Ecuador - Guayaquil Urban Development and National Low Income Housing Projects

Document of The World Bank FOR OFFICIAL USE ONLY MICROFICHE COPY Report No. 10238 Report No. 10238-.EC Type: (PPR) TWEDDLE, E/ X31707 / T9 073/ OEDD3 PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) AND NATIONAL LOW-INCOME HOUSING PROJECT (LOAN 2135-EC) DECEMBER 31, 1991 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. BVIATIONS AND ACRONYMS AID - U.S. Agency for International Development BEV - Banco Ecuatoriano de la Vivienda - Ecuadoran Housing Bank BCE - Banco Central de Ecuador - Ecuador Central Bank BP - Banco del Pacifico CENAPIA - Centro Nacional de Pequena Industria y Artesania - National Small Industry and Artisan Center CONADE - Consejo Nacional de Planificaci6n - National Planning Council GTZ - Gesellachalt Fur Techniche Zussamenarbet - (German) Society for Technical Cooperation JNV - Junta Nacional de la Vivienda - National Housing Board 4 FOR OFFICIAL USE ONLY THE WORLD BANK Washington. D C 20433 USA Office of Director-&eraI Opeawwnn [Valuatu.A December 31, 1991 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on Ecuador Guayaquil Urban Development Project (Loan 1776-EC) and National Low-Income Housing Project (Loan 2135-EC) Attached, for information, is a copy of a report entitled 'Project Performance Audit Report on Ecuador - Guayaquil Urban Developmenc Project (Loan 1776-EC) and National Low-Income Housing Project (Loan 2135-EC)" prepared by the Operations Evaluation Department. Attachment This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) AND NATIONAL LOW-INCOME HOUSING PROJECT (LOAN 2135-EC) TABLE OF CONTENTS Page No. Preface ......................................................... i Basic Data Sheets ................................................ ii Evaluation Summary ............................................... vi I. BACKGROUND ................................................. 1 A. Macroeconomic Context .................................. 1 B. Urban Sector Background ................................ 2 C. Government Structure and Responsibilities .............. 3 II. THE PROJECTS ............................................... 4 A. Issues in Project Preparation .......................... 4 B. Project Objectives ..................................... 6 C. Project Components and Institutional Responsibilities .. 7 III. PROJECT IMPLEMENTATION AND RESULTS ......................... 9 A. Factors Contributing to Component Cencellation and Execution Delays ..................................... 9 B. Financing and Costs .................................... 11 C. Performance by Component ............................ 13 D. Economic Evaluation .................................... 19 E. Financial Performance .................................. 20 IV. POINTS OF SPECIAL INTEREST ................................ 21 A. Poverty Impacts ....................................... 21 B. Urban Sector Policy .................................. 21 C. Housing Policies ..................................... 22 D. Institutional Development .............................. 23 V. ROLE OF THE BANK ........................................... 23 VI. CONCLUSIONS AND LESSONS ................................... 25 A. Conclusions ............................................ 25 B. Lessons Learned ....................................... 26 ANNEX 1. Comments from the Borrower and the Executing Agency ......... 30 This document has a restricted distribution and may be used by recipients only in the performance of their offieial duties Its contents may not otherwise be disclosed without World Bank authorization. i PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) AND NATIONAL LOW-INCOME HOUSING PROJECT (LOAN 2135--EC) PREFACE 1. This is a Project Performance Audit Report (PPAR) on two urban development operations in the Republic of Ecuador. The first, the Guayaquil Urban Development Project. Loan 1776-EC in the amount of US$31.0 million, was approved on December 11, 1979. It closed two years behind schedule, on June 30, 1987. Total disbursements were US$25.8 million and US$5.2 million were cancelled. The second is the National Low-Income Housing Project, Loan 2135-EC for US$35.7 million, approved on May 4, 1982. The closing date was June 30, 1989, one year behind schedule, with total disbursements of US$33.7 million and cancellation of US$1.9 million. 2. The PPAR has been prepared by the Operations Evaluation Department (OED), based on Project Completion Reports prepared by the Infrastructure and Energy Operations Division of Country Department IV of the Latin America and Caribbean Regional Office (LA4IE). Staff Appraisal Reports, project legal documents, correspondence files, interviews with project staff, and other relevant material. It is also based on interviews with various national and local authorities during an Audit mission carried out in March 1991. The kind cooperation of BEV/JNV and municipal personnel in the preparation of this report is gratefully acknowledged. The respective PCRs were previously submitted to the Board as OED Reports No. 8262 (Guayaquil), dated December 1989, and No. 9239 (Low-Income Housing), dated December 1990. 3. The PPAR generally agrees with the findings of the two PCRs but takes a combined perspective on the two operations and expands the conclusions relating them to the design of future projects in Ecuador and elsewhere. In order to give a wider perspective, OED has included additional information about the operations, the urban and economic environment of the country, relevant policy issues raised, problems encountered during implementation, and major outcomes. 4. The draft PPAR was sent to the Borrower and Executing Agency for comments on October 18, 1991. Comments received from the Ministry of Finance and Junta Nacional de Vivienda/Banco Ecuatoriano de la Vivienda have been attached as an Annex to the Report. PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) BASIC DATA SHEET KEY PROJECT DATA Appraisal Estimate Actual Total Project Cost 51,600 37,000 Loan Amount 31,000 31,000 Disbursed 31,000 25,784 Cancelled - 5,216 Date of Physical Completion 12/84 12/86 Time Overrun 24 months Economic Rate of Return 17-19% 22% (Sites and Services) CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ Millions) FY80 FY81 FY82 FY83 FY84 FY85 FY86 FY87 FY88 Appraisal 0.4 6.9 14.8 23.3 29.3 31.0 - - - Actual - 0.3 2.2 3.0 3.8 6.0 16.3 22.1 25.8 PROJECT DATES Planned Revit, 1 Actual Identification 08/77 08/78-12/78 Project Preparation 10/77-11/77 11/77 Appraisal Mission 11/78-12/78 Loan Negotiations 10/79-10/79 Board Approval 12/79 Loan Signature 05 '80 Effectiveness Date 10/80 Loan Closing 06/85 06/86-12/86 06/87 Loan Completion 12/84 06/86 12/86 iii OTHER PROJECT DATA Borrower - Republic of Ecuador Executing Agencies - Municipality of Guayaquil BEV/JNV - Ecuadoran Housing Bank and National Housing Board CENAPIA - National Small Industry and Artisan Center Education Ministry MOP - Ministry of Public Works Follow-on Projects - National Low-Income Housing Project (Loan 2135-EC) and Second National Low-Income Housing Project iv PROJECT PERFORMANCE AUDIT REPORT ECUADOR NATIONAL LOW INCOME HOUSING PROJECT (LOAN 2135-EC) BASIC DATA SHEET KEY PROJECT DATA Appraisal Estimate Actual Total Project Cost 100.4 87.1 Loan Amount 35.7 33.7 Disbursed - 33.7 Cancelled - 1.9 Date of Physical Completion 06/87 06/88 Time Overrun 12 months Economic Rate of Return 18% 20% CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ Millions) FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 Appraisal 1.2 4.4 12.5 25.8 33.9 35.7 - - Actual - - 1.4 6.5 17.6 26.5 31.9 33.8 Actual as 2 of estimate - - 11.2 25.2 52.0 74.2 89.4 94.7 PROJECT DATES Planned Revised Actual Identification 05/80 Project Preparation 03/80-07/8 Appraisal Mission 07/81 Loan Negotiations 03/82 Board Approval 05/82 Loan Signature 10/82 12/82 Effectiveness Date 03/83 12/83 Loan Closing 06/87 12/37 06/88 Loan Completion 12/87 12/84 12/88 V OTHER PROJECT DATA Borrower - Republic of Ecuador Executing Agencies - BEV/JNV - Ecuadoran Housing Bank and National Housing Board Follow-on Projects - Second National Low-Income Housing Project vi PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) AND NATIONAL LOW-INCOME HOUSING PROJECT (LOAN 2135-EC) EVALUATION SUMMARY Introduction 1. The present report reviews the performance of two urban development projects in Ecuador whose implementation initiated in the early 1980s and proceeded throughout most of the decade. They are the first two urban projects in the country and were followed by municipal development and housing finance operations which incorporated significant advances derived from their implementation experience. The evaluation highlights the contrasts in performance of the two projects whose results generate interesting insights into issues of design, effectiveness, and the institutional dimensions of this type of poverty alleviation effort. ObJectives and Description 2. The Guayaquil Urban Development Project was conceived as an integrated program addressing multiple aspects of urban poverty and including both economic assistance and shelter improvement components (para. 2.12). The project was restricted to Ecuador's largest city, Guayaquil, and emphasized development of the municipal government's capability to deliver local services on a sustainable, cost recoverable basis. Project components were to include: the provision of about 6,400 loans to small-scale enterprises to be implemented by a private bank (Banco del Pacifico) and supported by training; squatter upgrading, comprising land regularization, landfill, and basic infrastructure for two existing neighborhoods located in flood-prone areas; serviced plots for 3,600 families with off-site infrastructure and community facilities (primary schools, health centers and markets) to serve adjacent areas; and 9,700 reduced standard housing and home improvement loans to be granted by two national agencies, the Ecuadoran Housing Bank (BEV) and the National Housing Board (JNV) (paras. 2.15- 2.16). 3. The National Low-Income Housing Project (NLHP) was essentially a sector operation which supported a major change in national housing policy by shifting its focus toward the provision of shelter affordable to the low-income population (paras. 2.13-2.14). Its other main emphasis was on the reform of housing finance instruments through the application of variable real interest rates to the loans financed by BEV. The project was implemented in 15 large and medium-sized cities and was designed to include: the construction of 8,200 new housing solutions and 7,500 small loans to improve basic housing units already financed by BEV. It was also to contain urban upgrading and community facilities components, together with technical assistance to BEV/JNV to address various managerial and operational needs (para 2.18). vii Implementation Experience 4. Implementation of the Guayaquil project suffered from a variety of problems which affected both the upgra_'4ng components under the Municipality's responsibility and assistance to sma' Ale enterprises. The Municipality was never able to reach an agreement wit. '.eneficiaries of the upgrading projects concerning cost recovery. Funding and contract execution problems for a sites and services scheme caused long implementation delays, prompting neighboring families to invade the plots and leading the Municipality to abandon the works. The design of the credit program for small-scale enterprises, in turn, proved to be too complicated and rediscount rates set by the Central Bank were too low, discouraging the participation of private financial intermediaries. Both credit assistance and the upgrading components, r.s originally planned, were later abandoned, although the Municipality did carry out in-fill and emergency works in four neighboziLoods affected by flooding, aiIJ the project had to be reformulaced in mid-course. Restructuring was an attempt to save the operation by expanding the sites and services and core house components under BEV/JVN's responsibility, since the initial investments implemented by these two agencies had been comparatively successful. As a result, a total of 5,400 low-cost housing units was built and 7,900 home improvement loans were granted under the project (paras. 3.01-3.06). 5. Implementation of NLHP was much more satisfactory due to its simpler design and the better performance of its executing agencies. Physical targets for both low-cost housing and home improvement loans were exceeded and the project was implemented with a 17% cost underrun in US dollar terms, mainly because of local currency devaluation. It was also relatively successful in introducing housing finance policy changes and generating institutional advances within the participating sectoral agencies (paras. 3.07-3.08). Results 6. The poverty impact of the Guayaquil project, assessed either in terms of the number of direct beneficiaries or the potential effects of the introductl - of improved municipal service delivery and financing practices, was negligiblt. Early cancellation of the small enterprise credit component and the problems with squatter upgrading works essentially restricted the project's impact to families benefitted by BEV/JNV housing investments. The intended longer-run effects to be derived from institutional strengthening of the Municipality also failed to materialize (para. 4.01). In the case of NLHP, even though the number of beneficiary families exceeded the original target, no information exists on their breakdown by income levels. The operation, as implemented, however, had features which were incompatible with a strict poverty alleviation program. This opened it up to middle class families which have traditionally had priority in BEV/JNV programs. Loans were granted for the refurbishing (instead of just upgrading and completion) of housing units and procedures for verification of socio-economic conditions were not very rigorous. Given the liberal eligibility criteria and built-in incentives for BEV to favor the relatively wealthier families in the target population, the operation's poverty alleviation impacts can be assessed at best as marginal (para. 4.02). viii 7. The general failure of the Guayaquil project also limited any significant achievement of its stated objective of reinforcing the Municipality's capacity to serve low-income areas through the introduction of improved financial management and cost recovery practices. The initial working assumption that local governments should take primary responsibility for delivering and financing urban services was not borne out by the experience under the project. In fact, just the opposite occurred, with central agencies performing in a more responsive manner than the Municipality. While this does not disprove the general thesis that local governments should be the principal actors in urban development, it does highlight the need to address their weaknesses and establish appropriate incentives for improved local resource mobilization and service delivery (paras. 