Report No Public Disclosure Authorized Document of The World Bank FOR OFFICLAL USIE ONLY. Public Disclosure Authorized

Transcription

1 Public Disclosure Authorized Document of The World Bank FOR OFFICLAL USIE ONLY Report No Public Disclosure Authorized PROJECT COMPLETION REPORT MEXICO Public Disclosure Authorized HOUSING FINANCE PROJECT (LOAN 2947-ME) JUNE 13, 1994 Public Disclosure Authorized Infrastructure and Energy Division Country Department II Latin America and the Caribbean Region This document has a restricted distribution and may be used bv recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.

2 CURRENCY EQUIVALENTS (As of November 1, 1993) Current Unit = Mexican Pesos Nuevos (NP$) US$1.00 = 3.15 Mexican Pesos Nuevos Monthly Minimum Wage (Federal District) = Mexican Pesos Nuevos (as of 1/1/93) ABBREVIATIONS AND ACRONYMS BANOBRAS CETES CPP FONHAPO FOVI FOVISSSTE FSAL INFONAVIT SEDUE SHCP Banco Nacional de Obras y Servicios Pblicos S.N.C. (National Development Bank for Public Works and Services) Certificados de la Tesorerfa de la Federaci6n (Treasury Bills) Costo Promedio Porcentual de Captaci6n (Average Cost of Funds to the Banking System) Fondo Nacional de Habitaciones Populares (Govemment Low-Income Housing Fund) Fondo de Operaci6n y Financiamiento Bancario a la Vivienda (Housing Fund for Commercial Banks) Fondo de la Vivienda del Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado (Housing Fund of the Social Security System for State Service Worker Financial Sector Adjustment Loan Instituto del Fondo Nacional de la Vivienda para los Trabajadores (Institute of the National Housing Fund for Workers) Formerly the Secretarfa de Desarrollo Urbano y Ecologfa (Ministry of Urban Development and Ecology), the Secretarfa de Desarrollo Social (SEDESOL), the Ministry of Social Development Secretarfa de Hacienda y Cr6dito Publico (Ministry of Finance and Public Credit)

3 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C U.S.A. Office of Director-General Operations Evaluarion June 13, 1994 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on Mexico Housing Finance Project (Loan 2947-ME) Attached is the 'Project Completion Report on Mexico - Housing Finance Project (Loan 2947-ME)" prepared by the Latin America and the Caribbean Region. Part n was prepared by the executing agency of the Borrower. This was the first Bank assistance to the housing finance sector in Mexico. The US$300.0 million loan was to support housing finance reforms and to increase the supply of affordable housing for the low-income urban population. The entire loan was disbursed 26 months before the original closing date despite a one-year delay before effectiveness. The project was successfully implemented because "ownership" was ensured through local preparation, enabling regulatory environment, and ability to seek and find practical solutions to problems. The least favorable outcome was the sliding of the targeting from low- to higher-income groups endorsed by the Bank which waived eligibility ceilings. The most encouraging outcomes were the participation of the commercial banks in financing housing, and the deregulation (notably of the steep notary fees on real estate transfers) decided by the goverment in Sector studies led to measures to mobilize savings for housing and to promote rental housing investment. The PCR gives a satisfactory account of the project and the reasons for its success. Overall, the project is rated as satisfactory, its sustainability as likely, and its institutional impact as substantial. America. The project may be audited together with other large housing projects in Latin 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.

4

5 FOR OFFICIAL USE ONLY MEXICO HOUSING FINANCE PROJECT (LOAN 2947-ME) PROJECT COMPLETION REPORT Table of Contents Page No. PREFACE.... i EVALUATION SUMMARY ii PART 1: PROJECT REVIEW FROM BANK'S PERSPECTIVE... I Project Identity... 1 Background Project Objectives and Description... 3 Project Preparation, Design and Organization... 4 Project Implementation... 6 Project Results Project Sustainability Bank Performance FOVI Performance Project Relationship Consulting Services Project Documentation and Data Findings and Lessons Learned PART II: PROJECT REVIEW FROM THE EXECUTING AGENCY'S (FOVI'S) PERSPECTIVE PART III: STATISTICAL INFORMATION Table I Related Bank Loans.. 23 Table 2 Project Timetable.. 23 Table 3 Cumulative Estimated and Actual Loan Disbursements Table 4 Status of Major Loan Covenants..24 Table 5 Use of Bank Resources..26 Ths 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.

