Report No. 16736 Brazil Impact Evaluation Report Learning from Best Practice in Five Urban Projects Medium-Sized Cities (Loan 1"720-BR) Recife Metropolitan Region (Loan 21 70-BR) Preparation of Metropolitan Development Programs for Fortaleza/Salvador (Loan 21 93-BR) Parand Market Towns Improvement (Loan 2343-BR) Northeast Urban Flood Reconstruction (Loan 2645-BR) June 17, 1997 Operations Evaluation Departnent Document of the World Bank Currency Equivalents Currency unit = Real (R$) US$1.00 = R$1.07 R$1.07 = US$0.93 Abbreviations and Acronyms BNH National Housing Bank (Banco Nacional de Habitag5o) CBO Community Based Organization CNDU National Urban Development Council (Conselho Nacional de Desenvolvimento Urbano) CNPU National Committee for Urban Policy and Metropolitan Regions (Commisco Nacional de Politica Urbana e Regi6es Urbanas) ED Executive Director (World Bank) ICR Implementation Completion Report ID Institutional Development IDB Inter-American Development Bank IER Impact Evaluation Report IERR Internal Economic Rate of Retum IPPLAN-JF Juiz de Fora (Municipal) Planning and Research Institute (Instituto de Pesquisa e Planejamento) IPLANAT Natal (Municipal) Urban Planning Institute (Instituto de Planejamento Urbano de Natal) IPTU (Municipal) Urban Property Tax LAC Latin America and the Caribbean Regional Office (World Bank) MDU Ministry of Urban Development (Minist6rio de Desenvolvimento Urbano) NGO Non-governmental Organization OED Operations Evaluation Department PAR Performance Audit Report SAR Staff Appraisal Report SEAIN Secretariat of International Affairs (Secretaria de Assuntos Internacionais) of Ministry of Planning SUDENE Superintendency for the Development of the Northeast (Superintendencia do Desenvolvimento do Nordeste) UPP (Bank) Urban Policy Paper Director General, Operations Evaluation Robert Picciotto Director, Operations Evaluation Department Roger Slade for Elizabeth McAllister Divison Chief : Yves Albouy Task Manager : Alain Barbu The World Bank Washington, D.C 20433 U.S A Office of the Director-General Operations Evaluation MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Brazil Impact Evaluation Report Learning from Best Practice in Five Urban Projects: Medium-Sized Cities Project (Loan 1720-BR) Recife Metropolitan Region Project (Loan 2170-BR) Preparation of Metropolitan Development Programs for Fortaleza and Salvador (Loan 2193-BR) Paranf Market Towns Improvement Project (Loan 2343-BR) Northeast Urban Flood Reconstruction Project (Loan 2645-BR) Attached is the Impact Evaluation Report (IER) prepared by the Operations Evaluation Department (OED) on the above mentioned urban development projects implemented in Brazil during 1979-1989. The projects had varied designs and covered all the major urbanized regions of the country. They all shared, however, common aims of improving the efficiency and equity of urban service provision. The aim of the impact evaluation study was to identify and assess sustained impacts of these projects, more than five years after their completion, sufficient time for sustainable benefits to be manifest. The study focused upon three areas of impacts: (i) living conditions for the urban poor; (ii) decentralizing urban service provision; and (iii) participatory development with beneficiary communities. The study concludes that the projects helped place municipalities at the forefront of urban service provision and demonstrated that well designed physical investments can lead to sustained improvement in urban living conditions for low-income residents. Important side effects were strengthened community-based organizations which enabled local residents-especially women-to learn the rules of urban service provision in Brazil. Among the study's recommendations: (i) continue project efforts of this kind to improve the living conditions of the poor; (ii) disseminate best practice experiences of successful project cities; (iii) encourage institutional reform at the municipal level; (iv) use robust materials for physical solutions where maintenance is typically wanting; (v) support urban policy and research institutes; (vi) ensure participatory assessment is always part of project evaluation; and (vii) incorporate base-line data and monitoring indicators into the design of new projects to facilitate subsequent monitoring and evaluation. Attachment  CONTENTS PREFA CE . ..... ............................................ ................................................. v EVALUATION SUMMARY.................................................................................. vil 1. MAIN CONCLUSIONS AND RECOMMENDATIONS .........................................1 Conclusions..................................... ............. 1 Recommendations .............................................. 3 2. STUDY OBJECTIVES AND DESIGN ............................................................... 7 Objectives and Scope of the Study............................... 7 Impact Domains, Indicators and Evaluation Hypotheses ........... ..... 8 Focus upon Best Practice. .......................... ......... 10 Data Collection Methods ........................... ........ 11 Beneficiary Participation in Evaluation............................ 11 3. THE PROJECTS AND THEIR SETTING ............................................................ 15 Poverty Focus of Early Projects (1972-1983) ............... ....... 15 Executive Directors' Voice.................................. 16 Efficiency Focus of Formal Strategy (1983-87).................. 16 Shift to Decentralization (1988-92) ................................ 17 Results of Completed Urban Projects........................... 18 On-going Sector Operations and Priorities ........................... 20 4. IMPACTS ON LIVING CONDITIONS.............................................................. 21 Framework of Evaluation............................... ..... 21 Beneficiaries' Perspective ....................................... 21 Infrastructure and Services: Expert Assessment ............... ..... 24 Housing Improvements..................................... 28 Local Economic Development ........................... ..... 30 Social Development....................................... 31 This report was prepared by Mrs. Alcira Kreimer (Task Manager) and Mr. Roy Gilbert (Consultant). Specific contributions were made by Claudio Volonte, Paulo de Goes and Sonia Rocha (Consultants). Ms. Stacy Ward and Mr. Arthur Anger provided administrative assistance. 5. DECENTRALIZING URBAN SERVICE PROVISION....................................... 35 Framework of Evaluation..................................... 35 Helping Municipalities Develop a Project Culture................ ..... 36 Urban Policy Making at the Local Level. .................... ...... 39 Enabling a Bigger Municipal Role.............. ................ 40 Decentralization as a Factor of Performance.. .......................... 41 6. POOR RESIDENTS LEARN THE RULES OF THE GAME ............................45 Framework of Evaluation................................... 45 Strengthening Community Based Organizations ...................... 46 Contribution of Women................................ ..... 48 Learning the Rules of Urban Service Provision ......... ............ 48 Holding Local Government Accountable. .................... ..... 50 BIBLIOGRAPHY ................................................................................................. 53 ANNEX I -Tables .............................................................................................. 57 ANNEX 2 - Comments by the Borrower........................................................... 83 MAPS - IBRD 28458 through IBRD 28465 II LIST OF TABLES Main text: 2.1 - Impact Domains, Indicators and Hypotheses ......................9 2.2 - Indicators of Living Conditions in Case Study Cities 1980-91................ 13 3.1 - Project Summaries.............. ................ ..... 18 4.1 - Beneficiaries' Perspective on their Living Conditions.................. 22 4.2 - Infrastructure Impact Expert Assessment..................... 25 4.3 - Environmental Conditions: Expert Assessment ............ ..... 27 4.4 - Housing Improvements ......................... ........ 29 4.5 - Impact upon Economic Development................................ 31 4.6 - Impact upon Social Conditions ................................ 33 5.1 - Indicators of a Municipal Project Culture ........................ 37 5.2 - Indicators of Project Ownership by Municipalities................ 42 6.1 - Indicators of Participation: Community Perspective......... ...... 52 Annexes A. I - Brazil 1975-96: Benchmark Urban Sector Events ............... 57 A.2 - Brazil 1975-96: Government Policy relating to the Urban Sector........................ ............... 61 A.3 - Brazil 1975-96: Bank Non-lending Activities relating to the Urban Sector................................... .......... 65 A.4 - Brazil 1975-96: Bank Lending Operations relating to the Urban Sector...................................... ...... 73 III 8 PREFACE This Impact Evaluation Report (IER) presents findings of a study designed to identify and assess lasting impacts of five urban development projects financed by the Bank in Brazil during 1979-1989. This was when both government policy and Bank non-lending activities in the sector underwent major transformatiors. OED's chose Brazil for this pioneering study, as the country with the Bank's largest and most varied urban lending and non-lending program in the Region. Beyond the projects themselves, the IER looks at sector issues that are topical today in Brazil, such as the decentralization of urban service provision, for instance. All five urban projects evaluated here were completed at least seven years ago, sufficient time for sustainable benefits to be manifest. The projects had varied designs and covered all the major urbanized regions of the country. They all shared, however, common aims of improving the efficiency and equity of urban service provision for lower income urban populations. In order to focus on lessons of successful experiences, the impact evaluation study looks closely at best practice cases among the cities that participated in the projects. OED launched the study in May 1995 and benefited from valuable and practical support of SEAIN in Brasilia throughout. Three missions visited eleven cities throughout Brazil between October 1995 and July 1996. Responses among stakeholders in these cities-where the project experiences are firmly ingrained in the collective memory-far exceeded our best expectations. The team appreciates the enthusiastic participation of all who helped in this work. Two features of the study methodology are worthy of note. First, it used case studies of "best practice" project cities to extract lessons from experiences on the ground that can be applied elsewhere. Cities that improved their infrastructure and housing over the project period were chosen for evaluation. Second, the IER was a participatory exercise: it relied upon inputs from final project beneficiaries themselves. The evaluation team met with more than one hundred beneficiaries in group discussions held in project areas of each of the cities visited. The enthusiastic participation of stakeholders in the exercise far exceeded the study team's expectations. To discuss the preliminary findings of the study with Brazilian officials and sector experts, a workshop was held in Brasilia in April 1997. Workshop discussions benefited from the active participation of some forty sector experts from all regions of Brazil. Important conclusions from those discussions are incorporated The five IES projects are: (1) Medium-Sized Cities (Ln1720); (2) Recife Metropolitan Region (Ln 2170); (3) Preparation of Metropolitan Development Programs for Fortaleza/Salvador (Ln 2193); (4) Parani Market Towns Improvement (Ln 2343); and (5) Northeast Urban Flood Reconstruction (Ln 2645). v into the text of this IER. OED wishes to thank workshop participants for their valuable contributions. Special acknowledgments go to SEAIN for organizing the event as well as collaborating enthusiastically with this study since its inception. vi Evaluation Summary MAIN CONCLUSIONS AND RECOMMENDATIONS 1. The projects helped place municipalities at the forefront of urban service provision, and demonstrated that well designed physical interventions can lead to sustained improvement in living conditions for low income residents. Future efforts at providing urban services in Brazil should focus upon replicating demonstrated successes of best practice project municipalities. 2. The report's key recommendations are: * Continue efforts to improve living conditions of the urban poor through the kinds of project interventions evaluated by the study, a point that was strongly endorsed by participants at the Brasilia impact evaluation workshop. * Disseminate best practice experiences of cities that have excelled in improving urban service provision through learned articles, the mass media, the Internet, and material produced by urban institutes and professional associations. * Encourage institutional reform at the municipal level. Through the projects, municipalities demonstrated that they could manage urban service provision and local residents increasingly expect municipalities to do so. * Urban improvements should apply robust and comprehensive physical solutions that can best withstand the neglect of maintenance to which they are likely to be subject. * Support urban policy and research institutes-the study found that they helped to provide continuity and innovate in service provision at the municipal level. * Participatory assessment-especially involving women-should always be part of project evaluation. Final project beneficiaries not only have a voice to be heard, but are often the principal store of information about a project's history. * Base-line data and performance indicators should be incorporated into the design of all new projects to facilitate effective performance monitoring. This recommendation was unanimously endorsed by the Brasilia workshop. vi STUDY OBJECTIVES AND DESIGN 3. This study seeks to learn how best practice project municipalities succeeded in improving living conditions of the poor, and from these cases, to make practical recommendations for policy makers in the sector. The evaluation focuses upon a few key impact domains, notably living conditions, decentralization and beneficiaries' voice in improvements. Data collection from both secondary sources and during missions to eleven cities throughout Brazil enabled the study team to test hypotheses of project impact within each of these domains. Direct participation of final beneficiaries in the evaluation gave important insights into project performance. 4. The study focused upon three impact domains: (i) living conditions of the urban poor, (ii) decentralizing urban service provision; and (iii) users' voice and learning about urban service provision. For each domain, the evaluation team identified performance indicators and formulated evaluation hypotheses for testing during field visits to eleven cities throughout the country. These field visits including key informant interviews, site inspections and most importantly, group interviews with beneficiaries in the project areas. 5. The evaluation looked at best practice cases at the municipal level, in the search for examples of excellence that might be replicated elsewhere. Best practice project cities were those that improved their urban services most between 1980 and 1991, according to census data. THE PROJECTS AND THEIR SETTING 6. All five projects were prepared without the benefit of country sector work. Policy dialogue in Brazil-with early emphasis upon poverty alleviation through priority investments in the poorer regions of the country-evolved only as part of project preparation itself. The federal National Council for Urban Development (CNDU)-responsible for projects-was the main interlocutor of the Bank during 1979-85. In approving these early projects, Bank Executive Directors emphasized efficiency goals later endorsed by Bank sector papers on Brazil from 1983 onwards. 7. Implementation Completion Reports (ICRs) and Performance Audit Reports (PARs) considered all but one of the operations satisfactory, ratings endorsed by this study. Although varied in scope, all projects aimed to both improve living conditions of the urban poor and to help institutional development. IER findings focus upon these two aspects and from the lessons learned, the study recommends key actions to help improve urban service provision across many municipalities throughout Brazil. 8. Decentralization of service responsibilities toward municipal government was the strategic setting for the projects. The federal government had been responsible for sector policy when the projects were prepared, but fiscal constraints forced a federal withdrawal. After some short-lived efforts to retain an viii urban sector responsibility during 1985-88 when Brazil briefly had a Ministry of Urban Development, the 1988 Constitution gave municipal government primary responsibility-and resources-for managing urban development. 9. Despite a cutback of Bank lending to the sector, three ongoing Bank financed projects continue to pursue the aims of efficient urban service provision at the municipal level. At the same time, the Inter-American Development Bank (IDB) has substantially increased its own lending to Brazil in this sector. IMPACTS ON LIVING CONDITIONS 10. As intended, the projects improved the living conditions of the urban poor, especially through better access to their areas and reduced flooding risks for local residents. In many improved areas, buses and service vehicles were able to enter the neighborhood for the first time. Areas that had been prone to flooding before the projects reported no major problems since the completion of project works. 11. Typical lack of maintenance meant that infrastructure solutions using more robust materials had more sustainable impacts. Low-cost sewerage solutions often failed to operate through blockages and ruptures, while street paving of cobblestones broke under heavy loads of buses and garbage trucks. Project beneficiaries and expert observers had strikingly similar viewpoints on these issues, especially about the poor maintenance of infrastructure in these areas. 12. Impacts upon housing conditions were significant. While there had been little enlargement of housing units after the projects, better quality materials were used and property values increased. Impacts upon the local economy were modest with some additional local commerce following the projects. More important, perhaps, was that better accessibility allowed residents to commute to jobs elsewhere, thanks to the projects. Social impacts were stronger, especially in strengthening community life within these neighborhoods. DECENTRALIZING URBAN SERVICE PROVISION 13. Thanks to the projects, municipalities strengthened their own management of project preparation, evaluation and procurement. Many introduced economic evaluation of proposed investments for the first time. Most cities set up special procurement units and now manage most of their bidding and contracting work in-house. They also simplified the regulatory framework for project approvals. In developing the basis of a "project culture" in this way, municipalities seized the opportunity to play a bigger role in decentralized urban service provision as the federal government withdrew from the sector. Still more has to be done to consolidate this new project culture, however, so that municipalities can give more attention to project maintenance which remains deficient in most cities to this day. ix 14. There were only modest impacts upon urban policy formulation at the local level, which still today has to fill the gap left by federal retrenchment. Some cities nevertheless established urban planning and research institutes, using project personnel to staff them. These institutes have been responsible for important policy innovations and updating master planning efforts. Most cities adopted a more rigorous policy stance on cost recovery policy. 15. Municipalities developed strong ownership of projects that were originally the products of federal government agencies. This decentralization to municipalities after the reduction of federal activities in the sector augurs well for the long-term sustainability of urban improvements in Brazil. Municipalities were willing to take on greater management responsibilities. Local ownership went deep enough in some cases to ensure smooth transfers of the projects between adversaries at times of political transition at the municipal level. Transitions at the municipal level occurred three times during project implementation. POOR RESIDENTS LEARN THE RULES OF THE GAME 16. The projects had important side effects in strengthening community-based organizations (CBOs), thereby enabling residents to learn about urban service provision. As convenient interlocutors for local govemment and useful monitors of project progress on behalf of residents, CBOs were empowered through the project actions. This effect was felt strongest in medium-sized cities. Metropolitan administration is inevitably complex and the physical remoteness of project sites from city hall undermines effective CBO: local government communication. CBOs in small towns did not flourish, but for opposite reasons. Small town residents reckon that they can approach local officials directly, making community level intermediation unnecessary. 17. Women played an important role in community strengthening. They were among the most active participants in group discussions with the study's evaluation team and the most knowledgeable about urban improvements in their areas. Men present at these meetings were generally more reticent, and were uneasy about the constant questioning that community participation involves. 18. Residents learned a lot about urban service provision through the activities of their CBOs. They learned about who was responsible for service provision and what the roles and relationships of key parties were. None of the local communities visited knew about the projects during their planning phase, so that learning took place during project implementation. City councilors (vereadores) played a big participatory role at that time. 19. Residents' newly acquired knowledge pointed to municipal government as being ultimately responsible for urban service provision. In this way, the projects indirectly encouraged residents to hold municipalities accountable for these services. The project experience thus gave little incentive for self-help efforts by communities themselves. Nevertheless, revitalized community participation in x direct partnership with local government should enhance future project performance and sustainability, thanks to more stakeholder control and commitment. xi  1. Main Conclusions and Recommendations The projects helped place municipalities at the forefront of urban service provision, and demonstrated that well designed physical interventions can lead to sustained improvement in living conditions for low income residents. Future efforts at providing urban services in Brazil should focus upon replicating demonstrated successes of best practice project municipalities. CONCLUSIONS 1.1 Better living conditions for the poor: Project documentation estimates that together the five operations improved living conditions for some 2.0 million urban poor in 464 municipalities throughout the country. Better paving allowed improved access to project areas and resulted in sustained improvements in the eyes of both beneficiaries and also sector experts. Street paving also reduced atmospheric dust. Becoming more desirable residential areas, property values in improved neighborhoods increased substantially. [Chapter 4] 1.2 Reduction of urban flood risks: Proper drainage eliminated accumulation of polluted water in ditches, a serious health hazard for children in particular before the projects. Prevention works-through the Northeast Flood Project especially-helped diminish the risk of flood waters invading homes and property. Properly working drainage in fact reduced the incidence of flooding in improved neighborhoods. 