percent of formal workers' wages). It could be argued that these funds could be allocated to the contributing population (the formal workers) rather than to low- and very low- income households of mostly informal workers who do not contribute to FGTS. In this context, one could further argue that on the basis of equity, aid to the very poor and funding to assist the informal sector in their transition to the formal sector could be supported under public sector intervention policies. Public Housing and Urban Spending 5.44 Data on housing and urban spending are not very accurate and much less detailed. In addition, because expenditures on urban development and housing are aggregated, a detailed view of actual municipal spending on low-income housing is not possible. From available data, however, the following considerations can be drawn: (a) Public sector expenditures on housing are low in Brazil. During the last five years, the Brazilian public sector spent an annual average of R$12 billion, or only 1 percent of GDP on housing and urban development. This is equivalent to R$75 (US$30) per capita.. (b) Municipalities finance the bulk of housing expenditures. They contribute about 70 percent of the total housing expenditures, whereas the states' and the federal government's share was only 20 percent and 10 percent respectively. In 1998 housing accounted for I percent of federal social expenditures, compared with 1.2 percent in 1996. In the same year, education accounted for 6 percent of social expenditures and public health, 11 percent (Table 38). While education and health are (and should be) priority sectors in all public expenditure programs, a marginal increase in housing expenditures targeted to the poor would have a considerable positive effect on the sector especially if one takes into account the importance of housing in productive capital formation and household asset accumulation. In fact, Brazil is spending proportionately less on housing than countries at a comparable stage of development. (c) At the municipal level, housing accounts for only about one-fifth of the total municipal spending, compared with 35 percent for education and 29 percent for health. (d) Municipalities invest different proportions of their budgets in housing. Roraima and Amapa seem to make the largest effort, with 38 percent invested in housing and urban development, indicating the pressure rapid urban growth places in these relatively young states. Brazil: Progressive Low Income Housing Pane 73 Table 37. Basic Infrastructure and Housing Needs of the Urban Poor Urban Southeast Urban Northeast Urban Total no. of households ('000) 7,737 6,914 2,021 4,711 31,200 Physical Deficits of Services % without in-house water connection 4.0 6.3 17.3 18.4 14.0 % without sewerage 22.6 14.6 59.3 76.1 51.0 % living in favelas 7.2 8.3 5.7 n.a. n.a. % living in nondurable houses 1.1 n.a. 4.0 n.a. 4.6 Investment Required (mil R$) Water connection 93 131 105 260 1,310 Sewerage 1,311 757 899 2,689 11,934 Access road and drainage 446 459 92 215 1,622 Subtotal: Basic infrastructure 1,850 1,347 1,096 3,164 14,867 Permanent house (32 sq. m.) 331 296 315 734 5,590 Total investment 2,181 1,643 1,411 3,897 20,457 Source: Goncalves 1997, 1998. Table 38. Expenditures in Urban and Housing R$ MiUlion in 1999 Prices 1996 1997 1998 1996 1997 Structure % Federal 972 1,147 1,180 8 10 States 2,142 2,562 2,463 18 21 Municipalities 9,078 8,345 n.a. 74 69 Total 12,192 12,055 100 100 Per Capita As % of GDP Federal 6.1 7.2 7.4 States 13.4 16.1 15.4 Municipalities 57.0 52.4 n.a. Total 76.5 75.7 1.1 1.1 Expenditures in Social Sectors Structure % Federal Education 9,441 8,987 9,208 6.4 5.4 5.5 Health 17,903 21,439 18,631 12.2 12.9 11.2 Social security 55,959 60,674 66,413 38.1 36.5 39.8 HousingfUrban 972 1,147 1,180 1.2 1.1 1.0 Total 120,966 132,108 137,626 100 100 100 Municipalities Education 13,812 13,424 36.2 34.9 Health 10,692 11,202 23.8 29.2 Social security 4,540 5,432 11.9 14.1 HousingfUrban 9,078 8,34 23.8 21.7 Source: Govemment statistics, IPEA. 5.45 While no accurate data are available, anecdotal evidence indicates that municipalities are able to control the growth of slums and informal settlements. This means that the above calculations are investments made to accommodate newcomers. Realistically, it would take in the area of R$22 billion to solve the persistent housing deficit. How could this be financed? 5.46 A 10-year plan in which the financial scheme would require an annual mobilization of R$2 billion for low-income housing (Table 39). The breakdown of the costs would be 40 percent Page 74 Brazil: Progressive Low Income Housing on housing construction, 20 percent on land, and 40 percent on infrastructure. A simple scheme would call for municipalities to contribute the land; the states and the central government to finance the infrastructure on a shared base; and the 40 percent for construction to come from residential savings and microcredit (leveraged on very small up-front capital subsidies if needed). Table 39. Illustrative Scenario for Financing a 10-Year Low-Income Housing Program Annual Values (R$ million) Source of financing Purpose Total annual budget Additional required for housing amount Federal government Infrastructure/ Some up- 1,100 front capital subsidies 400 State govemments Infrastructure development 2,400 400 Municipalities Land supply 8,600 400 Residents and credit House construction FGTS: 1,500 or Microcredit: 8.2 800 Individual contribution $2,500 per household Total 2,000 Notes: Assuming that the total cost of housing is $7,200 as in Table 32; and that the share of financing is 20% central govemment, 20% states,20% municipalities, and 40%lo credit and resident savings. For a total of 3 million households, a 10- year program would provide 300,000 units per year. The cost to be bome by each household would be $2,500, which seems compatible with the income level of low-income households. 5.47 Possible elements include: (a) The federal government would increase its presence in the sector to facilitate the improvement of shelter and housing services for the very poor. At present, the central government spends less than I percent of its social budget on housing. This is less than one-fourth of the housing expenditure of the majority of other countries. The outlined program would require an additional 40 percent to the housing budget, which nevertheless would bring the budget to R$ 1.4 billion only (compared with R$66 billion for social security and R$9 billion for education). Even with this addition, housing expenditure would still account for a low 1.4 percent of the government's social expenditures. (b) The states' responsibility in housing would be one of coordination and intermediation. They could play a key role in gathering information and disseminating practices - sharing information on costs, prices, and collaborative efforts. They could also provide municipal associations with technical support and kits for housing improvement, and facilitate regularization and registration of new plots. In financial terms, the plan would require states to mobilize an additional 20 percent of what they currently allocate for housing (i.e., R$400 million compared with the R$2,300 they spent in 1997-98). (c) Municipal governments would need a comprehensive package of technical assistance, freedom to use the instruments created by the Constitution, incentives to promote popular participation, authority to use flexible systems of expropriation and usocapido to improve land titling and the liquidity of local land markets, and flexible standards of land, construction, and services provision. In financial terms the plan would require an Brazil: Progressive Low Income Housing Page 75 increase of about 5 percent in their budgets. This would be relatively simple to finance, especially if progressive property taxation is employed. (d) Residents could contribute directly from their resources (in kind or in cash) and with borrowed money. Given the small size of the loan (R$2,400), a poor household (earning less than five minimum wages) would be able to afford the down payment and repay it at commercial terms. (e) Financial intermediation is central to this plan, with FGTS, NGOs, and microcredit institutions playing major roles. 5.48 Experience in Brazil shows that there would be sufficient resources that could be mobilized, especially, considering the high multiplier effect of housing investment (100 percent). Resources for municipalities could come from: (a) higher and better administered property taxes. Part of the potential increase in revenue could come from property tax payments by the regularized plot-owners; (b) sale of serviced land. Municipalities would recover part of the costs of infrastructure provision through these sales to the private sector for commercial use; (c) fees for permit of densification, like the system used in Parana; and (d) community savings in the form of mutirdo or materials banks. 5.49 General Concerns. Despite the overwhelming magnitude of the low-income housing problem in Brazil, national and international experiences have demonstrated convincingly that there are several alternatives to addressing it. The mobilization of microfinance, participatory programs, targeted subsidies, and flexible development regulations provide realistic ways to equate limited government resources with the vast extent of the housing demand. These key elements can coalesce into a new, more responsive national housing policy, while also taking into account other important concerns, viz.: (a) Better use of the FGTS. Part of the FGTS funds could be used to leverage actions at the local level, notably with credit associations of low-income households, as well as with state agencies that can provide serviced land and infrastructure. While FGTS should not be viewed as a public fund for budgetary purposes, a special line could be created to encourage the formation of residents' savings associations and subsequently small-scale house financing programs for their members. (b) Better focus on the deficit regions. As mentioned earlier, the federal programs benefit the Southeast region of Brazil much more than the Northeast. This is explained by the fact that even though the Northeast has a larger housing deficit, most of its population cannot not make the down payment required to be eligible for federal programs nor can they service FGTS loans. Special programs therefore should be designed to cater to the particular needs of households in the Northeast. (c) Central government subsidies rmay be necessary to reach the poorest households. A careful assessment of subsidy program alternatives drawing on successful experiences in other countries should be conducted before such policy is put in place. A system of up- front capital subsidies (as in Chile) would be preferable than below-market interest rates or rent controls. (d) Promoting a tripartite model of housing for the poor based on cost recovery and self- help principles - the poor contributing their savings and labor; the municipality offering the use of public land, the state providing infrastructure, and the central government financing up-front capital subsidies determined in consideration of the poverty level and Page 76 Brazil: Progessive Low Income Housing the housing needs of the municipalities. The construction of an institutional model that takes into account the "personality" of different levels of government and guarantees the collaboration of each level into the program is a major challenge that needs to be worked out by the authorities in conjunction with private sector and the inhabitants themselves. (e) Allocating central fund proportionate with the needs and income base of states and municipalities for the development of infrastructure. (f) Reforming laws and regulations to make them more flexible and allow municipalities to pursue land and housing development more vigorously in accordance with the provision of the Constitution. Brazil: Progressive Low Income Housing Page 77 References Agbola T (1990) Affordability and Cost Recovery in Shelter Projects: The Case of Nigeria. Third World Planning Review, 12(1), 59-74. Alberdi, B. and Serra, V. (1999) Finished and Unfinished Housing. Washington, D.C., The World Bank. May. Aldrich BC and Sandhu, RS (1995) Housing the Urban Poor: Policy and Practice in Developing Countries. New Delhi, Sage. Amis P and Lloyd P (eds) (1990) Housing Africa's Poor. Manchester, Manchester University Press. Angel S, Archer R, Tanphiphta S and Wegelin E (eds) (1983) Land for Housing the Poor. Singapore, Select Books. Angel S and Pomchokchai S (1990) The Informal Land Subdivision Market in Bangkok. In Baross and van der Linden J (eds) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. Asthana S (1994) Integrated Slum Improvement in Vishakhapatnam, India. Habitat International 18(1): 57-70. Baross P and van der Linden J (eds) (1990) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. Boonyabancha S (1997) Community Enablement, Poverty Alleviation and Integrated Savings and Credit Schemes in Bangkok. In Burgess R, Carmona M and Kolstee T (eds) The Challenge of Sustainable Cities. London and New Jersey, Zed Books. Buckley RM, Faulk D and Olajide L (1993) Private Sector Participation, Structural Adjustment and Nigeria's New National Housing Policy: Lessons from Foreign Experience. Central Bank of Nigeria Economic and Financial Review 31: 53-67. Bueno Laura Machado de Mello. Polfticas Publicas para o Manejo do Solo Urbano e a Questao Habitacional, in Polis, Politicas Publicas para o Manejo do Solo Urbano: experi8ncias e possibilidades, Sao Jos6 dos Campos, Agosto 1996. Burgess R, Carmona M and Kolstee T (eds) (1997) The Challenge of Sustainable Cities. London and New Jersey, Zed Books. Cheema GS (1987) Urban Shelter and Services: Public Policies and Management Approaches. London, Praeger. Cherkezian H (2000). Informe Sobre a Politica de Financiamentos e Subsidios Habitacional, mimeo Chogull CL (1994) Sustainable Housing Programmes in a World of Adjustment. Habitat International 18(2): 1-11. Page 78 Brazil: Progressive Low Income Housing Chogull CL (1995) The Future of Planned Urban Development in the Third World: New Directions. In Aldrich BC and Sandhu RS (eds) Housing the Urban Poor: Policy and Practice in Developing Countries. New Delhi, Sage. Devas N (1989) The Evolution of Urban Housing Projects in Lesotho: The Table of Five Schemes. A land Use Policy July: 203-216. Farvacque, Catherine and McAuslan, Patrick: Reforming Urban land Policies and Institutions in Developing Countries, Washington 1992. Ferchiou R (1982) The Indirect Effects of New Housing Construction in Developing Countries. Urban Studies 19(2): 167-72. Ferguson, B., Schor G. and Buchenau "Microfinance for Housing, Lessons from Practioners" Washington, 2000. Ferreira, H.G, and Barros, Ricardo Paes (1999) de The Slippery Slope, Explaining the Increase in Extreme Poverty in Urban Brazil, 1976-1996, Mimeo, April Funda,co Joao Pinheiro (2001) Deficit Habitacional no Brazil 2000, Projecto PNUD BRA-00/19, Belo Horizonte, December Gilbert A (1992) Third World Cities: Housing, Infrastructure and Servicing. Urban Studies 29: 435460. Goldberg M (2000): Housing Micro-credit Schemes, a contribution to the Mexico report on Low Income Housing.. Gon,alves (1997, 1998). "Situacao da infra-estrutura no Brasil" Mimeo Gugler J (ed) (1997) Cities in the Developing World: Issues, Theory, and Policy. Oxford, Oxford University Press. Hardoy JE and Satterthwaite D (1989) Squatter Citizen: Life in the Urban Third World. London, Earthscan. Ikejiofor U (1999) The God that Failed: A Critique of Public Housing in Nigeria, 1975-1995. Habitat International 23(2): 177-188. Informe Sobre a Politica de Financiamentos e Subsidio Habitacional, Epoca, March 12, 2001 pp. 78-82. IPEA (1998) Gestdo do Uso do Solo e Disfun,ces do Crescimento Urbano. Kessides C (1997) World Bank Experience with the Provision of Infrastructure Services for the Urban Poor: Preliminary Identification and Review of Best Practices. Washington, D.C., The World Bank. January. Lee M (1995) The Community Mortgage Program: An Almost Successful Alternative for Some Urban Poor. Habitat International 19(4): 529-546. Brazil: Progressive Low Income Housing Page 79 Lee-Smith D and Memon P (1988) Institution Development for Delivery of Low Income Housing: An Evaluation of Dandora Community Development Project in Nairobi. Third World Planning Review 10(3). Ludwig R and Cheema S (1987) Evaluating the Impact of Policies and Projects: Experience in Urban Shelter and Basic Urban Services. Regional Development Dialogue 4(8): 190-229. Malpezzi S and Sa-Aadu J (1996) What Have African Housing Policies Wrought? Real Estate Economics 24: 133-160. Mathey K (ed) (1990) Housing Policy in the Socialist Third World. London and Oxford, Mansell. Mathey K (ed) (1992) Beyond Self-Help Housing. London and Oxford, Mansell. Mayo SK and Gross DJ (1987) Sites and Services - and Subsidies: The Economics of Low-Cost Housing in Developing Countries. The World Bank Economic Review 1(2): 301-335. Mayo SK, Malpezzi S and Gross DJ (1986) Shelter Strategies for the Urban Poor in Developing Countries. Research Observer 1(2): 183-203. Mehta B, Nientied P and Mitra BC (1989) Building Regulation'and Low Income Housing: A Case Study from India. Cities February: 50-58. Morais, Maria Piedade (2000) Breve Diagn6stico Acerca da Questdo Habitacional no Brazil. CGPUR. May. Moser C and Mcllwaine C , 1998 Urban Poor Perceptions of Violence and Exclusion in Colombia. World Bank, Mimeo Motta Diana (2000) Os custos de Regula,co do uso do solo no Brasil: Um Paper Introduct6rio, Brasilia, June 1999. Neri MC., Amadeo E.J; Carvalho, AP Nasciemtno MC, 2000 Assets, Markets and Poverty in Brazil. Ensaios Economics da EPGE no 374. Rio de Janeiro, FGV/EPGE. Nientied P and van der Linden J (1985) Approaches to Low-income Housing in the Third World: Some Comments. International Journal of Urban and Regional Research 9: 311-329. Nientied P and van der Linden J (1988) The 'New' Policy Approach to Housing: A Review of the Literature. Public Administration and Development, 8: 233-240. Ogu VI (1999) Housing Enablement in a Developing World City: The Case of Benin City, Nigeria. Habitat International 23: 321-248. Payne G (1989) Informal Housing and Land Subdivision in Third World Cities: A Review of the Literature. Oxford, Centre for Development and Environmental Planning, Oxford Polytechnic. Payne G (1984) Low-income Housing in the Developing World. Chichester, Wiley. Page 80 Brazil: Progressive Low Income Housing Polis (1996) Politicas Publicas para o Manejo do Solo Urbano: experiencias e posibilidades, Sao Jose dos Campos. August. Porto, E., Diana Motta, M. Piedade Morais, G. Blank, J. Pupin (2001) Relat6rio Nacional Brasilero, Assembleia Geral das Nap5es Unidas, Centro das NaV5oes Unidas para Assentamentos Humanos. CNUAH-Habitat, Istambul + 5, Nairoby . February. Pugh C (1990) Housing and Urbanisation: A Study of India. New Delhi, Sage. Pugh C (1992) Land Policies and Low-Income Housing in Developing Countries. Land Use Policy 47-63. Pugh C (2000) Squatter Settlements: Their Sustainability, Architectural Contributions, and Socio-Economic Roles. Cities 17(5) 325-337. Pugh C (1997) The Changing Roles of Self-Help in Housing and Urban Policies, 1950-1996. Third World Planning Review 19(1) 91-109. Rocha S (2000). Pobreza e Desigualdade no Brasil: o Esgotamento dos Efeitos Distributivos do Plano Real. Texto para Discussao Interna no 721, IPEA. April. Rodell MJ (1984) Sites and Services and Low Income Housing. In Skinner RJ and Rodell MJ (eds) (1984). People, Poverty and Shelter. London and New York, Methuen. Rolink R (ed) (2000) Zonas de Especial Interesse Social in 125 Dicas. Ideias para a Acdo Municipal. Sao Paulo, POLIS. Schubeler P (1996) Participation and Partnership in Urban Infrastructure Management. Washington, D.C., UNDP/UNCHS (Habitat)/World Bank Urban Management Programme. Secrataria Especial de Desenvolvimento Urbano (SEDU) (2000) Relatorio Especial de Acompanhamento de Avalia ao dos Programas. Siddiqui TA and Khan MA (1990) Land Supply to the Urban Poor: Hyderabad's Incremental Development Scheme. In Baross and van der Linden J (eds) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. Siembela W and Moreno (1997) Expanding Housing Choices for the Sector Populate: Strategies for Mexico. Housing Policy Debate Volume 8, Issue 3. Fannie Mae Foundation. May. Silas J (1992) Environmental Management in Surabaya's Kampungs. Environment and Urbanization 4(2): 33-41. Skinner RJ, Taylor JL, Wegelin, EA(eds) (1987) Shelter Upgrading for the Urban Poor. Manila, Island Publishing House. Steinberg F (1990) Cairo: Informal Land Development and the Challenge for the Future. In Baross and van der Linden J (eds) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. Brazil: Progressive Low Income Housing Page 81 Swan PJ, Wegelin EA and Panchee K (1983) Management of Sites and Services Housing Schmes: The Asian Experience. Chichester, John Wiley. Tipple AG (1994a) A Matter of Interface: The Need for a Shift in Targeting Housing Interventions. Habitat International 18: 1-15. Tipple AG (1994b) The Need for New Urban Housing in sub-Saharan Africa: Problem or Opportunity. African Affairs 93:587-608. UNCHS (Habitat) (1996) An Urbanizing World: Global Report on Human Settlements. Nairobi, United Nations Centre for Human Settlements (Habitat). UNCHS (Habitat) (1990) The Global Strategy for Shelter to the Year 2000. Nairobi, United Nations Centre for Human Settlements (Habitat). UNCHS (Habitat) (1993) The Urban Poor as Agents of Development: Community Action Planning in Sri Lanka. Nairobi, United Nations Center for Human Settlements (Habitat). UNDP (1990) Human Development Report 1990. Oxford, Oxford University Press. van der Linden J (1986) The Sites and Services Approach Reviewed. Aldershot, Gower. van Westen ACM (1990) Land Supply for Low-income Housing in Bamako, Mali: Its Evolution and Performance. In Baross and van der Linden (eds) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. von Amsberg J (2000) Attacking Urban Poverty in Brazil. Washington, D.C., The World Bank. June. Wakeley P (1988) The Development of Housing through the Withdrawal from Construction: Changes in Third World Housing Policies. Habitat International 12(3): 121-131. Wakeley P (1999) Urban Housing: The Need for Public Sector Intervention and International Cooperation. Cities 6(3): 195-201. Ward P (ed) ( 1982) Self-Help Housing: A Critique. London, Mansell. Ward P (1990) The Politics and Costs of Illegal Land Development. In Baross and van der Linden J (eds) The Transformation of Land Supply Systems in Third World Cities. Aldershot, Avebury. Wegelin E (1995) Squatter and Slum Settlements in Pakistan. In Aldrich BC and Sandhu RS (eds) (1995) Housing the Urban Poor. New Delhi, Zed Books. Werlin H (1999) The Slum Upgrading Myth. Urban Studies 36(9): 1523-1534. World Bank (1995) Better Urban Services: Finding the Right Incentives. Washington, D.C., The World Bank. Yap KS (2000) Sustainable Strategies for the Provision of Low-income Housing in Cities in Developing Countries. City Development Strategies Initiative May Issue. Brazil Low-Income Housing: Alternatives for the Urban Poor Page I Annex A: Urban Low-income Housing in the Third World: A Review of the Literature by Elizabeth G. Mehta and Dean Cira 1. The literature on low-income housing in the Third World abounds with case studies and evaluations of slums and squatter settlements transformations, "low-cost" public housing provisions, "self-help" housing and their locally defined variations. It embraces a wide range of approaches adopted by developing country governments in their efforts to find solutions to the housing problems of the urban poor. A more accurate understanding of the performance of these approaches is an essential precondition for the development of more responsive policies and action towards low-income housing. What are the nature and extent of the low-income housing "solutions" adopted by developing country govemments? Whose housing needs do they meet? Do they make a positive contribution to the urban poor's housing stock? What lessons can be learnt from their experiences? 2. This review of the literature attempts to contribute to find answers to the above issues by bringing together the perspectives of many researchers, practitioners, observers, analysts, and students of housing and urban development. Focusing on the various approaches to urban low- income housing in the Third World, the intention is to study the topic from publications and data from case studies available from 1980 to the present. Within these limitations, the study attempts to describe these approaches, to provide tentative explanations as to why some are more successful than others, to analyze the critical issues they face, and to identify promising approaches which provide new options and directions for the future. 1. Urban Growth and the Issue of Housing the Urban Poor 3. More than one and a half billion people now live in the cities and towns of developing countries. Each year these cities absorb some 12 to 15 million new households. Within the next 25 years, more than half of the total population of developing countries will be urban dwellers. As cities in developing countries expand, so inevitably do the numbers of the urban poor. This continuing long-terrn urban growth makes the challenge of housing the urban poor greater and more urgent. Because this growth is taking place against a background of low incomes, it has outstripped the developing countries' capacity and resources to cope with increasing demand for housing and related services. The burgeoning slums in many large cities of the developing countries are a testimony to the inability of governments'to address deficiencies in shelter and service provision. 4. While it is difficult to get reliable estimates of the number of people living in slums and squatter settlements some attempts have been made. A 1990 estimate suggests that at least 600 million urban dwellers live in life-threatening or health-threatening homes and neighborhoods because of poor quality shelters, dangerous sites, and inadequate provision for safe water supplies, sanitation, drainage or health care (Hardoy and Satterthwaite, 1990). Another indicates that irregular or informal settlements house from 20 to 80 percent of the urban populations of various cities, affecting 15 to 70 percent of the urban populations of developing countries (UNCHS, 1996). 5. In terms of supply, the number of dwellings constructed annually in the Third World is usually between 2 and 4 percent per 1,000 inhabitants while the population is expanding at between 20 and 35 and the urban population at between 25 and 60 persons per 1,000 inhabitants a Page II Brazil: Progressive Low Income Housing year (Hardoy and Satterthwaite, 1997). The many people swelling the ranks of urban dwellers, i.e., the low-income majority, therefore present a major challenge to developing country governments tasked with the responsibility of meeting increasing demands for housing and services. II. Evolution of Low Income Housing: The Main Approaches 6. Most governments have recognized, to a greater or lesser extent, the gravity of the housing situation in the cities and have responded to the housing crisis in a variety of ways, ranging from the violent destruction of slum communities to the tacit support of organized invasions, from the construction of highly subsidized flats to the introduction of basic services in squatter settlements, and from direct housing provision to housing "enablement." These types of responses were reflected in the three most widely adopted approaches that have emerged during the last 40 years: (a) the earlier public housing construction which was implemented in many countries from the 1950s through to the 1960s and 1970s, (b) the later government-assisted "self- help" programs which were initiated in some countries in the 1960s and were widely adopted in the 1970s and 1980s, and the "enabled" approaches which have started in the mid-1980s and are still continuing. The considerable experience gained with the first two approaches since the 1950s makes it possible to assess their contribution and their future potential. Several lessons can be drawn from its experience so far. In reviewing these approaches, it may be useful to place them in a historical context. How did governments perceive the housing problem of the urban poor? What prompted them to adopt these approaches as "solutions" to the problem? A. 'Official' Approach: Public Housing 7. Many developing country governments faced tremendous pressures when rural-urban migration first triggered unprecedented urban population growth throughout the 1950s and early 1960s. The conventional view then, inherited from Great Britain and other developed countries, was that it was the responsibility of public authorities to meet the housing need. The latter was assessed in terms of housing "deficit," that is, the number of housing units required to meet future demand as well as the existing backlog of families living in substandard accommodation. The almost universal initial response of governments was to adopt the housing "solution" of developed countries, i.e., construction of "low-cost" public housing projects either for direct sale or for rent. Following Western planning principles and techniques, public housing meant highly subsidized blocks of flats with high standards of construction and infrastructure, and zoning and building regulations discouraging production of lower standard housing. In many cases, it involved destruction of slum areas and squatter settlements (Mayo and Gross, 1987; Mehta et al., 1989; Nientied, 1988). 8. Public housing as a "solution" became popular owing to the general attitude then prevailing among government planners and decision-makers. The latter recognized that the lack of affordable housing in the formal market had engendered slums and squatter settlements. By flouting planning legislation concerning standards and procedures, they viewed these settlements as representing a threat to planned urban development and even to the rule of law itself. To many policymakers and planners, these popular "informal" settlements developed without official approval and built in violation of zoning and/or building codes were illegal and therefore needed to be demolished and replaced with new, modern buildings (Payne, 1984; Ward, 1982). 9. That public housing did not reach the low-income urban population for whom it was intended has been widely reported in the literature. While dwellings in public housing projects conformed to high standards of construction and service provision, they were far too expensive Brazil Low-Income Housing: Alternatives for the Urban Poor Page Im for the targeted households and required large subsidies (Payne, 1984). Typically located on cheaper sites in the urban periphery, public housing often went unoccupied for long periods - a result of poor location, inadequate infrastructure, or rents that, even with subsidies, proved to be far too costly for low-income households (Mayo et al., 1986; Ward, 1982). The irregular nature of the incomes of the majority of the poor meant that they simply could not afford the monthly repayments or rent of even the least expensive public housing unit. Often the result was massive defaulting on repayments or nonpayment of rents, and / or eventual sale of the units to better-off households. Subsidy levels were high enough to create serious budgetary pressures on governments, virtually guaranteeing that public housing could not be made available to the low- income urban population (Ward, 1982; Payne, 1984; Nientied, 1988; Hardoy and Satterthwaite, 1989). 10. Because of the high costs of constructing completed housing, the scale of public housing provision invariably fell far short of meeting the quantitative demand for low-income housing. Unit costs were so high such that very few units were built relative to need; indeed, in many instances, less than 20 percent of the units originally planned were built (Hardoy and Satterthwaite, 1989). The reality was that the resources required were simply not available, nor were likely to be available (Wakely, 1988; 1989). 11. In Latin America most governments which for many years sought to reduce their national housing problems principally through public housing have gradually began to realize that this approach was doomed to failure. They simply could not build enough homes to respond to the large numbers of families requiring better accommodation. The unwelcome consequence was that the number of families living in informal housing was increasing over time (Mathey 1990; Valenca, 1992; Gilbert and Gugler, 1992). 12. In addition, public housing also failed to satisfy the low-income households' actual needs. Devoting more attention to issues of sound building design and adequate sanitation, it ignored the importance of location, i.e., proximity to workplaces, for low-income households. Moreover, although the usually medium- to high-rise public housing blocks were offered to improve the quality of life of squatters and slum-dwellers, the completely different circumstances in the countries where they were modeled after (in general developed countries) reflected an attempt on the part of governments to change the way the urban poor were accustomed to. For example, many poor households were engaged in hawking at the ground level and resettling them to high-rise blocks denied them of this income-generating opportunity. A typical consequence was that low-income households who were allocated public housing sold or informally rented the units to others because the size, design, and location ill-matched their needs (Hardoy and Satterthwaite, 1989). 13. Very few governments now have large public housing programs although many tried such programs during the 1960s, 1970s, and early 1980s. It is curious to note that despite having been discredited, this approach continues to be the thrust of public housing policies in Nigeria (Ikejiofor, 1999). B. Self-Help Housing: Sites and Services and Upgrading Projects 14. By mid-1960s, disillusion with public housing went hand in hand with the more positive attitude emerging in governments towards slum and squatter settlements. The earlier aversion to the squalid and disorderly conditions of popular settlements gave way to a growing appreciation of the ability of squatters and slum-dwellers to improve their settlements over time largely Page IV Brazil: Progressive Low Income Housing through their own remarkable ingenuity and resources. Early proponents saw the positive value of this "self-help" approach by low-income households to solve their housing problems and successfully advocated that governments capitalize on and support these initiatives. 15. By the end of the decade, there began a gradual withdrawal of government involvement in direct public housing provision and greater reliance on self-help housing with government assistance. Two approaches have since been widely utilized - slum upgrading and sites and services. * Sites and services involved the provision of tracts of vacant land for housing, and where basic supporting services were set aside and low-income residents constructed their own dwellings or were provided "core houses." * Slum upgrading complemented sites and services projects but they focused mainly on improving existing housing sites of the urban poor rather than on developing undeveloped land. 16. The key to making these government-assisted self-help projects work was to price sites and housing at a level that would make them affordable to low-income households, allow costs to be fully recovered, and generate a moderate surplus that could be recycled to replicate projects on a larger scale. Affordability meant lowering planning and engineering standards vis a vis density, living space, building materials, and services, thereby lowering costs. Cost recovery and replicability were seen as necessary to ensure that projects could be scaled up with the expected modest initial surpluses to be used for financing new schemes. 17. Along the cost-recovery-affordability-replicability principles have evolved a growing number of sites and services and upgrading projects adopted by governments and largely supported and financed by international institutions such as the World Bank and the U.S. Agency for International Development. The World Bank alone had actively pursued both approaches from 1972 onwards. The major part of its loans were absorbed by sites and services, slum upgrading, and integrated urban development projects which usually included a sites and services and/or urban upgrading component (Nientied and van der Linden, 1988). 18. After more than a decade of experience with sites and services and upgrading projects, several studies reviewed progress and looked closely at how they had worked so far (Angel et al., 1983; Nientied and van der Linden, 1988; Payne, 1984; Skinner et al., 1987; Ward, 1982). Broadly, many sites and services projects were found to have suffered from problems similar to those experienced in public housing projects, i.e., too costly for poorer groups, badly selected locations, and with plot sizes and site lay-outs which ill-match the needs and priorities of the intended beneficiaries (Payne, 1984; Rodell, 1984). The World Bank in its review found good examples in El Salvador, the Philippines, Senegal, and Zambia (Box 1). Another example of its success occurred in Madras (now Chennai), India, in its sites and services experience in the 1977- 86 period (Pugh, 1990) (Box 2). However, several shortcomings of the Bank's sites and services were identified, including claims that the shelter provided is not affordable; benefits accrue disproportionately to better-off households; cost recovery is poor; subsidies persists at unsustainable levels; and because of the above, it is not budgetarily possible to reproduce projects as they are now designed and implemented on a large scale (Keare and Parris, 1982; Mayo and Gross, 1987). 19. With upgrading programs, perhaps the two most serious difficulties are how to sustain the initial impetus and how to expand them to the point that they reach most if not all of those in need. In an evaluation of slum upgrading projects, it was suggested that the capacity of local governments to invest in and maintain new infrastructures and services and to continue the upgrading processes was often limited. Upgrading, it has been noted, temporarily made up for Brazil Low-Income Housing: Alternatives for the Urban Poor Page V this deficiency but the basic institutional deficiency was not removed. In most upgraded areas, there were serious problems with maintenance within a few years after the initial period of investment in housing and infrastructure was completed (Hardoy and Satterthwaite, 1989). Another serious shortcoming is the inability of governments to reach sufficient people through this approach, largely because it did little to address the more fundamental constraint of land, housing finance and building materials, and the provision of infrastructure and se rvices (UNCHS, 1996). A rare exception is the Kampung Improvement Program, especially the program in Surabaya, Indonesia, which is now widely considered to be one of the most successful large scale upgrading programs (Box 3). 20. Even in successful upgrading schemes such as the Hyderabad in India (Box 4), it has proved difficult to sustain community interest in maintaining services particularly in 'softer' areas such as health, education and social development (Asthana, 1994). The scheme's community based approach was for the most part dependent upon self-financing but was seldom able to mobilize more than a modest volume of resources. The external supporting agent was a government agency but because the community was the main actor there was litle difficulty in demarcating responsibilities and tasks. However, its progression from small-scale measures of aided self-help to a large-scale upgrading and housing program financed through credit facilities led to a weakening of community directed activity. Eventually the program became increasingly supply-driven and paternalistic. 