6

7 i MEXICO HOUSING FINANCE PROJECT (LOAN 2947-ME) PROJECT COMPLETION REPORT PREFACE 1. This is the Project Completion Report (PCR) for the Housing Finance Project for which Loan 2947-ME in the amount of US$300.0 million was approved in June The loan was closed in April 1992 approximately 26 months ahead of the appraisal estimate. It was fully disbursed and the last disbursement was in April The PCR was jointly prepared by the Infrastructure and Energy Operations Division on the Latin America and Caribbean Regional Office (Preface, Evaluation Summary, Parts I and III), and the executing agency (FOVI), which prepared its own extensive project completion report that the Bank reviewed when preparing the PCR. The FOVI report is summarized as Part II of the PCR, and the full text is available in LAC files. 3. Preparation of this PCR was started in August 1993, and is based, inter alia, on the Staff Appraisal Report; the Loan, Project and Guarantee Agreements; supervision reports; studies carried out under the project; correspondence between the Bank and the Borrower; interviews with Bank staff and with Mexican government officials; and internal Bank memoranda.

8

9 ii MEXICO PROJECT COMPLETION REPORT HOUSING FINANCE PROJECT (LOAN 2947-ME) EVALUATION SUMMARY Objectives 1. The primary project objectives were to: (i) support housing finance reforms allowing commercial banks to achieve full cost recovery on mandatory housing loans; (ii) stimulate the introduction of savings schemes geared to households and institutions; (iii) increase the supply of affordable housing to the lower income segments of the population, especially through the testing of a new lower price housing product; (iv) provide a strong incentive to state governments to reform speedily mortgage registration procedures necessary to register mortgage liens which fully cover interest capitalization; and (v) widen the sector dialogue to include the role of the public sector, coordination of public housing programs, and policies related to investment in rental housing and mobilization of savings by households and institutions (para. 3.01). Implementation Experience 2. The loan was approved in June and signed in September However, the loan was not made effective for more than a year until November This long delay occurred because the Government introduced a new system for transferring foreign loan funds to project implementing agencies. Accordingly, the proposal had to be examined carefully by Bank staff. The Loan, Guarantee and Project Agreements had to be formally amended and, as the amendments were intended to serve as a model for modifying the legal documents of other previously approved Bank loans, Board approval was sought. Despite this long delay, the loan was fully disbursed in April 1992, some 26 months earlier than expected as FOVI made up for the significant drop in commercial bank lending for lower cost housing when the mandatory lending requirement was removed (paras and 5.02). 3. Due to a shorter disbursement period than expected and a steady increase in housing costs during the project implementation period, fewer lower income houses were financed than expected during appraisal. Moreover, the commercial banks, which were expected to finance a

10 significant number of lower income houses with their own resources, significantly reduced their lending for such housing, preferring to lend for higher priced housing. This was the direct result of Government action in March 1989, as part of a plan to restructure and liberalize the financial system, to eliminate interest rate controls and mandatory lending requirements (paras and 5.03). iii 4. During project implementation, limitations on the capitalization of interest on housing loans (which prevented full cost recovery) were eliminated and the costs of registering mortgages were significantly reduced by defining housing credits in multiples of the minimum wage. In May 1989, FOVI introduced an auction system for allocating its funds. The auctions have been highly successful, and permits builders to plan their construction and directs funds to the most efficient builders (paras and 5.05). 5. Two waivers were made to loan covenants during project implementation. Due to an increase in the cost of houses FOVI financed, the Bank authorized on an exceptional basis the financing of a somewhat higher priced unit. It also approved a waiver to allow for increases in the maximum sales price of the FOVI-financed houses (in terms of minimum wages) throughout the year, in line with the anticipated inflation rate for that year, to compensate for the lag between inflation and the adjustment in the minimum wage. The policy dialogue was advanced during project implementation by discussions between the Government and the Bank on the rental and savings mobilization studies and on the findings of the pilot program for marketing lower cost houses (paras to 5. 11). Results 6. Although the amounts of housing finance going to the lowest housing solutions financed by FOVI declined from the mid-1980s level, when the commercial banks were forced to lend for this type of housing, overall housing finance increased as credit subsidies were eliminated and the banks adopted financing mechanisms similar to FOVI's and charged market interest rates on such lending with their own resources. As a result, commercial bank lending for housing with their own resources increased from US$450 million in 1989 to some US$5.2 billion in 1992, compared with an average of US$880 million between 1983 and 1987 (para. 6.01). 7. The auctions of FOVI funds facilitated the entrance of new promoters and gave an advantage to the more efficient promoters and builders. They also provided an incentive for the states to reduce administrative burdens and costs on housing in order that promoters could develop projects which could