1.3 Robust materials and sustained impacts: Only physically robust infrastructure can survive the neglect of maintenance typically faced by improved areas and other poorer neighborhoods of Brazilian cities. Lack of maintenance shortens the useful life of infrastructure using low cost technologies. Thus, in the long run, low cost technologies and lower quality construction materials may not lead to the sustained improvements sought. Low-cost sewerage applied to basic sanitation in a number of project cities, for instance, suffers serious operational failures-particularly blockages-after five or more years. Street pavement using Irregular cobblestones, although less costly at the outset, soon breaks up when regularly used by heavy vehicles such as buses and garbage trucks. [Chapter 4] 1.4 Efficiency, an increasing priority: The Bank's approach to urban development in Brazil since 1972 has progressively given more importance to efficiency over poverty reduction goals. Urban projects-such as the ones reviewed by this study-were left adrift from a strategic context. Initially inspired by equity concerns of reducing poverty, sector work focused upon urban poverty alleviation, recommending affordable urban services and housing. Following the international debt crisis of 1982 and the weakening of the Brazilian national economy, though, Bank sector strategy embraced efficiency goals more 1 explicitly. It called for direct cost recovery and allocation efficiency, regardless of project beneficiaries' income levels. At the same time, fiscal constraints led to federal government retrenchment from the urban sector. This resulted in an unplanned but de facto decentralization of responsibilities to municipal governments. [Chapter 3] 1.5 Project culture within municipalities: As intended, the projects succeeded in strengthening management systems and procedures within local government. Local officials acquired greater skills in project preparation, appraisal and evaluation. New units for the procurement of goods and works were established within a more project-friendly regulatory environment at the municipal level. Able project teams were created in cities that were able to replicate lessons learned in good management practice. One important caveat: this new culture rarely extended to project maintenance, which still remains inadequate in most cases. [Chapter 4] 1.6 Decentralized urban service provision: Through the project culture impact, the projects unintentionally allowed municipalities to play a bigger role in decentralized urban service provision as the federal government withdrew from this function. Furthermore, the projects firmly anchored responsibilities for urban service provision to municipalities. They demonstrated that local governments could make sustainable improvements in chosen areas given adequate resources. [Chapter 5] 1.7 Urban planning and research institutes: Local urban planning and research institutes were important contributors to best practice cities and to the conti uity of urban policies and investments. Project management teams them elves form the nucleus of entities created in the mid-1980s at the muni 'pal level as project activities themselves neared completion. Municipal urban planning and research institutes of these kinds are still in full operation today in the cities of Juiz de Fora and Natal. [Chapter 5] 1.8 Revitalized community organizations: Monitoring project interventions gave a new importance to community-based organizations (CBOs) in the eyes of local residents. Local governments treated CBOs as useful interlocutors during project implementation. Prior to the projects, most low-income areas already had some form, albeit weak, of community organization. Residents-especially women-see that CBOs can help ensure improvements are made in accordance with their own interests. This effect is most strongly felt in medium-sized cities. In larger metropolitan centers, residents have less confidence in successful CBO intermediation with remote and complex local governments. Residents of small cities, who have easy direct access to local officials as needed, feel community level intermediation to be unnecessary. [Chapter 6] 1.9 Communities learn rules of the game: Through project interventions- which amounted to an on-the-job civic education program for beneficiaries-low- income residents learned about procedures and responsibilities for urban service provision. Residents also held local government more and more accountable for these services. Project implementation enabled residents to know, for instance, 2 who was responsible for which urban services and where to lobby for their interests. This was no mean feat, given the complexities of three tiers and many overlapping responsibilities in Brazil's federal system of government. Residents show little interest in self-help schemes for urban services of the kind that are widely used in the urban kampungs of cities in Indonesia. Unlike their counterparts in Indonesia, Brazilian residents feel that such schemes allow local governments to shirk their own official responsibilities for urban service provision. [Chapter 6] 1.10 Best practice cities as case studies: For their outstanding performance, best practice cities provide good case studies, whose evaluation can help unveil some of the secrets of success. In the state of Paran6, for instance, best practice municipalities Almirante Tamandar6 and Colorado reduced their basic sanitation deficits by 41.8 and 50.5 percentage points respectively over the 1980-91 period. The average for Paranc as a whole was only 3.9. In the Northeast region of the country, the best practice project city of Propri! reduced its housing deficit by 14.1 percentage points during the 1980-91 period, while the average for other cities in the state of Sergipe was only 8.2 percentage points. Similarly, AracajO reduced its basic sanitation deficit by 22.2 percentage points over the same period, against an average of 5.7 for other cities in the state. [Chapters 2 and 4] RECOMMENDATIONS 1.11 Continue efforts to improve living conditions of the urban poor: In view of the satisfactory long-term physical impacts of the projects (details Chapter 4), local governments especially should try to continue to find ways of replicating these experiences. Of all IER conclusions and recommendations, this one enjoyed the greatest consensus among participants of the Brasilia workshop. To meet new demands-especially in Brazil's poorer Northeast region where urban service deficits are most pronounced-requires stronger fiscal effort by municipalities and creative solutions for the participation of the private sector in these activities. According to workshop participants, Brazil today needs to clarify more precisely the role of urban policy in poverty alleviation. 1.12 Disseminate best practice experiences: This could be through learned articles, feature stories in the mass media, on-line databases on the Internet and material produced by urban institutes and professional organizations. Better knowledge of a municipality's administrative efficiency, financial performance and results on the ground that would help, among other things, assessments of the solidity and creditworthiness of individual municipalities, as future borrowers in financial markets. Wide dissemination of these results-formalized through prize giving for outstanding achievements-can encourage competition between municipalities. Knowledge of these success stories is raw material for reforming weaker municipalities and helping them emulate their more successful peers. Brasilia workshop participants, however, offered words of caution about dealing with best practice cases. They stressed that access by weaker municipalities to 3 best practice solutions should be as broad as possible, lest existing inequalities among cities in Brazil be made worse. 1.13 Municipal reform: The project experience demonstrates that it is feasible to improve municipal management of urban service provision. Moreover, low-income urban communities increasingly hold local government responsible for these services. For these reasons, federal and state government policies, supported by Bank urban sector work, should aim to facilitate municipal reform in this area. This can encourage, among other things, innovative financial intermediation to mobilize private sector funding. Such funding could be used for project interventions to improve the living conditions of the urban poor-a job that municipalities can do well, but not on a sufficient scale for lack of resources. According to participants at the Brasilia workshop, reform at this level should focus upon fiscal strengthening. Workshop discussions made it clear that reform programs have to adopt realistic time-frames for poorer municipalities to effectively assume greater responsibilities under decentralization. 1.14 Robust and comprehensive physical solutions: To have sustainable benefits over the long term, projects should apply complete and proven technical solutions that use durable materials. These can better survive maintenance that is typically lacking. Without maintenance, initial cost savings of cheaper technologies can be more than wiped out by additional investments necessary to replace prematurely obsolete infrastructure. As pointed out during the Brasilia workshop, low-cost solutions are not always in the end low-cost. For street paving to have sustainable impacts, for instance, project interventions should include sidewalks and drainage so that the infrastructure can function properly. Works should use robust materials that can withstand the neglect, which they are likely to be subject to over many years. The importance of robust solutions was recognized during the workshop, but with a key caveat that maintenance itself needs to be improved. 1.15 Support policy and research institutions: To ensure long-term commitment to sector policy development and urban planning at the city level, it is essential to have a solid organizational structure in which these activities take top priority. The best practice project cities provide many examples of this. To help innovate in urban policy development, future project interventions should both support the establishment of new urban research and planning institutes in the private and public sectors, and consolidate existing ones. 1.16 Participatory assessment as part of evaluation: This should be a routine at the ICR and IER stages. Final project beneficiaries--especially women- -have a very special role to play in project evaluation. As well as providing a user perspective upon project outcomes, they constitute the principal-and sometimes the only-repository of project information essential for evaluation purposes. Consulting beneficiaries during evaluation also helps give final project users and beneficiaries a strong voice as stakeholders in the improvements made. Beneficiary participation in evaluation emerged as a priority theme during discussions at the Brasilia workshop and later written comments that 4 emphasized the strategic importance of local communities in the search for solutions to urban problems. 1.17 Base-line data and performance indicators: Participants at the Brasilia workshop agreed that project designers should record data that portray base-line conditions-i.e. without the project-and also identify performance indicators at the inception of all new operations. The IER fully endorses a recommendation that, among other things, provides a firmer basis for future project monitoring and evaluation. In the meantime and without such data for older projects, evaluators will have to continue using base-line surrogates. These may include secondary data-such as census information used by this IER-that best fits ideal temporal or spatial references for an impact evaluation exercise. Performance indicators themselves-both quantitative and qualitative-should reflect changes in base-line conditions brought on by project interventions and be central to such evaluations. Written comments by Brazilian experts emphasized the need to resume urban research-abandoned in the 1980s in Brazil-to develop a better understanding of baseline conditions. 5  2. Study Objectives and Design This study aims to learn how best practice project municipalities succeeded in improving living conditions of the poor, and thereby make practical recommendations for policy makers in the sector. The evaluation focuses upon a few key impact domains, notably urban living conditions, decentralization and beneficiaries' voice in improvements. Data collection from both secondary sources and during missions to eleven cities throughout Brazil enabled the study team to test hypotheses ofproject impact within each one of these domains. Direct participation offinal beneficiaries in the evaluation gave important insights into project performance. OBJECTIVES AND SCOPE OF THE STUDY 2.1 Objectives: This impact evaluation study (IES) has four principal aims: " To learn how best practice municipalities address and reconcile efficiency and poverty reduction equity goals in urban service delivery. " To identify and assess: (i) sustainable direct and indirect project impacts upon the living conditions of the urban poor in Brazil; (ii) specific project actions associated with Brazilian municipalities that succeeded in improving these conditions efficiently; (iii) innovations introduced by the projects that can be replicated elsewhere. " To draw lessons about project design and implementation features that most help improve the living conditions of the urban poor. " To provide practical and focused recommendations for Brazilian and Bank policy makers on ways of bringing sustainable benefits to the poor efficiently. In short, the objectives of this study are to determine if and how the original goals of the projects (Table 3.1) are still being met, to identify unexpected side effects and to make recommendations for future action in the sector. Participants in the Brasilia seminar, held April 22-23, 1997 at SEAIN, considered that project evaluators themselves are responsible for helping to ensure that recommendations are integrated into new project preparation. 2.2 Projects with a broad scope but similar aims: Despite geographical diversity and varied nomenclatures, all five projects evaluated by this study aimed to tackle both efficiency and equity aspects of urban service 7 improvements. This was through selected investments in urban infrastructure and technical assistance to executing agencies in cities throughout Brazil. The study highlights questions such as the reconcilement of efficiency and equity goals in urban development, which are at the heart of efforts to define a strategic response to the development challenges of the sector in Brazil today. For this reason, the IER also looks at how more recent projects incorporate the lessons from the Bank's earlier experience in determining the relative priorities of increased efficiency in service delivery and financing and meeting the needs of the urban poor through project and non-lending work. IMPACT DOMAINS, INDICATORS & EVALUATION HYPOTHESES 2.3 Domains: The study focuses upon three possible impact domains: (i) living conditions of the poor; (ii) decentralizing urban service provision; and (iii) users' voice and learning about urban service provision. The five projects may have impacts in many other areas too, but budget and time constraints require the study to focus only upon the most important ones. Potential project impacts upon municipal finances, for instance, are not covered by this study. 2.4 Initial research questions: The three domains were chosen as priority for this study by the desire to learn answers to key sector questions still topical in Brazil today. Questions raised by the evaluation team included: " What do permanent improvements in living conditions of the poor in best practice cities owe to the projects themselves? m Are improvements in living conditions for the poor significantly better in project cities than in other cities, which did not receive project support? " What is the impact of Bank non-lending activities in Brazil, if any, upon decentralizing urban development responsibilities on to the shoulders of municipal administrations? " Do projects have institutional development (ID) impacts that help cities meet both efficiency and equity goals in service delivery? " Is participation by final beneficiaries a factor in project success? 2.5 Evaluation methodology: Within each impact domain, the team identified indicators to test evaluation hypotheses formulated about possible project impacts. In the case of the 'living conditions' domain, for instance, the study used quantitative indicators of households served at the municipal level, using census data of 1980 and 1991. Hypotheses about physical improvements 2 References in this study are to technical efficiency, namely the production of a given output of urban services for fewer (i.e. less costly) inputs. The concept of allocative or demand efficiency is referred to briefly the report's discussion of project's impacts upon decentralization in Chapter 5. In using the term equity, the study refers to project goals of equitable allocation of the benefits of urban services to low income residents, commonly called poverty targeting in project documentation. 8 on the ground were tested through direct observation by the study team during site visits to the areas improved. For other impact domains, indicators were mostly qualitative and hypotheses tested through contacts with key informants and community groups in the project cities themselves. Table 2.1 summarizes key details. 2.6 Quantification: To measure project physical impacts upon living conditions, the study constructs quantifiable impact indicators to measure changes in service provision before and after the projects at the municipal level. The evaluation also conducts with and without project evaluations, through comparing the performance of project and non-project cities. The expectation, of course, was that progress toward meeting demand for infrastructure would be greater in project cities than in unassisted cities. Table 2.1 Impact Domains, Indicators and Hypotheses - Domain 1: Living Conditions of the Poor Indicators: Key Evaluation Hypotheses: 1. households with adequate Greater improvement during 1980-91 in project cites than other cities. water supply and sanitation 2. households with adequate Greater improvement during 1980-91 in project cities than other cities. housing quality and space 3. physical improvements on the Project interventions led to sustained improvements in living conditions ground of intervention areas, subject to maintenance constraints. 4. local economic and social Project improvements triggered new local economic and social activities activities. - Domain 2: Decentralizing Urban Service Provision Indicators: Key Evaluation Hypotheses: 5. project ownership Local project ownership increased as responsibilities shifted to municipalities. 6. local legislation Project municipalities adopted laws and regulations to simplify project approvals, planning and execution. 7. municipal administrative Project interventions led municipalities to focus upon project evaluation, procedures procurement, improved methods of cost recovery. 8. municipal administrative Project municipalities made permanent changes to facilitate project structure planning and execution. 9. nature of decentralization Decentralization was within the public sector, primarily devolution of responsibilities to municipalities. - Domain 3: Users' Voice and Learning about Urban Service Provision Indicators: Key Evaluation 1iypotheses: 10. strength of formal community Projects gave community organizations-especially among organization women-a new role, mainly in monitoring implementation. 11. learning the rules of urban Projects allowed residents to learn about formal responsibilities and service provision procedures for urban service provision. Residents learned to hold local government accountable for these services. 12. residents' participation in Users had little or no voice during the planning stage, but this did not project planning undermine their support for interventions that were obviously necessary. 13. NGO participation Insignificant role because few urban NGOs exist in Brazil and their participation was not factored into project design. 9 2.7 Qualitative assessment: To evaluate project ID impacts the study compiled time lines of key ID events related to projects. From sector reviews and other literature, these events were mapped chronologically along four vectors: (i) Brazilian government sector policy; (ii) Bank policy for the sector as a whole; (iii) Bank policy for the sector in Brazil; and (iv) Bank financed project events. Each vector was disaggregated into a more detailed time line (Annex Tables A.1-A.4). The study team uses the evolving framework of government policy and Bank sector strategy to judge whether the occurrence of these events at particular times can be attributed to interventions by the projects evaluated. 2.8 Priorities: Study definitions of impact domains, indicators and hypotheses are based upon the priorities set by the projects themselves as well as themes of sector strategy topical today, notably the decentralization of urban development responsibilities. During the course of the study, the evaluation team's discussions with project beneficiaries, key informants and finally Brasilia workshop participants helped confirm that the themes selected were highly relevant to key project stakeholders. FOCUS UPON BEST PRACTICE MUNICIPALITIES 2.9 Reasons: Understanding how the most successful project municipalities were able to improve living conditions so much can provide a key to replicating the mechanisms and instruments used elsewhere. Looking at selected cases such as these among more than 464 municipalities benefited by these projects also kept the study itself manageable. It is important that evaluation should be cost effective and expeditiously completed if it is to provide useful inputs for future work. The focus at the level of the municipality is for a number of reasons. Since the 1988 Constitution, municipalities have been more responsible for urban service provision in Brazil. They also inherited responsibility for the projects evaluated here. Secondary data for this study is most readily available at the municipal level. Looking at best practice cities was stimulated by the evaluation team's concern for the plight of other weaker municipalities. By elucidating some of the features of best practice, the study hopes to disseminate solutions that can help bring less successful cities up to form. 2.10 Identification: From census data, the study identifies best practice cities among project municipalities as being those making the greatest progress in improving the urban infrastructure and housing conditions of their residents between 1980 and 1991 (details Table 2.2). The evaluation case study work focuses upon two best practice cities each from the Medium-sized cities project (Loan 1720), the Paran. Market Towns Project (Loan 2343) and the NE Flood Project (Loan 2645). It also includes the three metropolitan cities that hosted projects in the Northeast region. Some adjustments in selection were made for mission logistical reasons. The study also covers Juazeiro as the twin town of 10 best-practice Petrolina. In addition it considers Natal, a city that has achieved very high levels of urban service and housing. DATA COLLECTION METHODS 2.11 Secondary data: Before collecting primary data during site visits, the evaluation team reviewed a considerable body of information from secondary sources. These included Bank project documentation itself, evaluation documents, Bank sector papers and government urban policy documents referring to the 1975-1996 period. This review was particularly important to help ensure that possible exogenous-in other words non-project-causes of impacts could be isolated from project effects themselves. Finally, the study assembled 1980 and 1991 census data on urban service levels to portray the before project and after project situations of the cities studied. 2.12 Primary data: OED fielded three missions to eleven cities in Brazil in September and December 1995, and July 1996 to collect primary evaluation data by the following means: " group interviews with beneficiaries in the areas improved under the projects. " key informant interviews with present and former local government officials, others with knowledge of the projects and NGO representatives. " expert on-site observation and assessment of the cost, quality and condition of project infrastructure and living conditions, using a standard checklist for all cities. The guidance of federal authorities knowledgeable of the project experiences- notably SEAIN in Brasilia and SUDENE in Recife-helped ensure the success of regaining the institutional memory of these important experiences. BENEFICIARY PARTICIPATION IN EVALUATION 2.13 Excellent response: Eleven group meetings with 8-16 project beneficiaries in each of the cities visited gave important evaluation insights into participatory development of poorer areas of Brazilian cities. Responses were very good and residents most willing to help with the study's inquiries at very short notice. Even for project events happening more than ten years ago, residents' memories today remain very firm about what were watershed 3 As a result of this selection, the cases evaluated cover a wide range of city types throughout the country. They include: (i) the three metropolitan regions of the Northeast, Fortaleza, Recife and Salvador; (ii) five medium-sized cities, Petrolina, Juazeiro, Natal and Aracaji in the Northeast and Juiz de Fora in the Southeast; (iii) two small market towns, Propri6 and Colorado; and (iv) a dormitory suburb of the metropolitan region of Curitiba, Almirante Tamandar6. 11 interventions in their neighborhoods. Most areas visited have undergone no further improvements since the original project interventions. 2.14 Element of surprise: To avoid possible stage-management of community responses to its inquiry, the evaluation team selected the beneficiary groups to interview only upon the missions visiting the selected cities. After arrival, the evaluation team identified the project areas with the help of local officials. One of these areas was randomly selected and the team got to work immediately contacting and setting up a meeting with the neighborhood group. 2.15 Group interviews: In some cities, community social workers or the local community based organization (CBO) itself helped locate residents to make up the community group for interview. In other cases, members of the evaluation team itself went from door to door to invite residents to participate in the discussions. In no cases did the local authorities hand-pick participants. In larger metropolitan cities, it proved more difficult to mobilize residents for such meetings than in smaller towns. 