21. Irrespective of the internal successes of some projects, neither sites and services or slum upgrading had by any means catered to the increase in low-income housing demand. Very seldom had they gone beyond pilot or demonstration phase and therefore the scale on which the approach was followed remained grossly insufficient. In this respect, upgrading compared somewhat favorably with sites and services (van der Linden, 1986). By the early 1980s, the World Bank and other agencies' lending gradually shifted in relative terms from sites and services toward slum upgrading (Mayo and Gross, 1988). Later, it reduced support for self-help projects and redirected its strategy towards a more programmatic approach with broader and deeper institutional reform and development. Its new approach focused on building capacity and appropriate policies for municipal management, fiscal reform, and resource mobilization. 22. Because neither sites and services or slum upgrading yielded sufficient numbers or other benefits as anticipated, the need for alternatives became more evident. Several research studies and the 'learning by doing' experience with self-help approaches pointed towards change and suggested a new policy framework, replacing narrow project-to-project approaches with a more comprehensive framework based upon new political economies of market-state roles. This shift was necessitated not only by the fact that previous programs (including the direct construction of public housing units, sites and services schemes, and slum and squatter upgrading) failed to achieve their intended objectives (Buckley et al., 1993; Tippleb, 1994b; Malpezzi and Sa-Aadu, 1996), but they also failed to recognize the contributions and potentials of the private and informal sectors, which collectively account for 80-95 per cent of housing production in developing countries (Gilbert, 1992; Tipple, 1994a; Ogu, 1999). C. New Partnerships: The Enabling Approach 23. As above mentioned, the large-scale learning process in the previous period revealed the limnitations of the sites and services and upgrading approaches and portended some rethinking of the shelter agenda in cities. At the same time, during the 1980s, economic stagnation and the debt crisis facing developing countries necessitated cuts in public expenditures, including for Page VI Brazil: Progressive Low Income Housing housing, infrastructure, and services (UNDP, 1990; Choguill, 1994). The attendant falling incomes experienced by many people and hence less capacity to pay for housing, and growing pressures for government action from those with inadequate housing prompted a radical change from the earlier predominantly "supply" or "provider" approach to the "enabling" approach which gained strength throughout the 1980s. To some extent arising from government's obvious failure to provide enough housing to meet the demand, attention has also been focused on enabling the private sector (which includes many households building for their own needs) to provide dwellings. 24. The endorsement of the enabling strategy to shelter provision came in 1988 following the adoption of the Global Strategy for Housing to the Year 2000 (GSS) by the General Assembly of the United Nations (UNCHS, 1990). The GSS urges governments to adopt strategies that will improve access to land, finance, infrastructure and services, removal of restrictive regulations, introduction of realistic building and land use regulations and strengthening of institutional frameworks that foster people's participation in the housing process. It marks a redirection from the early project-by-project approach of most governments to a whole sector basis for housing development. It encourages new partnerships that allowed a combining of resources from national government agencies, municipal authorities, private sector, NGOs, international funding agencies, voluntary or community sector resources, and individuals/households (UNCHS, 1996). 25. Over the past decade and continuing up to the present, a series of innovative, "enabled" approaches have been adopted. Several promising practices emerge from a review of some of these initiatives. They include community mortgage schemes to provide the poor with access to urban land, community funds which offer housing finance with a collective guarantorship, and community construction contracts that enable the poor to build their own infrastructure. 26. The Million Houses Program in Sri Lanka (Box 5) represented a reduction in the scale of national government investment in housing and an active policy to stimulate and support the investments of individual households, community organizations and municipal authorities. One of the most innovative aspects of the urban sub-program of the Million Houses Program was the system of community construction contracts that were awarded to community organizations who then undertook the construction that had previously been done by a contractor. This process lowered costs, often improved the quality of the product, made maintenance more likely and generated employment within the community (UNCHS, 1993). The fact that urban infrastructure (such as roads, markets and drainage networks) could be developed using local labor, as demonstrated by the community contracts in Sri Lanka, had prompted governments in Zambia and Bolivia to also try similar experimentation with "community contracts." Sri Lanka's program sought to mobilize individual and community resources and combined them with support from national and provincial governments to benefit a significant proportion of the population. Its method of community action planning constituted an effective process for building functionally- oriented community-based organizations (UNCHS, 1996). 27. In the Philippines, the Community Mortgage Program (Box 6) was set up to help low- income households to acquire title to the land that they occupied illegally. This provides them with the possibility of incremental investment that both improves their homes and adds to the value of their assets, and also to improve infrastructure and services. In the program, most of the loans are for land because it is land that the communities find it hardest to acquire. But five years after acquiring land, the estimated total value of the investment in the residential area is about three times the size of the loan for land purchase (Lee, 1995). 28. The Urban Community Development Office in Thailand (Box 7) started in 1992 illustrates a qualitative change in the knowledge and experience of community organizations in Brazil Low-Income Housing: Alternatives for the Urban Poor Page VII the country and a qualitative increase in their capacity over the last 10 years. Its program provides an integrated credit system for community development with different loans available, including those for income generation, small revolving funds, and housing. The program charges minimal rates, an average of 7 percent, with differential interest being charged on community revolving funds, housing and micro-enterprise credit. Its approach responds to the multiple and related needs of low-income communities. The changes it introduced in low-income communities brought new possibilities including the prospect of large-scale impact for the program (Boonyabancha, 1997). 29. The experience of the Khuda Ki Basti in Hyderabad, Pakistan (Box 8) demonstrates that self-financed incremental improvements are feasible even among very low-income communities. Carried out by the Hyderabad Development Authority mainly as a result of the efforts of its farsighted management, the highly innovative program illustrates how building the capacity of government institutions for mediation between community level organizations and regular management processes may be achieved by scaling up smaller pilot projects originated by either government agencies or NGOs. The more comprehensive management responsibility assumed by the community in a collaborative approach generally encompassed self-financing of the facilities concerned. Cost recovery therefore was seldom a problem (Schubeler, 1996). Latin American Experiences 30. In Mexico, the government established the National Fund for Popular Housing (FONHAPO) which provided community-based finance (Box 9). FONHAPO offered subsidized loans, specifically directed at low-income households, which was managed at the level of the household or community. Households were only expected to repay some 40 percent of the loans obtained from FONHAPO. By providing subsidy through the loan, FONHAPO ensured that it could be used in a way that was most effective to the community. For example, they could either self-build or negotiate with local contractors for building services. A further advantage_is better value for money. An evaluation at the end of the FONHAPO program demonstrated that the house constructed had cost about half of what they would have cost if the government had built a house of equivalent size. By delivering the subsidy direct to the people who were responsible for managing the housing construction or improvement process, the full value of the subsidy reached those for whom it was intended (UNCHS, 1996). 31. FONHAPO also provides an example illustrating contemporary efforts to learn lessons from the obvious ability of the private sector developers to provide affordable housing. One of its most innovative aspects was the attempt to make private sector produce solutions that were acceptable to a social system of production. They achieved this through competitions in which commercial firms bid for tenders to build a certain quality of house at the lowest price. This is also a method used in several countries in the North where government agencies (national or municipal) or social housing institutions put to tender the "low-price" housing they want and rely on competition between building firms to produce the best-value final product. Smaller informal construction companies have benefited from such programs because they increase local demand for building materials and building skills (UNCHS, 1996). 32. In El Mezquital in Guatemala, the strong community participation mobilized to plan, execute, and finance investments in housing and community services has significantly improved the quality of life in informal settlements. By setting up cooperatives and revolving funds, participating communities provide and manage housing loans and community services in these Page VIm Brazil: Progressive Low Income Housing settlements. With strengthened management capacity and high cost recovery of investments, the sustainability of community projects is ensured (Box 10). 33. In Chile, the government's housing policy and programs are widely regarded as a success. Its Basic House Program and its direct housing subsidy system have facilitated the active engagement of the private sector, both formal and informal, in housing construction. Rather than being the direct provider, the State acts as a facilitator and financier, channeling substantial resources to families with demonstrated savings capacity. Other success factors include an integrated approach that incorporates the housing needs of all income groups, and an efficient and transparent system of subsidy allocation that allows beneficiaries free choice as to the type of units they prefer to access in the housing market (Box 11). 34. The case of ARGOZ in El Salvador presents a unique example of how responsive regulatory adjustments and loan mechanisms can enable progressive subdivision development to take place, thereby significantly improving access to housing by the poor. ARGOZ is one of the country's biggest private developers which has been subdividing and selling plots, and providing loans for both house improvements and expansion. It offers lease contracts to its buyers, with an option to buy. It also assists new settlements in organizing for community projects to install or improve essential services and facilities. In its 15 years of operation the company has developed 250,000 plots in 500 subdivisions throughout El Salvador. Catering mostly to low-income households has not prevented ARGOZ from becoming a highly profitable company (Box 12). 35. What all the above government-enabled approaches have in common is that they build on collaborative efforts of various actors rather than external resources, and that they bridge the gap between the so-called informal and the formal sectors of urban society. Although government- enabled housing programs provided through partnerships varies widely - and plays an important role in meeting the needs of many low-income communities - the degree of success achieved by these programs is yet unclear. This is also because the approach is relatively new and in-depth evaluations of its impact particularly on the poor's housing stock are yet to be conducted. However, a few interesting pattern has emerged, pointing to an enormous potential for change if governments build on the rich though yet poorly documented source of experiences from which they can learn much. M. Critical Issues A. Political / Economic 36. Several evaluation studies cited a number of political / economic factors which had seriously hindered public housing, sites and services and slum upgrading from reaching the low- income households for whom they were intended. In the earlier public housing projects, it is common to find that units ostensibly constructed for the poor had been appropriated by lower to middle income families whose demands for housing are also unmet by either the public or private sectors - this in spite of the fact that the building of designs and specifications were not exactly what they would desire. This takeover of lower standard housing than a household can afford to pay, or downward 'raiding' as it is sometimes known was a common feature of public sector housing provision in which there is a backlog of demands at all levels. Attempts may be made to restrict or prohibit transfers of title, but in practice these were easily short-circuited (Ward, 1982). 37. In sites and services projects, a recurrent issue was the mass dislocation of original beneficiaries through gentrification. Physical improvements in terms of better infrastructure and services tended to make neighborhoods attractive to higher income households who eventually displace the initial settlers. In a citywide upgrading program in Lusaka, improvements were Brazil Low-Income Housing: Alternatives for the Urban Poor Page IX evenly spread over all slum areas of a city so that no one area was singled out as a potential middle-income neighborhood. Gentrification was thus avoided (Mathey, 1997). Conversely, in Mexico City and Tunis, one-third of the original beneficiaries were displaced by higher-income households (Ferchiou, 1982). 38. Public housing, sites and services, and upgrading projects also suffered from another common problem, i.e., poor location. Rather than developing housing solutions suited for low- income groups, pressure to reduce cost prompted governments to locate projects on cheap land sites on the urban periphery. Since most jobs were found in city centers, a peripheral location did little more than add transport cost to already high housing and site development costs poor households had to bear. Even where transport was subsidized, low-income households could not readily afford the time lost by several hours of public transport ride in from the suburbs. As a result, many recipients either defaulted on payments or sold out to middle-income households who could afford to meet these extra hidden costs. 39. It is also argued that many projects had concentrated upon issues of "sound" standards and design while ignoring the importance of location and the needs of low-income households. Experience in Africa has shown that in a bid to meet these set standards, public housing delivery inevitably became capital intensive while the final product was mostly beyond the means of the low and middle income groups. In Nigeria, data from the defunct Nigerian Building Society and the Federal Mortgage Bank demonstrate that houses built by the government were outrageously more expensive than similar houses built by NGOs and private individuals (Ikejiofor, 1999). Such costly projects were implemented and made affordable only through heavy reliance on' subsidies. The latter limited the scope of programs which in turn gave rise to problems in targeting and eventually in upward leakage of units (Cheema, 1987; Agbola, 1990). 40. In the World Bank-assisted Dandora sites and services project in Nairobi, Kenya, incidence of benefit leakage has also been noted where about half of all occupied plots were rented out to non-allottees and some units had been sold presumably to members of higher- income groups (Chana, 1984). An earlier evaluation of other Bank-assisted sites and services projects found that beneficiaries tend to be more representative of median income groups than of the poorest households (Keare and Parris, 1982). 41. Since governments were unable or unwilling to meet the increasing levels of subsidies required to address the immense and increasing demand effectively, the alternative is to reduce subsidies or to do away with them altogether. However, if subsidies are to be reduced, someone else must pay the bill. The focus then shifts back to full cost recovery. This simple conclusion clashes with the fundamental problem of the target group's limited ability to pay. Even in the days of substantial subsidies, so called low income housing tended to be too expensive for the poor. The problem was usually resolved by simply redefining the target group, effectively pricing the lower-income households out (Mathey, 1997). 42. In addition to failures for financial reasons, the allocation process ensured the poorer households did not get the benefits of low-income housing projects. In Thika, Kenya, elected town councilors prepared lists of people who did not belong to low-income households but who were to receive plots in sites and services projects in exchange for substantial payments (Andreasen, 1989). Housing administrators were also noted to extract large commissions or bribes for approving the allocation of a unit to households not officially eligible, or wishing to jump the long waiting lists (Payne, 1984). Page X Brazil: Progressive Low Income Housing 43. In some cases, political support for self-help projects proved hard to find especially from groups who benefited or hoped to benefit from public housing projects. These included local politicians who used subsidized, high quality public housing projects as a means of dispensing favors to their supporters, mostly investors and speculators. They therefore opposed any proposals which met a higher proportion of demand and which would correspondingly reduce their powers of patronage (Payne, 1984; Ikejiofor, 1999). In yet other cases, opposition to both public housing and self-help projects came from public authorities who derived many benefits from illegal housing. The latter provided these authorities with votes and graft; rid them of burdensome demands and responsibilities, and offered them opportunities to exercise political patronage by making the urban poor view whatever facilities bestowed upon them as a favor and not as a right (Keare and Parris, 1982; Swan, 1983; Nientied, 1988;). B. AdministrativelInstitutional 44. For many developing country governments, self-help and enabling approaches to low- income housing represented a radical change from their experience with public housing projects, not only in terms of technical standards but in terms of planning and implementation processes as well. Sites and services and upgrading projects required the cooperation and coordination of more agencies than what was the case with the earlier approach. It also entailed a number of new elements to which conventional bureaucratic structures and systems could not easily adjust: the acceptance of a major role to be played by participating people, of an increased role of the private sector (including households building for their own needs), of lower standards and untidy-looking neighborhoods. There was thus ample scope for conflict among all actors involved. 