4.03-4.05). 8. A major policy objective in the shelter sector was to promote a shift in emphasis from middle to low-income housing and solutions affordable to the urban poor. Home improvement loans were one way of reaching the population living in sub-standard housing and on untenured land. These objectives were fully met and the corresponding lines of financing now represent the mainstream of BEV operations and JNV construction activities. As concerns the introduction of more sustainable housing finance practices by BEV, which meant moving from below-market fixed interest rates to inflation-hedged variable rates, success has been only partial. Variable interest rates have been introduced, but only on loans financed with external credits, while rates established by the "Junta Monetaria Nacional" continue at below market levels. The housing finance system as a whole still depends heavily on central government budgetary allocations and the bulk of its funding derives from foreign borrowing (paras. 4.07-4.09). 9. Institutional development constituted an explicit objective of both projects. This was to be promoted through the adoption of upgraded financial and managerial practices by project executing agencies supported by direct technical assistance. In the case of Guayaquil, efforts to strengthen municipal operational capacities by introducing cost recovery practices and improving internal organization failed dramatically. As a result, the municipal government continues to be characterized by weak administration and deficient service provision and is in constant state of fiscal crisis (para. 4.10). Institutional strengthening efforts under NLHP were far more effective. BEV has begun to employ financial planning practices encouraged by the Bank. General improvements in accounting and cost control are a direct result of technical assistance funded by the project. With respect to JNV, which acts as a major developer and construction company, advances in operational planning and control are largely due to long-term in-house assistance provided by foreign experts (GTZ) partially funded by the Bank. However, JNV still needs to streamline its structure and more extensively adopt planning procedures which, at present, are largely limited to externally financed programs (para. 4.11). Sustainability 10. Cancellation of various components of the Guayaquil project speaks for itself with respect to its impact on municipal service delivery. Failure of both sites and services and neighborhood upgrading schemes is a reflection of the inadequate implementation arrangements adopted by the Municipality. Furthermore, the project could do little to bring about any significant change in local ix financing practices. As a consequence, the city continues to suffer from chronic infrastructure shortages. Abbreviation of the small-scale enterprise component, in turn, highlights the pitfalls of overly complex implementation mechanisms, but does not negate the merits of the basic approach to supporting small businesses, which continues to be followed on a more pragmatic basis by private banks in Guayaquil. 11. For NLHP, an estimated 83% recovery of BEV loans and an average rate of return of 18% for the three main components (17% for housing, 21% for home improvement loans, and 41% for upgrading) provide one set of indicators of project sustainability. Home improvement loans, as a modality of assisting the poor, proved to be a successful approach and were both adopted in subsequent operations and incorporated as a major line of activity by the housing finance system in Ecuador. Likewise, the simplification of housing design standards, making units affordable to lower income segments, opened up the system to a larger share of the urban population. As in many other countries, however, public housing provision continues to be dependent on fiscal subsidies. 12. As a direct result of these two projects, follow-on operations for municipal development and housing finance have incorporated lessons derived from the early experiences. Shelter projects now focus on the sustainability of the housing finance system while following basically the same approach (i.e., home improvement and low-cost housing) taken in NLHP. The most recent urban sector (i.e., municipal development) loan to Ecuador, in turn, addresses the major political issue of intergovernmental transfers, is designed as an on-lending operation open to a broad range of municipalities, and emphasizes technical assistance to strengthen local finances. These design features were incorporated partially in response to the flaws in the Guayaquil project. Findings and Lessons 13. The two operations under review had sharply contrasting execution experiences and results. Much of the Guayaquil project, as originally appraised, was never implemented because the Municipality proved unable to carry out many of its initial responsibilities. As a result, the loan had to be restructured in 1984 and several important components (eg. small business support, slum upgrading, community facilities) were dropped. The project likewise failed to achieve most of its institutional, policy, and financial objectives, particularly in terms of strengthening municipal finances and service delivery and recovering investment costs from project beneficiaries. Inadequate cost recovery, together with the persisting weak executing capacity of the Municipality, in turn, has effectively limited both possibilities of project replicability and the likely sustainability of many of the benefits provided under the operation. Only the expansion and comparatively much more efficient implementation of the components under the two national housing agencies, JNV and BEV, saved the project from an even more unsatisfactory outcome. 14. NLHP, in comparison, was'a generally successful operation. Even though some initial implementation delays occurred and project investments had to be redirected to support reconstruction efforts following major natural disasters in 1982-83 (floods) and 1986 (earthquake), all of the operation's original objectives were substantially met. Unli.ke the Guayaquil experience, x moreover, institutional development goals were mostly achieved, contributing both to the satisfactory implementation of the project itself and preparation of a follow-on operation (the Second National Low-Income Housing Project, approved in January 1988), which is currently under execution. NLHP's success can be largely attributed to a combination of: (i) adequate Bank/Borrower application of both the positive and negative lessons of the Guayaquil operation; (ii) clear and simple project design, which focused on two complementary institutions (i.e., JNV and BEV) that had performed well in the first operation; (iii) supporting Government policies and programs; (iv) good Bank supervision which appears to have provided valuable assistance in problem identification and resolution; (v) effective technical and financial assistance to JNV and BEV more generally; and (vi) flexible and timely responses from the Bank in dealing with two natural calamities ar.d inflation-induced local cost increases and currency devaluations. 15. Various lessons of a general nature can be derived from these two projects. The Guayaquil operation confirms the need to avoid excessive dependency on institutionally fragile and highly politicized local governments and for the prior assessment of their administrative and financial capabilities. A second lesson, which is illustrated in different ways by both operations, is the desirability of limiting the sectoral and institutional complexity of urban development projects, particularly when a multitude of executing agencies and two or more different levels of government are involved, as was the case in Guayaquil, but not in NLHP. A third lesson is the importance of attaining sufficient cost recovery in urban shelter operations to ensure project replicability and sustainability, as well as the frequent difficulty of doing so. Beyond these general conclusions, several more specific lessons can be summarized below. 16. Prolect Design. The contrasting performance of the two projects raises issues both of design and the effectiveness of urban development assistance. By restricting the project to a single municipality, as in the case of Guayaquil, potential local development and policy effects were limited, while the introduction of competition among cities was impossible. Likewise, its multiple components and institutional interdependencies compromised project performance. Sector operations such as the Low-Income Housing Project, on the other hand, often can have only a limited urban development impact since they normally focus on a comparatively narrow range of problems and do not address many basic local development, financing, and management issues. However, their relatively straightforward design and dependence on fewer institutions appears to greatly enhance their probability of success. The key is to find a proper balance between sector policy goals and implementation prospects. Sector operations can be designed in such a way as to involve local agencies in their execution as long as the latter are willing to adopt internal managerial and service delivery changes (paras. 6.04-6.06). 17. Implementation Arrangements. Even more important than design considerations, the experience of these operations reveals that institutional pre-conditions often play a decisive role in project outcomes. Institutional commitment to project goals and a viable implementation set-up, which includes adequate funding and administrative autonomy of the main executing agencies, are among the key requirements. It is, thus, important to conduct careful preliminary institutional assessment, giving particular attention to the degree xi of commitment of agency leadership to project goals, actual institutional capacity to carry out the operation concerned, and the possible legal, financial, and political constraints to which it or they are subject. These factors seem to have been largely overlooked in the Guayaquil project. Implementation arrangements in both operations, furthermore, included special project units. This worked well in the case of BEV/JNV, which established such units in its branches for implementation of NLHP and later absorbed them into their regular structures, but not in that of Guayaquil where the project unit was given only limited authority and was bounded both by the legal constraints applying to all local governments and restrictions imposed by centralizing and distrustful hayors (paras. 6.07-6.08). 18. Shelter Solutions. bEV/JNV's highly successful home improvement loans proved to be an appropriate response to market demand and a useful instrument for addressing low-income shelter needs. In moving away from the more expensive and highly subsidized finished housing previously offered by BEV, it was possible to extend the benefits of public sector shelter investments to a lower income bracket and to a larger number of beneficiaries. Considering that many Ecuadorans were living in substandard housing, either in plots which they already owned or in areas which they had informally occupied, home improvement loans helped to provide them with access to more adequate shelter within their economic possibilities. In the future, however, beneficiary income levels should be better scrutinized to avoid problems of mistargeting (paras. 3.27-3.28). 19. Approaches to Institutional Development. The refusal of the Municipality of Guayaquil to accept technical assistance demonstrates that the availability of resources alone is not sufficient to foster institutional development. The projects also contain several lessons with respect to institutional strengthening efforts. First, as concerns the relative effectiveness of long-term capacity-building versus shorter interventions, GTZ's contribution through an in-house technical assistance program proved to have lasting effects on BEV/JNV. Second, technical assistance is best absorbed when in tune with changes desired and accepted by the institution involved. The Municipality's resistance to internal reforms and the adoption of new service delivery practices illustrates this point. Third, organizational stability is important in order for such efforts to be successful given the often crucial role of top management support for the introduction of meaningful institutional change. The success of such assistance when applied in structured and stable organizations like BEV/JNV demonstrates that technical support can be a worthwhile investment (para. 6.11). 20. Municipal Development. An inadequate initial understanding of the political environment in Guayaquil and the Municipality's attitude toward social development contributed to the failure of most of the components under local government responsibility. This suggests the need for the Bank to improve its ex-ante analysis of executing agency implementation capabilities and commitment to project goals. Despite specific clauses in the Loan Agreement with respect to cost recovery, these were never implemented. This stemmed directly from the Municipality's populist practices which local community organizations came to expect, leading them to resist change, particularly when introduced by "foreign" entities (in this case the Bank). As a result, cost recovery proved to be as much a political as a technical issue (para. 6.10). xii 21. Assistance to Small-Scale Enterprises. The Guayaquil project missed an opportunity to set in motion a joint public-private initiative which had most of the ingredients for success. However, the complexity of the component's implementation arrangements and its inadequate compensation to private banks led the later to abandon the program. In spite of this, future efforts to support small-scale enterprises could well follow the same basic lines proposed in the Guayaquil operation, provided certain adjustments are made. These include: (a) encouragement of private bank participation on the basis of rediscount rates which cover administrative costs and the provision of training and other out- reach activities; (b) avoidance of unnecessary bureaucratic procedures, thereby shortening the distance between borrowers and financial agents in order to reduce intermediation costs; and (c) restricting the involvement of public agencies to the identification, selection, and training of beneficiaries. For best results, such training should be coupled with on-the-job technical assistance rather than being delivered exclusively in a classroom environment (paras. 6.12-6.14). 