2.16 The discussions with beneficiaries: The venues of the meetings were varied, sometimes a community center, school, health post, neighborhood bar, the home of one of the beneficiaries, or even outdoors on a neighborhood sidewalk. In this way, fruitful discussions lasting around 90 minutes were held with some 105 final beneficiaries in the cities visited. Their fascinating stories led the team to look more closely at participatory urban development than foreseen in the original study design. When, at the very first group meeting, a local resident said that the project made the area 'heaven' compared with how it was before, she may have been exaggerating, but the evaluation team knew that it was on to something. 12 Table 2.2 Indicators of Living Conditions in Case Study Cities 1980-91 1991 Share of population with adequate services In terms 1980-91 percentage point change of share of population 1991 1980-91 of: with adequate services in terms of: (1) (2) (3) (4) (5) (6) (7) (8) Municipality Urban Urban pop. Water Basic Overcrow- Housing Water Basic Overcrow- Housing - State population Growth Supply Sanitation ded Quality Supply Sanitation ded Quality p.a. Housing Housing Fortaleza 1,768,637 3.6% 70.4% 39.6% 68.0% 78.4% +30.2 -26.9 +11.8 -9.4 -CONr stat 4,143,809 17% 506% 30.6% U8,e% 9010% +23,4 -15.0 +1311 +8.4 Recife 1,298,229 0.9% 80.7% 51.9% 74.3% 63.8% +22.4 +14.0 +13.0 -28.0 *Petrolina 125,273 2.4% 64.4% 44.0% 58.5% 96.4% +21.3 +6.6 +13.5 +8.2 -Pemmbucostate 5.028.263 27% 74.6% 370% 699% 854% +217 +14.9 +13.0 -3.9 Salvador 2,073,510 4.0% 88.3% 66.7% 90.0% 67.5% +15.7 +14.1 +11.9 -1.0 Juazeiro 102,266 1.3% 65.9% 17.1% 56.8% 99.7% +10.6 -9.2 +9.1 +23.2 -Bahia state 6,966,914 3.9% 700% 40.9% 6.5% 95,3% +12,7 +411 +10.5 +7,f *Natal 606,887 4.7% 85.9% 81.4% 70.7% 98.0% +16.7 +1.2 +11.3 +1.9 - o Norte state 1,66492 18% 87% 505% d66% 911% +21.3 +If-4 +.1+ *Aracaji 402,341 3.9% 86.0% 74.5% 73.9% 97.6% +17.1 +22.2 +13.4 +2.6 *Proprid 21,944 0.3% 85.2% 64.5% 64.1% 99.5% +14.6 +41.9 +7.7 +14.1 - Serglp. state 997,233 4.6% 8L4% 4&9% 696% 9&3% +21.4 +174 +13.4 +&7 *Almirante Tamandar6 59,080 7.4% 79.9% 54.7% 56.5% 98.4% +59.9 +41.8 +13.9 -3.4 *Colorado 15,616 -0.8% 91.9% 50.5% 84.4% 99.9% +33.2 +50.5 +19.9 -0.1 - ParanA state 6,150,567 3.1% 91,2% 5a4% 80.0% 97.1% +20.3 +10.4 +20.2 -2.1 *Juiz de Fora 380,249 2.4% 97.0% 93.2% 89.0% 98.2% +10.2 +10.1 +16.4 +0.3 - Mines Gerals state 11,786,893 2.6% 87,4% 73.5% 79.8% 98.9% +18,0 +178 +23.2 -1.7 -BRAZIL 110,990,990 3.1% 859% 06.8% 716% 94.% +14,0 +8.7 +15.6 -40.2 Notes: (1) Adequate water supply defined as households with internal water supply to house, whether supplied by main system, well, spring or other source. (2) Adequate basic sanitation defined as households with waste outlet collected to main sewerage system or septic tank, whether individual or shared. (3) Adequate housing with respect to overcrowding defined as households living at less than 1.85 persons per room. (4) Adequate housing with respect to quality defined as households not living in favela type units. (5-8) The share of households adequately served in 1991 minus the share adequately served in 1980. Source: IBGE Demographic Census 1980 and 1991 13  3. The Projects and Their Setting All five projects were prepared without the benefit of country sector work. Urban policy dialogue in Brazil-with early emphasis upon poverty alleviation-evolved only as part ofproject preparation itself The federal National Council for Urban Development (CNDU)- responsible for the projects-was the main interlocutor of the Bank during 1979-85. In approving these early projects, Bank EDs emphasized efficiency goals that were later supported by Bank sector papers on Brazil from 1983 onwards. ICRs and PARs consid- ered all but one of the operations satisfactory, ratings endorsed by this study. Although varied in scope, all projects aimed to improve living conditions of the urban poor and to help institutional development. IER findings focus upon these two aspects and from the lessons learned, the study recommends key actions to help improve urban service provision across many municipalities throughout Brazil. POVERTY FOCUS OF EARLY PROJECTS (1972-1983) 3.1 No sector work background: All five projects were prepared during 1979-83 without the benefit of country specific Bank sector work. This meant that each operation had to find its own policy priorities. Sector work for Latin America began only in 1983, and specifically for Brazil only in a 1985 paper. Without sector work, project preparation was inevitably protracted; it had to include a policy dialogue too. Preparation of these projects typically required Bank staff inputs of 75-100 staff weeks over a 3-5 year period, compared to 25 staff weeks over 1-2 years for operations today (World Bank 1994. Annex Tables). 3.2 Reasons: Until 1983, the Bank relied upon centrally conceived blueprints of sector policy that were not tailored to individual countries such as Brazil. There had been little progress in sector dialogue between the Bank and Brazil itself, owing to lack of support by the National Housing Bank (Banco Nacional de Habitagdo - BNH). At that time, BNH was at the center of an impressive array of sector policy instruments embodied in Brazil's Second National Development Plan (ll-PND 1975-79). BNH had little time-or need for additional resources-to engage the Bank in policy discussions.4 3.3 Urban lending and CNDU: Bank urban lending picked up in 1979 after an inter-ministerial committee in Brasilia, CNDU (National Council for Urban 4 Consequently, the results of the only Bank project with BNH as borrower-the Sites and Services and Low-cost Housing Project (Loan 1653-BR)-were marred by what the OED audit referred to as the lack of interest of an "unwilling borrower". 15 Development) was put in charge of federal government urban policy, and over- saw external lending for the sector. All five projects evaluated by this study were prepared through CNDU.5 The Bank engaged CNDU in sector policy discussions closely but still on a project-by-project basis only. The underlying Bank agreement with CNDU was that efficiency entailed the large-scale replication of project experiences. Equity concerns meant that Bank investments should concentrate, as a priority, upon Brazil's poorest Northeast region. EXECUTIVE DIRECTORS' VOICE 3.4 EDs' concerns with cost recovery: Policy direction also came, from Bank Executive Directors (EDs) at Board Presentations. For each of the five projects, EDs wanted to know how cost recovery would work. They saw cost recovery an instrument for mobilizing resources for additional urban investments needed in Brazil. Threats to the efficient implementation of the projects themselves were also of concern to EDs. They feared the operations were unduly complex and difficult to manage efficiently, involving too many components, cities and agencies.6 3.5 Skepticism over equity goals: The Board also supported the projects' poverty reduction equity goals. But EDs were keen to leam how the projects could create employment opportunities within intervention areas themselves and thus be instrumental in helping the urban poor help themselves. Also, EDs wondered whether the projects' equity goals were being thrust upon a reluctant Brazilian govemment, which might be unconvinced about the efficacy of this approach. They were concerned that urban improvements would further stimulate rural-urban migration, eventually overloading services and worsening living conditions. In short, EDs challenged Bank staff to ensure that these urban projects would deliver efficiently on their promises of sustained improvements to the living conditions of the urban poor in Brazil. EFFICIENCY FOCUS OF THE FORMAL STRATEGY (1983-87) 3.6 First attempts to spell out a strategy: From 1983, Bank sector strategy in Brazil began to take form, albeit in a piecemeal and ad hoc fashion. Brazil's increasingly volatile institutional environment following the debt crisis of 1982 made it necessary for Bank to spell out a country strategy in all lending sectors. For urban, the first attempt came in the Brazil chapter of the 1983 sector paper on Urban Development in LAC in the 1980s. This endorsed efficiency and equity 5 Initail identification and preparation of the Medium-sized Cities Project (Loan 1720-BR) began under CNDU's predecessor, the National Committee for Urban Policy and Metropolitan Areas (CNPU). * The Brasilia workshop highlighted cost recovery, its conception and measurement, as an priority element of future project design and monitoring. 16 goals as twin pillars of an urban strategy. Out of this paper came CNDU's agreement with the Bank that project investments in the metropolitan centers of the Northeast region would be instrumental to this strategy. 3.7 Support for macro-economic growth: The 1985 Urban Sector Strategy for Brazil--the first Bank sector paper devoted exclusively to Brazil-consolidated the earlier LAC approach. It offered a potentially resilient strategy that might survive the expected turbulence of transition from military to civilian rule in that year. Clearly, the way of doing urban business in Brazil was about to change radically. CNDU itself was abolished in 1985. The 1985 Strategy's answer was to look beyond specific organizational arrangements and favor urban projects with efficiency goals as a way toward faster economic growth.7 The latter, was agreed by all parties, to be a top priority. Poverty reduction was still in the strategy, but more as a constraint upon macroeconomic growth that had to be addressed, rather than as an end in itself. Perhaps for being too bold in shifting emphasis from equity to efficiency, the 1985 Paper had only limited dissemination (in green cover) and was not endorsed by the incoming civilian administration. 3.8 A short-lived Ministry: The new civilian government appeared to give urban development a high political priority in 1985, when it created a separate, but short-lived, Ministry of Urban Development (MDU). But BNH-still the principal financing agent of the sector-was unexpectedly abolished the following year. MDU itself, after internal reorganization and a change of name was itself made extinct in 1988. These events represented the final dismantling of direct federal involvement in the sector and were a prelude to the 1988 Constitution that invoked a major shift of power, funding and responsibilities to municipal government. Although the projects themselves made no direct contribution to these major political changes in Brazil, the study finds that they did demonstrate that municipalities could become effective partners in urban development programs after the federal government withdrawal (Chapter 5). SHIFT TO DECENTRALIZATION (1988-92) 3.9 First, project action: Decentralization seemed an appropriate strategic option in Brazil in the run up the 1988 Constitution, but really it was the only avenue still open to continue urban development programs following federal government retrenchment from the sector during 1985-88. For this reason, the Bank supported two municipal development projects in the southern states of 7 Interestingly, discussions during the Brasilia workshop focused upon the importance of considering the macro-economic context when designing interventions to improve urban service provision. One Bank Strategy victim of the dismantling of the federal apparatus was the 1987 Urban Development in the Northeast paper, which called for closer attention to economic growth in the region. The paper argued that poverty targeting per se was not a sufficient condition to justify Bank investment. Federal government retrenchment interrupted Bank urban activities, leaving the approach proposed by this paper untested. 17 Paran6 and Rio Grande do Sul in 1989. For such projects, decentralization promised a context where equity and poverty alleviation concerns could be incorporated into development programs driven by basic concerns for efficiency. 3.10 Later, policy support: At a global level, the Bank supported decentralization through municipal development in its 1991 Urban Policy Paper (UPP) which revived the sector policy debate and specifically endorsed munidipal solutions to urban service provision as an efficient solution. These tenets were incorporated into two sector papers for Brazil, the 1991 draft Medium-Term Strategy for the Infrastructure Sectors and the 1992 The Challenge of Municipal Development in the 1990s paper. Equity considerations took second place in both papers, although the municipal report reckoned that local governments had a key social role to play in poverty alleviation. This strategy drove the design of newer municipal development projects in Brazil-approved in 1993 and 1997 respectively-currently under way in the states of Minas Gerais and Bahia. RESULTS OF COMPLETED URBAN PROJECTS 3.11 Results thus far: Since 1978, the Bank helped finance altogether fourteen completed urban projects in Brazil-including four urban transport projects not covered by this IER-in an amount of US$1.2 billion. Table 3.1 summarizes details of the five operations covered in this study. 3.12 Ratings: Four out of the five operations reviewed by this study were rated satisfactory by OED (Table 3.1). The Medium-size Cities (Loan 1720-BR) and Paran. Market Towns (Loan 2343-BR) projects were particularly successful, with top ratings in all categories, including sustainability and institutional development. Only the Recife Project (Loan 2170-BR) gained an unsatisfactory rating, owing to its unduly ambitious design and difficult macro-economic conditions during implementation, according to the PAR. The IER reconsidered these ratings in light of new information available and concluded that they accurately reflect the performance of these projects as far as meeting original project objectives are concerned. 3.13 Project similarities: While the scope of objectives among the five operations varies, they did all aim to improve living conditions in project cities. Likewise, formal project descriptions were diverse, but always included components of upgrading low-income areas9 and local institutional development. For this reason, much of this evaluation focuses upon these two components. This narrow focus is for methodological reasons only, and does not infer that project components should not include complementary actions-such as a Although housing the urban poor, all project areas visited by the evaluation team had regular street geometry, subdivisions and tenure. For this reason, they cannot be classified as favelas. Nevertheless, written comments on this study from Brazil argued that urban policy should also address informally occupied settlements within the city, something this study fully endorses. 18 environmental campaigns and the mobilization of preventative "health agents"- recommended in written comments by Brazilian experts upon earlier drafts of this report. 3.14 Later projects: More recently completed projects in the Brazil urban portfolio include municipal development projects in the southern states of Santa Catarina, Paranc and Rio Grande do Sul.10 Their respective ICRs rated all three satisfactory. They are not covered by this IER, however, since they have not yet been completed for five years, the minimum time for sustainable impacts to be considered measurably. Table 3.1 Project Summaries Objectives Description Ratings/final cost/final loan Medium-size Cities (Ln 1720) Improve urban infrastructure and 0 Upgrading through infrastruc- Overall: Satisfactory services; strengthen federal ture in low-income areas Institut Dev. Substantial agencies responsible for urban 0 Community services Sustainability. Likely policy. 0 Improve municipal admin. - final cost US$150 million - loan amount US$70 million Recife Metropolitan (Ln 2170) Improve housing and urban 0 Upgrade low income areas overall: Unsatisfactory services; induce desired patterns 0 Metropolitan infrastructure Instftut Dev: Negligible of metro development; income 0 Income and employment Sustainabilitr Unlikely generation for the poor; components - inal cost US$216 million institutional development. Institutional development - loan amount US$108 million Fortaleza/Salvador Pilot Metropolitan (Ln 2193) Prepare development programs 0 Detailed engineerng and Overall: Satisfactory for Fortaleza and Salvador; investment studies and TA Institut Dev Partial strengthen planning agencies. 0 Pilot upgrading works Sustainabllity* Marginal - final cost US$12 million - loan amount US$$6 million Parana Market Towns (Ln 2343) Long term borrowing for municip- 0 Municipal infrastructure Overall: Satisfactory alities; systematic selection of 0 Community facilities InsUtut Dev: Substantial subprojects; fiscal recovery of 3 Equipment and vehicles Sustainability. Likely investment costs; better living 0 TA for cadastres, tax and - final cost US$108 million standards through infrastructure, municipal accounting - loan amount US$52 million NE Flood Reconstruction (Ln 2645) Rehabilitate flood-damaged 0 Reconstruction of housing Overall: Satisfactory municipalities; strengthen 0 Urban infrastructure Institut Dev: Partial planning for flood prevention. 0 Community facilities Sustainabillty: Likely ib Micro-enterprise credits - final cost US$166 million sh Studies - loan amount US$99 million Sources: Project ICRs and PARs. 10 Santa Catarina Small Towns Improvement Project (Loan 2623-BR), the Paran State Municipal Development Project (Loan 31200-R), and the Rio Grande do Sul Municipal Development Project (Loan 3129-BR). 19 ON-GOING SECTOR OPERATIONS AND PRIORITIES 3.15 Current projects: Recently, the Bank has reduced its urban lending portfolio in Brazil. Currently, three urban projects are currently under implementation as of early 1997.11 Two municipal development projects, one each in the states of Minas Gerais and Bahia, are geared towards efficiency goals through trying to improve productivity in urban service provision at the municipal level. The third project in the Northeastern state of Ceard, focuses more upon poverty alleviation through targeted investments in water resources in that drought-prone region. The reduction in urban lending comes at a time when Bank understanding of Brazil's urban sector has never been better founded, as far as sector work is concerned. 3.16 Inter-American Development Bank: Since receiving a mandate from the U.S. government to act as its representative in Latin America, the Inter- American Bank (IDB) has increased its lending for urban development projects in Brazil.12 In the case of the state of Paranc, an IDB operation continues reform efforts begun with earlier Bank-financed urban projects. 3.17 Current priorities: The reconcilement of efficiency and equity goals in reform efforts in Brazil remains at the center of Bank sector strategy in the country. As reform moves downstream with decentralization, a major challenge is how to develop multiple policy dialogues with many municipal interlocutors simultaneously. Clearly, the old-style one-on-one dialogue with an agency like CNDU is no longer appropriate. This IER aims to provide elements to help wholesale policy advice to local governments on a large scale through the dissemination of the lessons of best practice experiences. " (i) Minas Gerais Municipal Development Project (Loan 3639-BR); (ii) Cearc Urban Development and Water Resources Management Project (Loan 3789-BR); and (iii) Bahia State Municipal Management Project (Loan 4140-BR) with total Bank financing scheduled for US$390 million. 1 There currently three IDB financed urban projects under implementation in Brazil: (i) Urban improvement program of Rio de Janeiro; (ii) Favela upgrading program of So Paulo; and (iii) PARANA URBANO municipal development program, with scheduled IDB lending of US$580 million. 20 4. Impacts on Living Conditions As intended, the projects improved the living conditions of the urban poor, especially through better access and reduced flooding risks for local residents. Typical lack of maintenance meant that infra- structure solutions using more robust materials had more sustain- able impacts. Project beneficiaries themselves and outside expert observers had strikingly similar viewpoints on these issues. Impacts upon the local economy were modest, while social impacts were more significant. Based upon these findings, the study makes three recommendations. First, efforts should continue to improve living conditions of the poor. Second, infrastructure and service solutions should employ robust materials. Third, evaluation should always take into account the beneficiaries'perception ofproject outcomes. FRAMEWORK OF EVALUATION 4.1 The presumption that improvements to living conditions can be attributed to the projects is strong. These operations were the only significant interventions in recent years in the areas visited, according to local residents and government officials. Purely in terms of evaluation methodology, the projects' exclusiveness in the areas concerned meant that there was no chance that the improvements observed were outcomes of other operations. In terms of neighborhood improvements themselves there is a widespread belief-echoed at the Brasilia workshop and by written comments on earlier drafts of this report-that successive project interventions in one area are likely to lead to even greater improvements than a single operation. BENEFICIARIES' PERSPECTIVE 4.2 Better conditions: Final beneficiaries themselves rate project improvements as making living conditions better in seven and much better in three improved areas of the cities visited (Table 4.1). Only in Recife, whose project outcome was rated unsatisfactory, do local residents see no improvement in the project area. Nowhere do they consider that living conditions are worse. These findings are based upon residents' opinions about changes after the project in physical attributes such as access, health and education services, basic sanitation, drainage, street paving and maintenance. 4.3 Dramatic improvements over time: Within the groups interviewed, longer-term residents, who had lived in their respective area for 12-15 years or more, graphically recall the poor conditions of their neighborhoods before the projects. Perhaps the most striking case of improvements was in the city of 21 Colorado, where the project had literally saved the habitability of neighborhoods threatened by severe erosion. In Natal, for instance, the project area was known to them as the "sewer of Natal" before the improvements. Even people who had moved into these areas after the project leamed about the conditions prevailing before the projects from older neighbors. Perception of the improvements is very important in itself too. Residents in Petrolina, for instance, explained to the evaluation team how they feel that their area has become more 'respectable' as a result of the project. Table 4.1 Beneficiaries' perspective on their living conditions City: Rating of project outcomes: Most commonly cited examples of: - Project area Before:after With:without - satisfactory outcomes - unsatisfactory outcomes Fortaleza Better (no comment) Regular bus services Lack of sewerage. - Parque Potira More day-care centers Flooding is worse. Recife Same (no comment) Bus service right into Flooding is worse. - Casa area. Maintenance inadequate. Amarela More paved streets. Salvador Better Better Less flooding. Lack of health services. - Antonio Better vehicular access. Maintenance inadequate. Balbino *Petrolina Much better Better Better drainage. Blocked sewerage - Vila Moc6 Health posts and schools. system. Maintenance inadequate Juazeiro Better Better Better vehicular access. Flooding still occurs. - Alto do More school places. Maintenance inadequate Cruzeiro *Natal Better Same Better vehicular access. Flooding still occurs. - Quintas Access to nearby health Maintenance inadequate. posts *r Much better Better Less flooding/sewage none. - 18de Forte Bus service more reliable *Proprii Better Better Less flooding. Maintenance inadequate. - Bairro More schools. Am6rica *Almirante Better Better Better bus services. Side streets still unpaved. Tamandar6 Health post and schools. - Lamenha Grande *Colorado Much better Same Better vehicular access. none. - Colorado Less flooding. *Juiz de Fora Better Same New bus services. Flooding still occurs. - Santa Health post and schools. Maintenance inadequate. Efigrnia Notes: * best practice cases. "before:after outcomes' summarize beneficiaries' comparison of the physical characteristics of the their areas today with those of the unimproved pre-project condition of the same areas "with:without outcomes' summarize beneficiaries' opinion of the physical conditions of their areas today with those of similar, but still unimproved areas of their cities today. Source: Impact Evaluation Study Group Interviews with Beneficiaries 1995-96 4.4 Better conditions than neighbors': Local residents in six cities reckon their neighborhood to be better off today than other similar but still unserviced 22 areas of the respective cities. In three more cities-all best practice-residents consider their area to be the same as others with respect to the physical environment (Table 4.1). All residents, including shorter-term ones, made comparisons with and without the project using a control area explicitly identified for this purpose. In two cases, Recife and Fortaleza, local residents were reticent to give their opinions of other areas, however, for fear of offending their neighbors from those neighborhoods. In the three cases where residents report no difference-Natal, Colorado and Juiz de Fora-residents are not dissatisfied with their own respective area, but simply reporting that it has become the same as respectable neighborhoods of town. 4.5 Access: Project improvements most commonly cited by local residents- in all cities but Petrolina and Proprii-are related to better access to the area (Table 4.1). In some cases--Salvador, Juazeiro, Natal and Colorado-paved streets and improved drainage allowed regular vehicular access to these areas for the first time. Thus bus routes enter right into areas where they could not penetrate before. In Fortaleza, for instance, residents recall a half-hour walk to the nearest bus stop before the project. Today, buses ply right through the project area. In Aracaji, a new bus terminal was inaugurated in the improved area itself in 1986, shortly after the completion of the project. In 1985, new bus routes were added to serve project areas in Almirante Tamandar6 and Juiz de Fora, as soon as the project works were completed in those cities. In addition, paved streets allow access by garbage trucks helping to guarantee cleaner public areas in the project sites. 