11 iv obtain FOVI financing. Accordingly, this resulted in a much stronger FOVI as it enhlanced its reputation by taking a leading role in achieving agreemenits with the states to simplify regulatory procedures in housing. Moreover, the achievements of FOVI in designing tlhe new financing instrument for housing and refining it to produce positive market returns were not only important for financing by commercial banks but also for the programs of the payroll contribution funds which were also important in housing finance in Mexico (paras to 6.05). 8. The studies carried out under the loan identified measures needed to mobilize savings and promote investment in rental housing, and steps have been taken in both these areas. The pilot program for lower cost housing was instrumental in identifying needed reforms in local regulatory rules for housing. Experience under the loan was also instrumental in identifying the need for finding another more suitable index for adjusting mortgages than the minimum wage and a more effective index is being sought. Finally, the loan was important for advancing the sectoral dialogue and in identifying the policies needed to be addressed in the follow-up loan to FOVI, the Housing Market Development Project (Loan 3497-ME) (paras to 6.08). Sustainability 9. Two Bank loans have been made subsequent to the Housing Finance Loan, the Second Low-Income Housing Loan (3140-ME) to FONHAPO in 1989 and a Housing Market Development Loan (3497-ME) to FOVI in Both projects have extended and cemented the reforms initiated under the Housing Finance Loan. In October 1992, the federal and state governments undertook a program to deregulate the housing sector and to reduce the cost of building permits, service rights, title registration and taxes. Notaries have agreed to limit notary fees for title registration. In order to implement the sweeping land reform made in 1992, the Government has initiated land certification and titling programs in each state which are expected to eventually relieve a significant constraint to housing development. In late 1992 and early 1993, the housing funds for private and public sector employees respectively introduced new policies and mechanisms to improve their efficiency, financial self-sufficiency and the transparency of their operations. Finally, the Government is carrying out a study to develop a secondary market for mortgages to broaden the housing finance market and mak-e it more liquid (paras to 7.05). Finding and Lessons Learned 10. As spelled out in Part 11, the objectives of the project were achieved and exceeded in several areas. The project was well designed and timely and

12 v was supported by a sound and comprehensive set of sectoral policy objectives. Despite a significant delay in the making the loan effective it was rapidly and successfully implemented. When problems emerged, creative solutions were developed, for example, to adjust financing terms to make housing more affordable when housing costs increased in real terms relative to the minimum wage. Also, the FOVI auction system became a useful device to assist promoters to plan their home building projects, to channlel resources to the more efficient builders, and to provide state and local governments with an incentive to simplify and make more efficient their regulatory frameworks for housing and to reduce housing costs, as home builders could construct and sell houses at lower cost in those states. These achievements can be attributed in large part to entrusting project implementation to a highly competent trust fund which had a long record of achievement and which had the full support of its trustee, the Bank of Mexico, which provided it with intellectual leadership and support in achieving its objectives (para ). 11. Another important factor in project success was that the project was designed in Mexico, based on studies carried out by the government and reflected sector policies supported by the government. Bank staff played a supportive role in extending and defining more clearly the policy framework and widened the housing policy dialogue in the areas of housing program coordination and housing finance harmonization. This contributed to project success and the achievement of significant policy reforms in housing finance and in the operations of key sector institutions such as INFONAVIT, FOVISSSTE and FONHAPO (para ). 12. Key project objectives to reduce housing credit subsidies, stabilize housing finance flows and increase savings mobilization were fully consistent with the government's objectives to reform the financial sector. These reforms were fully supported by the Bank under the Financial Sector Adjustment Loan. Bank resources under the Housing Finance Project eased the impact on lending for low cost housing caused by the elimination of mandatory lending requirements. Interest rate liberalization increased lending to the housing sector. The harmonization of project objectives in Bank loans to the same or related sectors was important in achieving success of individual projects and in addressing broader sectoral issues (para ).