22. Targeting Social Programs. The problems that occurred with respect to the unanticipated direction of housing subsidies to middle class families under NLRP could have been avoided through a combination of more careful project resource allocation criteria and better supervision of the selection procedures used by the financial agents. In the future, such precautions should include: (a) income eligibility levels lower than those applied by the program which created an incentive for financial agencies to favor higher income families, who represented a relatively lower risk; (b) a rule preventing the same families from receiving subsidized loans twice, at least within a reasonable time interval; (c) pre-definition of the neighborhoods where assistance would be granted, perhaps by making them part of urban upgrading schemes in clearly defined low-income areas; and (d) improved supervision and out-reach social work both to ensure that household income surveys are properly carried out and to better design and implement complementary assistance programs (paras. 6.15-6.17). I PROJECT PERFORMANCE AUDIT REPORT ECUADOR GUAYAQUIL URBAN DEVELOPMENT PROJECT (LOAN 1776-EC) AND NATIONAL LOW-INCOME HOUSING PROJECT (LOAN 2135-EC) 1. BACKGROUND A. Macroeconomic Context 1.01 By the late 1970's, Ecuador was reaping the benefits of the oil price boom which was the main element supporting GDP growth of 5.2% a year between 1960-70 and 9.1% in the 1970-80 period. As a result, per capita income expanded at an average of 5.9% a year throughout the 1970s'. Crude oil exports increased their share in total exports from practically zero in 1970 to 70% in 1983. 1.02 Economic growth was accompanied by substantial expansion of the Ecuadoran public sector. Government expenditures grew by 9.6% in real terms between 1972 and 1983, relying heavily on oil revenues and external borrowing. A large proportion of these expenditures were employed in low-return investments or private consumption supported by government subsidies. The drawbacks of this course were later revealed when these two sources of fiscal income were drastically reduced during the following decade. 1.03 Two events in the 1980s negatively affected the Ecuadoran economy: the debt crisis starting in 1982 and the near halving of oil prices in 1985-86. Shortfalls in public expenditures during the 1970s were made up through heavy borrowing abroad, leading public sector debt to increase 13-fold between 1973 and 1982. Thus, when interest rates suddenly soared in 1982, the Government was faced with higher payments on external debt at the same time as commercial bank lending for new investments was rapidly shrinking. The main effects on the economy were a sharp rise in inflation, from 16% in 1982 to 48% in 1983, and abandonment of the fixed exchange rate of 25 sucres to the dollar, which had prevailed for more than a decade. 1.04 The sudden decline in oil prices in 1985 came at a time when the economy seemed to have regained stability following measures taken in the early 1980s to reduce the public sector deficit and which also caused a 2.8% decline in GDP in 1983. With oil prices falling from US$26 a barrel in 1985 to less than US$13 in 1986, export earnings plummeted by nearly 50%. The Government reacted by floating the exchange rate and freeing most interest rates. However, it failed to cut public spending, which rose by 4% of GDP in 1986 despite a severe drop in fiscal oil revenues. By the end of 1987, after the interruption of oil exports following a massive earthquake, total public spending had climbed to about 32% of GDP. As a result, 1987 marked the second year of negative economic growth during the decade (-3.2%). Inter-American Development Bank, 1990 2 1.05 Despite these problems, until the early 1980's, inflation of the sucre was moderate by Latin America standards. It reached the two digit level in 1972, but remained below 16% until 1982, averaging 11.7% a year during the 1970s. However, it jumped to 48% a year in 1983, to remain at around 30% throughout the rest of the decade. Domestic interest rates, tightly regulated until 1984, remained extremely low or negative in real terms. Passbook savings rates, which help finance the national housing system, continue to be unattractive to this day. As a result, the system still subsists on mandatory deposits imposed by the Government and the required maintenance of savings accounts at the national housing bank (BEV) in order for families to qualify for public sector housing loans. 1.06 In short, the first two Bank-supported urban development projects were implemented at a time of rapidly changing economic fortunes in a country which had enjoyed sustained growth during the two previous decades. Significant macroeconomic adjustment measures were implemented during the overlapping project implementation periods, at times imposing severe budgetary constraints on government agencies. For BEV, specifically, this implied a reduction in annual budgetary allocations and the need to rely increasingly on external borrowing to finance its operations. For the municipalities, in turn, it meant fewer subsidies from the national government. B. Urban Sector Background 1.07 Ecuador had a population of about 8.6 million in 1980, which grew to nearly 11 million by the end of the decade. At the time the Guayaquil project was appraised, the urban population was about 44% of the total, having experienced an impressive average annual growth rate of 5.1% between 1965-80 2. Urban growth was largely concentrated in the two principal cities, Guayaquil and Quito, the country's industrial/export and administrative centers respectively. By 1987, these two cities accounted for nearly half of the total urban population and 70% of national output. The combined population of the nine next largest (secondary) cities, which individually ranged between 100,000 and 500,000 people, was less than half the joint population of Guayaquil and Quito. Nonetheless, the secondary cities constitute the fastest growing class of urban centers, which exercise an important regional role in a country marked by clear geographical divisions. 2 World Development Report, 1989 3 Ecuador: Distribution of Cities by Size (thousands) Classes 1974 1987 No. Pop. No. Pop. More than 1 million - - 2 2,710 500 to 1 m. 2 1,423 - - 100 to 500 1 161 9 1,229 50 to 100 7 457 5 380 10 to 50 23 446 43 840 less than 50 76 262 79 306 Total Urban 105 2,699 138 5,465 Source: INEC 1.08 Quito, located in the Andean highlands, has a population of approximately 1.2 million. Its recent growth is associated with the oil boom mentioned in the previous section. By virtue of being the country's administrative and political center, it has capitalized on the tendency toward spatial concentration which frequently accompanies the expansion of government activity. 1.09 Guayaquil, on the coast, is Ecuador's principal port and industrial center. Throughout the 1960 and 1970s, it experienced rapid growth through immigration from rural areas which caused its population to double over a 10 year period, reaching 1.7 million in 1987. Largely as a result, the city is characterized by widespread poverty and a chaotic pattern of urban growth. Occupation of much of the southern zone along the Guayas River occurred through spontaneous "invasions" of public and private lands which were tolerated by successive local administrations. This established a pattern of "informal" urbanization that was subsequently repeated to the north and in other scattered areas around the city. These invasions were organized by local populist leaders who demanded improved infrastructure and services from the Municipality. C. Government Structure and Responsibilities 1.10 Ecuador is divided into 20 provinces, 162 cantons with constituted municipalities, and about 1000 other cantons without local governments, mostly in rural areas. Municipalities are headed by Mayors and have local legislative bodies (Consejos Municipales), which are elected for four year terms. 1.11 Under the Municipal Law, local governments are responsible for the provision of numerous urban services including water supply and sewerage, drainage, solid waste collection, street construction and maintenance, street lighting, parks and green spaces, markets, and slaughterhouses. They collect 4 their own revenues, particularly property taxes and license fees ("patentes"), but subsist essentially on central government transfers. Provincial governments take responsibility for rural roads and engage in a variety of complementary urban works, which are financed through formal agreements (convenios) with the central government. 1.12 The country is constituted as a unitary system. One implication of this is that municipal governments are regarded as divisions of the central government. As a result, they enjoy only limited administrative and financial autonomy. Local budgets are part of the national budget, so they must be approved by the central government. Municipalities are also subjected to the same treatment as central government agencies with respect to contracting, purchasing, personnel hiring, and other administrative functions. Accordingly, their administrative acts have to be approved by central agencies, often through lengthy bureaucratic procedures. II. THE PROJECTS A. Issues in Project Preparation 2.01 The Guayaquil project was the Bank's first urban sector operation in Ecador. At the time it was approved, the Bank was attempting new approaches to urban development, which had evolved from the early slum upgrading and sites and services programs to an increasing emphasis on employment generation and the institutional strengthening of municipal governments. This approach called for simultaneously addressing the physical infrastructure requirements of depressed neighborhoods and the economic conditions of the urban poor. As a result, numerous projects were designed as multi-sector and multi-component "integrated" urban development operations. 2.02 From this perspective, the choice of Guayaquil for the initial urban operation in Ecuador was justified both by the city's rapid growth and the significant portion of its population that was living in poverty. In addition, Guayaquil's local development problems, which included the uncontrolled and often illegal occupation of swampy, flood-prone areas to the south of the city, seemed to make it an appropriate setting for application of the Bank's prevailing concepts of urban assistance. 2.03 The Bank's concern with poverty alleviation called for a combined effort to address the economic and environmental situation of the urban poor. In Guayaquil, the existence of an on-going credit program for small-scale enterprises by a private institution (Banco del Pacifico) presented an opportunity for a project component that was seemingly uncomplicated and yet well within this general philosophy. Thus, about 21% of project resources were allocated to this component, which depended on a rediscount facility in the Central Bank. What appeared to be an ideal opportunity to combine public and private initiatives in a socially oriented program, however, later proved to be institutionally cumbersome and an uneconomic proposition for the private banks. 2.04 The principal implementing agency was to be the Municipality of Guayaquil. Its responsibilities included direct execution of squatter upgrading projects, provision of infrastructure and services for areas where 3EV would 5 later build low-cost dwellings, and various actions involving land regularization and cost recovery. In assessing its capacity to undertake such activities, however, the Bank gave insufficient attention to the widely perceived weaknesses of the municipal government and its populist tradition of freely supplying infrastructure to poor neighborhoods under pressure from local political bosses. 2.05 The Municipality's difficulties were not completely disregarded, however, as the project allocated roughly US$200,000 for technical assistance and included a loan covenant requiring preparation of a study on municipal finances and management. The Loan Agreement also required establishment of a special project unit which, so as to avoid the shortcomings of other municipal interventions, was to be granted financial and administrative autonomy. 2.06 Despite expeditious project preparation, the Mayor with whom the mission negotiated the Municipality's components was out of office by the time the operation was approved. This pattern continued throughout the project implementation period, with six different Mayors holding office between 1980 and 1986. 2.07 The National Low-Income Housing Project (NLHP), in turn, was designed with an eye toward avoiding the problems experienced by the Guayaquil operation. As its name suggests, it was essentially limited to shelter investments and the introduction of operational changes among the executing agencies. NLHP also attempted to build on the few successful aspects of its predecessor, namely the favorable experience with home improvement loans and BEV/JNV's satisfactory performance in Guayaquil. 2.08 Thus, the more general urban finance and management aspects of the Guayaquil project were largely abandoned in NLHP in favor of a more pragmatic concern with improving the effectiveness of project implementation. Working with central government institutions, better funded and enjoying greater administrative autonomy than local governments, was a key consideration in designing the follow-on operation. 2.09 The second project also sought to consolidate the notion of affordable shelter, extending the Guayaquil experience with low-cost construction to the national level and supporting the expressed desire of government authorities to shift the emphasis of housing policy toward the needs of the low- income population. Previous activities of the two principal housing institutions, BEV and JNV, had almost exclusively involved the construction and financing of finished houses, affordable only by middle-income families. In the Guayaquil project, the concepts of basic housing units ("piso-techo"), sites and services, and home improvement loans, accompanied by efforts to simplify municipal building standards and urban infrastructure requirements, had been successfully introduced. These became the central components of NLHP. 2.10 Government concern about the significant gap in population size between Quito and Guayaquil, on the one hand, and secondary and smaller cities, on the other, caused NLHP to focus on national urbanization issues, prompting the allocation of at least 57% of project funds to ten intermediate cities. The formal criteria for the selection of these centers included their relative importance in the urban hierarchy, their recent growth rates and sizeable unmet 6 housing demand and the availability of JNV property for the new housing components. All of the cities initially included in the operation, moreover, were provincial capitals. The other urban areas later added to the project included coastal towns affected by floods (1984) and earthquake damage (1987) which did not meet to the above criteria. 