4.6 Drainage: An important direct impact of the projects cited by residents of areas previously prone to annual flooding in four cities-Salvador, Aracajo, Propri. and Colorado-is that no serious flooding has occurred since the completion of project drainage works some 10 years ago. On the other hand, residents report that flash flooding in Fortaleza and Recife has actually worsened since the projects. In both cases, this is because of additional paving without the necessary drainage works. Formerly, the porous surfaces of unpaved areas absorbed much of the run-off water. Flooding still occurs in project areas in Juazeiro and Natal, but this is due to inadequate maintenance, which allows for garbage to block the drains. 4.7 Social services: Residents also value additional health posts, day care centers and schools in the improved areas. At the time of the visit of the evaluation team, however, local people were dissatisfied that health posts had almost no medicines due to a national supply crisis at that moment. The new facilities were not components of the project themselves, but were for the most part financed and implemented by municipal governments shortly after the completion of the projects. A important facility-the technical school in Propri6 for instance-was built by a private charitable foundation after improvements had been made in the area. This is an example of an indirect project impact triggered by initial upgrading and demand for additional improvements by the local community. 23 4.8 Inadequate maintenance: Lack of infrastructure maintenance is the commonest complaint by residents. It was mentioned in seven cities: Recife, Salvador, Petrolina, Juazeiro, Natal, PropriA and Juiz de Fora. Streets and sidewalks are not cleaned or repaired. Although solid waste is now regularly collected, much garbage still piles up in the streets and blocks drainage. Residents admit that careless disposal of garbage by local people undermines project benefits. The study gathered no evidence of self-help efforts to clean up neighborhoods as are sometimes organized in the urban kampungs of Indonesia (World Bank 1995). In Recife for instance, accumulated garbage even undermines the physical stability of steep slopes. Only in Colorado do local residents consider maintenance to be better today than before the project. That is because flooding and serious erosion prevented the municipal government from carrying out basic maintenance in conditions prevailing before the project. 4.9 Water supply and sanitation: Local residents also complain in most of the cities visited that, although water distribution networks had been built, water availability is irregular, sometimes only at the end of each day. Also, few of the improved areas have an adequate sewerage system. A low-cost 'condominial' network was implemented in Petrolina, but the success of this very cheap solution depends upon proper use and maintenance by users themselves. Residents there informed the evaluation team that their neighbors allowed the backyard sewer to block frequently, causing sewage to back up into their houses. INFRASTRUCTURE AND SERVICES: EXPERT ASSESSMENT 4.10 Pavement: Cobblestones are the most common type of street paving (Table 4.2). Only in Recife is the pavement condition of cobblestones rated good, according to the visiting engineers. They rate the condition as fair only in the other five cases. Improved access to project areas increases the flow of traffic- particularly heavy service vehicles such as buses and solid waste trucks-which break up pavement not designed for these kinds of loads. Asphalt paving is proving more resilient and is rated in good condition today in all four cases, even where maintenance is not good as in Almirante Tamandar6 and Juiz de Fora. 4.11 Drainage: All projects include drainage either in the form of covered ducts running underneath each side of the street (five cities) or open channels mostly in the center (five cities). Fortaleza was a special case that had no drainage component at all (Table 4.2). Overall, covered ducts are in better condition today, while the visiting engineers rated two of the open channels in Natal and Almirante Tamandar6 to be in poor condition. Where residents report less flooding, covered ducts are generally used. Areas where flooding is worse rely upon open channels which are frequently blocked by garbage and other debris, such as concrete linings which are detached from canal walls. 4.12 One final technical observation about an innovative drainage technology used for the first time at the project site in Salvador: there, concrete stairways for 24 slopes are designed in a modular fashion, and each piece includes a hollow U- shaped duct directly beneath the steps that carries rainwater to the main drainage channels at street level below. Despite no maintenance or repairs in ten years, these stairways continue to function well to the present day. Moreover, residents report fewer flooding problems in the area. 4.13 Water supply and basic sanitation: Visiting engineers rated the overall conditions fair or poor in most cities except AracajO and Colorado. Sewerage is generally the most problematic. In Juiz de Fora, for instance, project design allows untreated sewage to be discharged into storm drains, polluting open drainage channels. A higher standard of system is used in Colorado, separating household sewage from storm drainage with better results for the neighborhood environment. The engineers confirmed the blockages of the low cost sewerage system reported by residents in Petrolina. Table 4.2 Infrastructure impact: expert assessment City: Pavement Pavement type of drainage street basic overall - project area material condition drainage condition lighting sanitation maintenance Fortaleza Irregular fair Surface poor fair poor poor - Parque Potira stones Recife Cobblestone good Open fair excellent fair fair - Casa Amarela channel Salvador Cobblestone fair Open fair good poor fair - Antonio Balbino channel *Petrolina Cobblestone fair Covered good good fair fair - Vila Moc6 ducts Juazeiro Cobblestone fair Covered fair good fair fair - Alto do Cruzeiro ducts *Natal Cobblestone fair Open poor fair - fair - Quintas channel *Aracaj(j Asphalt good Covered good good good good - 18 de Forte ducts *PropriA Cobblestone fair Covered fair - - fair - Bairro Am6rica ducts *Almirante asphalt good Open poor - - fair Tamandar6 channel - Lamenha Grande *Colorado asphalt good Covered good good good good - Colorado ducts *Julz de Fora asphalt good open fair good poor fair - Santa Efigfnia channel Note: * best practice cases. Impacts are given one of four ratings: excellent, good, fair, poor. Source: Impact Evaluation Study's expert observations of project sites 1995-96. 4.14 Poor maintenance: Finally, expert assessments reveal many indications of inadequate maintenance such as poor street cleaning, inadequate removal of solid waste from public areas, broken street pavement and sidewalks, and very poor service to unblock drains and sewage pipes (Table 4.2). On this important issue there is unanimity of opinion between the study engineers and the local residents themselves (cf. Table 4.1). 25 4.15 Costs: Visits by engineers did not enable an analysis of project costs for three main reasons. First, cost records are naturally kept in current prices subject to past hyperinflation. Second, physical quantities in municipal records generally refer to several sites and works going on at the same time, and not just the one of the Bank financed project. Third, some municipalities extended project works at their own expense making it unclear where the physical boundaries of the project site lie. 4.16 Internal economic rates of return: With inadequate project cost data, accurate re-estimation of internal economic rates of return (IERR) of the majority of project investments is not possible. Nevertheless, local authorities in Salvador and Juiz de Fora made special efforts to verify project costs, allowing the evaluation team to calculate IERRs of 25% and 37% respectively for the projects' paving investments. In both cases, the IER uses increased local properties values reported by residents themselves to construct benefit streams for the evaluation.14 These values indicate that the investments are still worthwhile from an economic point of view. A broader comparative evaluation is made difficult, however, by the scarcity of historical benchmark estimates with which current performance could be compared. Completion reports do not include IERR estimates because of the data difficulties just mentioned. Local governments have the capability to make these estimates today (details: Chapter 5), but do not have records of earlier evaluations for which they were not responsible. Appraisal of the projects led to IERR estimates in only three cases: Medium- sized cities (Loan 1720-BR) - 12%; Recife Metropolitan (Loan 2170-BR) - 16%, and Parand Market Towns (Loan 2343-BR) - 24%. The figures given refer to average internal rates of return from paving investments across samples of cities or locations. 4.17 Environmental conditions: Visiting engineers rated the cleanliness of public areas according to whether garbage was strewn about and/or discarded in public places. Air pollution was assessed from direct observations of atmospheric dust particles, vehicle exhaust and offensive smells of sewage or solid waste. Fire risk was assessed on the likely ease of access to fire-fighting equipment to any part of the area and the risk of fire passing from house to house. Assessments of traffic conditions focus upon traffic congestion and likely vehicle:pedestrian conflicts to the disadvantage to the latter. Noise is assessed mostly through questioning local residents. Finally, the study assesses the quality of green areas and tree planting on the site (Table 4.3). 4.18 Cleanliness: The urban environment with respect to cleanliness of public areas is rated good in six cities, Fortaleza, Salvador, AracajiO, Almirante 13 The precise moment that costs were incurred-necessary to know in order to convert to real values-is not clearly discernible from project records. 14 These compare with a project-wide IERR estimate at appraisal for paving investment under the Medium- size Cities project of 12%. For being an engineering loan, no estimate was made for the Salvador under the Pilot Metro project at appraisal. 26 Tamandar6, Colorado, and Juiz de Fora (Table 4.3). This finding supports observations by local residents about improvements in solid waste collection services after the completion of the projects. Engineering inspections revealed that in project areas where cleanliness is only fair-Petrolina, Juazeiro and Natal, for instance-there are no public collection points equipped with containers where residents and others can bring garbage for collection. 4.19 Air quality and fire risk: Expert assessments rated this as excellent in five cities (Table 4.3). The most dramatic improvement, which results directly from project investment in paving, is the reduction of atmospheric dust particles. Conditions with respect to fire risks are also rated excellent in eight and good in the remainder. This finding confirms safety in these areas with respect to basic norms of land use and setback as far as the propagation of fire and access of fire fighting equipment is concerned. Table 4.3 Environmental conditions: expert assessment City: Cleanliness air fire risk traffic noise trees and - project area of public pollution conditions pollution green areas areas Fortaleza Good fair good good fair fair - Parque Potira Recife Fair fair good fair fair poor - Casa Amarela Salvador Good excellent excellent good fair fair - Antonio Balbino *Petrolina Fair excellent excellent good good fair - Vila Moc6 Juazeiro Fair excellent excellent good good poor - Alto do Cruzeiro *Natal Fair excellent excellent excellent good fair - Quintas *Aracajfi Good good excellent good good good - 18 de Forte *PropriF Poor good good good fair fair - Bairro Am6rica *Almirante Good good excellent poor good good Tamandar6 - Lamenha Grande *Colorado Excellent excellent excellent excellent excellent excellent - Colorado *Julz de Fora Good good excellent excellent excellent fair - Santa Efigdnia Note: * best practice cases. Impacts are given one of four ratings: excellent, good, fair, poor. Source: Impact Evaluation Study's expert observations of project sites 1995-96. 4.20 Traffic: Increased access to project areas could be expected to worsen traffic conditions, but in fact poor conditions prevail only in Almirante Tamandar6, where heavy vehicles now pass through an area that was formerly inaccessible (Table 4.3). Traffic management measures are necessary to avoid increasing 27 accidents in this area. In all other cities except for Recife, traffic conditions in project areas are rated good or excellent. Even though-traffic has grown in all these areas, it is still at levels far below busier parts of the respective cities. 4.21 Noise: Following the project interventions, noise levels from traffic and loud music have increased but still to acceptable levels (Table 4.3). In comparative terms for the residents, the changes may have been significant, however, given the lack of movement and extreme quiet of the pre-project conditions. 4.22 Green areas: Finally, with respect to tree planting and green areas, impacts are less than desired (Table 4.3). Only one area, in Colorado, was rated excellent in this regard, for having open green squares within the site and regular tree planting along all streets. Other cities whose areas were rated good were Almirante Tamandar6 and AracajO, where careful attention is paid to maintaining trees planted in public areas. Poor conditions are reported in Recife and Juazeiro, whose project areas remain almost totally devoid of greenery, either in the form of open spaces or trees. The study concludes that more attention needs to be paid to 'greening' neighborhood improvements of this kind. HOUSING IMPROVEMENTS 4.23 Enlargements: For most cities in the Northeast region, local residents informed that there had been little enlargement of individual dwelling units since the completion of the projects (Table 4.4). Apart from the slow growth of personal incomes in that region since 1989, the fixed geometry of what were fully occupied sites does not permit much expansion, except costly vertical construction. The exceptions were in Aracaj(j and Proprit, both cities of the state of Aracaj(j, a Northeastern outlier with the lowest incidence of urban poverty in that region. AracajO residents report that many houses have been enlarged, some even gaining an additional floor. In the richer South and Southeast regions of the country, local residents of all cities report enlargement of housing units on the sites. Interestingly, in both Colorado and Almirante Tamandar6, residents mentioned that a few newcomers first built new and larger houses soon after the improvements. Later, existing residents emulated these efforts, adding rooms to their own dwellings. 4.24 Building materials: The use of better building materials for housing was triggered by the projects, but with varied impacts across cities (Table 4.4). In best practice cities of the Northeast especially, more permanent materials are used today than before the projects. Thus traditional rustic materials-known in the region as 'taipa' which is a mixture of mud and wood-now give way to brick and concrete construction. A similar phenomenon occurred in Colorado in the South, but the traditional material used in that region was wood. No change of material is reported for Fortaleza, Juazeiro, Almirante Tamandar6 and Juiz de Fora where, according to local residents, houses have long used permanent materials. The two cases where traditional materials are still widely used and 28 where there has been little progress with more permanent materials are Recife and, surprisingly Natal. In both cases, severe unemployment and little growth of personal incomes were factors in the lack of house improvement observed. Table 4.4 Housing Improvements City: Size of housing Kind of building materials Price of housing units" - project area units used Fortaleza Same Same Increase - Parque Potira - permanent materials before - especially along main road Recife Same Same Decrease - Casa Amarela - rustic materials (taipa') - especially units on slopes used Salvador Same More permanent Increase - Antonio Balbino - especially along main streets *Petrolina Same More permanent Increase - Vila Moc6 - including reinforced - doubled since the project concrete Juazeiro Same Same Increase - Alto do Cruzeiro - many small houses - permanent materials before - significant increases *Natal Same Same Increase - Quintas - some permanent, some rustic *Aracaj6 Increase More permanent Increase - 18 de Forte - extra rooms, floors - rustic materials no longer - significant increases used *Propri Increase More permanent Increase - Bairro Am6rica - many houses - rustic materials no longer enlarged used *Almirante Increase Same Increase Tamandar6 - but some favelas - permanent materials before - along main road especially - Lamenha develop Grande *Colorado Increase More permanent Increase - Colorado - newer, larger - many wooden houses - dramatic increases houses before *Juiz de Fora Increase Same Increase - Santa Efigdnia - larger on average - permanent materials before Notes: * best practice cases. estimates based upon reports by local residents themselves. Source: Impact Evaluation Study's Group Interviews of Beneficiaries 1995-96. 4.25 Property values: Residents were unanimous-except in Recife-that the value of their houses is much higher today than before the projects (Table 4.4). The prices of well-located units along main roads increased most. Remembering past prices in Brazil's inflationary environment is impossible and residents' assessments of historical housing values are mostly based upon affordability. Data on house prices are easily accessible now, though, given the stable monetary unit of real and familiarity of most residents with recent real 29 estate transactions in their areas.15 The only reported case of falling house values was for the project area in Recife, where units on increasingly unstable slopes had devalued sharply in recent years. LOCAL ECONOMIC DEVELOPMENT 4.26 Employment creation: Only residents in one city-Juazeiro-report that employment opportunities within the project area have improved since the project interventions. This was not due to the projects, however, but to a program to stimulate micro-industries through a very large local community center called the Casa do Menor (Table 4.5). In all the remaining cities but one- Recife-local residents confirm that very little employment is available today in what still are and always were predominantly residential areas. In this respect, residents see little difference between the opportunities available in their area, from those available in other residential districts of town. In short, employment opportunities in these as well as other areas, are more dependent upon the overall conditions of the urban, regional and national economies. At the time of the evaluation team visits, employment prospects were only modest in all regions of the country, as reflected in the general pessimism of the responses of local residents. Project induced employment creation within project areas has nevertheless been negligible. 4.27 Access to employment elsewhere: Access to other centers of employment is reportedly easier today than before the projects in nine of the cities visited (Table 4.5). In Almirante Tamandar6, for instance, residents were particularly enthusiastic about better and more frequent bus services that made daily commuting to the center of Curitiba feasible for low income travelers for the first time. Previously, would-be commuters would have to walk more than 40 minutes along unpaved streets to the nearest bus stop. Improved access is important for local residents in Juazeiro for them to seize opportunities offered by a booming regional economy based upon irrigation by the Sao Francisco River Valley. Two exceptions to the improved access were Petrolina and Recife: in Petrolina, local residents always had good access to jobs before the project; in Recife, unemployed people among local residents expressed their pessimism about the downturn in the labor market of that city as a whole. 4.28 Commercial activity: Local commerce improved in eight cities following the project interventions notably in the form of local shops, bars and pharmacies (Table 4.5). In AracajiO, a new supermarket is now operating on the project site. Residents' recall that housing was more affordable-either it terms of rental or periodic installment payments-before the projects than today, concluding that housing today is currently more costly and valuable. While their assessment is what one would expect following project improvements, falling incomes making housing relatively more expensive, may have led some residents to infer rising house values when none may have in fact occurred. The IER recognizes this an a risk of error that inevitably goes with a historical perspective of values in Brazil's volatile hyper-inflationary climate until the recent past. 30 These effects cannot be attributed to the project alone, since commercial success depends primarily upon the buying power of the residents. Nevertheless, as improvements in the majority of the project areas occurred soon after completion of project infrastructure, the latter may be a contributing factor. In three cities, residents report no change to local commerce. In Recife, this reflects the overall lack of improvement in the area. In Juazeiro and Colorado, it reflects the proximity of the sites to the main commercial areas of the respective cities. Table 4.5 Impact upon economic development City: Employment opport- Access to jobs outside Local commerce - project area unities within area area Fortaleza Same Improved Improved - Parque Potira - still few jobs in area - in center of Fortaleza - more shops Recife Worse Worsened Same - Casa Amarela - fewer jobs than before - fewer jobs in general - nearest supermarket 6km Salvador Same Improved Improved - Antonio Balbino - still little work in area - buses make it easier - new bakers and bars *Petrolina Same Same Improved - Vila Moc6 - no new jobs in area - access was easy - many new small shops before Juazeiro Improved improved Same - Alto do Cruzeiro - micro-industries in area - to irrigation projects *Natal Same Improved Improved - Quintas - still no work in area - to other parts of city - more shops *Aracaji Same Improved Much improved - 18 de Forte - still difficult to find work - to industrial estate - supermarket opened *Proprid Same Improved Much improved - Bairro Am6rica - still no jobs in are- more bars and shops *Almirante Tamandar6 Same Much improved Much improved - Lamenha Grande - few jobs in area - by bus to Curitiba - new shops and pharmacy......... *colorado Same mproved Improved - Colorado - still no jobs in area - to outside of town - some new shops Juiz de Fora* Same Improved Same - Santa Efig6nia - still few jobs in area - easier journey to work - no new commerce Note: * best practice cases. Source: Impact Evaluation Study's group interviews of beneficiaries 1995-96. SOCIAL DEVELOPMENT 4.29 Before and after the projects: With respect to possible social impacts of the projects, the evaluation team discussed a series of questions with local residents. The team asked about community life in improved areas before and after the projects, not only as far as organized groups are concerned, but also with respect to the spirit of community and sense of belonging to the neighborhood. Questions about family life tried to elicit comments about the stability of the family unit and raising of children in the area. Inquiries about children's health probed for evidence of the reduced incidence of contagious infections typically associated with unsanitary urban conditions. Finally, 31 questions about public safety-a priority issue in Brazil's urban areas today- sought to clarify whether residents felt safer and knew of less crime occurring, and to gather evidence about the level of policing in these areas. 4.30 Community life: Formal community organization is stronger and the residents' sense of belonging is greater in seven of the cities (Table 4.6). In most of these cases, residents refer to more intense activities of their Community- based organization (CBO) which now acts like a pressure group on behalf of the community (details: Chapter 6). Community life is weaker in Recife however, where neighborhood neglect leads residents to suspect where the loyalties of the CBO lie. In Propri;, an existing CBO was disbanded when residents realized that they could easily gain direct access to municipal authorities in this small town. Colorado interestingly never had a CBO, and organized community life practically does not exist in the project area, the only one visited by the evaluation team that does not have a specific neighborhood name. The group discussion with the evaluation team itself had to be held on a sidewalk for lack of an other meeting place. The small scale of the town of Colorado appears to allow individual contacts to work well, removing the need for formal community level arrangements. 