13 MEXICO HOUSING FINANCE PROJECT (LOAN 2947-ME) PROJECT COMPLETION REPORT PART I: PROJECT REVIEW FROM BANK'S PERSPECTIVE 1. Project Identity Name Loan Number RVP Unit Country Sector Subsector Housing Finance Project 2947-ME Latin America and the Caribbean Mexico Infrastructure/Finance Housing 2. Background 2.01 Mexico's Housing and Housing Finance System. Mexico has suffered from chronic housing shortages for many years. In 1988, at the time the project was being prepared, the country's housing deficit was estinated at 5 million units, and was expected to grow over time. Statistical data is imprecise, but it appears that recent efforts, including Bank projects in the housing sector, have slowed growth in the housing deficit. The housing problem was compounded by the population's uneven and generally low capacity to generate savings for housing financing and the scarcity of longterm housing finance. Mexico had a highly segmented structure of housing supply. During the decade prior to 1988, about two thirds of new housing was supplied by the informal sector, i.e. the households themselves (self-help construction), and the remainder had been financed equally by public programs and commercial banks. Starting around 1985, the informal sector and commercial banks began to reduce their voluntary participation in the sector while the relative importance of public programs increased in response to adverse economic developments. High inflation, which reached a peak of 170% in 1987, and declining real wages had a negative effect on the availability and affordability of housing finance, and on availability of savings for down-payments on homes. Also, at that time, Central Bank (Banco de Mexico) regulations required commercial banks to lend 6% of their total resources for "social interest" housing loans, directed at families earning between 2 and 12 minimum wages. Because of the large differential between mortgage portfolio yields and the average cost of funds to the

14 2 commercial banks (known as "CPP", Costo Promedio Porcentual de Captaci6n, a key reference rate in Mexico's financial system) foregone revenues of the commercial banks amounted to US$670 million equivalent in These revenue losses forced commercial banks to increase interest rates on their unregulated loan portfolio to offset, at least partially, these losses. The "Fondo de Operacion y Financiamiento Bancario a la Vivienda (FOVI), a trust fund of the Central Bank, was also providing finanicing for housing through the commercial banks but the banks were allowed to make a financial spread on such lending In order to address the severe credit subsidy problem resulting from higih and increasing levels of inflation and low fixed interest mortgage rates under the mandatory lending system, the Central Bank had authorized commercial banks to use a new variable rate mortgage in Because of expected resistance from borrowers, Central Bank regulations initially set interest rates on mortgages well below market rates, and commercial banks were unable to fully recovery their costs on such loans. As inflation continued to climb, the Central Bank made significant modifications to its regulations governing the variable rate mortgage in 1986 aiid 1987 to reduce subsidies, before issuinig in March 1988 new rules that would allow commercial banks to achieve full cost recovery on the variable rate mortgage loans (this was a Board condition of the Housing Finanice Project). The key features of the mortgage instrument approved in 1988 were: (a) interest rates were equal to CPP; (b) monthly loan payment defined as a percentage of the loan amount and adjusted in parallel with the minimum wage to ensure affordability; (c) no limit on capitalization of interest, in case the loan payment is insufficient to cover the interest due; and (d) a flexible loan maturity subject to a 20 year maximum. The achievement of full cost recovery was seen as a key step in sustaining housing finance flows after mandatory lending requirements were eliminated. Thlese were also seen as an important contribution to interest rate and financial sector rationalization The Bank's Prior Involvement and Strategy for Mexico's Housing Sector. The Bank began its direct involvement with Mexico's housing sector in 1985 when it approved Loan 2612-ME, the Low Income Housing Project. This loan was to be the first in a series of lending operations in the housing sector with the aim of opening a policy dialogue that would address issues such as capital recovery and the measurement, control and reduction of housing finance subsidies, while improving distributional equity in the allocation of housing credit. As agreed under this loan, a very comprehensive housing finance study was carried out by the Central Bank in 1986 in close consultation with the Bank to serve as the main vehicle for a thorough dialogue on housing finance. The study's analysis and findinigs contributed to the changes made in 1986 through 1988 in the lending