2.11 A significant policy dimension of the second project was the introduction of an approach to public housing provision consistent with sound financial management principles. This essentially involved a change in BEV's interest rate policy, which entailed a change from fixed to variable rates set at a level that would shield the bank portfolio against the effects of inflation. Achievement of this objective, however, was constrained by existing legislation which placed interest rate ceilings on government loans. These rates were established by the "Junta Monetaria" (Monetary Board) based on macroeconomic and social considerations and have traditionally been two or three percentage points below market rates. In compensation, private financial institutions were compelled to invest a variable portion of their reserves in BEV bonds, providing the housing bank with a guaranteed funding source. The idea of variable interest rates initially encountered resistance but was eventually adopted, if only for credits financed by external loans. B. Project Objectives 2.12 The Guayaquil Urban Development Project is described by the SAR as a multi-component, multi-institution effort to alleviate poverty. It was a pioneering operation in the urban sector in Ecuador, designed to initiate and/or expand specially targeted programs of employment generation and residential development for low-income families that were expected to directly benefit about 1OZ the city's population. Specific objectives were to: (a) provide financing for small-scale enterprises; (b) develop the Municipality's technical, financial, management, and cost recov2ry capabilities, enabling it to better serve low- income residential areas; and (c) assist BEV in financing appropriate and affordable housing for low-income families. 2.13 NLHP took a more narrowly focused sectoral approach, aiming essentially at providing new housing and improving shelter conditions of low- income urban dwellers. Its main policy objective was to support the introduction of low-cost housing finance mechanisms for the urban poor, a segment previously neglected by the official housing bank. It also sought to support national urbanization priorities by directing a significant portion of the loan funds to secondary cities to help reinforce their role in the urban hierarchy. 2.14 The promotion of affordable housing for the urban poor was an innovative approach in Ecuador. As elsewhere, it was to be achieved through the simplification of design standards -- both for infrastructure requirements and in terms of the size and degree of completion of housing solutions -- in favor of progressive or phased development based on self-help and mutual aid for shelter improvement. The other major innovation centered on the redesign of housing finance instruments making them better attuned to the financial capacity of the target population. This also meant changing the approach to shelter investment, moving away from land development and completed housing to sites and services and home improvement loans. 7 C. Proiect Components and Institutional Responsibilities 2.15 Even though they focused on the same general target populations, the two projects followed significantly different approaches in pursuit of their objectives. The Guayaquil operation consisted of an integrated effort to address poverty problems through income generation and squatter upgrading, urban infrastructure, and home improvement loans. The low-income housing project, in turn, was limited to shelter investments, but had a broader territorial scope, involving secondary cities throughout the country. While the Guayaquil project tried to combine the efforts of several institations in addressing the problems of specific low-income neighborhoods, NLHP attempted to develop an innovative approach to shelter finance, operating through two closely interlinked national institutions. 2.16 The components of the Guayaquil Urban Development Project were: (a) Suport to Small-Scale Enterprises (US$10.8 million). An existing line of credit operated by the Banco del Pacifico was to be expanded to enable it to provide an additional 6,400 loans to small artisans and petty commercial activities. The small-scale entrepreneurs to be benefitted with these loans would also be eligible for technical assistance from the National Center for Small Industry and Handicrafts, CENAPIA. Total beneficiaries were expected to include some 5,200 enterprises, or 16,000 individuals (owners and workers), in the city's low-income residential areas. (b) Services to Low-Income Areas (US$24.2 million). Following a multi- component approach aimed at improving living conditions in several of Guayaquil's poorest neighborhoods (Guasmo Norte, Masapingue, Floresta, and Alegria), sub-components included: (i) squatter upgrading, comprising land regularization, landfill, and basic infrastructure for Guasmo Norte and Masapingue to be developed with active community participation; (ii) serviced plots for a total of 3,600 families in Alegria and Floresta; (iii) off-site infrastructure to serve areas adjacent to these projects; (iv) community facilities, consisting of five primary schools, three health centers, and five markets to be built in the neighborhoods to be upgraded; (v) technical assistance, including consultant services for project design, engineering, tendering, and supervision, plus a study of the city's financial management; (vi) administration, including funds to support the operation of a special project unit in the Municipality and to partially fund computerization of the Property Registry Office, which was considered a major bottleneck in the city's chaotic land registration situation; and (vii) community development, consisting of grants to innovative programs for the production of locally demanded goods and services, with the assistance of social workers. (c) Housing Loans (US$16.6 million). This component, to be carried out by BEV, was to consist of 9,700 small loans for the improvement of existing homes and construction of new housing in the four 8 neighborhoods to be upgraded by the project and other low-income areas of the city. 2.17 The project was restructured in March 1984 in an attempt to expedite disbursements for slow or non-performing components. Reformulation involved: (i) a scaling down of the Municipality's components and an increase in BEV/JNV's responsibilities; (ii) elimination of most of the small-scale enterprise component; (iii) increased technical assistance to the Municipality to improve its financial management and administrative procedures; and (iv) allocation of a small amount of loan funds to rehabilitate infrastructure damaged by flooding. The loan closing date was also extended by two years, from December 1984 to December 1986. 2.18 The Low-Income Housing Project, in turn, was designed to support shelter and infrastructure construction in low-income areas in Quito and nine secondary cities. Its components were: (a) New Housing (US$60 million): construction of about 8,200 housing solutions, comprising serviced lots and two types of basic housing units (18 m2 and 36 m2 in area), together with associated infrastructure. (b) Home Improvement Loans (US$20 million): provision of about 7,500 small loans (ranging from US$170 to US$3,300) to finance new homes or improve basic units; the loans were targeted on families in low- income areas and had a 5 to 15 year repayment period at an initial interest rate of 15% (i.e., the standard BEV rate). (c) Urban Upgrading (US$5 million): provision of legal tenure and basic infrastructure to about 2,400 households in the target cities. (d) Community Facilities (US$8 million): financing for ten schools, seven health centers, nine day-care centers, fourteen community centers, and one police station to be built in the project areas. (e) Project Management (US$7 million): funding for project unit operational costs including salaries, vehicles, and equipment. (f) Technical Assistance (US$2 million): preparation of studies for project management and design, low-cost housing technology, and related concerns (210 consultant days), staff training, and computer equipment for BEV/JNV. 2.19 NLHP was first amended in December 1984 to expand its scope to include sub-loans for families affected by floods earlier that year and extend its coverage from 10 to 15 cities. It was amended a second time in August 1987 to allow for the reconstruction of residential dwellings and municipal buildings damaged by a severe earthquake. The municipal buildings in question were located in the towns of Quijos, Chaco, and Borja. Other rehabilitation activities, including the extension of grants and sub-loans to affected households, were directed to the provinces of Napo, Pichincha, Imbabura, Carchi, and Pastaza. 9 2.20 The principal implementation responsibilities in the Guayaquil operation are indicated below. JNV and BEV also had similar functions in NLHP. (a) The Municipalitl of Guayaquil was responsible for providing infrastructure, upgrading the four project areas, and establishing cost recovery mechanisms, in addition to coordinating both physical development and income generation activities in its jurisdiction. (b) CENAPIA, which was essentially a training agency, was made responsible for delivering technical assistance and processing the selection of small-scale enterprises to benefit from loans to be granted by the Banco del Pacifico, a private institution which had been successfully operating its own line of credit for small enterprises and artivans since 1973. (c) The Central Bank of Ecuador was in charge of coordinating execution of the Guayaquil project at the central government level and operating the rediscount window for the small-scale enterprise component. (d) The Ministry of Public Works was assigned responsibility for the construction of two storm drainage canals, upgrading 12 km of roads, and preparing a master sanitation plan for Guasmo. (e) The Ministries of Education and Health were placed in charge of project schools and health posts. However, the involvement of these two ministries never materialized due to their lack of interest in the project. (f) JNV, the National Housing Board, was responsible for setting public policy and carrying out housing construction, including new housing components. (g) BEV, the Ecuadoran Housing Bank, was responsible for the provision of home improvement loans and the financing of low-income housing. III. PROJECT IMPLEMENTATION AND RESULTS A. Factors Contributing to Component Cancellation and Execution Delays 3.01 The Guayaquil project, in addition to failing to achieve the goals of several of its components, was subject to delays that caused implementation to take eight instead of the five years originally programmed. Disbursements were extremely slow, having reached only US$6 million by the end of the sixth year of project implementation. By the final closing date, only US$25 million, or 83% of the loan, had been disbursed and the rest was subsequently cancelled. 3.02 The National Low-income Housing Project, while also presenting some delays, was implemented much more expeditiously than the Guayaquil operation due to the much stronger commitment of BEV/JNV management to project objectives and their adequate response to problems arising during execution. Overall, NLHP was completed with just a one year delay, having exceeded both original new housing 10 and home improvement loan targets. By the time of completion, 94% of loan proceeds had been disbursed. The rest were cancelled because of the Bank's reluctance to grant a further extension since a new operation with BEV/JNV was already underway. 3.03 The Guayaquil project was riddled with difficulties of various types, ranging from administrative and operational to financial and legal problems. The majority of these can be attributed either to institutional weakness or insufficient commitment to the project on the part of the Municipality. Political instability and discontinuity in local administration -- Guayaquil had six different Mayors between 1980 and 1986 -- were a significant cause of implementation delays. 3.C4 Start-up problems in Guayaquil included a delay in setting up the Municipality's project unit, which was a condition of loan effectiveness, as well as in defining the legal status and degree of administrative autonomy of this unit. After these initial problems were resQlved, others arose in the execution of contracted works. This was the case, for example, with the Alegria sites and services scheme, where a dispute between the contractor and the firm hired to supervise construction disrupted execution of the subproject. The latter systematically blocked approval of disbursements based on alleged contractor non- compliance with technical requirements for the infrastructure works. This situation dragged on for nearly two years, leading to a legal dispute and causing the contractor to eventually give up the construction contract altogether. In the end, infrastructure works at Alegria were never completed and the site was invaded by nearby families. 3.05 Legal problems resulting from the nature of the municipal regime in Ecuador, which allows only limited autonomy to local governments in matters such as contracting, budgeting, and human resource management, also posed serious difficulties for the project. Municipalities are subjected to administrative control by both the "Contraloria General de la Republica" and the "Procuraduria General de la Republica." 3 This means that all contracts over a certain amount must be submitted for central government scrutiny Rrior to being issued by the municipalities. This routinely caused the contracting process to last over a year, which, compounded by internal disagreements within the municipal government itself, delayed procurement even further. According to the PCR (para. 37), such problems were not adequately foreseen by the Bank at the time of appraisal. 3.06 The political climate in Guayaquil, moreover, was never favorable to the project. Constant rivalries between the municipal executive and the local Council prevented even the simplest administrative acts from proceeding smoothly. The Council not only exercised its authority to review each contract, but created its own parallel procurement committee. Agreement was even more difficult when it came to cost recovery covenants, which were strongly opposed. The combination of these problems, plus the inability of the Municipality to negotiate a cost 3 Equivalent to the Central Comptroller and Attorney General's Offices, respectively. 11 recovery agreement with final beneficiaries, eventually led to cancellation of the upgrading subprojects and most of the community facilities under the local government's responsibility. 