4.31 Family life: According to local residents, experience of changes in family life varies among the cities visited (Table 4.6). They considered it better or much better in Salvador, Petrolina, Juazeiro and Colorado, with less violence, less divorce, and fewer abandoned street kids. In short, family life has become more stable. On the other hand, in Natal, AracajO and PropriA, local residents reported almost exactly the opposite, with more divided families and juvenile delinquency. With so many factors determining the quality of family life-notable income and employment, level of schooling among them-it is not possible to isolate impacts due specifically to the projects. Interestingly, residents in Fortaleza and Recife were reluctant to talk about this issue with the evaluation team, feeling that this was not a matter to discuss with strangers. 4.32 Children's' health: Overall, conditions are better, with dramatic improvements in two cases (Table 4.6). In Petrolina, the incidence of skin and intestinal infections dropped rapidly after the paving and servicing of the area. In Almirante Tamandar, there have been fewer respiratory ailments among children since street paving reduced atmospheric dust in the area. In two exceptional cases, children's' health is reportedly worse. In Natal, domestic effluent is disposed directly into a local river where children play. In Proprii, inadequate street cleaning leaves garbage to pile up in open areas, where local children easily acquire worm infections. 4.33 Public safety: Notwithstanding a national trend of increasing urban violence, residents of six areas believe the neighborhoods are safer today than before the projects (Table 4.6). Among reasons given are the existence of fewer firearms in Petrolina and fewer known cases of violent crime in Salvador and Juazeiro. In other cities, residents simply report feeling safer, without specifying precisely why. On the other hand, residents from Fortaleza, Recife, Almirante Tamandar6 and Natal report a worsened security situation, reflecting increasing 32 crime in metropolitan areas generally, as well as a specific pocket of drug-related crime within the project area in Natal. The impact upon Almirante Tamandar6, where the project ended the former isolation of the target area, was to introduce the neighborhood to typically metropolitan problems of Curitiba not only in terms of crime, but also worsening traffic accidents. In this context, the reported reduction of crime by residents of the project area in Salvador was somewhat unexpected. Table 4.6 Impact upon social conditions City: community life family life Children's' public safety policing - project area health Fortaleza Better (no comment) Better Worse Same - Parque - CBO more - general - violent crime up - practically Potira active improvement absent Recife Worse (no comment) Same Worse Same - Casa - CBO weaker - respiratory - slope accidents; - non-existent as Amarela andpatemalistic allergies common regular homicides before Salvador Much better Much better Better Better Worse - Antonio - CBO acts as -family life more - diarrhea less - less crime today - non-existent Balbino pressure group stable common *Petrolina Much better Better Much better Better Better - Vila Mocd - CBO acts as a - now more stable - fewer skin and - fewer firearms - frequent patrols; pressure group intestinal infect- today none before ions Juazeiro Much better Better Better Better Better - Alto do - community ties - fewer street kids - diarrhea less - violent crime is - some police Cruzeiro stronger common less patrols *Natal Better Worse Worse Worse Same - Quintas - stronger sense - more family - intestinal - area is more - non-existent of community violence disease common dangerous *Aracajo0 Much better Same Better Better Much better - 18 de Foite - CBO more - juvenile delin- - diarrhea less - crime is less; - police post active quencycommon area feels safer within area *Propr:- Worse Worse Worse Better Much better - Bairro - CBO disbanded - families more - more cases of area is safer - police post Am-itca divided worms within area -;Amlant Mu B bete Worse Much better Worse Same Tamandaril - CBO set up for - lack of respect - fewer intestinal - more crime and - still very thin on - Lamenha the first time for older and respiratory traffic accidents the ground Grande --- - --------- members -- problems---- - -------- - --------- *Colorado Same Wetter Better Same Much better Colorado - no CBO, nor - area is still safe - regular policing name of area *Ulm;ire Fora Same Same B Better eter Better - Santa - CBO still weak - area is now - some policing Elgnia safer than before now; none before Note: * best practice case Source: Impact Evaluaion Study's group interviews of beneficiaries 1995-96. 4.34 Policing: Changed public safety correlates with policing of these areas (Table 4.6). In five of the six cases with better security--Petrolina, Juazeiro, Aracaje, Propri and Juiz de Fora--residents report better policing of their areas. In Aracaji and Propri, for instance, permanent police posts have been operating in the project areas since project completion. This impact is less strongly felt in metropolitan areas such as Fortaleza, Recife and Curitiba (of which Almirante Tamandar6 is a dormitory suburb). 33 노 5. Decentralizing Urban Service Provision Thanks to the projects, municipalities strengthened their own management ofproject preparation, evaluation and procurement. They also simplified the regulatory framework for project approvals. In consolidating their 'project culture" in these ways, municipalities seized the opportunity to play a bigger role in decentralized urban service provision after the federal government withdrew fi-om the sector. On the other hand, there were only modest impacts upon urban policy at the local level, which still has to fill the gap left by federal retrenchment. Nevertheless, municipalities developed strong ownership of the projects, even where the basic concept was not their own. Local ownership went deep enough in some cases to ensure smooth transfers of the projects between adversaries at times ofpolitical succession at the municipal level. In the long run, decentralization such as this contributes to the sustainability of urban improvements in Brazil by helping to ensure continuity of service provision. FRAMEWORK OF EVALUATION 5.1 Why focus upon decentralization: The study considers decentralization of responsibilities for urban service provision to municipalities as a significant 16 dimension of the institutional development (113) impacts of the projects. ID impacts by themselves had no direct bearing upon the pace and nature of Brazil's decentralization-which marked a watershed in the country's political history-but the projects did help municipalities become effective partners in decentralization. An intended outcome of these operations was to help municipalities develop effective and efficient management of urban service provision within a project friendly local regulatory framework. Effective decentralization of key responsibilities for urban service provision-particularly with less federal and state government activity in this field-is crucial for the sustainability of urban improvements. 5.2 Broad concepts of institution and ID: It is important to clarify the broad concepts of ID and institutions used in this study, as there is still no consensus among authors about the meaning of these terms. According to one definition, "institution encompasses entities at the local or community level, project management units, parastatals, line agencies in government, and so on"; in This study uses a broad concept of 11), as a process of changing rights, obligations and incentives of the principal players-governments, private sector, NGOs and beneficiary communities--in urban service provision. 35 other words, institution is a synonym for an organization or agency (Israel 1987 p. 11). Another author defines institutions as: "factors affecting the relationship among actors-especially government and private parties-and the structure of incentives" (Kessides 1993 p. ix). Most commonly in Bank literature, institutions are treated as organizations, government departments in particular (Brinkerhoff 1994 p. 137). This study nevertheless uses a broad definition and treats ID as changing "rules of the game" and strengthening the actors' ability to deal with them.17 5.3 Intended and unintended impacts: The IER evidence of such strengthening within municipal administrations, which enabled local governments to fulfill greater responsibilities for urban service delivery as decentralization progressed. This indirect project impact was unintended and could not have been foreseen at the time the projects were appraised when the federal government looked set to remain the principal player in urban service provision in Brazil. The subsequent federal withdrawal from the sector (paras. 3.7-3.9) provided an unexpected opportunity for stronger but indirect impacts at the local municipal level. HELPING MUNICIPALITIES DEVELOP A PROJECT CULTURE 5.4 Definition: The projects helped participating municipalities develop a project culture, understood here as effective and efficient management of urban service provision in a project friendly regulatory framework. Project culture helps create specificity and organizational effectiveness within a municipality, enabling it to formulate urban service goals more precisely and increase the intensity and timing of its project efforts (Israel 1987. pp. 47-73). The study found evidence of a project culture in three areas: (i) improved project preparation and evaluation skills; (ii) more attention to procurement; and (iii) a project friendly regulatory framework (Table 5.1). Unfortunately, project culture did not extend to maintenance, which was found wanting in nearly all cities. 5.5 Evidence: Participants in the Brasilia workshop expressed some skepticism about whether a project culture phenomenon is a permanent and sustainable feature of municipal practice. With this in mind, the IER only reports evidence of changes in practice sustained since the completion of the projects more than seven years ago. In metropolitan cities of the Northeast, for instance, local governments undertook economic evaluation of projects for the first time and still apply it today. The municipality of Recife aimed to reduce urban service deficits in different parts of the city in a systematic way, as done at project appraisal. Both approaches enhance specificity. In a wide range of municipalities 1 The evaluation team's own expectations of local ID impacts were modest at the outset of the study. An earlier review found that only one third of urban projects worldwide substantially met their ID objectives (World Bank 1994 p. 37). Also, municipalities typically lack the specificity and competition characteristics necessary to stimulate dynamic ID, according to one author (Israel 1987 p. 20). 36 in all regions, municipal management improved through the creation of in-house technical teams for the first time, some of which later formed the nuclei of permanent urban planning and research institutes mentioned earlier. These results coincide with the declared project objectives of strengthening the management capacities of the executing agencies. Table 5.1 Indicators of a municipal project culture project preparation Procurement project friendly and evaluation skills management regulatory framework Fortaleza economic evaluation Training of personnel 1979 and 1985 updating used for the first time masteplans Recife aimed solving service deficits systematically Salvador economic evaluation Special unit set up simplification of project used for the first time approvals *Petrolina project mgt team set up up update and simplification for the first time of land use and tax code. Juazeiro Specalun-it set up 1983 ate of1 building code Natal project mgt teamsetup - 1989 new environmental which later became protection laws IPLANAT *Aracajo (still dependent upon rem m consultants) has its own staff for first L time *PropriA city staff acquired own Special unit set up - engineers for first time *AlmIrante Tamandar6 (still through the state) all work no ne in- house *Colorado city undertakes its own (still through the state) 1989 simplification of tax engineering work in- law house for the first time *Julz de Fora project mgt team set up; all work now done in- simplified land use law. later became IPLAN house Note: * best practice cases Source: Impact Evaluation Key Informant Interviews 1995-96 5.6 Procurement In most cities-seveni out of eleven-in-house capacity to manage procurement was enhanced as a result of the projects. Four of these cities set up and staffed special procurement units within their municipal administrations for the first time. After the projects, nearly all cities now manage their own procurement affairs. Before the projects, most were dependent upon state or federal agencies for these services. The evaluation expected that transparent and locally accountable procurement practices would ensure more efficient use of resources through lower contract prices for urban investments. Concrete evidence of the results achieved by municipalities vis e vis alternative Salvador, Petrolina, Juazeiro, AacajO, Propft, Almirante Tamandar, and Juiz de Fora. 37 arrangements proved difficult to gather. Brazil's past hyper-inflation makes historic cost data opaque as far as real values are concerned. In hindsight, there clearly is a need for unit values of contracts to be among monitoring indicators of future municipal performance. 5.7 Regulatory framework: As a result of the projects, eightl9 of the eleven municipalities simplified regulations and local laws for project approval and implementation. The municipalities of Fortaleza, Petrolina and Aracaji updated city-planning laws. Petrolina and Colorado enacted entirely new property tax laws as part of project implementation. Rather than pursue legislation through a divided city council, however, Salvador preferred to simplify regulations for approving urban development projects. Whichever path they now follow, most project municipalities now have local regulatory frameworks that make it easier to approve urban development projects. 5.8 Use of computers: In eight citieS20 key informants reported intensified application of computers to project management and other tasks since the project interventions. In addition to computerizing municipal payroll and tax administration-common today among local governments in Brazil-information technology is typically applied to the financial management and implementation supervision of urban investment projects in general. This is particularly for cost and quantity controls during project executions as well as computer assisted design for new projects. In larger cities such as Fortaleza and Salvador, sophisticated digital mapping was used for the first time within the scope of the projects themselves. 5.9 Sustainability: Through consolidating their project preparation and evaluation skills, in-house management of contract bidding and awarding and providing a project-friendly regulatory environment, municipalities are better able to contribute to the sustained provision of urban services. This bodes well for the future project performance. In the before-project context of centralization of urban service responsibilities in the hands of the federal government, municipalities had neither the opportunities nor the need to develop such skills. Decentralization also encouraged the dissemination of project experiences. Officials from Natal and Petrolina, for instance, have received many visitors from other municipalities seeking advice and guidance on project management. Municipalities in the southern regions of the country do not report similar city-to- city exchanges, which nevertheless take place through state level agencies in that part of the country. 1 Fortaleza, Petrolina, Salvador, Juazeiro, Natal, Aracaji, Colorado and Juiz de Fora. 2s Fortaleza, Petrolina, Salvador, Juazeiro, Natal, Almirante Tamandar6, Colorado and Juiz de Fora. 38 URBAN POLICY MAKING AT THE LOCAL LEVEL 5.10 Urban institutes: The projects helped strengthen urban planning and research capabilities within local government. Project management teams set up in four cities-Petrolina, Natal, Colorado and Juiz de Fora-to handle the projects, were the first experiences of staffing of this kind in these municipalities. Urban planning and research institutes were established in Natal (IPLANAT) and Juiz de Fora (IPPLAN-JF) as a direct result of the projects. Project staff in those cities form the core personnel of these urban institutes. Both institutes are still very active today, having produced city master plans as recently as 1994 and 1996 respectively. IPPLAN-JF, using the project experience, was able to plan and implement a series of follow-on environmental and sanitation programs. INPLANAT was instrumental in introducin? participatory budgeting in Natal, as one of the first such experiences in Brazil. 5.11 Modest impacts on urban policy making: After retrenchment at the federal level during 1985-88, municipalities contacted by this study have done little to fill the resulting urban policy gap. Written comments on this study from Brazil confirmed that more needs to be done to simplify urban development policy and make urban planning effective.22 In some cities, it is still not clear what the underlying purpose of urban development policy should be; whether its priority should be to support a strategy to address either efficiency or equity goals in the terms discussed earlier in this report (details: Chapter 3). 5.12 Significant impact on cost recovery: At the level of project management, however, there is evidence of a more aggressive local policy stance toward cost recovery. In Fortaleza, Natal, Almirante Tamandar6 and Colorado, for instance, IPTU (property) tax is levied on some of the areas improved under the projects if not for the first time, then at a higher rate than previously. The collection of user charges for street lighting and cleaning is now routine and rigorous in the cities of Petrolina, Aracaji and Almirante Tamandar. But in none of the cities is betterment tax charged to beneficiaries of the areas improved. Incentives for municipalities to increase their own revenues through these taxes may have been undermined by the provisions of the 1988 Constitution which guarantee a greater share of all public revenues for local government. 21 Participatory budgeting is devolving budgetary authority to locally elected neighborhood councils, through a process of consultation at various levels throughout the city (Peterson 1997 p. 17). 22 Suggestions include the simplification of land use legislation, dealing with land tenure problems in urban areas, policies that focus upon the existing urbanized areas of cities and measures to halt environmental and social deterioration. 39 ENABLING A BIGGER MUNICIPAL ROLE 5.13 New responsibilities for municipalities: The projects succeeded not only in helping municipalities manage their existing functions better, but also in enabling them to fulfill new responsibilities that came with decentralization. When the federal government canceled or curtailed many locally financed programs in response to the fiscal crisis from 1982 onwards, the projects survived, albeit with counterpart funding constraints, because of loan agreements with the Bank. At a very difficult moment for the sector in Brazil, these operations therefore allowed municipalities to demonstrate their ability to provide urban services in a more autonomous way. Although important, this impact was unintended and was achieved indirectly through the impacts upon a municipal project culture that were discussed earlier. 5.14 Decentralization providing more services: Despite slower economic growth, greater progress in urban service and housing provision was achieved in Brazil during the 1980s-when municipal participation was also greater-than during the 1970s. This is consistent with the paradigm of demand efficiency of decentralized provision as reported in the literature, which contends that local government can respond better to local preferences than central government (Dillinger 1995. pp. 30-31). On the other hand, basic service needs of poor areas of Brazilian cities are so obvious that a precise validation of demand may not be necessary.23 The study cannot establish the validity of the pros and cons of decentralization in these terms, but does note that the projects demonstrated the municipalities' ability to be effective suppliers of urban services in the new decentralized arrangements adopted in Brazil. 5.15 Countrywide impacts: There are significant ID impacts in both the poorer and the richer regions of Brazil and also in larger and smaller cities. Despite considerable income disparities between the Northeast and Southeast where the share of the poor among urban households is 42.6% and 18.8% respectively, Northeastern Natal and Southeastern Juiz de Fora are both able to operate successful urban institutes that were established under the projects. Even small project cities such as Propri. (pop. 21,705) and Colorado (pop. 15,725) continue to benefit from in-house project preparation and evaluation capacity established the first time, thanks to the projects. 23 This kind of situation led one critic of decentralization to argue that: "In most developing countries, the problem (of demand efficiency) is not to reveal the fine differences in preferences between jurisdictions but to satisfy the basic needs, which are, at least in principle, well known, and need not be revealed. The potential welfare gains (of decentralization) associated with a better match of supply and demand are not large" (Prud'homme 1993. p.7). 40 DECENTRALIZATION AS A FACTOR OF PERFORMANCE 5.16 Dimensions of ownership: Decentralization also gave municipalities the opportunity to demonstrate stronger ownership than was originally expected for these federally inspired operations. This section reviews project ownership in terms of the four usual dimensions: (i) locus of initiative--whether local or elsewhere; (ii) level of intellectual conviction of local policy makers; (iii) expression of political will by local leaders; and (iv) efforts towards consensus building among stakeholders (Johnson and Wasty 1993). 5.17 Willingness to assume ownership: The varied ways that municipalities took over ownership depended more on local political circumstances than project design. Local governments were rarely the originators of the ideas behind the projects. The study therefore asks how ownership developed at the municipal level, following the withdrawal of the federal government, the original owner. Study findings indicate that municipalities are willing and able to assume a strong commitment to project ideas that are not theirs, perhaps because the projects addressed basic needs for low-income populations, needs which were understood and unquestioned by local officials. 5.18 Consensus building: At the municipal level, the projects fostered consensus building efforts in six of the cities visited (Table 5.2). In four of these-Fortaleza, Petrolina, Juazeiro and Juiz de Fora-intellectual conviction was first forged at the technical level through small but cohesive professional project teams. Written comments on an earlier draft of this report from Juiz de Fora confirmed that project teams had to make great efforts to win over not only local politicians, but also the staff of traditional municipal departments who felt they were bypassed by the project. The team in Juazeiro, on the other hand, included the city mayor himself. With a united front, teams such as these would try to convince their local politicians of the merits of their project. Especially in large cities prone to local political conflicts, the forging of such a technocratic conviction is key to ensuring the successful implementation of such programs. In remaining cities-all but one in the Northeast region-consensus building was not considered necessary by key informants. In these cases, there already existed what the literature calls a "dominant coalition" of powerful city mayors and their allies.24 Natal, nevertheless, tried to build consensus with local communities by setting up a series of district councils to oversee urban investments in different parts of the city. 5.19 Transfer between administrations: During implementation, responsibility for the projects was transferred between succeeding municipal administrations three times, in 1982, 1985 and 1989. Most cities report smooth transfers of project files and databases, and the retention of key project staff 24 This was the case of Natal and AracajOi where powerful city mayors could effectively impose solutions without having to convince their weak adversaries. In the case of Salvador, there was conflict between city and state level governments, but both parties shared a conviction about the need for the project and worked to ensure that it was completed as planned. 