45. Experiences with participatory housing demonstrate a wide variety of implementation arrangements ranging from placing prime responsibility with local governments (as in Karachi, etc.) to implementation through separate national (Manila) or metropolitan (Calcutta) authorities (Devas, 1993). At the national level central government housing or public works ministries were primarily in charge. Implementation of sites and services and upgrading projects often required the creation of executing agencies capable of coordinating investments across sectors in a single site. This typically involved the creation of special units, separate from the existing institutional structure, and were often organizational branches of a national (or in large federal countries, state) government (World Bank, 1995). 46. Extensive evaluation of self-help projects undertaken by independent researchers and the World Bank points to lack of coordination among concerned government agencies, and unwieldy institutions delaying implementation (Keare and Parris, 1982; Nientied and van der Linden, 1985, 1988; van der Linden, 1986; Skinner et al., 1987; Payne, 1989; Mathey, 1990, 1992). The fundamental weaknesses identified were the lack of coordination and integration of multiple tasks being implemented simultaneously by different sectoral or specialized agencies. This was so given the fact that sites and services and upgrading projects involved the installation of basic infrastructure (water, sewerage, roads, electricity) and basic services (garbage disposal, health and sanitation), apart from housing construction and/or improvement. Functional fragmentation and problems in inter-agency coordination persisted as services were often provided along sectoral lines thereby hampering project execution and completion. 47. In Karachi, for example, the designing of trunk infrastructure was not coordinated with the planning of individual schemes as a result of which schemes could not be provided with water and access. The result were delays in the allocation and occupation of serviced sites which had adversely affected the cash flow of the implementing agencies working, at least in theory, on a 'no profit no loss basis' (Wegelin, 1995). In World Bank-assisted projects in El Salvador, Senegal and the Philippines, delays in land acquisition and preparation held up construction, which in an Brazil Low-Income Housing: Alternatives for the Urban Poor Page XI inflationary situation resulted in time overruns leading to higher materials costs. With the exception of Senegal, however, the programs operated relatively smoothly and encountered few additional problems once a certain momentum was built up (Keare and Parris, 1982). 48. Another frequent issue is the split responsibility between project execution and operations and maintenance. While sites and services and upgrading projects were typically executed by national or by provincial/state agencies or special project units created within them, operations and maintenance after project completion were usually handed over to municipal or local governments. Often, the latter lacked the technical and financial capability to assume responsibilities over the new facilities. No wonder that in most upgrading projects, serious problems with maintenance were encountered within a few years after completion of the initial period of housing and infrastructure investment (Hardoy and Satterthwaite, 1989). A notable exception was the Kampung Improvement Project in Surabaya, Indonesia, where the inhabitants in the "areas" to be upgraded were involved in project decision-making, contributed their resources, and ensured much better maintenance after the project was completed (Silas, 1992). 49. Many deficiencies in the current administrative arrangements and operations were a result of inadequate capability and lack of sufficiently qualified personnel involved in project implementation, especially given the novelty of participatory self-help projects (Doebelle, 1983; van der Linden, 1986). Where government agencies responsible for project execution might have adequate technical skills, their ability to work with low-income communities was extremely limited (Cheema, 1987). Even in what is considered to be one of the most successful upgrading programs, the Kampung Improvement Program in Surabaya, Indonesia was regarded as too slow largely because of the time needed io develop agreements between public authorities and the kampung inhabitants (Silas, 1992). Conversely, the success of UCDO in Thailand was predicated on the qualitative increase in its capacity obtained from experience in community organization and participatory techniques over the last ten years (Boonyabancha, 1995). C. Technical The issue of land 50. Access to land is a major issue that has received considerable attention in the literature. By and large, public housing and sites and services projects in the past did not reach their intended beneficiaries partly because of unavailability of land in sufficient quantity and in appropriate locations. As earlier mentioned the government's choice of project sites was invariably dictated by cost considerations, peripheral lands being cheaper to acquire. The result is the eventual appropriation of the usually highly subsidized low-income housing projects by better-off households who could afford not only the extra commuting or transport cost to and from employment sources, but also the costs of housing and service improvements. 51. Another difficulty is in acquiring land through expropriation along with the long delays in doing so. In many countries, responsibilities for land registration and for land records rest with two or more public agencies, resulting in lengthy delays associated with land transfer and acquisition. This hindered low-income housing development. Some evidence points to the continuing failure of measures to acquire adequate and suitably located land because of political opposition. In fact, those in positions of power and influence were more likely to own vacant lands, and could usually ensure that any proposed changes in land policy would leave their landholdings intact for speculative activities in the future. Problems with land acquisition were Page XII Brazil: Progressive Low Income Housing one of the main reasons why many governments which tried to scale up sites and services projects into a continuous program could not implement and finance this expansion (Payne, 1984). 52. In Africa, a related problem is that the record of cities with a high proportion of land owned by public authorities is often poor. For instance, in Tanzania and Zambia, a refusal to countenance a market in urban land and a failure to provide a responsive and efficient land allocation system had simply led to large unofficial markets with all the problems these posed for urban authorities in terms of haphazard development, insecure tenure and lack of records for ownership or tenure. In Dar es Salaam, on the other hand, the problem was more related to the fact that the demand for land for housing outstripped the administrative capacity to allocate them. Urban development and capital authorities in Pakistan, Nigeria, Tanzania, and Brazil also have a poor record in providing affordable land for the majority of urban dwellers (UNCHS, 1996). 53. In yet other cases, the problem is not so much the inability of governments to set aside land for housing the poor but their unwillingness to do so. One example of this is provided by the Delhi Development Authority which had marked 20,000 hectares of land for compulsory acquisition but failed to substantially increase the supply of land for low-income housing developments. It released a substantial part of the land to high-income groups with less than 10 percent going to low-income groups (UNCHS, 1996). 54. The recognition of access to land as a key issue has led to an increasing interest in, and research into, the processes by which the poor obtain access to land (Angel et al., 1983; Payne, 1989; Amis, 1990; Baross and van der Linden, 1990). In the past the poor were able to obtain access to land in a variety of ways. In much of Africa communal land was allocated by traditional chiefs in accordance with need. In Latin America and South Asia, squatting on public and private lands, and invasions of land were common. In many parts of the world, one of the main ways in which people obtained access to land was through informal, private subdivisions, particularly of fringe land. 55. Whatever the mechanisms though, it appears that it has now become progressively more difficult for the poor to obtain land (Baross and van der Linden, 1990). The problem of land provision for low-income housing is becoming more acute because of continuing growth of cities and many public lands have in many cases been exhausted. Since there appears to be a marked trend worldwide towards greater commercial articulation of urban land, it is becoming clear that many of the 'no cost' (or minimal cost) options are no longer open. Thus in many parts of Africa, traditional land allocations have effectively become market transactions (Devas, 1989; van Westen 1990, 1998; Amis and Lloyd, 1990); occupations of public and private land, and unauthorized settlements are increasingly organized by land grabbers and illegal subdividers, usually supported by politicians (Ward, 1990; Siddiqui and Khan, 1990); and private subdivisions of land are now largely controlled by commercial property developers (Steinberg, 1990; Angel and Pomchockchai, 1990). 56. Thus, while squatting may be the only affordable form of housing for the urban poor in the past, in today's cities, land is too valuable such that pure squatting is no longer an option. The urban poor have to resort to the land market just like all the rest as even "informal" lands have become commercialized. Building a house is no longer possible without having to seek permission of someone in a position of authority over the land. The poor no longer pay for occupancy rights but rather for protection against eviction at least for a certain period of time. However, since this security is not based on the rule of law such protection may fail (Yap, 2000). Brazil Low-Income Housing: Alternatives for the Urban Poor Page XmI Issues concerning standards 57. Evidence from many developing countries demonstrates the unsuitability of public housing units for low-income households owing to its high delivery costs and limited impact. Part of the reason for this is the high standard of design and construction often adopted in such projects. Central in understanding the high building and planning standards used in public housing is governments' conception of a house as a product that has to be completely finished before use. In this conventional definition, the housing package is viewed as comprising a decent house and all services and infrastructure required in it, each meeting certain minimum physical standards of design and construction determined by the government. Acceptance of a house depends, therefore, on whether or not it provides a set of comfort and hygiene. This definition conflicts with the traditional conception of housing by the poor as a process which starts from a minimum package and grows over time, under their control and within their means. Earlier experience with public housing has shown that in a bid to meet a set of standards all at once, housing delivery inevitably became too costly and relied heavily on large amount of subsidies (van der Linden, 1986; Mayo and Gross, 1987; Hardoy and Satterthwaite, 1989). 58. Sites and services projects attempted to rectify the above by lowering the standards thereby reducing costs. To match the ability to pay of target beneficiaries, the latter were to obtain only unfinished "core houses" which they could complete over time. In some not so unusual cases the starting solution may consist of a sanitary. unit plus a kitchen only, or even less (sanitary unit or "roof plan"). In upgrading schemes which typically concentrate on improving infrastructure, the affordable improvement may be modest. There were extreme cases, however, as in a particular KIP in Indonesia where all that could be afforded was the provision of four toilets and nine water tabs for 12,000 residents (Ludwig and Cheema, 1897). 59. While they are necessary for public safety and economy, one of the problems with building codes and standards in many Third World countries is that they usually are outdated vestiges of colonial administration or are patterned on conditions applicable only in more developed countries. Whereas developed countries may well be able to afford high standards in construction and layout to achieve perfectly understandable aims, it does not necessarily follow that a developing country should adopt these Western standards which might be totally inappropriate to its own climatic, cultural, and economic circumstances. The imposition of high standards constrains the provision of even basic housing for the urban poor. In other words, as mentioned above, high standards mean high cost which in the end price the poor out of the market (Baken and van der Linden, 1993). 60. Another major technical consideration which has received considerable attention is the degree of efficiency with which land is used. Since land usually represents a substantial portion of total project costs and is universally scarce in locations where it is most needed, it is obviously vital to put each square meter'to the best possible use. In many sites and services projects, it is the land standards that seem to cause one of the most serious cost problems. Such overprovision of land for circulation and public open space is difficult to justify since somebody somewhere has to pay to develop and maintain it. There also seems to be a tendency for the land provision to be designed not only for low-cost structures being built at the time, but for some imaginary standard twenty years in the future. As a result too much land is used and land prices are high. Although the cost of land as a percentage of total housing cost per unit varies widely, it can be substantial (Choguill, 1995). Page XIV Brazil: Progressive Low Income Housing 61. Other problems sometimes occur if land standards are set too high. In central city areas, pressure on the poor on existing land resources is usually the greatest due to their understandable desire to be near employment sources. Projects in such areas which provided such excessive plot sizes often resulted in the residents subdividing the plots by themselves and renting or selling them to others. This is particularly dangerous when full economic costs are charged for the plot. It leads to overcrowding and possibly even slum creation which the authorities were attempting to avoid (Choguill, 1995). 62. Few countries have made a concerted effort to revise their building regulations so as to allow more appropriate standards. In some countries and cities, the building regulations are more than 50 years old and are, in theory, still in operation. For instance, in Madras, the refusal to permit cheap and unconventional building materials and practices in sites and services schemes ensured that most of the low-income population could not afford to take part. An insistence on inappropriate standards for, for instance, plot sizes and infrastructure standards have increased prices to the point where a high proportion of all land developments take place illegally. There are a few examples of governments revising standards, although many may have permitted particular projects, including government projects, to use more realistic standards (UNCHS, 1996). 63. There were also cases when project execution was hindered because of disputes between governments and their beneficiaries over design standards. In Dandora project in Nairobi, Kenya, for example, such disagreement greatly delayed the implementation of phase 2 of the project and the construction of community facilities. This resulted in cost overruns owing to both required changes in the designs and the increase in the cost of construction due to inflation and price escalation (Lee-Smith and Memom, 1988). 64. Standards may be introduced for good reasons but their enforcement usually compounds problems for low-income self-build households. Governments that seek to support households and communities developing their own housing need to modify housing standards. The Mexican experience had shown that minimum standards soon became maximum standards for the poor. Within FONHAPO, the concept of minimum standards for both improvement and newly built housing was replaced by that of incremental housing improvement. In the Sri Lankan government's Million Houses Program, urban communities were allowed to propose new standards for development within their residential areas which were then accepted by government authorities (UNCHS, 1996). 65. In the literature, there is little disagreement about the need to lower standards to make housing affordable to the poor. While inappropriate and invariably high standards characterized the earlier public housing and sites services projects, there are now a few examples of governments revising standards or permitting particular projects to use more realistic standards. IV. Directions for the Future 66. While past experiences in public housing and self-help program provide many useful lessons, and often quite contrasting results, available literature has also made it clear that these approaches have substantial shortcomings. The recent "enabling" approach to housing provision which has demonstrated a number of successful models provides new sense of direction but still one of considerable variety and can mean different configurations in the field. Today the magnitude of the housing problem of the poor worldwide and the limitations of conventional approaches to housing leaves no alternative but to turn to the knowledge, resources and capacities of the tens of millions of low-income communities who are already the most active in the process of incremental house-building and site improvements for themselves. There is increasingly Brazil Low-Income Housing: Alternatives for the Urban Poor Page XV poorly documented examples of governments supporting a process by which these communities are enabled to realize their own goals and in a manner which is satisfactory to them. These innovative actions implemented in the last decade may yet have little impact in relation to the scale of the problem but they give substance to the idea that there are more effective ways in which government can work with low-income groups and their organizations. The enabling approach does not suggest a diminution of the role of government but their role should be transformed to a more activist, decentralized and supportive of citizen efforts and of community initiatives. Reviewing experiences of the last four decades, there can be found a number of areas where government can intervene to help those whose housing needs and priorities are not met by the "market." Evidence from literature suggests that these are areas which are crucial if efforts toward housing the poor are to succeed: 67. Access to Land: What can be said with much confidence is that individual projects, no matter how well planned and executed, are unlikely to make a significant or sustained contribution to housing the urban poor unless sufficient land, in appropriate locations and at affordable cost, are made available to the poor. The broad issue of land supply to effectively reach the lower income groups therefore should be addressed directly and responsibly. The development of strategies and capabilities to acquire and develop land on a scale sufficient to accommodate needs should thus be a prime concern of an effective policy approach to housing the poor. The experience of the Community Mortgage Program in the Philippines demonstrates success in enabling low-income households to acquire land and develop their plot and housing incrementally. Efforts at housing development, however, should be geared towards opening up land supply not only on a project-to-project basis but at the city level. 