3.07 In the case of the Low-Income Housing Project, delays attributable to the executing agency included the longer than expected period required to comply with loan covenants requiring preparation of a draft financial policy statement. This statement, which was to be adopted by BEV's Board of Directors, involved both a new emphasis on low-income housing and adoption of real interest rates for sub-loans financed with Bank loan proceeds. Covenants also covered the need for an agreement between BEV and JNV regarding their respective responsibilities in project implementation, the requirement that JNV report to the Bank on its construction program during the project period, and an evaluation of sites for new shelter construction under the operation. Given the absence of a Planning Department in JNV and the agency's general lack of long-term planning for construction operations, compliance with these requirements took more time than anticipated, causing a delay of about six months in loan effectiveness. 3.08 Other relevant factors mentioned by the PCR in the case of NLHP were protracted bureaucratic procedures between BEV and JNV, delays in the production of final designs by JNV, and difficulties in the internal technical decision making process in both institutions. By mid-1985, under Bank insistence, a Project Manager with seniority in the BEV hierarchy and a Senior Planning Assistant, to organize planning functions within JNV, were appointed, thus solving many of the project's internal management problems. B. Financing and Costs 3.09 Financing of both operations was affected by endogenous and exogenous factors, resulting in cancellation of parts of the corresponding Bank loans. In the Guayaquil project, negotiated at US$31.0 million, US$5.2 million was cancelled by the time of closing despite the attempt to expedite its implementation by expanding BEV/JNV's components and eliminating or reducing those under the Municipality's responsibility. 3.10 NLHP, in turn, was appraised at US$35.7 million. It closed one year behind schedule, with total disbursements reaching US$33.7 million and the eventual cancellation of US$1.9 million. The following table summarizes programmed and actual costs for the two operations by component. 3.11 A major factor in the significant cost underruns in US dollar terms in both projects was the substantial devaluation of the sucre during their implementation periods. The exchange rate, which was s./ 25 - US$1 in 1978, had reached s./ 114 - US$1 by 1988. In the case of NLHP, this did not affect the achievement of project physical goals, but did put a strain on local counterpart funding, forcing BEV to seek additional central government budgetary assistance. The Central Bank also assumed the foreign exchange risk for the operation. 12 I. Guayaquil Urban Development Project Estimated and Actual Costs (US$ millions) Components Planned % Actual I Small-Scale Enterprises 10.8 20.9 0.2 0.5 Service to Low-Income Areas 24.2 46.9 22.7 61.4 Housing Loans 16.6 32.2 14.1 38.1 Total 51.6 100.0 37.0 100.0 II. National Low-Income Housing Project Estimated and Actual Costs (US$ millions) Components Planned % Actual % New Housing 58.6 58.4 24.5 48.8 Home Improvement Loans 20.9 20.8 40.2 46.2 Upgrading and Community Facilities 10.9 10.9 0.8 0.9 Technical Assistance 2.3 2.3 2.9 3.3 Project Management 7.7 7.7 0.8 0.9 Total 100.4 100.0 87.1 100.0 Average Cost Per Solution (US$) Planned Actual % Change New Housing 7,146 8,862 +24.0 Home Improvement Loans 4,976 2,083 -58.2 13 3.12 Inflation played an important role in local cost increases, causing project costs to more than double in local currency terms. This was particularly painful for the Municipality of Guayaquil, whose revenue sources were limited. BEV, however, was able to reallocate funds internally from other programs in order to maintain loan disbursement schedules. C. Performance by Component 1. Small-Scale Enterprises 3.13 Despite the well-known problems of implementing informal sector support programs, including difficulties in reaching intended beneficiaries, high administrative costs in operating such credit lines, and frequently inappropriate technical assistance arrangements, Guayaquil had in its favor the positive experience already underway at the time the first urban project was prepared. This consisted of a line of credit operated by the Banco del Pacifico (BP), a national bank based in Guayaquil, for a program to support small-scale enterprises and artisan production. This program dated back to 1973 and, in addition to the credit component, provided training and other forms of technical assistance to participants. The special credit program, which charged only slightly below market interest rates, was regarded by BP both as a way of promoting economic development in poorer areas of the city and of building up future clientele. 3.14 In short, some of the most common problems encountered in attempts to establish informal sector support programs, particularly finding appropriate institutions to carry them out, were already resolved. The only requirement for the project was to provide sufficient funding to expand existing BP operations and attract other local banks to the same activity. Yet, the component was a resounding failure, being suspended after just 97 loans (for a total of US$60,000) were made and only a small number of dubiously effective training courses had been delivered. 3.15 The main reasons for this outcome were: (a) The design of the prograr. was overly complicated. Processing of each loan required the participation of three institutions. CENAPLA, as the government agency in charge of beneficiary selection and training, had the role of identifying individual enterprises and assisting them to make loan requests. These were then sent to BP, which reviewed and submitted them to the Central Bank for approval. The Central Bank, which operated the rediscount window through its Guayaquil office, insisted in examining the complete documentation for each loan before returning it to BP for the actual lending. In addition, participants in the program were required to attend training seminars by CENAPIA as a precondition for receiving the loan. (b) The interest rate spread was inadeauate. The spread permitted by the lending facility, representing the difference between the interest rate charged to borrowers and the rediscount rate offered 14 by the Central Bank, was only 2%. BP, however, deemed this insufficient to cover administrative costs. (c) Selection of beneficiaries was made by CENAPIA based on criteria which did not always meet the requirements of the Bank of the Pacific. This was one of the principal reasons why BP decided to give up the program. (d) The training provided by CENAPIA was inadequate both because it was delivered in a classroom environment rather than as on-site technical assistance, which the agency was probably incapable of delivering, and because the training provided was often incompatible with the educational level of the trainees. (e) The involvement of other private banks did not materialize. Despite efforts by the Central Bank to involve other banks in Guayaquil, they opted not to participate because of their lack of experience in this area, the unattractive rediscount conditions, and their reluctance to engage in a program that, in addition to the normal difficulties of lending to the informal sector, was perceived as entailing too many bureaucratic requirements on the part of the public agencies involved. 3.16 By July 1982, two years after loan effectiveness, the Bank of the Pacific officially communicated to the Central Bank its decision to stop utilizing the rediscount line for small-scale enterprise loans in connection with the Guayaquil project. CENAPIA continued its training program, despite severe funding problems, but no significant impact of this effort has been reported. 3.17 Above all, this failure exemplifies how overly complex institutional arrangements and excessive governmental bureaucracy can frustrate an operation which attempts to built on successful existing experience. Had the Bank loan been channeled by more direct means to Banco del Pacifico, allowing it to continue operating its on-going line of credit with minimal disruption, and had the interest rate spread been more favorable, the component could have worked satisfactorily. To this day, BP continues its own credit assistance program for small-scale enterprises. 2. Neighborhood Improvement 3.18 Responsibility for execution of this component of the first operation was assigned to the Municipality of Guayaquil, which was to be responsible both for direct implementation of the infrastructure segments of the upgrading program and coordination of the other agencies involved in this part of the project. The Municipality was to take charge of the pilot upgrading projects in Guasmo Norte and Masapingue and the sites and services scheme at Alegria, as well as to build the associated community facilities. 3.19 A variety of implementation setbacks led to cancellation of most of these programmed improvements. The proposed upgrading of Guasmo Norte and Mapasingue depended both on prior land regularization (i.e., lot definition and tenure arrangements) and the community's agreement to bear part of the cost of 15 the works. Land regularization was delayed by legal problems involving both national legislation concerning land expropriation for social purposes, which was under debate in Congress at the time, and difficulties in improving the filing systems of the Property Registrar Office (Corte Superior de Justicia) in Guayaquil, which was responsible for land registration matters. 3.20 For three years, the city attempted to negotiate an agreement with community leaders on the basis on which the upgrading activities at Guasmo Norte and Masapingue would be carried out. This implied an agreement by beneficiaries to pay back part of the cost of the works to be undertaken. However, according to a participant observation study conducted by a Bank consultant who lived in the project area for about 18 months, the Municipality dealt almost exclusively with community leaders who "misrepresented the project to their followers due to their wish to obstruct any governmental undertaking which they saw as jeopardizing their own hold on the population and as inimical to the interest of their own party." 4 The consultant suggests that the Municipality should have tried to circumvent these leaders by talking directly to the community. Because of the lack of agreement, the Municipality concluded (incorrectly according to the Bank consultant) that the community was not interested in the scheme and, thus, suspended the Guasmo Norte project. 3.21 The Alegria sites and services component was also a disappointing experience. The Municipality was responsible for developing and implementing the infrastructure part of the project, with housing to be constructed later by JNV. Works started in June 1985 with an expected execution period of 16 months. Insufficient counterpart funding on the part of the Municipality and a dispute between the contractor and a supervising company ultimately caused abandonment of part of the project. Even though BEV/JNV did eventually build low-income housing on another part of the site, much of the original area was invaded by nearby families. Some of these were intended project beneficiaries who are now living on the site with neither infrastructure nor land tenure. 3.22 At the same time, another upgrading component was being developed at La Floresta under the responsibility of BEV/JNV. This was a sites and services project having a minimal housing solution consisting of flooring, columns, and a roof (from which its name, "Piso-Techo," was derived). This was a design innovation for both institutions, which were accustomed to building and financing finished houses. This subproject had only minor implementation problems, even though many units remained unoccupied for some time due to the higher expectations of their comparatively well-to-do beneficiaries. In this case, furthermore, there was no need for lengthy negotiations with the community since beneficiaries recognized that BEV was a financial institution that operated exclusively through loans. 3.23 The repeated implementation delays in the upgrading components under the Municipality's responsibility, together with the aforementioned non- performance of the small-scale enterprise component, led to a restructuring of 4 See Salmen, Lawrence F., "Participant-Observer Evaluation of Urban Projects in La Paz, Bolivia and Guayaquil, Ecuador," World Bank Discussion Paper, Report No. UDD-36, November, 1983. 16 the loan in 1984 through which the Guasmo Norte and Masapingue schemes were dropped and projects under BEV/JNV responsibility were expanded to include Sauces VI and VII, both consisting of slightly larger versions of the "Piso-Techo" design introduced at La Floresta. 3.24 According to the PCR, a total of 3,200 families, out of the 6,400 households initially projected, were benefitted by upgrading works. On the other hand, 5,400 households, as compared with an appraisal estimate of 3,700, received sites and services solutions. The much larger number in sites and services plots were accommodated in the areas added after project reformulation. 3. Home Improvement Loans 3.25 This component can be considered the most successful and innovative aspect of both operations. The key objective of this form of financing was to circumvent land tenure problems in addressing shelter problems, while offering assistance to households whose incomes were too low to allow them to qualify for regular shelter loans. Thus, loans were extended to families living in sub- standard conditions to finish or expand small or unfinished dwellings, many of which were originally supported by BEV through its low-income housing credit line. 3.26 The component's success can be illustrated by its quantitative achievements: a total of 7,900, out of an expected 9,200 loans were granted in the Guayaquil project, and 19,300 loans, compared with a programmed 4,200, were made through NLHP in 15 cities around the country. The latter increment was funded through the reallocation of resources from other components of the operation (especially community facilities, but also new housing and even project management), as suggested in the table comparing estimated and actual project costs above. In addition, as the cost table also indicates, the average size of home improvement loans in US dollar terms under NLHP was only 42% of that anticipated at appraisal, while the actual unit cost of new housing solutions was 24% higher than estimated at appraisal. More generally, home improvement loans proved to be an effective means of reaching a segment of the population formerly marginalized from public sector shelter programs. BEV adopted a policy of backing such loans with mortgages (for those able to prove ownership of their properties) or providing funds on an individual basis (for those who could not), in which case two guarantors were required. 