41 (Table 5.2). Only three of these cases, however-Propri6, Almirante Tamandar6 and Juiz de Fora-involved political adversaries. Evidently, smooth transfers between adversaries are a sign of strong local ownership. On the other hand, conflictive transfers occurred in the three metropolitan cities of the Northeast as well as Juazeiro. No conflict led to a complete breakdown of dialogue over the projects. More typically, conflict built upon incoming authorities' distrust of their predecessors and questioning of basic project objectives-initially treated as a compact between the Bank by their adversaries. For their part, outgoing authorities might withhold project information from their successors. Project technical teams can be disbanded at the initiative of either the incoming or outgoing administration. Conflictive transfers of this kind in Recife delayed project implementation by at least six months. Table 5.2 Indicators of Project Ownership by Municipalities Consensus building Transfers between Origin of project ideas efforts administrations Fortaleza Yes Conflictive Federal government (through project technical team) Recife None Conflictive Federal govemment Salvador None Conflictive Federal government (conflict between state and municipality) *Petrolina Yes Smooth Federal government (technical team (same political group) convinced politicians) Juazeiro Yes Conflictive Local (led by city mayor who was an engineer) Natal None Smooth Federal govemment (powerful city mayor and (same political group) dominant coalition) *Aracajo None Smooth Federal government (powerful city mayor and (same political group) dominant coalition) *PropRi - Yes Smooth Local (through the city council) (between political adversaries) *Almirante Tamandar6 Yes Smooth Local (with the assistance of (between political state_government) adversaries) *colorado None Smooth Local (unnecessary: existence (same political group) of dominant coalition) ruiz do Fora YesSmooth Local (through project (between political technical team) adversaries) Note: * best practice cases. Source: Impact Evaluation Key Informant Interviews 1995-96 42 5.20 Origin of project ideas: Six cities in the Northeast region correctly report that the project concept came from the federal government.25 By contrast, key informants from cities in the southern regions consistently confirmed local participation in the formulation of the project ideas. They understood that the lending operations allowed their own ideas to be implemented. The lack of local origin to the project ideas in the Northeast did not make municipalities there less committed to the projects themselves. Local political leaders there are willing to commit themselves to the projects of "neutral" third parties such as the federal government or the Bank, but they can be hostile to implementing those associated with their local adversaries. 25 The two Northeastern cities reporting the local origin of the project ideas-Juazeiro and Proprid--are both remote from their respective state capitals and their officials had less contact with outsiders involved in project preparation. 43  6. Poor Residents Learn the Rules of the Game The projects had important side effects in strengthening community- based organizations (CBOs), thereby enabling residents to learn about urban service provision. As convenient interlocutors for local government and useful monitors of project progress, CBOs became empowered through the projects. Women played an important role in this. None of the local communities knew of the projects during their planning phase, so that learning the rules of urban service provision occurred during implementation. City councilors played a big participatory role in this. Residents' newly acquired knowledge pointed to municipal government as being ultimately responsible for urban service provision. In this way, the projects indirectly encouraged residents to hold municipalities accountable for these services with little incentive for self-help efforts. Nevertheless, revitalized community participation in Brazil augurs well for future project performance and sustainability, through increased project control and commitment by primary stakeholders, local residents themselves. FRAMEWORK OF EVALUATION 6.1 A project side effect: Was community participation a factor in enhancing project impacts, or a side effect of the projects themselves? Participation was not a specific design feature of these projects, nor was such participation a formal Bank requirement for appraising these operations. Since then, however, Bank sponsored studies have strongly recommended including the participation of final beneficiaries in all stages of project development to improve performance and enhance commitment. In this way, greater community participation can lead to the long-term sustainability of urban improvements. From an impact evaluation perspective, however, this study looks at evidence of participation as an unintended side-effect of the projects, rather than as an input to project performance itself. 6.2 Participation as a process: This study follows the World Bank Participation Sourcebook in treating participation as "a process through which stakeholders influence and share control over development initiatives and the decisions and resources which affect them." (World Bank 1996 p. xi). For the IER, stakeholders are mostly local residents organized as a community-based organization (CBO). Their participation is voluntary. Both the theoretical 26 See the following references, for example: Baum 1985 pp. 479-480; World Bank 1992 pp. 93-97; World Bank 1994a p.71; Sch0beler 1996; World Bank 1996 pp. 3-8. 45 framework and systematic empirical evidence of likely development benefits from beneficiary participation are weak (Piccioto 1996. p. 5). Sometimes participation is still treated-erroneously in the view of this study--as an end in itself rather than a means to address the complex issues service provision to low income communities. 6.3 Participatory research: The study incorporates participatory research into its design. From its experience of the direct contacts made with beneficiary communities in all the cities visited by the evaluation team, OED endorses the conclusions of a recent Bank-sponsored study that this kind of research allows: "the field worker to gain access (often for the first time) to a wealth of facts, perspectives and insights from the community's and the staffs own store of wisdom and experience." (Narayan 1996 pp. 141-2). Despite their varied names27, CBOs contacted by this study have similar structures organized bottom-up and community leaders elected by their members. Most hold regular meetings to discuss collective problems and aspirations about their neighborhood, such as street paving and local bus services. 6.4 Key concepts used: The study regards urban service provision as having many characteristics of a public good, whose benefits are available jointly to a large number of people and from whom individuals cannot be excluded. It uses Hirschman's concepts of exit and voice. Exit is understood as the deliberate withdrawal of residents from their relationship with a development process. Voice is seen as a direct and informative way of alerting (usually senior) partners in that relationship to problems that arise (Hirschman 1986. pp. 78-79). The study also refers to empowerment, as the greater use of the beneficiary's voice in decision-making processes (Sch0beler 1996. p.34). STRENGTHENING COMMUNITY BASED ORGANIZATIONS 6.5 Prior weakness: Before the projects, CBOs existed in most of the areas visited but they were very weak and inactive according to local residents interviewed by the team. Residents complained that pre-project CBOs had no agenda of activities or regular meetings and that leadership was volatile and untrustworthy. In these unimproved neighborhoods there was a lack of support for CBOs. Except every four years at election time, local government tended to ignore them. The sorry state of CBOs echoes earlier studies of other countries, which found that both authorities and residents themselves gave less importance to community organizations of urban beneficiaries of projects, who act merely as consumers of services. This contrasts with the closer attention generally given to extension and other community groups in rural areas, regarded by officials and project managers as organizations of producers (Moser 1987). 27 Among those met were: Clube de Maes e Amigos de Alto de Cruzeiro (Juazeiro); Associagdo Comunitaria de Corrego de Areia (Recife); and Associagao de Proprietarios, Amigos e Moradores do Distrito Colonia Lamenha (Almirante Tamandar6, Paran6). 46 6.6 Sudden strength: Project implementation represented a sharp break with past neglect and helped revive and strengthen CBOs in these neighborhoods. Engineers, workmen and community workers from local government and contractors had to establish an "official" presence in these areas for the first time. Residents from Fortaleza, for instance, recall the arrival of engineers-whether from the municipality or a private firm, they did not know-inquiring about the best place to establish a site depot to safely store their equipment and materials. First contacts such as these are naturally directed toward existing CBOs.28 6.7 Empowerment: Local residents see empowered CBOs as a forum to discuss interests affecting the neighborhood as a whole and an instrument to pressure relevant authorities for solutions that satisfy these interests. Residents in Petrolina and Natal were clear about the sudden usefulness of their local CBO once the projects were underway. They noted that CBO meetings became more regular, better attended and most importantly, "there was something to talk about at last". In Almirante Tamandard, local residents explained how the neighborhood CBO came to life with the project works. Residents in Salvador explained how CBOs now act more like pressure groups seeking improvements to their areas. Unintended impacts of these projects are therefore to encourage esteem for and confer credibility upon CBOs. This is a way for local residents to find a voice not only in the implementation of the project, but in many other decisions affecting their area. 6.8 Modest impact in metropolitan cities and small towns: Project side- effects affecting CBOs were less strongly felt at the two ends of the spectrum of urban demographics in Brazil, namely large metropolitan areas and small towns. Metropolitan scale administrations are inevitably complex with numerous potential points of community:government contact, making a dialogue with CBOs a complicated process. Residents of peripheral areas located 10-12 kilometers from respective city halls in Recife and Fortaleza understandably felt their voice to be remote from centers of decision.29 In Salvador, where the project neighborhood was located less than four kilometers from the city center, local residents felt less alienated. CBOs do not thrive in the small towns of Colorado30 (pop. 15,600) and Proprit (pop. 21,900), but for different reasons. In both places, 28 Local residents esteem for their CBO in Petrolina grew when engineers first came to a community meeting to seek residents' views about the best sequence of streets to follow in improving pavement in the area. 29 An added difficulty in Fortaleza is the peculiar location of the project site across an unresolved border demarcation between the neighboring municipalities of Fortaleza and Caucaia, so that neither administration assumes full responsibility for the area. This remains an important institutional bottleneck that, in hindsight, was not properly addressed by the project. as The project area in Colorado was the only one visited that did not have a specific neighborhood name. Residents there found that many of the participation issues raised by the team to be fascinating, but not relevant to their area or town. 47 residents did not see the need for a formal community organization, when the access by individuals to local officials was easy. CONTRIBUTION OF WOMEN 6.9 Women's predominant role: Women play the principal role in strengthening community organizations in their areas. In all group discussions with beneficiary groups, the leading participants were women, even though men were often present. The women were generally more able and willing to provide information, particularly about the project impacts upon family finances after extra local tax liabilities, for instance. In Juazeiro, Salvador and Almirante Tamandar6, working women-often the main breadwinners where unemployment among men is high-particularly value child-care services provided by many CBOs after the projects. 6.10 Men's reticence: Many local men were openly skeptical about the usefulness of their involvement with CBOs. In the Northeast some men informed the evaluation team that community participation meetings were essentially a "women's business". This conclusion appears to come from the discomfort some men said they felt about questioning that goes on in such meetings, tantamount to prying for some of them. This was especially true in connection with questions about family income that come up when meetings tried to fix the prices of CBO services and subscriptions at affordable levels for the members, for instance. This view was shared by women. In Fortaleza, women affirmed moreover that men were uninterested in home and neighborhood improvements. For these reasons, women take the lead and assumed financial responsibility for involvement in CBOs and become the principal force behind the empowerment and voice of community organizations. LEARNING THE RULES OF URBAN SERVICE PROVISION 6.11 First steps: Searching for information from local government about urban service provision through CBOs was at the heart of the early stages of participatory development that in the project areas. Typically in the cities visited, CBO leaders sought more information directly from city hall as soon as they learned about the projects, generally only at start-up. Through local meetings called by community leaders, residents discovered more details about the projects themselves. In this way, they learned about the rules and responsibilities of urban service provision in Brazil through a process of de facto civic education instigated by their CBO. Neither residents nor local officials reported cases of 31 Between 1980 and 1990, the proportion of Brazilian women who were economically active rose from 26.6% to 35.5%, giving women a much bigger economic role in all regions of the country. Childcare services are an incentive for women to support CBOs. Without them, many women would be unable to work. 48 significant changes in project design or execution, however, as a result of community participation and consultation. 6.12 Moments of learning: In none of the cities visited did longer-term residents know about the projects during the planning stage (Table 6.1). In Aracaji, Almirante Tamandar6, Colorado and Juiz de Fora, beneficiaries told the evaluation team that they only learned about the projects when laborers and equipment moved on-site to begin work.32 Of course, the projects were planned under Brazil's military government, when community consultation was not considered a priority. This meant that most learning for residents took place during the implementation phase of the projects. 6.13 Means of learning: Apart from meetings, a remarkably wide range of means facilitate the flow of information between the community and local government (Table 6.1). The most common method, used in eight of the cities visited--all but Recife, Petrolina and Fortaleza--is through the participation of the formal representatives of the residents, elected vereadores as city councilors are known in Brazil.33 Democratization in Brazil in the 1980s favored a more participatory approach by local politicians. CBOs themselves, as they channel project information between residents and local governments, are an equally important instrument of communication.34 In six cities residents cited election campaigns as key moments of participation. Other means, such as newspapers, fliers, loudspeaker vans and opinion polls were also used, but more sporadically. The study found no evidence, from a beneficiary perspective, of the use of non- governmental organizations (NGOs) or TV and radio as means of local government:community communication over urban improvements.3536 6.14 Key things learned: By following project progress through their CBOs, residents for the first time learned details about: 32 Interestingly, municipal officials consistently reported more participatory activities than local residents did in all cities visited. In Aracaji and Recife, they referred to major public meetings in project areas which longer-term residents could not recall having taken place. 3 Strengthening CBOs does not appear to have a substitute effect with respect to traditional electoral responsibilities. On the contrary, solid CBOs become new instruments of articulation between political representative and their electors. At the same time democracy and local elections-held in the state capital cities for the first time in 1984--encourage CBO activities, especially during election campaigns. The IES learned of cases where vereadores began their political careers through the activities of CBOs. 3 The approach taken by CBOs toward local government is generally not conflictive. In no city visited did the evaluation team find evidence of a breakdown of communications, resulting in protests, strikes or other demonstrations by local residents. aS Fortaleza hosts an important experiment in house-building and improvement through an association of a Brazilian and a French NGO and the city government that involves significant community participation. Infrastructure and service provision, however, is through conventional municipal and utility channels. 36 Bank guidelines called for NGO participation in project preparation only from 1988 onwards (OMS no. 5.30), after these projects were appraised. 49 * urban service responsibilities of municipal, state and federal governments. * departmental responsibilities at the municipal level. * the relationship between private contractors and municipalities. * works implementation and materials used. * responsibilities for maintenance and operation. It was especially for local residents to know where and to whom among officialdom to address their claims for further attention. Thus residents learned the rules of urban service provision through a de facto process of civic education 37 among beneficiaries. HOLDING LOCAL GOVERNMENT ACCOUNTABLE 6.15 Community claims: Having learned about municipal responsibilities, local residents held local governments accountable for urban service provision to their areas. Some felt their neighborhood had as much a right to these services as wealthier serviced parts of the city. Community claims for improved services figured in elections for councilors (vereadores) and city mayor. Candidates would generally trade promises for better services in return for the votes of residents. In group discussions with the team, more vociferous residents were able to articulate their claims for better services clearly. Their frequent use of project terms in discussions-such as procurement and maintenance, for instance- indicates that they acquired, or at least consolidated, this skill through the project experiences. 6.16 Lack of incentive for self-help: Placing urban service responsibility on the shoulders of municipal government provides no incentive for communities themselves to seek self-help solutions-called mutir.o in Brazil-for improving their urban living conditions. Local residents who had gained service improvements, after all, did so without any direct inputs of their own. For these reasons, the study found no evidence of self-help provision of urban services stimulated by the projects. Mutirio solutions are applied in Fortaleza, but principally for the provision of private housing goods. Infrastructure in these cases remains a local government responsibility. 6.17 Contrast with Indonesia: In this way, the Brazilian experience contrasts with community provision of infrastructure-for which residents provide their labor free-in low income areas of cities in Indonesia, known as kampungs (World Bank 1995. pp. 52-61). In addition to the projects' own lack of incentives, 37 A major study of 800 households in Bangalore, India, concluded that an understanding of the rules of the game by beneficiaries is a key toward better service levels (World Bank 1994a. p. 71). ' The relationship between promises of concrete results and political allegiance at the local level is called clientelism in Brazil (Gay 1990 p. 657). 50 self-help do not flourish in higher income Brazil where urban improvements are larger scale.39 Also Brazil does not have Indonesia's intricate hierarchy of government that includes neighborhood and block units of local government which are primarily responsible for organizing self-help efforts in that country. 3 A CBO approach to initiating cooperative self-help development activities is "applicable in the most poorly served, marginal low-income settlements" (Schobeler 1996. p. 49). This criterion would exclude project areas covered by this IER. Although poor and inadequately served before the projects, they nevertheless had regular street and subdivision patterns and were accepted as parts of their respective cities. Table 6.1 Indicators of Participation - community perspective Timing: Means: City Planning; Execution City City CBOs NGOs Election TVlradio Newspapers Filers Loudspeaker Opinion - Project area S * stage mayor councilors campaigns vans polls Fortaleza -ParqueP ** P . 0 * -------- -+----------------I------------------------------------------- F------ -Case Amarela Salvador* -Anton. Balbino -Pero-n ---I---.-- .--- I - -- -- -- -- --1- - -- -- -- -- -- --- - woluc* -Ato Cruzeiro -rkiii --------------F-------F-----------------------F--------------F-------- 4------- -uintas * 0 1* 0 -Wi ------+ ------------I------- F---F ---- i------- F-----I------------I--------I1------- -soore* :* :* * ~------- *Proprf6-- - -- - -B.Am6dca * * *Amirante -- ---------------- -------------------------------- Tamandar6 - Lamenha G. *Colorado - Colorado * * *Juizde Fora ---------------F------ -------- ----------------------- -Sta Efgdnia Note: Blank cells indicate that there was no evidence of timing or means of participation. Source: Impact Evaluation Study Group Interviews with Beneficiaries and Key Informant Interviews 52 BIBLIOGRAPHY Baum, W.C. and Tolbert, Stokes M. 1985. Investing in Development: Lessons of World Bank Experience. The World Bank, Oxford University Press, New York Brinkerhoff, D.W. 1994. 'Institutional Development in World Bank projects: analytic approaches and intervention designs' in: Public Administration and Development. Volume 14, Issue No. 2, May 1994. Cernea, Michael M. 1988. Nongovernmental Organizations and Local Development. World Bank Discussion Paper No. 40, Washington D.C. De G6es, Paulo 1996. Rural Urban and Regional Migration in Brazil. Background paper commissioned for this study, Rio de Janeiro. Dillinger, William 1994. Decentralization and its Implications for Urban Service Delivery. Urban Management Program, World Bank, Washington D.C. Dillinger, William 1995. Better Urban Services: Finding the Right Incentives. A World Bank Publication, Washington D.C. Estache, Antonio (ed.) 1995. Decentralizing Infrastructure: Advantages and Limitations. World Bank Discussion Paper no. 290, Washington D.C. Gay, Robert 1990. 'Community Organization and Clientelist Politics in contemporary Brazil: a case study from suburban Rio de Janeiro' in: International Journal of Urban and Regional Science. Vol. 14, pp. 648-666. Heaver, R. and A. Israel Country Commitment to Development Projects. World Bank Discussion Paper No. 4, Washington D.C. Hicks, James F. 1992. Decentralization and Reform of the State: Issues and Options for Bank Assistance Strategy. EMENA Technical Department Regional Discussion Paper, World Bank, Washington D.C. Hirschman, Albert 0. 1986. "Exit and Voice: an Expanding Sphere of Influence" in: Albert 0. Hirschman Rival Views of Market Society and Other Recent Essays. Harvard University Press, Cambridge. Inter-American Development Bank 1994. 'Fiscal Decentralization: the Search for Equity and Efficiency' in: Economic and Social Progress in Latin America - 1994 Report. IDB, Washington D.C. Israel, Arturo 1987. Institutional Development: Incentives to Performance. John Hopkins Press for the World Bank, Baltimore. 53 Johnson, John H. and Sulaiman S. Wasty, 1993 "Borrower Ownership of Adjustment Programs and the Political Economy of Reform" World Bank, Washington D.C. draft. Kessides, Christine 1993. Institutional Options for the Provision of Infrastructure. World Bank Discussion Paper no. 212, Washington D.C. Morra, Linda and Patrick Grasso Draft Guidelines for Impact Evaluation Studies. World Bank Operations Evaluation Department, Washington D.C. Moser, Caroline 1987. "Approaches to Community Participation in Urban Development Programs in Third World Countries." in Michael Bamberger (ed) Readings in Community Participation. Economic Development Institute of the World Bank, Washington D.C. Musgrave, R. and P. Musgrave 1984. Public Finance in Theory and Practice. McGraw Hill, New York. Narayan, Deepa 1996. Toward Participatory Research. World Bank Technical Paper No. 307, Washington D.C. Peterson, George E. 1997. Decentralization in Latin America: Learning through Experience World Bank Latin America and Caribbean Studies Viewpoint, Washington D.C. Piccioto, Robert 1992. Participatory Development Myths and Dilemnas World Bank Policy Research Working Papers, Washington D.C. Prud'homme, Remy 1994. On the Dangers of Decentralization. TWURD Policy Research Paper, Washington D.C. Rocha, Sonia 1996. On Urban Poverty: Policy Experiences in Brazil in the 80's. Background paper commissioned for this study, Rio de Janeiro Rondinelli, D. 1990. Decentralizing Urban Development Programs: A Framework for Analyzing Policy Options. U.S. Agency for International Development, Office of Housing. Washington D.C. Sch0beler, Peter 1996. Participation and Partnership in Urban Infrastructure Management. UNDP/UNCHS/World Bank Urban Management Program. World Bank, Washington D.C. World Bank, 1983. World Development Report: Management in Development. Oxford University Press, New York. World Bank, 1991. Brazil: Medium-Term Strategy for the Infrastructure Sectors. World Bank, Washington D.C. World Bank, 1992. World Development Report: Development and the Environment. Oxford University Press, New York. 54 World Bank, 1992. Brazil: The Challenge of Municipal Development in the 1990s. World Bank, Washington D.C. World Bank, 1993. Annual Review of Evaluation Results 1992. World Bank Operations Evaluation Department Report 12403. Washington D.C. World Bank, 1994(a). World Development Report Infrastructure for Development. Oxford University Press, New York. World Bank, 1994(b). Twenty Years of Lending for Urban Development 1972- 92: An OED Review. World Bank Operations Evaluation Department Report No. 13117, Washington D.C. World Bank, 1995. Indonesia Impact Evaluation Report Enhancing the Quality of Life in Urban Indonesia: The Legacy of the Kampung Improvement Program. World Bank Operations Evaluation Department Report No. 14747-IND, Washington D.C. World Bank, 1996. The World Bank Participation Sourcebook. World Bank Environmentally Sustainable Development Publication, Washington D.C. World Resources, 1996. World Resources 1996-97: A Guide to the Global Environment Washington D.C. 55  Annex 1 Table A.1 Brazil 1975-96: Benchmark Urban Sector Events Non-Lending Activities Lending Activities (details in Table A.3) (details inTable A.4) Calendar Year Government Policy Bank Policy (Bankwide) Bank Policy (Brazil) Bank financed projects (details in Table A.2) 1973 National Low-Income Housing Program The Bank was involved in two key (Piano Nacional da Habitacao Popular, aspects: (1) provision of badly needed PLANHAP I). services and employment opportunities to the urban poor; and (2) strengthening Nstbined. iof the institutions at the federal and local level. 