68. Access to Finance: Access to credit is a vital element of a supportive approach to housing the urban poor. The need is for fostering innovative arrangements for low-income households to secure access to microloans. These can be in the form of small scale mutual credit institutions and alternative lending instruments attuned to the needs of the households which build their house incrementally. The experiences of the CMP and FONHAPO have shown that many low-income people could afford to purchase plots in legal subdivisions if they were given access to long-term credit. Loans for the purchase of land and housing can be provided using the plot of land as collateral. Such institutions can mobilize savings, can lend competitively to appropriate borrowers, and can rely on a high level of member supervision through peer pressure. 69. Appropriate Standards: In many countries, standards are identical for all income groups even though the poor can be satisfied with lower standards and can never afford to purchase costly public housing or private developer-built standard housing. Experiences with sites and services have shown that one of the reasons why low-income households sell out to higher incomes is that they can not afford the construction and on-plot development standards required of them. There is therefore a need for governments to adjust standards to the poor's levels of affordability. Standards for low-income housing projects should be different from standards applied to private sector developers although the extension of lower standards to private residential development can increase the supply of land and housing for the poor. The example of Sri Lanka's Million Houses Program illustrates a practical framework for target beneficiaries to arrive at flexible design standards, allowing them make their own tradeoffs between convenience and affordability. 70. Basic Infrastructure Provision: For low-income households basic infrastructure remains a major problem. Because of the effect of inadequate sanitation and lack of clean water on health, access to basic facilities is more important to low-income households than the quality Page XVI Brazil: Progressive Low Income Housing of their dwellings. The urban poor often pay more for water from private vendors than other income groups who have access to public networks. For efficient housing development for the urban poor, there is a need for coordinating the agencies responsible for provision of basic infrastructure (roads, drainage, water, sewerage, and electricity) to focus on servicing their existing or potential undeveloped settlements. 71. Targeted Subsidies: A major pitfall of public housing and sites and services is the existence of subsidies which are simply too high to allow for the large-scale replicability of these projects. As a result the available public funds could only finance a rather small number of pilot or demonstrated projects whose benefits often vs. eventually accrue disproportionately to better- off households. If it is fiscally possible for govemments to provide subsidies for equity or efficiency purposes, these subsidies should be targeted specifically at low-income households, devising means to minimize leakage to ineligible income groups. FONHAPO's experience suggests that govemment housing subsidies must be allocated directly to low-income communities if limited funds are to be used most effectively. 72. Inclusion and Partnership: The need is to innovate ways to strengthen the voice of the poor and build local capacity for partnership, broad based participation, and leaming from results. The govemment's new role would be to help build the infrastructure of partnership by bringing together all the concemed public agencies, private sector, NGO representatives and community- based organizations of low-income households, and by giving the latter a chance to lead in devising their own housing solutions along with their partners. By giving this new partnership a chance to exert leadership, the governments will be better positioned to assist in building long- term capacity in concert with the principles of enablement. They can also help them develop and enforce agreement that all parties will accept mutual responsibility and distinct accountability for outcomes. 73. In considering future directions for low-income housing, we need to refer to the above elements. Some or most of them have already been incorporated in successful programs cited earlier - in the case of FONHAPO in Mexico, the Urban Community Development Office in Thailand, the Community Mortgage Program in the Philippines, and many other similar examples. The new range of initiatives represents a newly emerging development paradigm, focusing on partnerships and the attendant advancement of collective know-how. The "enabling" approach holds promise as evidenced by its various applications to these programs. However, it has yet to mainstream its application across the board or across cities, "customized" to suit differential needs and localized variations, and to build on local processes and experiences. For their part, to equip themselves to implement the enabling framework effectively, govemments will have to continuously examine the results of these ongoing applications and experimentation, and help consolidate a process by which, year by year, the initiatives of the "new partnerships" are encouraged, supported and built upon. 74. Ensuring that the above elements are incorporated in future low-income housing initiatives requires important new directions necessary to further facilitate the enabling approach. Reforms are advocated in the areas of policy, institutions, and regulations; moving away from limited, project-based approaches and adopting the principle of enablement. This new approach may not have a monopoly on wisdom, nor does it have all the answers needed for the complex and difficult problem of housing the urban poor but it does provide the most promising way forward if the problem is to be addressed at a scale commensurate with its magnitude and adequate to improve substantially the housing conditions of the poor. Brazil Low-Income Housing: Alternatives for the Urban Poor Page XVII Box 1. Participatory Sites and Services and Area Upgrading: The World Bank-assisted Lusaka Project The Lusaka Project in Zambia is one of the earliest large-scale attempts at participatory urban housing and infrastructure development within a government-sponsored housing development project. Implemented between 1974 and 1981 by the Lusaka City Council with funding from the World Bank, the project provided physical and social infrastructure for 17.000 dwellings in one of the major squatter settlements of the city. In addition, some 12,000 new plots were prepared and serviced, either fully or with water supply alone. These were made available to households displaced by the upgrading program as well as by new residents. The project aimed to promote participation in certain types of planning decisions as well as the provision of investment and labor. The extent of participation was limited due to the planners' concern for rapid project implementation, a reluctance to involve the people in many technical decisions, and the fear that rising expectations might be frustrated by project delays and lead to social discontent. A Road Planning Group, which was established to design portions of the site layout, enabled a certain degree of interactive participation. Community representatives did not, however, include women. For example, while the resulting grid-iron layout gave each plot considerable privacy, the design made it impossible for women to simultaneously do homework and watch their children play, or to have the neighbors watch the house. In the implementation phase participation was more intensive as the project aimed to use household and community labor to the maximum. However, project managers stipulated that contractors would be hired if the self-help activities began to interfere with the implementation schedule. Project evaluation shows that self-help labor significantly lowered the cost of individual housing for the beneficiaries and for the government. There were no cost-savings with regard to civil works, however, and the attempt to mobilize collective labor inputs was not justifiable on economic grounds. The lesson seems to be that participatory strategies which aim only at cost-savings through labor mobilization - with no capacity-building objectives and only minimal participation in the planning phase - are not likely to produce community organizations which are sufficiently strong to actually produce cost-savings through community level inputs. Other cases which attributed much more importance to community development as an objective in itself, such as the FUNDASAL in El Salvador and Orangi Pilot Project in Karachi, have achieved far better cost-savings and cost-sharing performance (Sources: Moser 1987; from Schubeler 1996). Page xvm Brazil: Progressive Low Income Housing Box 2. The Madras Urban Development Project in India Housing reform in Madras occurred in the 1977-86 period under the aegis of the World Bank loan programs but with coordinated authority and financing from the Government of India and the Tamil Nadu state-level government. The Bank arrangement included some policy-based persuasion or conditionality. In effect this has led to the Tamil Nadu Housing Board (TNHB) redistributing its allocations in social housing down the scale of household income distribution and using the self-help approach in low-income housing. And in squatter upgrading, the Tamil Nadu Slum Clearance Board (TNSCB) redirected its emphasis from redevelopment to in situ rehabilitation. The TNHB and the TNSCB worked under the coordination and authority of the Madras Metropolitan Development Authority (MMDA) in Madras. The MMDA held a reformist position, in congruity with the World Bank's social aims; its chief executive was high ranking in the public service; and ways of doing things in Madras were by consultation and collaboration in pursuit of the general public interest. Accordingly, the scene was set for a striking innovation in sites and services programs, achieving greater effectiveness than similar World Bank schemes in other countries. The major innovation was in the methods used for cross-subsidization in housing and land allocation. Allocations of land and services reached down to the ninth percentile of the distribution of household income; contract conditions on title deeds reduced profiteering from sales in tight housing markets; housing improvement occurred progressively and within household affordability and choices; and cost recovery reached a high 95 percent level. The cross-subsidization was effected in differential land pricing; higher- income groups paid five times more than the lowest income groups for a unit of the land, and they purchased larger plots. Nevertheless the prices were below those in the broader markets in Madras. Repayments for the lowest income groups were within 10 percent of their incomes. Additionally, cross-subsidization flowed from commercial to social infrastructure uses and from commercial uses to housing. Welfare improved for all income groups, though in time housing mobility altered the original allocation, and some segregation represented a compromised inequality in pursuit of a broader cross-subsidized equalizing framework. The projects had limitations. They were localized and not capable of broader private-sector application (owing to building regulations). They did not fall within a wider framework for urban economic development, and policies in housing finance and general land acquisition were underdeveloped and sometimes restrictive (Source: Pugh 1997). Brazil Low-Income Housing: Alternatives for the Urban Poor Page XIX Box 3. Sustaining Improvements in Housing and Living Conditions: The Kampung Improvement Program in Surabaya, Indonesia The Kampung Improvement Program in Surabaya, Indonesia's second largest city, is regarded as one of the most successful government initiatives to improve housing and living conditions in Asia. Unlike many upgrading programs, the improvements made in water supply, drainage and other aspects have been sustained with households and community organizations making major contributions both to maintenance and to future investments. It was conceived not as a public works program but as a partnership with community organizations with the kampung inhabitants involved in determining priorities and preferences. The Kampung Improvement Program sought to avoid taking over what inhabitants, households or community organizations can do for themselves. The scale of the improvement program in terms of the number of people reached has helped to avoid the process by which housing and land prices rise in improved kampungs with poorer groups moving out. In the early 1960s, housing and living conditions in cities were put under further strain as city populations grew rapidly and public open space was invaded. In the late 1960s, local governments in Surabaya and Jakarta began to evict the people who had settled on land not designated for housing but also to improve conditions in existing kampungs. The first post-war Kampung Improvement Program in Surabaya, known as the W.R. Supratman KIP, was introduced in 1968. With limited public funding available, the city government of Surabaya only provided basic construction elements for infrastructure in the kampungs, mainly pre-cast concrete slabs and gutters. The inhabitants of the kampungs had to request these and take responsibility for installing them and constructing paths and drains. From 1976, funding from the World Bank became available to support the upgrading program in Surabaya. The main objective of the assistance was to increase the scale, scope and coverage of the program. The financial support was a soft loan of around 65 per cent of the project cost with repayment over a 20-year period. National and provincial government also provided a counterpart fund that represented around one-fifth of the total cost. The kampungs had to provide the land and space needed for the project which meant the removal of front fences and use of house frontage and, on occasion, the removal of houses, with provisions made to re house those who lost their house. After the project was completed, many people improved their own front yard and houses and community infrastructure and services also improved - for instance playgrounds and community halls (that were often used for kindergartens), street lighting, guard houses (to improve security). In the second stage, after 1974, a fixed sum was made available in the annual city development budget to respond to improvement requests from kampungs. The requests were prepared by kampung development committees with technical assistance from PWD staff and these had to conform to municipal standards and provide a detailed plan and cost estimate. After approval, the inhabitants also had to provide matching funds; these varied from more than half the total costs in kampungs with a high proportion of high income groups to a third for settlements where most of the inhabitants had low incomes. Interest in the program has been sustained and it is still widely used for local infrastructure. Each year, not all applications for funding can be met and there is some variation in the total funding allocated by the city government to this program. For instance, more funding was available between 1984 and 1986 as part of a program to prevent flooding. Other international agencies have also provided support for different aspects - for instance, UNICEF helped support improved standards of water provision and garbage collection and support for children's facilities (Source: UNCHS 1996). Page XX Brazil: Progressive Low Income Housing Box 4. The Urban Community Development in Hyderabad, India India's Urban Community Development Program (UCD) began in 1958 with a pilot project in Delhi supported by the Ford Foundation. Based on rural community development experiences of the early 1950s, the program aimed at improving the extremely poor living conditions and infrastructure services in low-income urban communities through supported self-help endeavors. The program was subsequently extended to Ahmedabad (1960), Baroda (1965), Calcutta (1966), and several other cities. Institutionally, the UCD was conceived as a part of municipal government which acted as a bridge between the govemment and the people. Its main aim was to strengthen community organization in low-income residential areas and to enable community groups to plan, finance, and implement self-help projects. The approach was based on the postulate that every community, no matter how poor, has some capacity for self-help improvement, and that communities should, in principle, seek outside assistance only after they have exhausted their own resources. The UCD Program in Hyderabad began in 1967. Water supply, sanitation, waste disposal services, and housing conditions in Hyderabad, a city of about two million inhabitants, were then woefully inadequate. The municipal government - strapped with a large budget and weakened by tax evasion and arrears - was hardly able to maintain existing facilities, let alone expand service coverage. The first program activities targeted a low-income neighborhood of about 7,000 families. Community groups were assisted in the identification of needs, planning, and implementation of small self-financed improvement projects. The first projects were quite modest, involving such items as street light repair and provision of waste collection bins. By 1974, activities expanded to include water and sanitation, health and nutrition programs, pre-school education, income generation, and shelter improvement. The second phase (1977-1982) focused on shelter improvement. After motivation and survey activities, house designs were prepared in consultation with concemed households. The UCD obtained land titles and the project was submitted to HUDCO (Housing and Urban Development Corporation) for loan approval. Work groups of about 20 households were then formed to undertake construction under the supervision of a municipal engineer. Material provision was facilitated by the program, while members provided labor at below-market rates. From 1981 onwards; the program was significantly expanded with ODA financing. The target was to provide housing for 150,000 families over a ten-year period. With the scaled up approach, the role of the people in planning and implementation was considerably reduced; the UCD acquired the more directive role of project manager. In the early years of the program, when UCD operated at a relatively small scale, significant impacts were achieved through the improved coordination between govemment agencies and community organizations, and significant volumes of community resources were generated for physical and economic improvement programs. The community organizations that were developed proved capable of a variety of project activities. The CBO also contributed to the reduced tensions between Muslim and Hindu communities in periods of communal strife. The approach of the UCD has been criticized as being somewhat patemalistic and, for the advancement of self-help activities, many communities remained dependent on constant inputs from the community organizers. In general, however, mobilized community organizations did significantly increase people's capacity to play an active role in local development processes. Evaluation of the large-scale housing program shows modest cost-savings of about 15 per cent through beneficiary labor inputs and material salvaging. Cost recovery performance in this program - estimated at about 45 per cent - has not been satisfactory, however. Many beneficiaries appear to have considered the loan as public funding that will ultimately be written off due to political pressures. Poor cost recovery and dependence on outside financing (UNICEF and ODA) thus raise questions regarding the program's replicability. On the positive side, a considerable private investment multiplier effect was achieved; beneficiary contributions of labor and money amount to over 90 per cent of the total public sector investment. This seems to suggest that the people were willing and able to pay for services or components where no govemment input was to be expected. Where the govemment was prepared to invest, however, and the people perceived that they would suffer no negative consequences from non-payment, they tended not to pay. Ex post facto cost-sharing in a govemment directed project in which people have a minimal role in planning and decision-making is shown to be problematic (Sources: Cousins and Goyer 1979; Shah 1989; from Schubeler 1996). Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXI Box 5. The Million Houses Program in Sri Lanka The launched in 1984 represented a shift in the role of govemment within housing from 'provider' with a high level of govemment expenditure to enabler and supporter of individual or household efforts with a much lower level of govemment expenditure. In the early 1980s, Sri Lanka's economic performance began to deteriorate, after several years of very rapid economic growth. There was also an escalation in costs of the direct construction program and of building materials in general; the direct construction program of the '100,000 houses program' had reached only around a third of its target before being suspended in 1983. The government had to reduce the allocation of public funds to housing and urban development. A Task Force set up to examine the results of the 100,000 Houses Program and to develop a new five-year program for housing identified various weaknesses including the high cost and unrealistically high standards implicit in a conventional public housing construction program, the failure to recognize the constraints on the key components of housing (e.g. finance, labor, building materials) and the failure to recognize the fact that most housing was constructed by households without government support and such housing was generally cheaper and more suited to their needs than govemment produced units. The recommendations of this Task Force contributed to the change in approach when a new housing program, the 'Million Houses Program' was announced in 1983. The Million Houses Program was based on 'minimal intervention but maximum support by the state; maximum involvement of the builder families.' The core of the program was small housing loans to low income rural and urban households to enable them to build or improve their houses. The Million Houses Program was made up of six sub-programs. Two were the responsibility of the National Housing Development Authority: the Urban and Rural Housing Sub-Programs. The other four were the Plantation Housing, the Mahaweli Housing, the Private Sector Housing and the-Major Settlement Scheme Housing Sub-Programs. In both the rural and the urban sub-programs, a considerable range of 'housing options and loans packages' were available for upgrading or repairing an existing house, for developing a new house where land is already available, for the provision of utilities (for instance water, sanitation, electricity) and registration of deeds, or where land shelter and services were needed. The rural sub-program was essentially loans to individual households; the urban sub-program for individual house-builders was far more organized around particular projects, because a loan program for individual house-builders was not sufficient to improve conditions. Most low-income urban households live in settlements without formal land tenure, so land tenure has to be regularized to create sufficient security of tenure to make the investment in housing worthwhile. Most urban settlements lacked roads, drains, piped water, and a system to dispose of human wastes and these could not be left to individual initiatives. Housing in urban areas is also subject to complicated planning and building regulations that need to be waived for low income housing projects. Thus, the project-based approach was indispensable. There was also a decentralization of decision-making, planning and implementation to local authorities and low- income communities themselves, largely through elected community development councils within each project area. The rural sub-program reached 258,762 families between 1984 and 1992; the urban sub-program reached 38,125. By 1991, the Urban Housing Division had started the implementation of over 300 low-income housing projects, more than 100 of which were in Colombo and these included 75 slum and shanty settlement regularization and upgrading projects and 21 sites and services scheme. The urban sub-program also pioneered Community Action Planning and Management - a methodology for poor communities to mobilize themselves and agree on what is to be done. The Million Housing Program was then superseded by the 1.5 million houses program that seeks to reach 1.5 million households between 1990 and 1995 with one of II sub-programs (Sources: Yap 1993; Hosaka 1993; UNCHS 1993; from Schubeler 1996). Page XXII Brazil: Progressive Low Income Housing Box 6. The Community Mortgage Program in the Philippines The Philippines government launched a Community Mortgage Program in 1988 to help poor urban households acquire title to the land they occupy and develop the site and their housing in 'blighted and depressed' areas. The Program was placed within the National Home Finance Corporation. It focused on the bottom 30 per cent of households to obtain access to housing especially those living illegally on land. The Program provides loans to allow community associations to acquire land on behalf of their members, improve the site, develop individual titling of the land and provide individual housing loans for home improvement or house construction. It is innovative in that it emphasized the importance of non-govemment groups in developing a solution to housing in the Philippines and its focus on achieving land security. To acquire the loan, the residents have to organize themselves into a community association which becomes responsible for collecting repayments and for ensuring that the loan continues to be serviced. If the development has to take place within an existing residential settlement, 90 per cent or more of the residents have to agree to be a party to the loan and improvement program. The land is purchased on behalf of the members and initially remains under common ownership of the association. It is the association which is responsible for collecting monthly rentals and amortization.from member beneficiaries until the community loan has been individualized. Both community based organizations and NGOs (and municipal governments) can take out loans and provide assistance in organizing member-beneficiaries and informing them about loan availability. Loans are available for up to 90 per cent of appraised value of property. Where residents were already on the site by February 25t' 1986, this valuation is not longer required if an agreement has been reached between the land owner and the community association as to the value of the land - although this must not exceed ten million pesos. Loans are provided for 25 years at 6 per cent interest. As of April 1994, the Program had already assisted 330 communities (39,992 households) with a total loan value of more than 835 million pesos. The average loan for each households have begun the process of obtaining a loan. Repayments rates are about 65 per cent. An essential part of the program is that either a government agency or an NGO acts as an intermediary, helping the residents form an association and supporting them while they negotiate for official title to the land, apply and secure the land, and begin site development and the financial operation of the loan. Just over half of the intermediary institutions which have acted as originators for loans are NGOs although some have been undertaken by local govemment units, national government agencies and financial institutions. Originators receive a small payment of 500 pesos per households from the government for this service (Source: UNCHS 1996). Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXIII Box 7. The Urban Community Development Office (UCDO) in Thailand In March 1992, the Thai government approved a budget of the equivalent of US$50 million to initiate an Urban Poor Development Program within the Seventh National Economic and Social Development Plan. The approval of this budget led to the establishment of a new organization, the 'Urban Community Development Office' with responsibility for implementing the Program nationwide. By the end of June 1995, loans totaling some 340 million baht (around $13.6 million) had been approved, benefiting some 140 communities with around 11,500 families. Loans had been made both in Bangkok and in the regional cities. By the end of 1996 the Office aims to have more than 850 'saving and credit organizations' in urban communities and more than 120,000 households belonging to them. It is also intended that at least 80 per cent of the urban poor who are members of 'savings and credit organizations' will have secure jobs and reasonably secure incomes. A considerable range of regional NGOs, federations of community organizations, municipalities and professionals are also involved in these initiatives. The overall objective of the Program is to strengthen the capacity of the urban poor and those living in illegal settlements to obtain increased and secure incomes, appropriate housing with secure rights, an improved environment and better living conditions. The people are considered to be the main actors in the community development process. The Office considers that local authorities, NGOs and other agencies external to the community, both governmental and non-governmental, also have an essential role but as supporters of the people's development process. Based on this understanding, there are six main objectives: I. To assist in the setting up of community savings groups and support for the operation of community saving and credit activities. 2. To support and enhance the managerial and organizational capacity of the saving groups. 3. To support the formation and development of community savings and credit federations and other forms of community networking. 4. To provide integrated development credit for communities (including credit for housing construction or improvement and micro-enterprises). 5. To support community groups in planning for an integrated community development process. 6. To support those already working in development activities to work together and enhance their working capacity. The Urban Community Development Office is a special organization within the National Housing Authority, but its Board has the power to make decisions autonomously. The concept of partnership is fundamental to the composition of the Board, which has nine members. Three are drawn from government organizations (the Bank of Thailand, the Finance Ministry and the National Economic and Social Development Board), three are community leaders, two come from NGOs, and the final member is from the private sector. The Governor of the National Housing Authority is the Chair of the Board and the Managing Director of the Development Office acts as Secretary to the Board. A major aim of the Urban Community Development Office is to strengthen communities' managerial capacity. It seeks to encourage and enable, in all possible ways, different forms of community organizations and federations to be set up and be involved in the Program. 'Credit' is used as a mechanism to strengthen the capability of communities to collectively deal with their own development issues. To be eligible for a loan, each community must operate a savings program, with a clearly defined management structure, for at least three months. Loans can be granted for up to a maximum of ten times the amount saved. All loans are conditional on there being a community management structure and operating mechanism as well as clearly defined beneficiaries and a project development management process. All committee members are required to sign their names as guarantors of the loan. The Development Office considers that the resources it provides should be supplementary to the financial resources of the communities themselves. Through the careful use of credit, the role of the community as initiator, organizer, planner and manager is enhanced. At every stage, the role of the community as the main actor in the implementation process as well as the main actor in development is reinforced. The credit system and associated management activities seek to gradually change the quality of the community organization to enable it to become a strong development unit able to more efficiently deal with forces extemal to the community as well as becoming a nucleus for an intemal integrated development process. The Program provides an 'integrated credit system' for community development with different loans available, including those for income generation, small revolving funds and housing. This approach responds to the multiple and related needs of the communities. The Office charges an average rate of 7 per cent interest a year with differential interest being charged on community revolving funds (10 percent), housing (3 per cent) and micro-enterprise credit (8 per cent). A further 2-5 per cent may be added by communities to provide their organizations with a small fund for collective projects. This enables the savings and credit organization to increase and strengthen its capacity. By mid- 1995, around two-thirds of the credits had gone to 15 new housing projects developed by slum community organizations that registered themselves into housing co-operatives and around 20 housing improvement projects. One- fourth had gone to income-generation but since income-generation credits were on average much smaller than credits Page XXIV Brazil: Progressive Low Income Housing Box 8. Reaching the Poor: Khuda Ki Basti, Incremental Housing Scheme, Hyderabad, Pakistan In Karachi, Pakistan, the Karachi Development Authority developed about 165,000 new plots on the outskirts of the city between 1978 and 1983 to meet the demand of low-income households. By 1987 about 20 percent of these plots had been provided with infrastructure services; however less than I percent of the plots were actually occupied. Overall it was estimated in 1986 that there were a total of 380,000 developed but vacant plots in Karachi, covering an area of over 4,850 hectares. In Hyderababd, a city of about one million inhabitants, the situation was similar. Of the 56,000 plots that had been developed by 1986, about 52,000 remained vacant. This anomaly arose in the face of an extremely high demand for urban housing, as demonstrated by the rapid expansion of illegally occupied katchi abadi in both cities. In Karachi, for example, new katchi abadi were estimated to have added about 217,000 dwelling units to total housing stock from 1972 to 1986. In 1986, the director of the Hyderababd Development Authority (HAD) and his colleagues set out to duplicate the success of the informal/illegal housing development process (without its shortcomings) within the HAD sites and services scheme at Gulshan E Shabaz. They began by identifying the basic reasons for the failure of the govemment programs and the relative success of the informal program. These concemed the targeting of the beneficiaries, affordability of the plots and services, administrative procedures, and timing of the process. The analysis and the response to the scheme, named Khuda Ki Basti, were as follows: Targeting: Govemment programs, though normally intended for low-income groups make it practically impossible for low-income households to participate, By announcing a scheme, collecting advance payments from plot buyers, and then improving the land, the state has in effect assumed the role of developer. The process may take up to ten years, however and is thus targeted mainly toward middle-income people looking for an investment hedge against inflation, and not low-income households with immediate housing needs. In Khuda Ki Basti, the essential criteria for plot allocation are the physical presence of the family and actual construction of a house on a site. Transfer of title is separated from physical possession of the plot. Only after a house has been constructed and occupied without interruption, and all installment payments are completed, is the actual transfer of title made. Affordability: Govemrnent schemes sell serviced plots with roads, drains, water supply, sewers, and electricity. However, the cost of land plus services is beyond the reach of low-income households. In Khuda Ki Basti, as in most katchi abadi settlements, the site is occupied first and infrastructure services are developed incrementally as the capacity of the inhabitants allows. At the outset, basic needs for water are met by tank truck or public water taps. Transport facilities are arranged through private informal mini-bus operators. Plot purchases make a down payment of Rs. 460 (later raised to Rs. 1000 or 10 percent of the total cost) and pay the remainder over a period of eight years. Procedures: The allocation of govemment-sponsored plots is generally accomplished through a complicated balloting procedure requiring application fomis and formalities, these procedures are beyond the capability of most marginally literate low-income households. In Khuda Ki Basti, the physical presence of the household is the crucial step in applying for a lot. To facilitate the process the scheme has built a "reception area" consisting of a series of one-room houses where applicants can stay while awaiting plot allocation and subsequent construction of the basic housing unit ("live and build"). Timing: The long wait between advance payment and land possession, common to govemment programs, is totally eliminated in the Khuda Ki Basti scheme; possession is immediate. Results: The experimental incremental housing scheme has been very successful. At the outset it was necessary to revoke the occupancy rights of some applicants with speculative interests who did not begin construction. Once people saw that the project meant business, the targeting procedure began to work as planned. By September, 1988 (after about two years) 3,241 plots had been allocated, 2,525 families were living on the site, and 2,481 pucca houses had been constructed. This contrasts with the remainder of the Gulshan E Shahbaz site, where all 11,600 plots were still vacant. A certain experimentation was also required regarding the provision of common facilities. When segments of the provisional nylon water pipe network began to disappear, the scheme stopped providing facilities centrally. Instead, inhabitants were organized at the lane-level to collect the necessary funds and replace the pipeline, which then remained undamaged. For the provision of other facilities such as sewerage, drainage, and intemal roads, inhabitants have been organized at the level of blocks of about 250 households. Improvements were only implemented after the residents had collected the required capital. Other facilities such as electric power were organized at the settlement level. By September 1988, self-managed infrastructure development had provided 10,970 meters of water sup[ply network, 90 public standposts, and 4,870 meters of sewerage line. About 60 percent of the housing units had soak pit latrines. There were eight mosques, eleven health clinics, fifteen schools of various levels, and twenty-nine regular teachers active in the settlement. A family enterprise scheme had created employment for 117 people, including 48 women. Cost recovery on the plot investment, reviewed in 1989, was reasonably good, with 60 percent of the residents paying regularly, 28 percent paying with some irregularity, and 12 percent not paying. By 1992, however, it appeared that the repayment rate had declined. In principle, land title is only granted after payment is complete. The slower rate of recovery means that the scheme will be running into cash flow problems in the planned "extemal" infrastructure investments. The development remains entirely self- financed. Only administrative overhead is paid for out of the HAD budget (Siddiqui 1988; Van der Linden 1989; TP O'Sullivan & Partners/PADCO 1991; Sultan 1993; from Schubeler 1996). Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXV Box 9. The National Fund for Popular Housing (FONHAPO) in Mexico FONHAPO is a national fund for low-income housing established by the Mexican govemment in late 1981. The fund sought to: * Double the low-income housing stock between 1980-2000; Finance housing construction by people with modest incomes who organized the development of their own housing; * Finance incremental housing development, not just new houses; * Redistribute national govemment funds in favor of people with low incomes to reinforce principles of equity and social justice; * support the development of democratic processes from the community upwards; * ensure that govermment support was used efficiently; * be decentralized; and institutionalize the fund within govemment. FONHAPO was designed by professionals with long experience of working with low-income communities and realistic assumptions were made about how much the people could afford and where credit was needed. It authorized credit principally to public sector organizations (such as housing or state-level real-estate institutions and municipalities) and to organized community groups seeking housing (such as housing co-operatives, civil associations, unions, and peasant organizations). All groups receiving credit had to be legally constituted and the credit was authorized to lend to the group as a whole, not to its individual members. When the Fund was fully operative, there were 800 accounts and 250,000 households receiving loans. Title to the land on which the housing units were developed or improved could not be individualized until the loan had been completely repaid. In some cases, communities maintained a collective land holding. Occasionally, a single loan was turned into two loans if there were divisions in the community. The most common reason for this was that one group wished to repay the loan faster than other groups. Credit could be used to fund any part of the housing process, with great flexibility for people to choose how the funding was used. Five main programs developed: site and services, incremental housing development, improvement and/or extension of existing housing, new finished houses and the production and distribution of building materials. The credit lines correspond to different phases of the housing process: site acquisition, technical assistance to develop projects, provision of basic services (water, drainage, electricity and roads); and either mutual aid construction processes or construction by third parties. The combination of possibilities gave the fund great flexibility to respond to the different needs, potentials and preferences of the participating groups. The program had very few technical standards but many standards for social processes. For example, there were no limits to the size of rooms but communities had to have regular meetings and fixed numbers had to participate when making critical collective decisions. The financial system was designed according to the ability of the people to repay. The maximum size of loan was 2,000 times the daily minimum wage of the region although the actual amount of credit was set according to the repayment possibilities of the soliciting group. A deposit of 10-15 per cent of the value of the loan was required before the loan was provided and the value of loan repayments could not exceed 25 per cent of the monthly income of the head of the household. Loans were given for 7-8 years. Communities that successfully repaid their first loan could apply for a second loan. Land was used as collateral for the loan. A subsidy was included in the program and half the total value of this subsidy was earmarked for those with an income of less than 1.5 times the minimum monthly wage. Up to 40 per cent of the total cost of a project could be subsidized; 25 per cent was given immediately and the remaining 15 per cent if the repayment schedule was maintained. Repayment did not follow a conventional form with the repayment of capital and interest. At the time the loan was taken out, the equivalent value of a monthly repayment (at nine per cent interest rate) was converted into a percentage of the minimum wage. As the monthly minimum wage changed, so did the value of the repayment. This system of calculating repayments was supported by the communities who recognized the need to maintain the capital value of the Fund and felt that their repayments were fair. An additional.charge equivalent to an interest rate of 2 per cent was later added and passed to the communities to cover administration costs. In 1983, when the program began operations, 25 per cent of all loans were given to communities and 75 per cent to other institutions such as local and state govemment departments. Within a few years, the structure of the program had succeeded in shifting a much greater share of resources to community development. Between 1985.and 1987, 66 per cent of the total number of loans were to community organizations. Administrative costs fell from 10 per cent of the total value of loans at the beginning of the program to 4 per cent after five years. The 4 percent govemment allocation to housing which was given to FONHAPO was soon producing 22 percent of all govemment Funded housing investment. In the first year, 5,000 units were completed and 15,000 started; by 1987, 63,000 units had been completed and a further 53,000 started. ('Units' does not mean a new house but a completed program of incremental improvement.) (Source: UNCHS 1996.) Page XXVI Brazil: Progressive Low Income Housing Box 10. El Mezquital El Mezquital is a collection of low-income informal settlements situated on the periphery of Guatemala City, Guatemala. Its total population is about 37,600, most of whom are very poor. For years the communities lacked even the most essential of urban services. The initial impetus to organize arose following a number of cholera outbreaks that adversely affected the settlements. In the early 1990s the World Bank, in collaboration with UNICEF and the Government of Guatemala, provided a large municipal reform loan to help improve the living conditions in these neighborhoods. By most accounts the project is a success. With an average investment of US$1,200 per household, the project financed improvements in the spatial distribution of these neighborhoods, removed families from high-risk and environmentally sensitive areas, improved pedestrian and vehicular access, and provided investments in new water and sanitation systems (including low-cost wastewater treatment plants), public lighting, drainage, and electricity supply, among others. For the residents of El Mezquital, the quality of life improved substantially as a result of the project (Table 27). In addition, they enjoyed greater security of tenure and had made significant investments and improvements in their dwellings. Crime rates had fallen, and so had the incidence of some diseases. Better living conditions had translated into higher real estate values: Land values had increased 11 times; the estimated economic rate of return was 13.4 percent. Indicators of Social and Environmental Improvements (El Mezquital Site) dici 0torA An 1Reporte Cdme, Rate 35*, a 'ordbPdityRate 48%: t 81' Cb6ildr 4.withDirhea....... -> 25%B *i>.-w* < ' .S1kijisaes < *-t 23% $ ' * J%~ -< hifayi iMortali Rate (children 1-5 yrs3 ~~,i,000Mving - . ~80 35 . gpac ;Sta ddsperFamilyl(m) . - .. -40' ,. °9.XW2 :Street hiliken from El Mezquitl * 15% - 12% W44V Sources: Mezquital Health Centers 1995 and CHILDHOPE Foundation 1992 and 1997. The most valuable lesson of El Mezquital is the importance of a strong community participation. In the two major project sites, Carolingia and El Mezquital, community-based organizations ensured that the costs of investments were recovered and that families paid for the consumption of urban services. These organizations took the responsibility for keeping a record of families who contributed, printed, and distributed bills; opened accounts for families; and made collections. For the community-run water supply and sanitation system, cost recovery was done through the collection of tariffs and each household was individually metered. A community water cooperative (COIVIESS) managed the water and sanitation systems. At the time of evaluation (1998) there were no delinquent payments. An urban improvement fund had been created to collect "voluntary" fees from the community (a proxy for betterment fees). These funds were used to cover maintenance costs and finance future investments. The community had accepted the principle of cost recovery since they knew that those funds would be used to address needs. COIVIESS had also formed a solid waste collection company and managed a revolving loan fund for investments in housing. Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXVII Box 11. Direct Housing Subsidy and Housing Financing Chile's approach to low-income housing is a system based on shared financing where families who contribute with their initial savings are provided with financing assistance in the form of direct subsidies. They are responsible for paying a mortgage credit. Both financing and house construction are provided by the private sector. Subsidies are granted on the basis of need, affordability, and initial savings. The state acts as financier and coordinator rather than provider. At present there are 14 such programs run by the Ministry of Housing and City Planning. The most important programs are the unified subsidy and the basic house programs, together accounting for 60 percent of the 100,000 houses built every year in Chile. * Unified subsidv: This program is available for families who can afford a small down payment and the payment of a mortgage loan. Depending on the size of the house and its price, the subsidy is granted up front. For example, for houses costing up to US$ 15,250, the subsidy is US$3,600 (more than 20 percent). For second and third class houses - priced between US$ 15.250 and US$ 30.500, and from US$ 30.500 up to US$ 45.750 respectively - the up-front subsidy is smaller. . Basic House Program: The benefits of the program are accessible to poorer families who do not own a house and who live as allegados (people without a residence of their own and who stay in other people's homes) or tenants. Under this program, the families must still contribute with initial savings and pay the optional mortgage credit. There are two modalities in this program: a) HUS modality - finances unfinished new houses with surface areas ranging between 38 and 42 square meters; with bathroom, kitchen, sitting room and two bedrooms; and serviced with water, sewerage, and electricity. The cost is about USS 7,010 to be financed by the household beneficiary's initial savings amounting to US$302, a State subsidy of USS 4,270, and an optional mortgage credit of US$ 2,440. b) Free choice: A certificate of subsidy is obtained to purchase a new or old social housing unit, or to build a house in the case of plot-owners. The maximum value of the house is US$ 12,200: USS 4,270 comes from State subsidy, US$ 610 from the beneficiary's initial savings, and US$ 3,050 from the optional mortgage credit. Chile's other housing progranis respond to specific contextual needs or to special circumstances of particular groups: subsidy for the elderly, progressive housing, estate densification, special program for workers, rural subsidy, urban renewal, and house leasing. Chile's direct housing subsidy program has been successful owing to several factors: families are free to choose the houses in the private market, whether old or new; the direct subsidy system has reached the lowest income group - accounting for 30 percent of the housing deficit; and the mechanism to apply for a housing subsidy and consequent allocation of benefits is transparent and credible (information on the beneficiary households' names are disclosed). Defaults have been minimal and housing savings accounts have increased considerably. There are now over a million housing savings accounts in different banks throughout the country. The program attracted private sector participation in project design, the use of new technology, house financing and production, marketing of social housing units, and mobilizing savings and granting of long-term mortgage credits for low-income housing. The system allows the state entity to focus mainly on the planning tasks, and granting of direct subsidies to families and development of new ways of attention within a framework of competition among families in similar conditions. Sustainability. The direct subsidy system has responded to the existing housing demand in the country through a regular and clear system of application, selection, and allocation of housing subsidies. It has also attracted the private sector in terms of investment and permanent production of social houses in response to a sustained and financed demand. The generation of a steady demand for housing has contributed to the stabilization of the traditionally cyclical housing market. Impact. The program contributed to easing the housing shortage. Since 1992, about 100,000 houses have been built annually, of which 70 percent are for low-income families. The program has also established the credibility of the direct subsidy system and has generated a savings habit among the low-income population. Page XXVIII Brazil: Progressive Low Income Housing Box 12. The ARGOZ Case: A Private Developer Adopts a Colonia Approach An example of a successful program offering "progressive subdivision development" to poor families is the private company ARGOZ. Since its establishment 15 years ago, ARGOZ has developed 500 subdivisions throughout El Salvador, with a total of 250,000 plots. The land developed by ARGOZ is unencumbered, arable land on city fringes. This allows the immediate legalization of the lease contract, which includes a purchasing option. In case of a late payment, 50 percent of the paid installment is retumed to the lessee. The lessee also has the option to reclaim his/her housing investment, and to negotiate directly the sale of his/her lot. ARGOZ is particularly appealing to its clients because of the security of its contract. The company signs an insurance policy for the debt balance in favor of the beneficiaries of the title holder, at no additional cost to the latter. ARGOZ also has a credit system for the lot holders, for both home improvement and expansion, as well as for other family needs. ARGOZ provides a viable, private sector altemative towards improving access to land and services by low-income families on a scale never seen before in El Salvador. Its privately funded land development program has expanded the country's chief municipalities by more than 56 percent and has helped curb squatting in those cities to a large extent. The success of ARGOZ stems from its ability to offer plot holders an efficient loan system in a program that generates both profits and social awareness. Furthermore, the company has demonstrated that being cost efficient pem-its it to direct generous profit margins to community-related projects, e.g., rehabilitating and / or upgrading facilities such as intemal electric power lines. The experience of ARGOZ underscores the key role of a sustainable housing program: access by low-income groups and a link with the formal housing sector. In the ARGOZ case, access depends on available financing and unit costs of the project. The developer, with a minimum investment, can secure a guarantee in his favor and attain a profit rate of close to 500 percent. In that way he can offer poor clients a medium-term loan (5-10 years) payable by relatively small, affordable installments. Other factors that contribute significantly to the success of ARGOZ are the informal, "user-friendly" setting for its operation (as opposed to that of a formal banking institution or a bureaucratic govemment agency which tends to alienate poor clients), a smooth-running institutional structure that clients find easy and expedient to work with, and a policy and approach that gives power to the communities and municipalities to provide full services over time. To conclude, as an altemative housing policy ARGOZ offers great benefits on a large scale: accessibility, easy financing, a non-bureaucratic treatment, as well as flexible progressive development. This example in El Salvador shows that private land development companies have a large potential to confront the housing problem of the urban poor successfully. Brazil Low-Income Housing: Altematives for the Urban Poor Page XXIX Annex B- Statistical Appendix Table 1: Population Trends and Demographic Densities in Brazil, 1940-2000 Years/Regions Brazil North Northeast Southeast South Center-West 1940 Rural 28,356,133 1,056,628 11,052,907 11,113,926 4,144,830 987,842 Urban 12,880,182 405,792 3,381,173 7,231,905 1,590,475 270,837 Total Population 41,236,315 1,462,420 14,434,080 18,345,831 5,735,305 1,258,679 Urbanization Rate (%) 31.24% 27.75% 23.42% 39.42% 27.73% 21.52% Demographic Density 4.88 0.41 9.36 19.97 10.2 0.67 1950 Rural 33,161,506 1,263,788 13,228,605 11,827,760 5,527,885 1,313,468 Urban 18,782,891 580,867 4,744,808 10,720,734 2,312,985 423,497 Total Population 51,944,397 1,844,655 17,973,413 22,548,494 7,840,870 1,736,965 Urbanization Rate (%) 36.16% 31.49% 26.40% 47.55% 29.50% 24.38% Demographic Density 6.14 0.52 11.65 24.54 13.95 0.92 1960 Rural 38,767,423 1,604,064 14,665,380 13,169,831 7,392,384 1,935,764 Urban 31,303,034 957,718 7,516,500 17,460,897 4,360,691 1,007,228 Total Population 70,070,457 2,561,782 22,181,880 30,630,728 11,753,075 2,942,992 Urbanization Rate (%) 44.67% 37.38% 33.89% 57.00% 37.10% 34.22% Demographic Density 8.29 0.72 14.38 33.34 20.91 1.57 1970 Rural 41,054,053 1,977,260 16,358,950 10,888,897 9,193,066 2,635,880 Urban 52,084,984 1,626,600 11,752,977 28,964,601 7,303,427 2,437,379 Total Population 93,139,037 3,603,860 28,111,927 39,853,498 16,496,493 5,073,259 Urbanization Rate (%) 55.92%. 