3.27 This proved to be a responsive financial instrument for BEV, which was thus able to multiply the number of beneficiaries of its programs with relatively low incremental administrative costs. The loans have a quick turnover and an excellent repayment record. The program, which was initially regarded as a small variant of its mainstream operations, has now became a significant part of BEV's portfolio. This can be considered a direct consequence of the project. Other international agencies, particularly USAID, actively support this modality, which provides additional testimony to its success. 3.28 There are strong indications, however, that a sizeable portion of tne loans were extended to middle income families, many of which occupied houses already financed by BEV. Evidence of this can be found in the quality of houses remodeled using home improvement loans, many of which display quite sophisticated 17 standards. Although technically not constituting a breach of the Loan Agreement, this suggests that a problem of mistargeting a socially oriented program may have occurred. Extending soft loans twice to comparatively higher income families is not consistent with the underlying philosophy of the projects. 3.29 BEV argues that, due to inflationary effects on nominal incomes and their comparative economic advancement, many families qualified both for the initial low-income housing and later also for home improvement loans strictly within the program rules. While this may be true, the lack of instruments to cross-check real family incomes and incentives to BEV regional managers to promote the loans suggest otherwise. In addition, local BEV agencies were granted considerable leeway in deciding which families should receive such loans and it appears that the project's standard eligibility criteria were not always rigorously followed. 4. Low-Income Housing 3.30 Construction and financing of affordable shelter for low-income populations was initially intended to be the main component of NLHP, an innovation which was successfully tested in the earlier Guayaquil project. The novelty in this approach lay in the introduction of the affordability concept. This essentially corresponded to a simplification of design standards and a lesser degree of completion of the housing solutions offered by BEV/JNV in order to reduce their prices to a level which low-income families could afford. The target group consisted of families between the 20th and 50th percentile of the urban household income distribution. 3.31 Through NLHP 8,796 families, as compared with an appraisal projection of 8,200, were benefitted with low-income housing loans. This figure includes a small portion of sites and services. Physical targets were achieved with a cost underrun in dollar terms of about 27%. This was largely due to accelerated devaluation of the sucre, especially after 1985. The scope of both the new housing and home improvement loan components was expanded twice, first in 1984 to enable BEV to assist the victims of floods in coastal cities and a second time in 1987 to address damage caused by a major earthquake. 3.32 NLHP had only minor implementation problems caused essentially by institutional changes at the beginning of the project. The causes of delay, according to the PCR, were protracted national procurement procedures, delays by local administrations in approving infrastructure designs, delays in land acquisition, and the above-mentioned natural disasters. Other than the need to simplify the design of its shelter solutions, this component did not pose major difficulties for the two executing agencies, which had been operating in the sector for many years. This was the principal reason for its expeditious implementation and suggests that the most effective strategy for introducing innovation is through moderate incremental change. 3.33 As concerns the component's social impact, questions must be raised about the rules of the program which allowed housing subsidies to be granted to relatively well-to-do families. In fact, the basic philosophy of the operation was to provide assistance to families living in sub-standard shelter whose incomes were insufficient for them to qualify for finished housing solutions. 18 Program regulations, however, were designed to accommodate other cases, such as funding the completion or expansion of unfinished ("piso-techo") or small houses also financed by BEV. 3.34 Physical evidence that dwellings with quite sophisticated finishing have been financed or that entire core houses have been torn down to be replaced by new, more luxurious ones are an indication that beneficiaries did not come exclusively from lower income strata. This was particularly evident in cities like Cuenca and Quito where the housing market is very active. One reason for this is the quick appreciation in the value of the land developed by JNV, thus making the price of the original units and the long-term conditions for payment a very attractive investment. BEV admits that an unspecified portion of the home improvement loans, varying from city to city between 20 to 50% of the total, was extended to families that &lready lived in BEV-financed homes. Beneficiaries frequently used these loans to expand their houses or to refine their exterior finishing. 3.35 While there is nothing improper about this practice (provided that the families concerned were in the expected income range), it clearly deviates from the social objectives of the public housing system, especially the segment to be supported with project proceeds. Families with incomes higher than those of the intended target group may have understated their real earnings in order to become eligible for subsidized loans. In the case of second loans to middle- income families, even though no clause prohibited this practice, it nevertheless constitutes a case of mistargeting since soft loans were granted to a better-off group, thus preventing needier families from receiving these benefits. 5. Technical Assistance 3.36 Technical assistance components in both projects aimed at building the capacity of implementing agencies and improving their internal planning and administrative procedures. In the case of Guayaquil, there was both a requirement in the form of a dated loan covenant and financing for a study on municipal financial management. However, the Municipality delayed contracting the study until there was not enough time in the project's implementation period to complete it despite the Bank's insistence on its importance. Although a study on the Municipality's urban development regulations was carried out, its recommendations were only partially adopted. Two other studies concerning specific local activities (a drainage plan for Guasmo and feasibility studies for sites and services and employment generation subprojects) were prepared, but their impact was limited because the respective components were not fully implemented. 3.37 The pattern seems to be that whenever the proposed studies or direct assistance proposals involved the core of the Municipality's operations, they were rejected or postponed. Thus, negotiations with a technical assistance firm to implement an administrative restructuring of the municipal Finance Department, which were actively encouraged by the Bank, were blocked by the Mayor in office at the time as being "too expensive" despite available financing through the project and the Bank's assessment that the proposed costs were reasonable. 19 3.38 Technical assistance in connection with NLHP, in turn, included consultant services, staff training, and improvement of the compute.: systems within the executing agencies. Specific studies in the areas of urban development programs, housing improvement, demand study methodology, social impact assessment methods, loan administration, portfolio financial management, and municipal development and neighborhood improvement methodologies were financed under this component. The majority of these studies have reportedly been instrumental in the preparation of more recent loan operations by BEV, as well as contributing to refine several aspects of project execution itself. The staff training sub-component was not implemented, despite the availability of funds, due to the resistance of top management to releasing personnel even for short-term training abroad (which was considered to be a "privilege"). Improvements in BEV's computer systems, however, were satisfactorily introduced. D. Economic Evaluation 3.39 According to the respective PCRs, both operations had relatively high re-estimated economic rates of return (ERRs). For NLHP, the ex-post ERR was estimated to be 20 percent, as compared with an appraisal calculation of 18 percent. There were some differences in the methodologies used for the two estimates, however, as a single calculation, covering all project components, was made by BEV for purposes of the PCR, while the appraisal report presented separate figures for the operation's three main components: new housing, home improvement loans, and urban upgrading. In both cases, however, benefits were estimated on the basis of imputed rental values for the types of shelter solutions provided under the project, with the ex-post calculation utilizing a weighted average of rent figures at the national level. 3.40 The PCR for the earlier Guayaquil project, in turn, indicates a re- estimated ERR of 22 percent, as compared with appraisal estimates in the range of 17-19 percent. However, it is important to point out that the ex-post figure refers only to part of BEV's investments (i.e., the completed parts of the sites and services component, excluding one subproject for which the final sales price of the units installed had not yet been determined), whereas the appraisal estimate referred to programmed project investments as a whole. No re-estimated ERR was presented for the Municipality's original components since they were never implemented, while part of its allocation was utilized for works (flood control) not initially contemplated in the operation. In addition, rapid inflation during the project implementation period reportedly made it impossible to accurately determine final costs, while the illegal occupation of a large part of one of the sites and services schemes means that there has been no cost recovery at all from this component. Accordingly, the actual rate of return on project investments as a whole may be substantially lower than the figure reported in the PCR. 3.41 The high re-estimated ERRs for project housing improvements are nevertheless consistent with the results of other Bank-assisted urban development operations for which a similar approach to estimating the net benefits of low- income shelter investments has been taken. In most of these cases, ex-post ERRs are well above the Bank's standard 10 percent threshold for acceptable rates of return, while, in many instances, the re-estimated figures are also higher than the initial appraisal projections. In Ecuador, as elsewhere, this appears to be 20 at least partially due to the relative scarcity of serviced land and adequate housing in rapidly growing urban areas, revealing the existence of considerable unmet demand -- and willingness to pay -- for such investments. In short, the economic "success" of many Bank-supported urban development projects, including the two operations presently under review, may be as much a function of the nature and speed of the urbanization process in its borrowing member countries as it is of the "goodness" of the investments they support per se. E. Financial Performance 3.42 Neither the PCRs nor the project files provide detailed information on project financial performance. Nevertheless, as with various other aspects, it is evident that there were substantial differences between the two operations. While virtually no cost recovery was obtained from a substantial part of the sites and services developed under the Guayaquil project, NLHP was considerably more successful in this regard. According to the PCR (pa=a. 30), only 4 percent of the mortgage loans extended by BEV under NLHP were reported to be delinquent at the time of completion, while just 10 percent of the beneficiaries receiving "signature" loans (i.e., loans not secured against real property) were in arrears. Overall, cost recovery was reported to be on the order of 83 percent of the total due, as opposed to an appraisal target of 90 percent. No information was presented in the PCR on the length of time the various families in arrears had been in this situation, their distribution among the participating cities, or what actions had been or were being taken by BEV to deal with this problem. 3.43 NLHP's much more favorable cost recovery performance when compared with the Guayaquil project reflects both different practices with respect to repayment on the part of BEV, on the one hand, and the Municipality, on the other, and changes in BEV's own financial policies over time. To some extent, it may also reflect differenceb in income levels of the target beneficiaries in each case. Unfortunately however, this hypothesis can not be fully tested with the data available since there is no breakdown of delinquent loan recipients by income. BEV's experience is nonetheless suggestive since it is likely that families receiving "signature" loans generally have lower income levels than those receiving mortgage loans. In any case, the reported delinquency rates are lower than the average for Bank-assisted low-income shelter projects. 3.44 Finally, the fixed interest rates applied to shelter loans under NLHP, which were adjusted twice during the project implementation period, were closer to market rates -- and thus less highly subsidized in conditions of rising inflation -- than those applied by BEV in the Guayaquil project. As a further step, variable, market-based rates have been introduced as a central feature of the follow-on Second National Low-Income Housing Project (Loan 2898-EC), approved in January 1988. The Bank's insistence on these changes in connection with the two low-income housing operations appears to have played a major role in improving BEV's financial situation more generally, although some degree of dependence on central government budgetary allocations still remains. 21 IV. POINTS OF SPECIAL INTEREST A. Poverty Impacts 4.01 Both projects had a strong social orientation. The first operation was described as a poverty alleviation project intended to "initiate large-scale programs of employment generation and residential development of benefit to low- income families in Guayaquil." In this connection, project impacts can be assessed either in term of the number of direct beneficiaries or the potential effects of introducing improved social service delivery and financing practices by the Municipality. On both grounds, the Guayaquil project failed to achieve meaningful results given the early cancellation of assistance to small-scale enterprises after only a small number of loans had been extended and of the upgrading components under the Municipality's responsibility. Likewise, attempts to improve municipal financial management and service delivery did not receive adequate attention from city government. 