1974 National Commission on Metropolitan Regions and Urban Policy (CNPU) created 1975-79 Second National Development Plan (II PND). 1975 National Low-income Housing Program II Housing Sector Policy Paper (PIano Nacional da Habitacao Popular, PLANHAP II). Urban Transport Policy Paper National Urban Development Fund (FNDU) 1976 National urban transport agency (EBTU) created. National urban development fund (UDF) created. Medium-Sized Cities program (MSC) began operations. 1977 1978 1979 CNPU Is abolished and replaced by the Sites and Services project approved National Council on Urban Development (Ln. 1654) 57 Annex I Table A.1 Brazil 1975-96: Benchmark Urban Sector Events Non-Lending Activities Lending Activities (details in Table A.3) (details In Table A.4) Calendar Year Government Policy Bank Policy (Bankwide) Bank Policy (Brazil) Bank financed projects (details in Table A.2) (CNDU) and an Undersecretariat of Urban Medium-sized Cities project approved Development (SDU). (Ln. 1720) Law 6766 on urban land subdivision 1980-85 Third National Development Plan (III PND). 1980 Law 6803 on urban Industrial land use Shelter Policy Paper 1981 1982 Recife Metro Dev Proj approved (Ln.2170) NE Metropolitan Development project approved (Ln. 2193) 1983 Learning by Doing Urban Development in LAC in the 1980s Parana Market Towns project approved (Ln. 2343) 1984 Sites and Services (Ln. 1654) closes 1985 CNDU abolished Urban Strategy Paper NE Metropolitan Development project Ministry of Urban Development and closes (Ln. 2193) Environment (MDU) created Santa Catarina Towns project approved (Ln. 2623) NE Flooding project approved (Ln. 2645) 1986 EBTU abolished; BNH abolished Urban Transport Policy Paper Medium-sized Cities (Ln. 1720) closes Caixa Economica Federal (CEF) assumed Salvador Metro Development project 58 Annex I Table A.1 Brazil 1975-96: Benchmark Urban Sector Events Non-Lending Activities Lending Activities (details in Table A.3) (details in Table A.4) Calendar Year Government Policy Bank Policy (Bankwide) Bank Policy (Brazil) Bank financed projects (details in Table A.2) responsibilities of BNH. approved (Ln. 2681) 1987 MDU became MHU (min. Urban Development in the Northeast housing/urbanism) 1988 MHU responsibilities curtailed as it Parana Market Towns project closes became Ministry of social welfare ministry (Ln. 2343). (MBES) NE Flooding project closes (Ln. 2645) National sanitation program (PLANASA) Rio moved from MBES to MOH (ministry of ergen proodpng health).2975) New Constitution put municipalities solely in charge of local urban development policy. 1989 Ministry of social welfare (MBES) Recife Metro Dev Project closes abolished (Ln.2170) Paran;§ Municipal Development project approved (Ln. 3100) Rio Grande do Sul Municipal Development project approved (Ln. 3129) 1990 Ministry of social action (MAS) created 1991 Urban Policy Paper Medium-term Strategy for the infrastructure Sectors 1992 Abolition of Ministry of Social Action (MAS) Challenge of Municipal Sector and of Secretariat of Regional Development 59 Annex I Table A.1 Brazil 1975-96: Benchmark Urban Sector Events Non-Lending Activities Lending Activities (details in Table A.3) (details inTable A.4) Calendar Year Government Policy Bank Policy (Bankwide) Bank Policy (Brazil) Bank financed projects (details In Table A.2) Development. Creation of Ministry of Social Welfare (MBES) with Urban Policy Responsibilties. 1993 Housing: Enabling Markets to Work Santa Catarina Towns project closes (Ln. 2623). Minas Gerais Municipal Development project approved (Ln. 3639) 1994 Ceart Urban Development/Water Company project approved (Ln.3789) 1995 Ministry of Social Welfare (MBES) Paran§ Municipal Development project abolished closes (Ln. 3100) Creation of Ministry of Planning and Rio Emergency Flooding Budget; its responsibilities included Urban 7construction project closes (Ln. Policy. Rio Grande do Sul Municipal Development project closes (Ln. 3129) 1996 Bahia Municipal Development Project approved (Ln.4140) 60 Annex I Table A.2 Brazil 1975-96: Government Policy relating to the Urban Sector Calendar Government policy approach Government policy recommendations Macro-economic events Year Population Indicators 1973 National Low-income Housing Program (Piano Nacional The Metropolitan Regions Law required the creation, by da Habitacao Popular, PLANHAP 1). respective states, of Advisory and Deliberative Councils Firs eiht etroolian ere reaed:Sao (DCs). The DCs had the responsibilities to prepare First eight metropolitan regions wereintegrated metropolitan regional development plans and Paulo, Belo Horizonte, Porto Alegre, Curitiba, Recife, programs and coordinate and execute programs and Salvador, Fortaleza and Belem. Rio de Janeiro was projects of metropolitan scope, including common added in 1974. services (integrated development planning, basic sanitation, metropolitan land use, transportation and the road system, the production and distribution of piped gas, and water resource use and pollution control). The law transferred former federal responsibilities of ensuring common services to the states. 1974 National Commission on Metropolitan Regions and The GOB introduced a fiscal incentive to all Urban Policy (CNPU) was created: inter-ministerial beneficiaries of SFH in the form of a rebate equal to the agency with advisory and prescriptive functions and average 12% of the beneficiary's actual loan responsibility for the coordination of national urban repayments in the previous fiscal year and credited development policy. against payments in the current year ("fiscal benefit"). 1975-79 Second National Development Plan (II PND). The urban development strategy formulated in the II PND was based on the proposition that recent urbanization trends (i.e., the rapid metropolization of Rio de Janeiro and Sao Paulo) had introduced serious distortions in urban structure and spatial organization. Two lines of policy were introduced: (1) the system of metropolitan areas was consolidated and measures were taken to prevent further concentration of population and economic activity in the SE region; and (2) medium-sized cities in all regions were promoted to reinforce an incipient trend that became evident in the 1960s by strengthening the second rank of the urban structure and supporting the development of the interior. Decision making and allocation of resources affecting local development remained highly centralized at the 61 Annex I Table A.2 Brazil 1975-96: Government Policy relating to the Urban Sector Calendar Government policy approach Government policy recommendations Macro-economic events Year: Population Indicators federal government level, to the detriment of the states and municipalities. 1975 National Low-income Housing Program it (Plano Nacional da Habitacao Popular, PLANHAP II). National Urban Development Fund (FNDU) 1976 National urban transport agency (EBTU) created. National urban development fund (UDF) created. Medium-Sized Cities program (MSC) began operations. 1977 1978 1979 CNPU is terminated and replaced by the National CNPU was terminated with the change in government Council on Urban Development (CNDU) and an and replaced by the National Council on Urban Undersecretariat of Urban Development (SDU). Development (CNDU). In addition to CNPU responsibilities, CNDU proposed annual and multi- Law 6766 on urban land subdivision annual urban development investment programs, together with the annual program for all state funds destined for urban development. The Executive Secretary of CNDU was also transferred to the Minister of Interior, where an Undersecretariat of Urban Development (SDU) was created in May 1979. SDU run the MSC. 1980-85 Third National Development Plan (1l1 PND). 1980 Share of urban population: 68% (82 million). During the 1980s there was a decreased in the urban population of large metropolitan areas. Mid-sized cities were receiving ("back migration") the population that left earlier and now was returning given the bad economic 62 Annex I Table A.2 Brazil 1975-96: Government Policy relating to the Urban Sector Calendar Government policy approach Government policy recommendations Macro-economic events Year: Population Indicators situation of the large metro areas. 1981 Rapidly declining of the national economic situation amongst other reasons as a result of the second petroleum price shock, that increased local pressures for the decentralization of investment decisions and sharply curtailed the resources available to the federal government to pursue urban development objectives. End of centralization. 1982 Direct elections of Govemors. 1983 1984 Direct elections of Mayors. 1985 Civilian administration took over the central govemment. CNDU abolished Ministry of Urban Development and Environment (MDU) was created to which BNH, the Executive Secretariat of CNDU, EBTU and SDU were transferred. 1986 EBTU abolished; BNH abolished Caixa Economica Federal (CEF) assumed responsibilities of BNH. 1987 MDU became MHU (min. housing/urbanism) 1988 MHU responsibilities curtailed as it became Ministry of social welfare ministry (MBES) National sanitation program (PLANASA) moved from MBES to MOH (ministry of health). New Constitution put municipalities solely in charge of 63 Annex I Table A.2 Brazil 1975-96: Government Policy relating to the Urban Sector Calendar Government policy approach Government policy recommendations Macro-economic events Year: Population Indicators local urban development policy. 1989 1990 integrated Regional Planning was put aside and several Share of urban population: 75% (82 million) of the regional planning agencies were closed. The municipalities were strengthened by letting them keep some of the taxes collected locally. Ministry of social action (MAS) created 1991 1992 Ministry of Social Welfare (MBES) abolished Creation of Ministry of Planning and Budget with responsibilities for Urban Policy. 1993 1994 1995 Ministry of Social Welfare (MBES) abolished Creation of Ministry of Planning and Budget; its responsibilities included Urban Policy 1996 64 Annex 1 Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1975 Housing Sector Policy Paper call for * sites and services projects *Late 70s: lower affordable standards * squatter upgrading projects support for GOB efforts to increase the productivity and Improve the * operations to be small demonstration living conditions of the poorest * support to housing finance institutions segments of the population, both rural and urban; *the promotion of institutional development policy reforms In order to rationalize resource allocation and enhance the coordination and control of public investments. ease foreign exchange constraint on Brazilian development by supporting projects to expand export capacity and promote domestic Import substitution and to provide part of the medium and long-term capital required by the country. Urban Transport Policy Paper call for * bus and rail transport solutions more attention to public transport solutions affordable by the urban pooe * cyclist and pedestrian facilities * lower cost traffic engineering solutions 1976 1977 1978 65 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year 1979 1980 Shelter Policy Paper emphasizing * priority for the urban poor shelter as a basic need of the urban poor. * programs to be affordable with cost recovery and community participation 1981 1982 1983 Learning by Doing focusing upon * efficient and equitable provision of Uan Development in LAC In the (only those specifically referring to inadequate urban services, Infrastructure urban services 1980s calling for Bank support in all Brazil) and institutional capacity for economic countries in the region (including Brazil) growth * Institutional strengtheningfor (I) stronger city and metropolitan r n anen ao * help private housing markets to management; (ii) better urban services work and Infrastructure for urban enterprises 0 help develop metro areas and and households; (ll) financially sound cities (management systems. * Bank to be wholesaler, not retailer housing programs for low Income services, production and shelter) of urban development finance families * develop national policy framework for housing, making existing programs and instruments more efficient and equitable a develop sound policies for urban transporth emphasizing city efficiency 1984 66 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1985 Urban Strategy Paper calling for urban * operations combining macro- policy: (I) with both macro-economic and economic objectives and inter- equity objectives; (ii) based upon states personal equity In urban and and municipalities' own priorities for regional development policies urban infrastructure and services; (ill) with incentives for local governments to * support for sound urban transport improve their urban management policies to improve city efficiency functions. * support for sound housing policies and effective housing finance mechanisms * strengthening urban management to ensure efficient administration * greater efficiency in provision of urban infrastructure and services. 1986 Urban Transport Paper highlighting * financial viability and economic The Bank decides not to support a need for efficient urban transport, efficiency to guide policy second national MSC project and to especially for the poor. essentially pursue a state by state, * improved bus operations metropolitan and municipal * examine other low-cost altematives development approach in much of its to expensive mass transit systems. urban lending in Brazil during the late 1980s. 67 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1987 Urban Development in the Northeast * greater local calling for urban lending to support (i) autonomylaccountability in urban economic growth within the constraints development programs of economic adjustment; (ii) sound resource management; (iii) directly * more emphasis upon economic targeted programs to benefit the poor; instruments of urban development (iv) efficient project Implementation and policy procurement. * urban policy should be more related to cities' capacity to mobilize resources than to parameters of city size * urban Investments in metro areas should focus more upon economic than social aspects * preventative rather than remedial action to divert migrants from metro, areas * Integration of urban and industrial development policies 1988 1989 1990 68 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1991 Urban Policy Paper calls for economic * strengthen management of urban Medium-term strategy for the (strategic priorities for urban development view of urban housing and infrastructure, infrastructure programs Infrastructure sectors (transportation, only) taking into account large numbers of telecom, water and urban) emphasizing * regulations to encourage private urban poor and deteriorating urban * improve city-wide regulations four efficiency goals: (1) contributing to sector involvement In municipal environment * strengthen municipal capacities non-inflationary growth; (ii) greater efficiency in service provision through services * "safety net" programs for the restructuring relative roles of public and 0 improved municipal finance poorest private sectors; (ill) poverty targeted policies, training and technical * raised awareness of urban basic infrastructure services; (iv) assistance environment supporting environmental management. tariff policies for municipal services aiming at cost recovery * review municipal environmental policies and make incentives to limit pollution * analyze housing/urban land use policies 69 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1992 Challenge of Municipal Sector * federal government to lead Development focusing upon: (I) more regulating 1988 Constitution efficient municipal services by removing changes affecting municipalities bottlenecks; (ii) shifting municipal development responsibilities to States; * federal govemment to lead in (ili) transparent financial transfers to urban pollution control municipalities; (iv) incentives to force * stronger role for States in seeking municipalities to maximize revenues; (v) equity among municipalities prices of municipal services reflecting production costs; (vi) municipal services * replace negotiated grants by that meet desirable performance formula driven transfers standards; (vii) efficiency through * strengthen municipal taxation economies of scale; (viii) increase equity capacity and incentives by allocating resources to municipalities in greatest need. * improve cost recovery in pricing * single set of books for local accounting * increase private sector role in service provision (parking, public transportation, toll roads, and solid waste collection) * better Information about municipalities 70 Annex I Table A.3 Brazil 1975-96: Bank Non-Lending Activities related to the Urban Sector Calendar Strategic approach (Bankwide) Recommendations (Bankwide) Strategic approach (Brazil) Recommendations (Brazil) Year: 1993 Housing: Enabling Markets to Work * Bank to discourage governments calls for governments use market from direct housing production enabling instruments instead of direct * But large scale infrastructure, sites production of housing. and services and upgrading programs for the urban poor. * sectoral approach by the Bank * regulatory and other institutional improvements 1994 1995 1996 71  Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1975 1976 1977 1978 1979 Sites and Services and Low- * promote lowest cost elements of (a) Support for existing COHAB ratings: satisfactory overall; uncertain existing squatter upgrading PROFILURB programs in Sao Paulo, sustainability and negligible institutional Cost Housing (Loan 1654-BR) programs and encourage lower Salvador and Recife: (i)31,800 development. housing and infrastructure urbanized lots; (0) 23,800 home standards improvement loans; (iii) 19,500 embryo costs: US$460.5 m. (SAR: US$200m) insituionllystrngtienthehousing units; (iv) complementary loan: US$93 m. (SAR: US$93 in.) institutionally strengthe li e infrastructure; (v) experimental squatter COHABs (state low-income upgrading in Recife. completion: 1985 housing companies) to increase delivery capacity (b) Funding to support PROFILURB PAR findings: programs in other states. & all three COHABs experienced delays in construction programs and their financial performance was unsatisfactory. u project suffered from BNH's weakened cash flow. h project was difficult to implement with a heavy demand on Bank staff time. * nevertheless, project contributed to Bank's lending strategy in Brazil. d chances of project success are very limited with an unwilling borrower in a sector susceptible to political change. 73 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. Medium-sized Cities (Ln. 1720) * promote economic activity through (a) Technical assistance and ratings: satisfactory overall; likely Improved physical Infrastructure, Infrastructure to promote economic sustainablity and substantial institutional urban services and municipal activity principally in low-income areas. development. administration in eight strategic secondary regions. (b) Building and staffing urban and costs: US$150m (SAR: US$200m) community infrastructure in low-income strengthen Institutional capacity of areas. loan: US$70 m. (SAR: US$70 i.) agencies responsible for coordination and execution urban (c) Improvement of municipal development policies, administration PAR findings: (the above in the cities of Campina project was innovative. Grande, Flodlanopolls, Jutz de Fore, Montes Clars, Naal Pelotas, too many project components and PeroiinaJuazeio, and Vitora) executing agencies. * future efforts should rely more upon municipal goveaments. ( degree of municipal involvement correlated with project success. 1980 1981 74 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1982 Recife Metropolitan Region * providelimprove housing, urban (a) Upgrading of 695 ha. for 160,000 ratings: unsatisfactory overall; unlikely (Ln. 2170) infrastructure and services within people; sustainability; and negligible institutional Recife. (b) Metro infrastructure, Ic: (1) 11km of development. * provide metropolitan scale BR1101 highway; (ii) 110 ha. of new casts: US$215.6m (SAR: US$347.8m.) infrastructure and services to Curado urban center (11i) 130 ha. of new induce desired patterns of metro. residential development at Curado; (iv) loan: US$108 m. (SAR: US$123.9 i.) development. JaboatAo Industrial park; (v) housing completion: 1990 ee enland development of 240 ha. in * generate and enhance income,river valley; (vi) especially for the urban poor. Improvements to solid waste system. 0 apart from providing localized shelter * strengthen the planning, (c) Income creation and employment, improvements, the project failed to administrative and financial Ic: (I) TA to micro-enterprises; (ii) achieve its objectives. institutions responsible for urban productivity centers to house 3,600 and metropolitan development, micro-enterprises; (111) 10 labor * project failure was due to an over- Intermediation centers; (Iv) Infrastructure ambitious design and adverse macro- and equipment for fisheries cooperative; economic circumstances. (v) shrimp breeding station; (vi) conservation of historic monuments and provision of tourist Infrastructure. (d) Institutional development, o (i) TA to strengthen metro agencies; ( ) sanitation education program. 75 Annex 1 Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. Fortaleza/Salvador 0 prepare programs and projects for (a) Institutional development, including ratings: satisfactory overall; marginal Metropolitan Development Fortaleza and Salvador for Bank detailed investment and engineering sustainability; and partial Institutional Programs Preparation (Ln. funding. studies, TA in financial and economic development. 2193-BR) analysis for project preparation, strengthen metropolitan planningfinancial agencies in Fortaleza and management; income and employment loan: US$5.6 m. (SAR: US$8.9 m.) Salvador. generation studies. (b) Pilot civil works for upgrading in completion: 1985 Salvador, CA i an, Fortaleza and PCR findings: Caucala. a cost effective engineering project Is a valuable Instrument to test borrower ownership and commitment. S Salvador demonstrated more ownership than Fortaleza. r implementation has to be flexible to adapt and adjust to changing macro policy imperatives in Brazil. 76 Annex 1 Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1983 Parand Market Towns (Ln. * develop long term borrowing and (a) Municipal infrastructure (street ratings. satisfactory overall; likely 2343) cost recovery for municipalities. paving, rural road rehabilitation, small sustaiability and substantial institutional bridges, power connections, water development * Introduce systematic criteria for the distribution, sanitation works, storm costs: US$105.2 (SAR: US$15.3m) selection of subprojects and for the drainage, street lighting, street furniture, allocation of investment resources. landscaping and site-works). loan: US$51.9 m. (SAR US$52.7 i.) * promote, develop and apply fiscal (b) Community facilities (rural schools, completion: 1988 recovery of investment costs by kindergartens, day-care units, social recipient municipalities. centers, health posts, bus terminals, PAR lindings: * improve living conditions in small slaughterhouses, markets, sport * project exceeded physical targets and towns, through the provision of grounds, gymnasium and recreation achieved Its institutional and financial urban Infrastructure and services. centers). objectives. (c) Equipment and vehicles for road 0 project improved subproject selection maintenance equipment, school buses, and resource allocation procedures. and ambulances (d) TA to municipalities for design, it is difficult to assess project impact tendering and supervision of s, on local economic development. cadastres, real estate assessment, tax * the operation helped to improve the administration, improved accounting, quality of life in small towns and cities town planning, and preparation of zoning throughout Paran. and circulation plans. 