45.13% 41.81% 72.68% 44.27% 48.04% Demographic Density 11.01 1.01 18.23 43.38 29.35 2.7 1980 Rural 38,566,297 2,843,118 17,245,514 8,894,044 7,153,423 2,430,198 Urban 80,436,409 3,037,150 17,566,842 42,840,081 11,877,739 5,114,597 Total Population 119,002,706 5,880,268 34,812,356 51,734,125 19,031,162 7,544,795 Urbanization Rate (%) 67.59% 51.65% 50.46% 82.81% 62.41% 67.79% Demographic Density 14.07 1.65 22.57 56.31 33.86 4.01 1991 Rural 35,834,485 4,107,982 16,721,261 7,514,418 5,726,345 1,764,479 Urban 110,990,990 5,922,574 25,776,279 55,225,983 16,403,032 7,663,122 Total Population 146,825,475 10,030,556 42,497,540 62,740,401 22,129,377 9,427,601 Urbanization Rate (%) 75.59% 59.05% 60.65% 88.02% 74.12% 81.28% Demographic Density 17.18 2.59 27.22 67.66 38.34 5.85 1996 Rural 33,993,332 4,249,174 15,575,102 7,176,774 5,356,639 1,635,643 Urban 123,076,831 7,039,085 29,191,749 59,823,964 18,157,097 8,864,936 Total Population 157,070,163 11,288,259 44,766,851 67,000,738 23,513,736 10,500,579 Urbanization Rate (%) 78.36% 62.36% 65.21% 89.29% 77.22% 84A2% Demographic Density 18.38 2.92 28.68 72.25 40.74 6.51 2000 Rural 31,847,004 3,914,152 14,759,714 6,851,646 4,780,924 1,540,568 Urban 137,697,439 9,005,797 32,919,667 65,410,765 20,290,287 10,070,923 Total Population 169,544,443 12,919,949 47,679,381 72,262,411 25,071,211 11,611,491 Urbanization Rate (%) 81.22% 69.70% 69.04% 90.52% 80.93% 86.73% Demographic Density 19.92 3.35 30.69 78.2 43.54 7.23 Source: Brazilian Institute of Geography and Statistics - IBGE. Page XXX Brazil: Progressive Low Income Housing Table 2: Annual Growth of Population (in percentage) BRAZIL North Northeast Southeast South Center- West 1940/1950 Total 2.39 2.29 2.27 2.14 3.25 3.41 Urban 3.91 3.71 3.51 4.08 3.88 4.65 Rural 1.60 1.84 1.84 0.64 2.97 2.94 1950/1960 Total 2.99 3.34 2.08 3.06 .4.07 5.36 Urban 5.15 5.04 4.63 4.91 6.44 8.90 Rural 1.55 2.37 1.02 1.06 2.90 3.89 1960/1970 Total 2.89 3.47 2.40 2.67 3.45 5.60 Urban 5.22 5.44 4.57 5.19 5.29 9.94 Rural 0.57 2.11 1.10 1.88 2.20 3.14 1970/1980 Total 2.48 5.02 2.16 2.64 1.44 4.05 Urban 4.44 6.44 4.10 3.99 4.98 7.69 Rural 0.62 3.70 0.53 2.00 2.48 0.81 1980/1991 Total 1.93 3.85 1.83 1.77 1.38 3.01 Urban 2.97 5.37 3.55 2.34 2.98 4.30 Rural (-)0.67 2.04 (-)0.28 (-)1.52 (-)2.00 (-)1.06 1991/1996 Total 1.38 2.43 1.06 1.35 1.24 2.22 Urban 2.12 3.58 2.56 1.64 2.09 3.01 Rural (-)1.07 0.69 (-)1.43 (-)0.93 (-)1.35 (-)1.53 1991/2000 Total 1.63 2.88 1.30 1.60 1.41 2.36 Urban 2.45 4.81 2.78 1.92 2.41 3.11 Rural -1.31 -0.54 -1.39 -1.03 -2.00 -1.51 Source: IBGE - Censos Demograficos 1940-91, Contagem 1996 e Dados Preliminares do Censo 2002 Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXXI Table 3: Population/Households, 1998 Population Households Density Regions North 11827 2607 4.54 Northeast 45702 10766 4.25 Southeast 68767 19315 3.56 South 24089 6933 3.47 Center West 10955 2980 3.68 Total 161340 42600 3.79 Areas Metropolitanas SAo Paulo 17054 4788 3.56 Rio de Janeiro 19346 2886 6.70 Belo Horizonte 3958 1048 3.78 Porto Alegre 3338 1016 3.29 Recife 3157 794 3.98 Salvador 2798 740 3.78 Curitiba 2585 6712 0.39 Fortaweza 2700 643 4.20 Beldm 1550 355 4.37 Total Areas Metropolitanas 47485 12932 3.67 Total Brazil 161340 42699 3.78 Sourcee: IBGE - Censos Demograficos 1940-91 Table 4: Brazil - Housing Deficit in 1991 Type of Deficit Metropolitan Other Urban Rural Areas Total Metropolitan Other Urban Rural Total Regions Areas Regions Areas Areas (In Thousands) (In Percentage of Total) ImprovisedlHomeless 33 59 55 147 0.7 1.2 1.1 2.9 Precarious Housing 230 434 1063 1728 4.6 8.7 21.3 34.5 Two or more fanmilies 980 1621 528 3129 19.6 32.4 10.6 62.5 Total Deficit 1244 2114 1646 5003 24.9 42.3 32.9 100.0 Source: Fundacao Joao Pinheiro 1995 and Morais, Piedade 2000. Page XXXII Brazil: Progressive Low Income Housing Table 5: Housing Deficit by Regions and Spatial Groupings, 1991 Regions Improvised/ Precarious Two or more Total Metropolitan Other Urban Rural Total Homeless Housing families Areas Areas Areas North 24 18 196 238 67 155 - 222 Northeast 32 1,268 929 2,229 313 767 1,149 2,229 Southeast 47 272 1,342 1,661 753 690 217 1,660 South 23 80 408 511 110 244 158 512 Center/West 21 91 253 365 257 107 364 Total Brazil 147 1,728 3,128 5,003 1,243 2,113 1,631 4,987 (In Percentagel North 0.5 0.4 3.9 4.8 1.3 3.1 0.0 4.4 Northeast 0.6 25.4 18.6 44.6 6.3 15.3 23.0 44.6 Southeast 0.9 5.4 26.8 33.2 15.1 13.8 4.3 33.2 South 0.5 1.6 8.2 10.2 2.2 4.9 3.2 10.2 Center/West 0.4 1.8 5.1 7.3 0.0 5.1 2.1 7.3 Total Brazil 2.9 34.5 62.5 100.0 24.8 42.2 32.6 99.7 Source: Fundacao Joao Pinheiro 1995 and Morais, Piedade 2000. *In thousands Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXXIHI Table 6: Type of Housing Occupancy by Region, Poor vs. Non-poor Urban Area Rent (%) Own Live in Live in All (%) (%) Ceded Invaded Property Property (%) (%) Metropolitan Regions, NE Fortaleza Poor 15 71 14 0 100 Non-poor 16 73 11 0 100 Recife Poor 17 66 16 1 100 Non-poor 18 76 5 0 100 Salvador Poor 10 74 7 9 100 Non-poor 20 73 5 1 100 Metropolitan Regions, SE Belo Horizonte Poor 14 70 11 4 100 Non-poor 17 72 11 0 100 Rio de Janeiro Poor 22 48 14 16 100 Non-poor 21 64 13 2 100 Sao Paulo Poor 22 52 21 5 100 Non-poor 19 72 7 2 100 Non-Metro Urban Areas Northeast Poor 14 69 14 3 100 Non-poor 18 72 9 I 100 Southeast Poor 14 67 18 0 100 Non-poor 17 73 9 1 100 All Urban Areas Poor 15 66 15 3 100 Non-poor 18 72 9 1 100 Source: Calculated from 1996/97 PPV Survey Page XXXIV Brazil: Progressive Low Income Housing Table 7: Selected Elements of Low-income Housing, 1993/1998 US$ 1993 1998 Average Rent 542 1759 Annual average income 4652 6207 Average income of rentees 4883 6724 Rent/income 0.11 0.26 1993 1998 Formal ownership Own 58.3 62.2 Being purchased 7.5 6.7 Tenencia 18.6 16.3 Invasion 5.7 5.7 Other 9.9 9.2 Total 100.0 100.0 Source: Morais, 2000 Table 8: Brazil: Average Monthly Urban Household Expenditure on Rent/Mortgage Consumption All Expenditures Housing Share of Housing (%) Decile (R$/month) Expenditure (R$/month) 1 236 51 21.8 2 347 71 20.4 3 395 83 21.0 4 471 108 22.8 5 596 153 25.7 6 647 161 24.8 7 836 189 22.6 8 992 265 26.7 9 1368 283 20.7 10 2522 526 20.9 Source( Calculated from 1996/97 PPV Survey. Table 9:Latin America Expenditures on Housing Improvement Average Income per Estimated Expenditures on Housing Expenditure as Month (in US$) Housing per month % of monthly Income (in US$)) <100 18 27 100- 150 35 26 151-200 48 27 201-250 62 27 251-300 70 25 301-350 76 24 351-400 106 28 >400 183 34 Source: Ferguson, op. cit. Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXXV Table 10 Cost Breakdown of Urban Upgrading, Recife 1998 Item R$ per Household Percent of Total Infrastructure Access Road 450 6.3 Drainage 350 4.9 Piped Water 300 4.2 Sewerage 750 10.4 Electricity 140 1.9 Subtotal 1,990 27.7 Land, Structure and Other Lot Clearing and Leveling 500 7.0 House 1/ 2,800 39.0 Land Acquisition 1,300 18.1 Miscellaneous 595 8.3 Subtotal 5,195 72.3 Total 7,185 100.0 1/ Brick, 32 square meters, including material and construction. Source: Estimated by COHAB based on Habitar-Brazil projects in Recife Metropolitan Region. Table 11 Poor Households in Metropolitan Areas, 1993/ 1998 Metropolitan 1993 1998 Annual Average Growth Regions Center Periphery Total Center Periphery Total Center Periphery Rio Janeiro 1564 1306 2870 1799 1372 3171 2.84 0.99 2.01 Poor 399 629 1028 289 401 690 -6.25 -8.61 -7.66 % 25.5 48.2 35.8 16.1 29.2 21.8 -9.49 Sao Paulo 2704.4 1602.3 4306.7 2730 1940 4670 0.19 3.90 1.63 Poor 812 668 1480 543 646 1189 -7.73 -0.67 -4.28 % 30.0 41.7 34.4 19.9 33.3 25.5 -5.82 Curitiba 394 205 599 410 304 * 714 0.80 8.20 3.57 Poor 74.3 82.5 156.8 41 77 118 -11.21 -1.37 -5.53 % 18.9 40.2 26.2 10.0 25.3 16.5 -8.79 Porto Alegre 390.5 517.3 907.8 421 576 997 1.52 2.17 1.89 Poor 54.90 106.3 161.2 34 66 100 -9.14 -9.09 -9.11 % 14.1 20.5 17.8 8.1 11.5 10.0 -10.80 Belo Horizonte 523.9 366.4 890.3 568 477 1045 1.63 5.42 3.26 Poor 155 172.8 327.8 104 176 280 -7.67 0.37 -3.10 % 29.6 47.2 36.8 18.3 36.9 26.8 -6.16 Salvador 511.9 104.8 616.7 600 120 720 3.23 2.75 3.15 Poor 229 67 296 205 48 253 -2.19 -6.45 -3.09 % 44.7 63.9 48.0 34.2 40.0 35.1 -6.05 Recife 319.00 380 699 353 438 791 2.05 2.88 2.50 Poor 170.4 222 392.4 139 213 352 -3.99 -0.82 -2.15 % 53.4 58.4 56.1 39.4 48.6 44.5 -4.54 Fortaleza 434 122 556 496 158 654 2.71 5.31 3.30 Poor 192.50 79.8 272.3 152 75 227 -4.61 -1.23 -3.57 % 44.4 65.4 49.0 30.6 47.5 34.7 -6.65 Belem 184.6 20.6 205.2 195 25 220 1.10 3.95 1.40 Poor 57.00 10.8 67.8 46 9 55 -4.20 -3.58 -4.10 % 30.9 52.4 33.0 23.6 36.0 25.0 -5.42 Brasilia 418.40 418.4 508 508 3.96 #DIV/O! 3.96 Poor 164.00 164 148 148 -2.03 #DIV/0! -2.03 % 39.2 39.2 29.1 #DIV/0! 29.1 -5.76 Total 6,920.80 4,258.00 11,178.80 7,512.00 4,933.00 12,445 1.65 2.99 2.17 Note: Date for Brasilis is limited to the Federal District and not to the Metropolitan Region Source: SerTa and Barralides 2000. Page XXXVI Brazil: Progressive Low Income Housing Table 12: Total Funds that could be mobilized by the low-income population Assumptions: Saving Period: 3 years; Credit Period: 5 years; Effective interest rate: 2% per month in US$ Average Income Monthly Payment Maximum Savings Maximum Credit Total Funds that per month (in US $) capacity (in US$) amount (in US$) amount (in US$) could be mobilized (in US$) <100 18 648 626 1,274 100- 150 35 1,260 1,217 2,477 151-200 48 1,728 1,669 3.397 201-250 62 2,232 2,115 4,387 251-300 70 2,520 2,433 4,953 301-350 76 2,736 2,642 5,378 351-400 106 3,816 3,685 7,501 401-500 112 4,032 3,893 7,925 501-600 137 4,932 4,762 9,694 Source: Ferguson, op. cit. Table 13: Hypothetical Micro Credit for Housing Loans Type of Loan Micro-enterprise Housing Complete Basic Improvement Solution Loan Amount US$100-5000 US$500 - 2,000 US$1,000 - 5,500 Maturity <2 years <3 years < 8 years Interest Rate Cost-covering, not Cost-covering not Cost-covering indexed indexed indexed Type of Collateral Little or no real Real Assets covering Real Assets covering assets, possible loss possible loss possible loss not covered Credit Analysis Cash-flow based Cash-flow based Cash-flow based Savings Requirement No No Yes TA to borrowers No Some More Standardization No No Yes Source: Ferguson, op. cit. Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXXVII Table 14: Summary of Main Programs in Housing, 1995-2000 in Millions of Reais Investment Structure Loan Counterpart % Benefic. Generated Inv. Per Inv. (%) (Families) Employ- /Family Empi., ('000) ment, in ('000 R$) thousand thousands Reais HabitarBrazil 1133 9.5 916 216 19.1 457 107 2.5 10.6 Pro Moradia 1101 9.2 757 343 31.2 332 105 3.3 10.5 Carta de Crddito 0.0 Individual 7217 60.5 6495 721 10.0 528 148 13.7 48.8 Asociafivo 2274 19.1 0.2 406 17.9 115 217 19.8 10.5 Apoio a producao 113 0.9 54 59 52.2 3053 10.8 0.0 10.5 Total 11932 100.0 10159 1773 14.9 1443 597 8.3 20.0 Fed. Budget 1132 9.5 917 216 19.1 458 108 2.5 10.5 FGTS 10800 90.5 9242 1557 14.4 985 489 11.0 22.1 Source:Secretaria de Estado de Desenvolvimento Urbano (SEDU). Table 15: Investment in Housing during 1995-2000 (million R$) 1995 1996 1997 1998 1999 2000 Habitar Brazil Subsidy .05 .67 1.48 1.40 1.0 - Pro Moradia .18 .51 .47 .11 Cartade Credito FGTS 17.3 81.3 82.8 168.2 144.5 37.5 Individual 143.1 Associativo 1.3 Apoio a produ 17 Total 17.4 82.0 84.3 169.5 145.4 Source: Secretaria de Estado de Desenvolvimento Urbano (SEDU). Page XXXVIII Brazil: Progressive Low Income Housing Table 16: FGTS Carta de Cr&Iito Programs in Housing by Region 1995- 2000 (in thousands R$) R$ Total Total Structure No. of InvJ Investment Credits Credit Families Familia Benefited Individual North 119,683 107,715 1.75% 6,852 17.47 Northeast 1,027,283 924,556 15.03% 92,902 11.06 Southeast 3,835,821 3,452,236 56.11% 242,822 15.79 South 1,318,468 1,186,621 19.29% 98,398 13.40 Center West 535,347 481,812 7.82% 48,324 11.08 Associativo North 28,077 24,100 1.37% 1,181 23.77 Northeast 195,833 166,049 9.47% 10,326 18.97 Southeast 1,665,632 1,361,590 77.62% 82,179 20.27 South 172,989 139,395 7.94% 8,372 20.66 Center West 78,380 63,148 3.60% 3,764 20.82 Combined North 147,760 131,815 1.67% 8,033 18.39 Northeast 1,223,116 1,090,605 13.79% 103,228 11.85 Southeast 5,501,453 4,813,826 60.88% 325,001 16.93 South 1,491,457 1,326,016 16.77% 106,770 13.97 Center West 613,727 544,960 6.89% 52,089 11.78 Total 8,977,513 7,907,222 595,121 15.08 Source: Secretaria de Estado de Desenvolvimento Urbano (SEDU). Table 17. FGTS - Average Credit Per Family by Region, 1995-2000 Region Ave. credit/Family (R$) Individual Associativo North 15,720 20,406 Northeast 9,951 16,080 Southeast 14,217 16,568 South 12,060 16,650 Center West 9,970 16,776 Source: Govemrnment statistics, Caixa Economica Federeal, SEDU Brazil Low-Income Housing: Alternatives for the Urban Poor Page XXXIX Table 18: Access to Urban Services (1992/1997) 1970 1992 1997 Water Connection 60 73.6 77.7 Sewerage Collection 20 38.9 40.8 Waste Collection 66.6 76.3 Electricity 88.8 93.3 Telephone 19 27.7 Refrigerator 71.5 80.3 Washing Machine 24.1 32 Radio 84.9 90.3 Television 74 86.2 - Source: Secretaria de Estado de DesenvolvimentoUrbano (SEDU). Figure 1. Income Distribution in Favelas ExlM of tbo Fav*l Prohlm, by Uitn Am 30.0% 1 2 3 S P. *pSqt_ Page XL Brazil: Progressive Low Income Housing Table 19: Federal Expenditures with Social Sectors (Constant prices, R$ Million, December 1999) Area de atuac,o 1993 1994 1995 1996 1997 1998 Education 58.3 10,368.5 9,440.7 8,987.3 9,208.4 Health 97.0 20,901.0 17,903.2 21,438.6 18,631.4 Food and Nutrition 3.7 941.8 665.7 1,157.0 1,025.6 1,290.3 Sanitation and Environment Protection4 6.0 482.5 230.4 880.1 1,005.4 897.5 Social Safety Nets (previdencia) 289.9 52,432.0 55,959.4 60,674.4 66,413.3 Social Assistance 1,351.9 1,533.4 1,596.6 2,668.1 2,862.1 Labor 3,011.0 2,891.1 4,419.9 5,023.5 5,146.5 6,436.0 Rural Organization 709.6 1,643.7 1,510.8 1,888.5 2,059.6 Science and Technology 345.4 451.8 458.6 425.6 507.7 326.9 Housing and Urban Sector 1,0973 858.3 404.7 971.8 1,147.1 1,179.6 Training and Human Resources 55.5 43.2 51.0 21.7 18.6 7.8 Public sector benefits 16,519.0 21,951.7 25,454.3 26,075.5 27,600.6 28,312.7 Total Social Expenditure4 93,247.9 100,197.6 118,563.2 120,965.9 132,108.4 137,625.6 As % of GDP 12.0 12.2 1iA 11.8 12.2 12.7 Structure of Expenditures (in percentage) Education 9.4 11.1 10.3 9.4 9.0 9.2 Health 15.7 16.4 17.6 14.8 16.2 13.5 Food and Nutrition 0.6 0.9 0.6 1.0 0.8 0.9 Sanitation and Environment Protection4 1.0 0.5 0.2 0.7 0.8 0.7 Social (previdencia) 46.9 42.9 44.2 46.3 45.9 48.3 Social Assistance 3.3 1.3 1.3 1.3 2.0 2.1 Labor/defesa do trabalhador 3.2 2.9 3.7 4.2 3.9 4.7 Rural Organization 0.6 0.7 1.4 1.2 1.4 1.5 Science and Technology 0.4 0.5 0.4 0.4 0.4 0.2 Housing and Urban Sector 1.2 0.9 0.3 0.8 0.9 0.9 Training and Human Resources 0.1 0.0 0.0 0.0 0.0 0.0 Public Sector Benefits 17.7 21.9 21.5 21.6 20.9 20.6 Total Social Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 Fonte: IPEA/DISOC, com base nos dados mensais dos sistemas SIAFI/SIDOR Brazil Low-Income Housing: Alternatives for the Urban Poor Page XLI Table 20: Brazil: Per Capita Expenditures with Housing and Urban Development, By State R$ constant prices of 1995 1982 1985 1990 1991 1992 1993 1994 North 11.56 5.78 64.23 15.34 7.50 4.96 431 Rond6nia 17.61 0.05 14.86 0.87 1.37 1.09 0.97 Acre 0.84 3.52 7.81 15.28 20.18 9.72 14.28 Arnazonas 24.35 7.01 8.25 3.33 1.12 3.55 5.42 Roraima 64.99 34.10 221.31 293.60 46.71 20.27 6.64 Para 4.25 3.79 1.65 3.21 1.75 1.57 1.23 Amapa . 19.84 58.25 184.38 39.35 46.78 23.55 24.08 Tocantins 541.50 52.51 33.07 19.45 11.26 Northeast 4.05 5.28 6.88 4.07 4.65 2.53 4.14 Maranhao 0.65 0.50 0.49 0.27 2.98 1.45 2.55 Piaui 0.60 0.02 1.38 0.05 0.02 0.00 0.00 Ceara 1.37 1.30 3.37 7.91 8.98 3.42 5.41 Rio Grande Do Norte 0.00 7.53 0.00 0.00 0.00 0.00 0.00 Paraiba 0.19 11.31 0.60 0.16 0.01 0.18 1.18 Pernambuco 7.88 5.54 18.62 7.99 5.09 3.63 7.31 Halagaos 0.67 1.27 26.97 10.59 1.46 1.73 3.19 Sergipe 4.94 5.68 30.65 10.87 7.33 5.45 14.14 Bahia 7.80 9.10 1.41 1.76 6.30 3.35 3.71 Southeast 2.06 1.57 5.06 9.18 11.64 8.56 10.04 Minas Gerais 3.42 1.29 2.11 0.76 6.81 1.84 0.72 Espirito Santo 1.50 0.67 4.28 10.08 2.35 1.66 3.00 Rio De Janeiro 3.36 2.48 5.32 4.02 5.58 7.40 8.42 Sao Paulo 0.81 1.39 6.49 15.39 17.24 12.91 15.84 South 2.60 1.94 2.32 1.96 4.49 2.56 5.26 Parana 1.21 1.52 0.45 2.72 6.71 3.44 6.56 Santa Catarina 0.25 1.26 0.88 1.48 4.50 2.45 3.42 Rio Grande Do Sul 5.06 2.67 4.76 1.48 2.44 1.82 5.00 Center West 10.27 11.67 16.69 22.10 27.80 33.80 30.09 Mato Grosso 10.35 15.52 4.24 1.25 0.64 7.15 3.00 MatoGrossoDoSul 0.90 0.24 2.15 0.00 0.15 0.88 6.39 Goias 0.72 2.32 0.76 1.68 7.08 9.07 5.73 Distrito Federal 51.83 ' 44.39 88.60 124.37 145.19 166.53 152.46 Source: Atlas das Desigualdades Regionais, IPEA Pane XLII Brazil: Progressive Low Income Housing Table 21: Expenditures in Housing and Urban Development, by State In Constant Prices of 1995 (Thousand of Reals, R$) States 1985 1990 1991 1992 1993 1994 Rond6nia 37.70 16,110.51 985.05 1,625.66 1,358.54 1,252.94 Acre 1,252.96 3,192.70 6,398.93 8,643.48 4,254.56 6,373.94 Amazonas 12,221.21 16,939.00 7,017.42 2,413.46 7,869.05 12,293.81 Roraima 4,902.13 45,849.63 64,247.67 10,750.09 4,887.72 1,671.15 Para 15,654.75 7,953.46 15,932.86 8,897.68 8,189.01 6,551.47 Amapg 13,312.51 51,777.40 11,429.03 14,021.43 7,273.19 7,647.36 Tocantins 485,361.32 48,269.49 31,260.26 18,876.19 11,151.83 Maranhao 2,231.39 2,398.63 1,354.87 14,939.72 7,392.79 13,154.14 Piaui 36.92 3,506.02 137.57 63.31 0.00 0.00 Ceara 7,544.83 21,212.32 50,409.03 58,031.15 22,396.05 35,870.90 Rio Grande do Norte 16,110.17 0.00 0.00 0.00 0.00 0.00 ParaAba 33,615.44 1,908.37 513.95 48.45 576.52 3,895.37 Pemambuco 36,611.70 131,325.55 57,051.75 36,770.00 26,510.12 53,878.04 Halagaos 2,841.63 66,738.30 26,683.46 3,753.58 4,513.20 8,447.94 Sergipe 7,410.47 44,917.85 16,253.37 11,167.40 8,453.09 22,319.03 Bahia 96,321.12 16,526.02 20,914.04 76,092.34 41,082.09 46,297.25 Minas Gerais 18,752.77 32,771.38 11,909.56 108,653.32 29,705.29 11,834.63 Espirito Santo 1,532.26 10,950.34 26,264.27 6,224.21 4,492.52 8,220.91 Rio de Janeiro 29,754.76 67,564.08 51,501.12 72,296.77 96,686.64 111,007.92 Sao Paulo 39,166.53 201,823.28 487,053.26 554,954.70 422,098.71 526,056.43 Parang 12,200.36 3,793.93 23,032.80 57,198.78 29,505.13 56,727.02 SantaCatarina 5,103.81 3,953.93 6,751.49 20,818.00 11,516.18 16,306.97 Rio Grande do Sul 22,435.74 43,017.52 13,534.84 22,614.52 17,041.46 47,401.68 Mato Grosso 24,108.15 8,310.38 2,548.29 1,344.97 15,564.42 6,730.70 Mato Grosso Do Sul 377.00 3,757.71 0.00 274.23 1,633.40 12,029.23 Goias 8,224.77 3,006.18 6,755.11 29,046.74 37,844.09 24,317.92 Distrito Federal 61,031.56 139,025.46 199,600.89 238,046.80 278,626.53 260,078.52 Source: National Secretary of Treasury, Ministry of Financeq Atlas de Desigualdades Regionais, IPEA Brazil Low-Income Housing: Alternatives for the Urban Poor Page XLmI Table 22. Brazil: Municipal Spending in Social Sectors, 1997 (R$ millions, 1999 prices) State Education Housing Health Social Total % of and and security Housing in urban sanitation Total Roraima 12 14 8 3 37 37.8 Amazonas 177 136 80 24 416 32.7 Sergipe 125 86 46 24 281 30.6 Pemambuco 442 350 247 195 1,235 28.4 Espirito Santo 296 183 115 54 648 28.3 Acre 35 25 19 9 88 28.1 Rio de Janeiro 674 905 1,109 678 3,366 26.9 Bahfa 709 490 390 248 1,837 26.7 Sao Paulo 4,701 3,452 4,200 2,293 14,646 23.6 Alagoas 101 53 68 17 239 22.1 Total Brazil 13,424 8,345 11,202 5,432 38,404 21.7 Mato Grosso Nonte 207 108 150 33 497 21.7 Goias 363. 204 322 112 1,001 20.4 Parana 922 434 484 322 2,163 20.1 Ceara 427 226 331 154 1,138 19.9 Tocantins 61 30 39 22 152 19.6 Piaui 148 83 171 26 428 19.3 Paraiba 189 83 126 39 438 18.9 Mato Grosso do Sul 138 53 70 35 296 17.9 Minas Gerais 1,587 738 1,525 454 4,303 17.1 Maranh5o 153 72 200 34 459 15.7 Rio Grande do Norte 138 66 185 55 444 14.9 Para 262 99 228 77 667 14.8 SantaCatarina 515 133 211 82 941 14.2 Rio Grande do Sul 978 306 805 437 2,527 12.1 Rond6nia 56 15 71 6 147 10.0 Amapa 8 1 1 1 11 8.4 Source: Govemment statistics, Ministry of Finance. Page XLIV Brazil: Progressive Low Income Housing Table 23. Rate of Urbanization and Human Development Index by State, Brazil Urbanization Rate Human Development WB 12201 Index LABrazil\Housing November 15 Anexes.doc November 4,2002 Distrito Federal 92.88 0.869 6:55 PM Rio Grande do Sul 78.87 0.869 Sao Paulo 93.1 0.868 Mato Grosso do Sul 83.22 0.848 Parana 77.88 0.847 Rio de Janeiro 95.43 0.844 Espirito Santo 77.64 0.836 Minas Gerais 78.42 0.823 Rond6nia 61.93 0.82 Roraima 70.52 0.818 Santa Catarina 73.13 0.816 Brasil (Average) 78.36 0.809 Amapl 87.12 0.786 Goiais 85.78 0.786 Amazonas 73.92 0.775 Mato Grosso 75.84 0.767 Acre 65.2 0.754 Para 53.51 0.703 Rio Grande do Norte 72.05 0.668 Sergipe 70.22 0.658 Bahia 62.41 0.655 Pemambuco 74.02 0.615 Ceara 69.21 0.59 Tocantins 70.66 0.587 Paraiba 68.43 0.557 Maranhao 51.92 0.547 Halagaos 63.11 0.538 Piaui 58.21 0.534 Correlation 0.683336414 Covariaton 0.858656518 1MAGWNO Report No.: 22032 BR Type: SR