4.02 The National Low-Income Housing Project, although attempting to address the shelter needs of the poor by reducing design standards and increasing accessibility to housing finance through home improvement loans, has features which are incompatible with a poverty alleviation program. Eligibility limits for both new houses and home improvement loans were set at the 50th percentile of the urban household income distribution. This opened up the program to middle class families which have traditionally had priority in BEV/JNV programs. As already noted, many such families have received two loans, one for a basic housing unit and another for its expansion or completion. Although the number of beneficiary families (28,096) exceeded original goals, there is no information on their breakdown by income levels. Given the excessively liberal eligibility criteria and the built-in incentives for BEV to favor the relatively wealthier families in the target population, poverty impacts can be assessed, at best, as marginal. B. Urban Sector Policy 4.03 The Guayaquil project had a clear goal of improving local urban policy formulation and implementation practices. One of its stated objectives was "to strengthen the Municipality's efforts to serve low-income areas by developing its technical capacity, improving its financial management and cost- recovery practices, assisting in the regularization of land tenure." Implied was the assumption that local governments should take the primary role in delivering and financing urban services rather than the national agencies that had traditionally prevailed in this domain. 4.04 Instead, just about the opposite was demonstrated and accomplished. Guayaquil's municipal government proved to be ineffective and unresponsive in terms of addressing local service needs, even in the presence of adequate funding for its projects. It resisted adopting either cost recovery practices or badly needed organizational changes that would have improved its financial performance and, thus, its ability to provide urban services. National agencies, in contrast, revealed greater competence in delivering public works, taking over implementation responsibility from the Municipality when it became evident that the latter would be unable to complete its projects. Against conventional 22 wisdom, the national housing agencies proved more responsive than the municipal government to the needs of the local population, adapting their output to the demands of low-to-middle income families. 4.05 The Municipality's poor performance in the first project does not disprove the general thesis that local governments of large cities should be the main actors in their own development. It does, however, clearly highlight the need to address the institutional weaknesses of municipal governments and to establish incentives for sound financing practices in relation to local service provision. This is in large measure a task for national policies implemented by sectoral agencies with a view toward decentralizing implementation and financing responsibilities. 4.06 As concerns the impact of the housing and infrastructuro built by BEV/JNV in the 15 cities where NLHP was implemented, a point should be made about site selection and the relationship between the two institutions and local governments. A majority of the project sites consisted of properties already owned by the housing institutions. In other cases, sites were selected more in response to economic, than to urban physical development, considerations. Furthermore, operational relations and cooperation between BEV/JNV and the municipalities were minimal given the tradition of self-sufficiency and the associated disregard for local government capabilities on the part of national agencies. Thus, local spatial development considerations were not a conscious concern in project design, nor can any significant impact be identified with respect to the improvement of municipal planning capabilities. C. Housing Policies 4.07 A principal policy objective with respect to the shelter sector was to promote a shift from the construction of middle class housing toward equity- based, low-income shelter solutions and home improvement loans. The latter was a way of reaching the population living in sub-standqrd conditions and facing the prospect of eviction for having settled on unsecured land. These two objectives were fully met, with the corresponding lines of financing presently representing the bulk of BEV's lending operations and JNV's construction activities. The decisive factors in this outcome were the Government's support of socially oriented policies and the commitment of the national housing institutions to these goals. 4.08 Another policy dimension of the second project was the introduction of more sustainable housing finance measures by BEV. Since the time of its creation, the practice of subsidizing housing loans by adopting below-market fixed interest rates had prevailed. Under conditions of low inflation and abundant oil-based annual budgetary transfers, the long-term financial implications of this approach did not pose a major concern for the housing bank. However, with the dramatic change in these conditions during the early 1980's, it quickly became evident that the financial solvency of the housing system would be compromised by such policies. 4.09 The proposal to change lending conditions, mainly through the introduction of variable interest rates, was resisted both inside and outside the BEV/JNV system. The two Bank projects were decisive in breaking internal 23 resistance, both through loan covenants requiring that "real" interest rates be applied to housing sub-loans and by educational efforts on the part of Bank supervision missions. Ne-ertheless, achieving positive real interest rates remained an elusive goal given regulations permitting adjustments only every six months and to a level determined by the "Junta Monetaria Nacional." Variable interest rates are nevertheless the rule today for loans funded with external credits, which constitute the majority of BEV's operations. Subsequent projects financed by international agencies contained similar conditionalities and, at present, there is an internal debate on adopting the practice for loans financed with passbook savings. Although the housing system still depends heavily on budgetary transfers and its lending rates continue to be below-market, an increasing portion of its funding derives from repayments of BEV's outstanding loans. D. Institutional Development 4.10 Institutional development constituted an explicit objective of both operations. It was to be promoted both through the more advanced internal operational practices implied in the design of project components and through direct technical assistance. In the case of Guayaquil, as mentioned above, efforts to strengthen the Municipality's operational capabilities by introducing cost recovery prsnti;e and financing technical assistance for internal reorganization failed dramatically. Partially as a result, the municipal government continues to be characterized by weak administration, has extremely deficient social and urban services, and lives in a constant state of fiscal crisis. 4.11 As concerns BEV and JNV, in turn, institutional strengthening efforts under NLHP were better rewarded. BEV has begun to employ financial planning practices encouraged and funded by the Bank in the analysis of its portfolio, in making long-term financial projections, and in other areas. The preparation of more recent externally funded operations has continually demanded such financial planning exercises, as well as the general improvement of accounting and cost control practices. With respect to JNV, which acts like a major developer and construction company, the focus of institutional development was on operational planning and control. Most of the achievements in this regard were the result of in-house technical assistance provided by a group of GTZ experts, partially funded by the Bank throughout the project implementation period. Their support was instrumental in improving internal organization, establishing operational planning procedures, which indirectly benefitted project implementation, and preparing follow-on operations. As acknowledged by its own staff, however, JNV still requires streamlining of its structure and more extensive adoption of planning procedures which, for the most part, are currently limited to externally-funded programs. V. ROLE OF THE BANK 5.01 Despite a period of nearly two years between identification and negotiation of the Guayaquil project, both the PCR and local officials seem to agree that appraisal was carried out prematurely and overlooked basic deficiencies in the design of various components and in the Municipality's implementation capacity. Final designs were not available prior to 24 implementation for the Municipality's components. Similarly, at appraisal, the Bank and the Municipality misunderstood military maps that showed the elevation of one of the upgrading sites (Guasmo), resulting in severe underestimation of the amount of landfill required to eliminate flooding. As a consequence, in addition to the difficulties encountered by municipal staff in convincing the community to pay for the works, it turned out to be impossible to eliminate flooding at the site on a cost recoverable basis. 5.02 Insufficient attention was given at appraisal both to the evident deficiencies of the Municipality, especially its chronic underfinancing, overstaffing, and lack of technical capacity, and to the limitations imposed by national law on its possibilities for improving tax assessment and collections. Furthermore, traditions regarding revenue collection and urbanization practices, conflicting relations between Mayors and the city council, and the frequent turnover at the top were characteristics of Guayaquil well known to local observers, which were either not fully detected or were insufficient to discourage the Bank from dealing with the Municipality. 5.03 Although supervision missions later identified the shortcomings of the Municipality's components, they were unable to correct them and were too cautious to take stronger measures such as transferring implementation responsibilities to BEV/JNV sooner and for all of the upgrading works. Even though part of its components were cancelled when the project was restructured, the Bank seemed confident that a technical assistance program for local finance and management, introduced at that time, would improve the Municipality's performance and still have a positive effect on the operation. The lack of interest of subsequent Mayors in this program resulted in its abandonment and only reinforced unfavorable local perceptions about the poor quality of municipal administration. 5.04 Supervision missions, moreover, either failed or did not have the power to impose changes in the operation of the small-scale enterprise component. This could have been done by negotiating simplified loan approval procedures with the Central Bank and removing CENAPIA from the critical path in program implementation. More decisive action on the Bank's part might have saved this promising component. 5.05 Problems in NLHP were less significant and resulted in time overruns which did not compromise overall project performance. Final designs were not available at the time of appraisal and offered inadequate solutions with respect to shelter design and materials; some had to be redesigned and required site adjustments. Insufficient attention was also given at appraisal to the financial difficulties of the Ministries of Education and Health. This compromised the community facilities component, a problem which has also occurred in other Bank projects in Ecuador, including the Guayaquil operation. 5.06 Close and timely supervision of NLHP, however, helped to correct managerial problems and influence BEV/JNV's internal policy decisions regarding interest rates, portfolio management, and internal organizational matters. According to local staff, the Bank's presence and influence have contributed decisively to the major policy shifts occurring in the Ecuadoran housing system over the last decade. 25 5.07 The Bank must also be credited for the decision to finance a fruitful participant observation study in the upgrading areas in Guayaquil. The study occurred over a three year period by a researcher who lived in the communities and was able to assess their perspective with respect to the project. By identifying the flaws both in BEV/JNV's housing designs, which were found to be incompatible with beneficiary expectations, and the Municipality's approach to local communities, (i.e., dealing unsuccessfully with the politically radicalized leadership instead of going directly to the population itself), the study generated a better understanding of the project context. The parallel experiences in Guasmo and Floresta in Guayaquil and a similar project in La Paz, Bolivia contributed valuable inputs for the preparation of future urban development operations. Among other findings, the study emphasized the need to work in close contact with beneficiary families from the outset in order to attain an adequate understanding of their needs and operations and to gather grass-roots support for project goals, improve the design of housing solutions and cost recovery mechanisms, and increase community involvement in the operation and maintenance of local facilities. VI. CONCLUSIONS AND LESSONS A. Conclusions 6.01 The two operations under review had sharply contrasting execution experiences and results. Much of the Guayaquil project, as originally appraised, was never implemented because the Municipality proved unable to carry out many of its initial responsibilities. As a result, the loan had to be restructured in 1984 and several important components (eg. small business support, slum upgrading, community facilities) were dropped. The project likewise failed to achieve most of its institutional, policy, and financial objectives, particularly in terms of strengthening municipal finances and service delivery and recovering investment costs from project beneficiaries. Inadequate cost recovery, together with the persisting weak executing capacity of the Municipality, in turn, has effectively limited both possibilities of project replicability and the likely sustainability of many of the benefits provided under the operation. Only the expansion and comparatively much more efficient implementation of the components under the two national housing agencies, JNV and BEV, saved the project from an even more unsatisfactory outcome. 