1984 1985 Santa Catarina Small Towns strengthen managerial and financial (a) Eligible infrastructure investments in ratings: satisfactory overall; likely Improvement (Ln.2623-BR) capabilities of municipalities 190 towns with less than 50,000 sustainabiity and partial Institutional Inaghterhnts, develts,mepor * improve local government resource Inaint.dvlpe. allocation (b) Technical assistance to improve costs: US$48.7 mn. (SAR: US$70 in.) provde oca ecnomi an soial managerial capabilities of municipalities, loan: US$24.1 m. (SAR: US$24.5 in.) p eo municipal associations and state infrastructure to stimulate agencies concerned with municipal completion: 19 production and exports development and veilsfrra tenderng andsuperRsifindiworks a t(to be completed -AK) 77 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. Northeast Urban Flood * rehabilitate flood-damaged (a) Shelter (I) rehabilitation of 63,000 ratings: satisfactory overall, likely Reconstruction (Ln. 2645-BR) municipalities units; (I) reconstruction of 50,000 units; sustainability and partial institutional (iiI) relocation and construction of 4,000 development strengthen planning for flood control new units; (iv) low cost sanitation and protection Improvement. (b) Urban infrastructure (i) repair, loan: US$98.6 m. (SAR - US$100 i.) improvement and extension of water completion: 1989 supply and sewerage systems; (ii) repair of streets and storm drainage; (iiI) flood PAR fndings: protection works; (iv) storm drainage * project exceeded target for works; (v) vehicles and equipment. Infrastructure. (c) Community facilities: (I) repair of flood protection works have already schools and health posts. averted major damage. (d) Micro-enterprise credits: (I) loans to 0 institutional strengthening objectives repair and replace machinery, equipment not achieved. & buildings. (e) Studies: (i) flood risk assessment; (Ii) flood management plans; (III) TA to for reconstruction and rehabilitation works; (Iv) evaluation of economic impact of reconstruction works; (v) equipment for Part E. 1986 Salvador Metropolitan n astrengthen the financial capacity of (a) Loans to finance urban infrastructure (prject nearing completion) Development (Ln. 2681-BR) metropolitan and municipal and municipal services (upgrading, agencies. access roads, solid waste collection and costs:(SAR US$151 in.) disposal facilities, maintenance program,(iii)eT suprtityc o ted e roo it productive activities).habilitation w (iv)utivit evalualoan: (SAR US$55 i c.) region. (b) Institutional development (fiscal and financial management, cadastres, D improved resource allocation, historic preservation, metropolitan * reduce the demand for federal and planning). state transfers. 1987 1988 78 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1989 Municipal Development * Improve the managerial and (a) Technical assistance to strengthen (project nearing completion) Project In Parand (Ln. 3100- Institutional capacity and fiscal urban management capabilities of BR) efficiency of municipalities and municipalities and state agencies costs:(SAR US$226.9 m.) state agencies. concerned with urban development completion: 1996 * provide basic economic and social (b) Execution of urban sub-projects loan: (SAR US$100 m.) Infrastructure in urban areas. (water, sewerage, solid waste, housing, community facilities, street paving, street * decentralize urban administration lighting, erosion and water pollution and financial management. control. * Improve targeting of urban programs to lower income population. * strengthen the State urban development fund (FDU). Municipal Development * improve the managerial and (a) Technical assistance to strengthen (project nearing completion) Project in Rio Grande do Sul institutional capacity and fiscal urban management capabilities of (Ln. 3129-BR) efficiency of municipalities and municipalities and state agencies costs:(SAR US$227 m.) state agencies. conceined with urban development. completion: 1996 * provide basic economic and social (b) Execution of urban sub-projects loan: (SAR US$100 m.) infrastructure in urban areas. (water, sewerage, solid waste, housing, community facilities, street paving, street * decentralize urban administration lighting, erosion and water pollution and financial management. control. * improve targeting of urban programs to lower income population. * strengthen the State urban development fund (FUNDOPIMES). 1990 1991 1992 79 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1993 Minas Gerais Municipal * strengthen municipal financial (a) Loans to finance TA to municipalities (project under implementation) Management and management in support of and grant financing of training programs. Environmental Infrastructure decentralization objectives. (b) Loans to finance urban Infrastructure costs:(SAR (S$331.1 m.) Ln. 3639()Lan-ofiacBubninrsrutr (Ln. 3639-BR) * improve environmental projects (water, sewerage, drainage, completion: (scheduled 1998) management. paving, street lighting). loan: (SAR US$150 m.) * study the feasibility of mobilizing (c) Loans to finance: (1) privatization pilot private resources for municipal program; and (II) micro-regional pilot services. program. * test provision of regional services through municipal associations. * transfer municipal management skills to less developed areas of Minas Gerais. 1994 Ceard Urban Development and * strengthen local government (a) institutional development through TA (project under implementation) Water Resources Management financial management and to municipalities and State agencies. costs:(SAR US$240 i.) (Ln. 3789-BR) institutional capacity of Ceard State agencies concerned with urban (b) urban Infrastructure (low-cost development and water resource sanitation, drainage, street paving and completion: (scheduled 1999) management lighting, community facilities, materials loan: (SAR US$140 m.) for self-help housing, and water supply). * improve living conditions in very poor neighborhoods of selected (c) water resources infrastructure cities component (earthen dams, water conveyance systems, rehabilitation of * increase efficiency of water use existing reservoirs). * provide safe, reliable and economic (d) pilot schemes for: (i) rehabilitating source of water to targeted cities in street children; and (ii) decentralized critical need, water management. 1995 80 Annex I Table A.4 Brazil 1975-96: Bank Lending Operations related to the Urban Sector Calendar Year: Project Objectives Description Ratings/Cost etc. 1996 Bahia Municipal Management * help Bahia state's program of (a) Institutional development (project under implementation) Project (Ln 4140-BR) reform for management of urban areas (b) urban infrastructure costs:(SAR US$222 m.) * improve the living conditions of (c) urban upgrading completion: (scheduled 2000) urban residents, primarily the poor. loan: (SAR US$100 m.) 81  ANNEX 2 Comments from the Borrower MINISTRY OF PLANNING AND BUDGET SECRETARIAT OF INTERNATIONAL AFFAIRS Fax No. 1462/97 Brasilia, May 30, 1997 For: YVES ALBOUY Chief, Infrastructure and Energy Division OED/World Bank Fax: No. 001 (202) 522-3125 From: STAEL BALTAR Acting Secretary for International Affairs of the Ministry of Planning and Budget Fax: No. (061) 225-4022 No. of pages: 7 (including cover page) NOTE: In case of a bad transmission, please call (061) 225-7185 Dear Mr. Albouy: Thank you very much for your correspondence of May 22 last, in response to our request to extend the deadline for submitting BrazilFs comments on the study to evaluate the impact of urban projects. After consulting the agencies involved in the projects, we agree with the preliminary version of the report prepared by the OED. We refer, in particular, to its more concise, logical, and objective format for presenting the results of that study, especially its conclusions and recommendations, which may be most useful in the formulation of new guidelines for the approval and implementation of future urban development projects in Brazil. Nonetheless, even with the extension of the deadline, we received comments from some agencies late; they are attached to this message, with a request that the Department also consider them in drafting the final version of the report. Sincerely, [signature] 83 ANNEX 2 ANALYSIS OF THE IMPACT EVALUATION REPORT ON URBAN DEVELOPMENT PROJECTS The actions undertaken by the World Bank evaluation project are essential to showing the value of the work done by the urban research and planning institutes in the cities where project interventions were undertaken in association with IBRD; to gauge the impact of the projects undertaken with the beneficiary communities; and to assess the ever greater responsibility that local governments are assuming for the provision of urban services to their communities. We note the following with respect to the conclusions presented in the report: 1. In general, improvements in the living conditions of the beneficiary populations were greatest in the areas targeted by the health, sanitation (water supply, refuse collection, and drainage), and road paving projects. These benefits were observed most tangibly in the areas that had more than one intervention. They were less evident in the areas that received only isolated projects. The multiplicity of problems and unmet needs in the communities involved diluted the impacts of the projects that were carried out in this fashion, as did the absence of parallel projects to sustain the official interventions, considering the scope of the planned objectives. For the purposes of evaluation, one should consider clustered and sustained interventions, i.e. the concomitant implementation of projects to address a given situation and component projects that facilitate and help to attain the pre-defined goal and objectives. For purposes of illustration, we can cite the following component projects: - environmental and sanitary education components for the sanitation and health projects (hygiene, etc.); and, - social assistance and psychological support components for the resettlement projects. 2. The experience of the CPM-IBRD in Juiz de Fora, although it contributed a new and advanced administrative model, unfortunately did not create a specific municipal culture. From the standpoint of evaluation, this project met with much resistance on the part of the implementing units, and the results were possible only because of the technical effort. The cyclical pace of government policies mitigates against the use and adoption of the resulting technologies in project design and implementation. The outsourcing of public services, neglecting or underutilizing installed technical capabilities, has hindered the application and development by a large number of public employees from the CPM-IBRD school of the procedures they have learned. Further work is still needed on the basic aspects of the cultural process, participation and communication. In the CPM the model was applied only in part, in project design, but not in project monitoring and ongoing evaluation. 84 ANNEX 2 3. We highlight the importance of having a research and planning institute in the municipal administration. One example is the case of Juiz de Fora. Soon after it was created, the Research and Planning Institute of Juiz de Fora took charge of designing and administering the CPM-IBRD projects. The technical team trained at that time learned a great deal. Soon thereafter the CPM was able to apply the lessons learned in projects such as the following: The Fund to Support Social Development (FAS), the Program for Social Priorities (PRO-MUNICIPIO), the Program for the Sanitation of Risk Areas (PROSANEAR), and the Program for Environmental Sanitation, Organization, and Modernization of the Municipalities (SOMMA). In addition to these projects, the office has been undertaking short-term as well as permanent activities in the area of urban planning, including drafting and monitoring laws relating to urban development. 4. The level achieved by Juiz de Fora in the document as a Best Practice City suggests to us a certain concern in the face of the new development prospects that lie ahead. The inevitable increase in the population, and with it the virtual growth in the demand for public services, requires significant investment to ensure that the quality of life remains at today's levels. [signature] Carlos Alberto de A. Zenkner Director of the Economic-Financial Department 85 ANNEX 2 IPEA/CGPU TECHNICAL NOTE Brasilia, May 27, 1997 TECHNICAL NOTE ON THE DOCUMENT "BRAZIL: IMPACTS OF URBAN PROJECTS" (May 2, 1997, draft) Diana Meirelles da Motta Based on the projects evaluated, the document appropriately sets forth the main objectives of the impact evaluation done by the Operations Evaluation Department of the World Bank, as indicated in the text distributed during the workshop held April 22 and 23 of this year. The suggestions now presented refer to item 1 of the document, entitled "Main Conclusions and Recommendations," and annexes Al - Brazil 1975-96: Benchmark Urban Sector Events and A2 - Brazil 1975-96: Government Policy relating to the Urban Sector. (1) ITEM 1. "MAIN CONCLUSIONS AND RECOMMENDATIONS" The following recommendations are aimed at highlighting issues that may arise in programs that support reforms by improving services to address locally determined priorities. It is hoped that this will help consolidate project evaluation efforts. Improving the Living Conditions associated with Urban Poverty When evaluating project impacts, special emphasis should be placed on urban poverty, and the need to ensure it is addressed by the urban policies of the federal, state, and municipal governments. In the peripheral areas of the metropolitan regions, urban problems are often associated with land use and poverty, and are generally aggravated by the steady incorporation of new areas that are parceled out clandestinely. For the most part, such informal human settlements are located in areas governed by legislation that places restrictions on occupation and construction, and they have become an alternative means of housing for the poorest sector of the population. Drafting suggestion: - The informal areas should be recognized as a key element in the city's economic and social development. Conventional planning approaches should be revised so as to take account of situations in which housing is provided by the informal sector. - Simplifying and adapting urban development plans and legislation on land use as a way to apply specific instruments for solving urban land titling problems, as well as developing 86 ANNEX2 environmental strategies specially tailored to the cities, in order to serve the occupied areas and contain the social and physical deterioration of the urban environment. Institutional Strengthening and Reform at the Municipal Level Mitigating urban problems includes recognizing that urban policy should be seen from an intersectoral perspective, in the macroeconomic and social context, with a view to better applications of urban investments, expanding social equity, and fostering economic growth, as well as improving the quality of life for the population. Drafting suggestion: - Evaluate urban management and planning from the institutional, financial, legal, and political-administrative perspectives, especially in relation to land use, housing, infrastructure, and public services, so as to contain the social and physical deterioration of the urban environment. - The regulatory, institutional, and legal environment should be enhanced with a view to eliminating the obstacles to greater access to land, housing, and urban infrastructure. It is recommended that new types and sources of financing for investment be found, based on expanded cooperation between the public and private sectors. - The institutional stability of the federal agencies involved in urban development, which have seen successive political-administrative reforms over the last 12 years, is fundamental to ensure the design and implementation of federal urban policy, and to attain specific goals. Support for Urban Policy and the Research Institutes Drafting suggestion: - Urban policy strategies should be based on a new model of state intervention, supported by administrative decentralization and an effective partnership with society and the private sector, considering the local communities to be strategic, and fostering the mobilization of local resources and the participation of the population. Under such a model it would be up to the federal government to regulate, articulate, and foster coordinated action to address the main urban problems. - Urban management can be improved through the activation and application of appropriate urban policy instruments (for urban development, as well as legal, institutional, financial, and other instruments) and by involving different actors, both formal and informal, in the urban development process. The lack of appropriate methods of analysis to obtain information about the city, urban planning practices, and the design, development, and application of its main instruments also stands in the way of satisfactory urban management. - As less urban research was done in the 1980s, an up-to-date empirical, analytical, and conceptual basis is needed concerning urban problems. Studies should be carried out on important 87 ANNEX2 issues to be considered by the various levels of government in order to support urban policy analyses and proposals. The planning and research institutes need support to perform their role of promoting urban planning and designing policies and policy instruments. (II) TABLE Al - BRAZIL 1976-96: BENCHMARK URBAN SECTOR EVENTS It is proposed that under the item government policies, 1992, the decision to abolish the Ministry of Social Action and the Secretariat for Regional Development be indicated, as well as the creation of the Ministry of Social Welfare (with responsibility for national policy in the areas of housing, sanitation, internal migration, and occupation of lands), and of the Ministry of Regional Integration (with responsibility for urban development, among other areas). It is proposed that mention be made, under government policies, 1995, of the creation of the Ministry of Planning and Budget, with responsibility for urban policy. (III) TABLE A2 - BRAZIL 1975-96: GOVERNMENT POLICY RELATING TO THE URBAN SECTOR It is proposed that under the item government policies, 1992, the decision to abolish the Ministry of Social Action and the Secretariat for Regional Development be indicated, as well as the creation of the Ministry of Social Welfare (with responsibility for national policy in the areas of housing, sanitation, internal migrations, and occupation of lands), and of the Ministry of Regional Integration (with responsibility for urban development, among other areas). It is proposed that mention be made, under government policies, 1995, of the creation of the Ministry of Planning and Budget, with responsibility for urban policy. 88 ANNEX 2 Brasilia, 5 June 1997 To: Dr. Roberto Bastos Carreiro Coordinator of Performance Evaluation - SEAINIMPO From: Jorge Guilherme Francisconi Ref: Comments on draft version of Impact Evaluation Report on Urban Development Projects prepared by the Operations Evaluation Department of the World Bank 1. These observations deal with the most important aspects of the partial knowledge that I have of the problem and also respect the spirit of the Report. The excellent quality of the work done should be highlighted, and thus the importance of its compatibility with present urban problems. Here follow some modest suggestions to complement that effort. 2 I suggest that the document makes clear the differences and/or similarities between: * objectives that originally guided the elaboration of the projects analyzed; * objectives that guide the Impact Evaluation Report. This comment aims at establishing clearly for the reader, if (i) the original objectives were achieved and if (ii) those objectives are the same and the ones that guided the evaluations of the IER. In the debates at SEAIN this question remained unclear. If I remember well, the Medium-sized Cities Project-in whose preparation I participated while still at CNPU, an entity that does not feature in the IER-and the Recife Metropolitan Project were guided by the need for a multidisciplinary action, geared toward economic and social questions, the strengthening of the system of medium-sized cities, and the reduction of regional inequalities-in the case of the Recife project. 3. The IER's understanding in item 6 (page vi) about the "Policy Dialogue in Brazil" appears to be incomplete, when it suggests that poverty alleviation and improving the conditions of the urban poor were among the objectives of the loans made at that time. This is partially true. The directives of the Second National Development Plan (II-PND) to reduce injustice and social inequalities, to strengthen the equilibrium of Brazil's urban network and reduce regional inequalities included-without being the central point-poverty alleviation and improving the conditions of the urban poor. As observed in the IER, the Housing Finance System (SFH) had too much money, but the FGTS system did not permit a broader participation of the poor in housing policy. On the other hand, if the IER is wanting in offering a more complete picture of that moment, the energy crisis should be remembered as an important causal factor in the loans requested by EBTU for Metropolitan Corridors, for the Porto Alegre suburban train (TRENSURB)-the Bank's first in this sector-and the corridors of Urban Agglomerations (AGLURB). Paving of streets for bus routes, in favelas, arises at this time, always geared toward a poorer population. It 89 ANNEX 2 achieved a great success, and was repeated in other projects and remains among the components with the most successful results. That is all, and congratulations for the Methodology adopted! Sincerely (signed) Francisconi 90 IBRD 28458 기 긔 긔 l ■ ■ ■ ■  IBRD 28459 VENEZUELA V ,.- J_ y BRAZIL COLOMBIb: Ocan JUIZ DE FORA MINAS GERAIS STATE IMPACT EVALUATION STUDY URBAN DEVELOPMENT PROJECTS B RAZI L SITE LOCATION: STA. EFIGENIA 7O (MEDIUM-SIZED CITIES PROJECT, LOAN 1720) --LU Belo Horizonte J BOLIA, CEr 1:.x F ,.:: 71~ - i -. '-- AND STREETS T...ULbo -. 1 . . . .s -- - -1-'-.i.ir1.' 'DARIES ,.1 . .. 3 IIl +a on d ,a p any judgment on the legal status of any tero or any endorsement or acceptonce of such Roý houndones. To Rio de Janeiro JUNE 1997  BRAZIL FORTALEZA A f e a n CEARA STATE IMPACT EVALUATION STUDY F.OPESTA URBAN DEVELOPMENT PROJECTS SITE LOCATION: POTIRA (NE PILOT METRO DEVELOPMENT PROJECT,LOAN 2193/ A. rd---ni, k t CITY CENTER SELECTED PROJECT SITE BUILT - UP - AREA MAIN ACCESS ROADS AND STREETS .» ALDEOTA RAILROADS PARANGABA RIVERS MUNICIPAL BOUNDARY pp.PPT STATE BOUNDARY es INTERNATIONAL BOUNDARIES BOM JARDIM -.I CASTELAO BRAZIL nr... :e-roor acepace of such C ~- a.,,a-. R,er.6,r To Juazeiro do Norte boundaries. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 0  IBRD 28461 k- TnP.12-a., ...- -35 C,I 3 ,-ed n - kl :r . o,-rr The O.Z...O -3. G o -T. G-.rt' '91 R,ver Ocani To Mc :2 ko To Rioc. ToMor- Altg,÷ g VEZUA .. BRAZIL COLOmBA arioni.c D._-n NATAL RIO GRANDE DO NORTE STATE IMPACT EVALUATION STUDY URBAN DEVELOPMENT PROJECTS BRAZIL SITE LOCATION: QUINTAS PERU (MEDIUM - SIZED CITIES PROJECT, LOAN 1720) 1BOVIA CITY CENTER | RAILROADS SELECTED PROJECt SITE RIVERS BUILT - UP AREA STATE BOUNDARY ARGENTIN . MAIN ACCESS ROADSAND STREETS ---- INTERNATIONAL BOUNDARIES Atlnwi,: 0:dan JUNE 1997 u IBRD 28462 BRAZIL Ro Seo Francisco JUAZEIRO BAHIA STATE ATr-m. i EV/AlLU ATION STUDY N DEVElOPMENT PROJECTS -¯-TiO.N: ALTO DO CRUZEIRO -''M SIZED CITIES PROJECT, LOAN 1720) iT BUILT UP AREA ® CITY CENTER SELECTED PROJECT SITE - MAIN ACCESS ROADS AND STREETS KILOMETERS -To Salvador BRAZIL SALVADOR BAHIA STATE I\IMPAGT EVALUATION STUDY 11 RB^ Nl f -n A DVELOPMENT PROJECTS - c.- SITE LOCATION· ANTONIO BALBINO AN rDEVOPMNT PROJECT, LOAN 2193) . EJillT UF I EA C. C1. CITY CENTEF bi-¯riEiTE ¯ - )1 I SELECTED FP-DJEK¯T SITE - /I.lN 1CCESS R .D$ AND STUEETz - a .--B R A Z l L ,7å-. -;j.r|, 5 '997 5 1 IJUNEr'997  IBRD 28463 BRAZIL ARACAJU SERGIPE STATE IMPACT EVALUATION STUDY URBAN DEVELOPMENT PROJECT SITE LOCATION: 18 DE FONTE E' R-AC- SIOUBIRA (NE FLOOD RECONSTRUCTION PROJECT, LOAN 2645) RuLT UP .1PEA A..~~~ En ,mPSC-Ci CENTEP I|SELECTED PK-51ECT SITE - mAIN ACCESS KDADS åtJ $TPEETS BANCA R10 BAI p RD SETUUIO VARC -AS BINDUJSTRIAL MKPA ?'FT EP f:ENTRO T. in.' 6.3,.k Rio Sergipe -L.ued Ca C-.:- lg- -.-1a BRAZIL PROPRIÁ BUILT UP AREA PRPRÅ CITY CENTER SERGIPE STATE CT C PC,ECT 51TE IMPACT EVALUATION STUDY - M.å N ACCESS S -s 5PT URBAN DEVELPOMENT PROJECT SITE LOCATION: BARRIO AMERICA (NE FLOOD RECONTRUCTION PROJECT, ' LOAN 2645) % 1a das Ped,-,nha BR A Z, L RU sOLIVIA L \V "AR ARGENTINA METERS \ - JUNE 1997  IBRD 28464 Sm BUILT UP AREA BRAZIL 0 CITY CENTER PETROLINA SELECTED PROJECT SITE PERNAMBUCO STATE To Recife - MAIN ACCESS ROADS AND STREETS IMPACT EVALUATION STUDY URBAN DEVELOPMENT PROJECTS JOAC, DE DEUS SITE LOCATION: VILLA MOCÓ (MEDIUM SIZED CITIES PROJECT, LOAN 1720) COSME E DAMIA0 DOH '!ELAP - -,NTC t.NDRE J4RDIM sip.IRO T 54, PALO JC.>E E PMFE PEDPO FC.RTAL DA PÅALMUNDO CIDADE APE JARD ouo COMINHO APEIA WPERIAAMALS PRETO DO 50L BPANCJ SA ONCA LD C0HAB v & v H AlUIl EDUAFf-O RIO COPPENJTE COHAB I PALHINHAS AASM ,ILA DOC E,A B4NCAIN PDRTA D05 BRAZIL RECIFE PERNAMBUICO STATE IMPACT EVALUATION STUDY JRBAN DEVELOPMENT PROJECTS SITE LOCATION: CORREGO DA ARE1A RECIFE URBAN DEVELOPMENT PROJECT, LOAN 2170) ATLANTIC OCEAN .. ,i.. ENC-ENH [. 1 .3 OA8, - ~ - PF -E-.- - U-lIT'-JFMA EA IJ - JKjIF;-L ECUND,R-N - ® C.ITY CFNTEP S ¶ELECTE PFO.ECI EITE 8 R A Z iFEF I N -CE5 PG DE FN V'D Fil TNj 1' 'J j h.x oL ci , d!i,E i... i j fv , I. . rI, u 1 - JUNE 1997 t IBRD 28465 1一他 一 JUNE 1997   IMAGING Report No.: 16736 Type IER