6.02 NLHP, in comparison, was a generally successful operation. Even though some initial implementation delays occurred and project investments had to be redirected to support reconstruction efforts following major natural disasters in 1982-83 (floods) and 1986 (earthquake), all of the operation's original objectives were substantially met. Unlike the Guayaquil experience, moreover, institutional development goals were mostly achieved, contributing both to the satisfactory implementation of the project itself and preparation of a follow-on operation (the Second National Low-Income Housing Project, approved in January 1988), which is currently under execution. NLHP's success can be largely attributed to a combination of: (i) adequate Bank/Borrower application of both the positive and negative lessons of the Guayaquil operation; (ii) clear and simple project design, which focused on two complementary institutions (i.e., JNV and BEV) that had performed well in the first operation; (iii) supporting Government policies and programs; (iv) good Bank supervision which appears to 26 have provided valuable assistance in problem identification and resolution; (v) effective technical and financial assistance to JNV and BEV more generally; and (vi) flexible and timely responses from the Bank in dealing with two natural calamities and inflation-induced local cost increases and currency devaluations. B. Lessons Learned 6.03 Various lessons of a general nature can be derived from these two projects. The Guayaquil operation confirms the need to avoid excessive dependency on institutionally fragile and highly politicized local governments and for the prior assessment of their administrative and financial capabilities. A second lesson, which is illustrated in different ways by both operations, is the desirability of limiting the sectoral and institutional complexity of urban development projects, particularly when a multitude of executing agencies and two or more different levels of government are involved, as was the case in Guayaquil, but not in NLHP. A third lesson is the importance of attaining sufficient cost recovery in urban shelter operations to ensure project replicability and sustainability, as well as the frequent difficulty of doing so. Beyond these general conclusions, several more specific lessons can be summarized below. 1. Project Design Strategy 6.04 The contrasting performance of the two projects raises issues of design and strategic approach to urban development assistance. As the first urban operation in Ecuador, the choice of a municipality with a record of unreliability, over-political administration, and institutional weakness in retrospect clearly seems to have been a strategic blunder. By limiting the project to a single municipality, potential local development policy effects at the national level were limited, while incentives for eventual competition among cities for project funds were non-existent. Likewise, the multiplicity of components and interdependencies among institutions severely compromised project implementation. On the other hand, the narrow shelter focus of NLHP also restricted its policy impact and did not address more basic urban service and infrastructure financing and management issues. It had the advantages, however, of being simple and of depending essentially on a single institution, thereby greatly enhancing the project's prospects of success. 6.05 The key to this dilemma is to find the proper balance between ambitious policy goals and realistic implementation possibilities. Sector operations can be designed so as to effectively involve local agencies under the condition that they adopt desirable policy measures. As concerns municipal development projects specifically, it is useful to introduce an element of competition, opening them up to broad categories of cities and favoring those that demonstrate good performance or are ready to accept assistance in implementing improved management and service delivery practices. 6.06 The lessons of the early projects have not been lost to the Bank. Some of the design recommendations mentioned above can be found in more recent operations in Ecuador and elsewhere. The latest municipal development project in the country, for example, is centrally administered, but open to all municipalities. It relies on on-lending arrangements which are conditioned to 27 fiscal policy adjustments and the adoption of acceptable cost recovery practices. It also places considerable emphasis on institutional strengthening, a chronic problem for Ecuadoran municipalities. Similarly, housing projects in the country presently give much greater attention to the financial solvency of sectoral agencies, while pushing for more significant organizational changes. 2. Implementation Arrangements 6.07 Even more critical than design features, the experience of these two operations suggests that institutional preconditions may play a decisive role in project outcomes. Institutional commitment to project goals and a viable implementation set-up, which includes adequate funding and administrative autonomy of executing agencies, are among such key pre-conditions. Choosing suitable implementing agencies requires careful ex-ante institutional analysis, with attention given both to their leadership's commitment to project goals and their capacity to carry out the project, as well as to the legal, financial, and political constraints to which they are subjected. This seems to have been largely overlooked in the Guayaquil project. 6.08 Implementation arrangements in both operations included the creation of special project units. This worked well in the case of BEV/JNV which established such units in its branches in all of the cities where NLHP was implemented and later absorbed them into their regular structures. However, in Guayaquil, the project unit was given only limited autonomy, bounded both by existing legal constraints on the administrative acts of all local governments and by further restrictions imposed by centralizing and distrustful Mayors. 6.09 The Guayaquil experience also highlights the importance of having a correct understanding of the socio-political environment in which the project is to unfold. In hindsight, it is clear that the local pattern of informal urbanization and the previously complacent attitude of city government toward this phenomenon posed a serious threat to the Guayaquil project. These constraints, however, were underestimated by the project team, resulting in failure of the upgrading components. This experience also illustrates that preconditions for cost recovery may often be inherently political and, thus, need to be addressed at that level. 6.10 A related issue concerns community participation, especially in upgrading schemes. The city tried for three years to obtain the acceptance of project conditions by local leaders, ultimately to no avail. More direct channels of communication with intended beneficiaries might have saved the aborted Guasmo project. 3. Approaches to Institutional Development 6.11 The projects present several lessons with respect to institutional strengthening efforts in urban development operations. First, concerning the greater effectiveness of long-term capacity-building, as opposed to short-term interventions, GTZ's contribution through an in-house technical assistance team proved to have lasting effects on BEV/JNV. Second, technical assistance is best absorbed when in tune with changes desired by and acceptable to the beneficiary institution. The Municipality's resistance to internal reforms and the adoption 28 of new service provision practices in Guayaquil illustrates this point. Third, organizational stability is important, given the crucial role of continuous support from top management for the implementation of meaningful institutional change. Again, the BEV/JNV case, representing a stable organization that was able to incorporate several institutional advances derived from technical assistance, contrasts with that of the Municipality of Guayaquil with its extreme instability, which di4 not. 4. Assistance to Small-Scale Enterprises 6.12 In the case of this normally difficult component, the Guayaquil project missed an opportunity to set in motion a joint public-private initiative which had most of the ingredients for success. However, the complexity of its implementation arrangements and insufficient compensation to private banks caused the latter to abandon the effort. Future efforts in channelling governmental support to small-scale enterprises should, nevertheless, follow the basic lines of this component, provided minor adjustments are made. 6.13 Private banks should be encouraged to participate in the implementation of soft loan credit facilities, since this reduces overall administrative costs, as compared with the alternative of their provision exclusively by public agencies. Savings derive from the greater specialization of banks in financial activities and the possibility of spreading costs across a variety of different types of operations. But such participation requires reasonable compensation in the form of rediscount rates which cover administrative costs and, in some cases, direct subsidies for training and other outreach activities. 6.14 Precautions should be taken to avoid unnecessary bureaucratic delays and minimize interference in actual loan processing, as these ultimately represent additional costs. In cases where public agencies are involved, preferably in beneficiary identification and training rather than direct lending activities, their support should be complementary to the strictly financial side of the operation. This would avoid the error made in the case of Guayaquil where CENAPIA was made responsible for the selection and training of beneficiaries and whose poor performance compromised the entire component. 5. On Targeting Social Programs 6.15 The problems of mistargeting in social programs are well known, but difficult to avoid. In cases similar to the home improvement loans component, under NLHP, where significant deviation of project assistance away from the intended low-income target groups appears to have occurred, more careful design, together with better supervision, could have avoided the problem. 6.16 As concerns household income levels, upper limits should be lower than those in NLHP which contemplated families with household earnings of up to the 50th percentile in the local income distribution. This allows too wide a margin and does not address the problems of the most needy, whose requirements are for the improvement of sub-standard dwellings, not second loans to embellish already subsidized housing. A condition to prevent higher-income families from 29 receiving two consecutive loans should be introduced in future projects to avoid an excessive concentration of subsidies. 6.17 An alternative approach would be to pre-select low-income neighborhoods where home improvement loans would be extended, thus making them part of urban upgrading schemes in clearly defined problem areas. In this way, the target population could be better identified a Priori and the social distribution of benefits can be more readily controlled and evaluated. In either case, a great deal of supervision and community organization are required, both to ensure that household income surveys are adequately carried out and to better identify and assist needy families. - 30 - Annex Page 1 of 3 Comments from the Borrower (Working Translation) 22449 MINFIN ED QUITO-ECU-91-11-19 SCP-MSG-298 MR. YVES ALBOUY CHIEF, INFRASTRUCTURE AND ENERGY DIVISION OPERATIONS EVALUATION DEPARTMENT WORLD BANK 1818 H STREET, N.W. WASHINGTON, D.C. 20433 USA THIS IS IN REFERENCE TO YOUR LETTER OF OCTOBER 18, 1991, RE FINAL REPORT OF LOAN 2135-EC "NATIONAL LOW-INCOME HOUSING PROJECT" TO WHICH I AM PLEASED TO INFORM YOU THAT THERE WE HAVE NO COMMENTS ON THE REPORT. WE TAKE THIS OPPORTUNITY TO EXPRESS OUR APPRECIATION FOR THE SUPPORT PROVIDED BY THE BANK IN THE EXECUTION OF THESE PROJECTS. SINCERELY, LCDO. WILSON ANDRADE PAREDES, ACTING UNDERSECRETARY OF PUBLIC CREDIT WORLDBANK 64145 22449 MINFIN ED -11191023 - 31 - ANNEX Page 2 of 3 COMMENTS FROM THE EXECUTING AGENCY (Working Translation) Junta Nacional de la Vivienda Banco Ecuatoriano de la Vivienda Presidencia November 29, 1991 No. 2520 Mr. Paul Meo Chief, Commerce, Finance & Industry Division Country Department IV Latin America and the Caribbean Region World Bank Washington, D.C. U.S.A. Attn: Mr. Yves Albouy Operations Evaluation Department Dear Mr. Meo: This is to acknowledge receipt of your letter dated October 18, 1991, transmitting the ex-post evaluation report on the Bank-financed National Low-Cost Housing Project (Loan 2135-ED) and the Guayaquil Urban Development Project (Loan 1776-EC). We have read the report carefully and, in general, agree with the findings of the audit which reflect the difficulties encountered during implementation as well as the achievements of the projects which have greater impact. Needless to say, the contents of the report as well as its conclusions will be of great value and will be taken into account by our Institutions, JNV and BEV, so that in the near future these lessons be learned and the institutional procedures improved. Notwithstanding the fact that the results of the evaluation is highly favorable to the Junta Nacional de la Vivienda and the Banco Ecuatoriano de la Vivienda, I wish to make the following clarifications: - 32 - ANNEX Page 3 of 3 1. Paragraph 4.11, Chapter IV, Institutional Development. Both the Junta Nacional de la Vivienda and the Banco Ecuatoriano de la Vivienda are governmental agencies with specific functions and the responsibilities of each entity are clearly defined. Among its various functions, the Junta Nacional de la Vivienda is responsible for developing national housing policies, carrying out feasibility studies and executing social sector housing programs. Thus, the statement made in para. 4.11 that JNV acts as a construction company is not correct. 2. Chapter III, Implementation and Project Results (page 13), states that the lack of a Planning Department at the Junta Nacional de la Vivienda and the general lack of long-term planning had a bearing on the delay experienced in loan effectiveness. In this regard, we would like to say that at the time of project execution, the Junta Nacional de la Vivienda did indeed have an established organizational structure as well as a long-term planning policy; in fact, proof of this is that follow-on projects were financed by the World Bank (Sacond Low-Cost Housing Project - Loan 2898-EC) and the IDB (Urban Development Project - Loan 805-SF-EC). The Planning Department was established in 1980 to carry out studies, plans and programs. At present it has enough in-house technical expertise that make it possible to make projections for the next 10 to 15 years. In view of the above, I would appreciate your incorporating these clarifications in the final evaluation report. Thanking you for your kind attention, I remain, Sincerely yours, (Signed) Arq. Patricio Mena Durini President