15 3 terms of the mandatory system of housing finance through the commercial banks to increase cost recovery (para. 2.02), and provided important input to the Government's formulation of its Housing Policy Statement issued in February The Bank's overall strategy in the housing sector was to support the Government's policies and reforms aimed at expanding the efficient supply of low and middle income housing through improvements in cost recovery and further streamlining of regulations related to housing financing. Continued Bank assistance, through the provision of financial resources, technical analyses and policy support was considered to be critical to achieving further progress in, and the deepening of, housing sector reforms in Mexico. 3. Project Objectives and Description 3.01 The primary project objectives were to: (i) support housing finance reforms allowing commercial banks to achieve full cost recovery on their mandatory housing loans; (ii) stimulate the introduction of savings schemes for housing geared to households and institutions; (iii) increase the supply of affordable housing to the lower income segments of the population, especially through the testing of a new lower price housing product; (iv) provide a strong incentive to state governments to reform speedily mortgage registration procedures necessary to register mortgage liens which fully cover interest capitalization; and (v) widen the sector dialogue to include the role of the public sector, coordination of public housing programs, and policies related to investment in rental housing and mobilization of savings by households and institutions The project was to help finance construction and individual mortgage loans' made by commercial banks during 1988 to 1993, targeted to households with incomes up to 4.0 minimum wages (about US$5,250 p.a. in March 1988 prices), defined as Type A housing. The loan would fund only those housing loans which incorporated the March 1988 Central Bank regulations that allow full cost recovery. It was estimated that 300,000 Type A houses would be constructed during the period with about 19% being financed by the loan. Bank financing was limited to 80% of the selling price of houses for construction subloans and to 90% for purchase subloans. Also, 1 Although the report states that mortgages were financed with Bank resources, they were financed only in those cases where the housing would not have been built absent a prior mortgage financing commitment to assure that the "productive purposes" requirement of the Bank's Articles of Agreement are respected.

16 4 to facilitate the test marketing of a new lower priced house the loan provided financing for such a pilot test. The project also included studies related to savings mobilization and investment in rental properties. 4. Project Preparation, Design and Organization 4.01 Project identification was closely related to the review in November 1985 of the preliminary findings of the Housing Finance Study carried out by the Central Bank under the Low Income Housing Project, Loan 2612-ME. A project brief was prepared by the Bank in March 1986 for a second low income housing project which identified the need to improve cost recovery and introduce new resource mobilization schemes such as contractual savings schemes and housing bonds to be sold to long term investors to generate increased resources for housing finance During project preparation, the Bank and Mexican authorities agreed on the need for the government to develop a housing policy statement outlining its objectives with respect to strategies for institutional development, cost recovery, mobilization of funds, public/private sector roles, spatial distribution and the coordination of housing sector activities. The government hesitated to give a specific timetable for policy reforms and strategy implementation during the 1987/88 political transition period, but issued a sectoral policy statement in February 1987 which formed the basis for longer term measures. The Bank concluded that it still needed a policy letter in support of the project to cover additional measures such as liberalizing the housing market, underlining the importance of improving cost recovery, expanding family and institutional savings for housing, addressing the issue of rent control and the development of clearer policies to stimulate investment in rental properties and providing a clearer statement of government policy on the payroll funds to assure that they would not be used to increase housing subsidies as those provided through the banking system were reduced. The additional policy measures identified areas in which the government has subsequently taken important actions (e.g. beginning to eliminate rent controls, reforming the payroll funds, establishing mechanisms to augment savings, etc.) 4.03 During project preparation, Bank staff initially recommended that the loan be tranched to assure progress in achieving full cost recovery by requiring that housing loans on average bear the average cost of funds of the commercial banks (CPP). At the same time, Mexican authorities requested, and the Bank agreed, that the loan amount be increased from $200 million to $300 million to increase Bank support for sectoral reform. Bank staff also proposed measures to mobilize savings for housing through the introduction

17 of a system for allocating mortgages according to individual's savings performance Prior to appraisal, the Central Bank revised its regulations to increase interest rates on mortgage loans and then increased them again after appraisal to CPP to reduce the interest rate subsidy even further. At the same time, it revised other regulations to eliminate any limitation on the amount of interest that could be capitalized on mortgage loans (para. 2.02). During negotiations, it was agreed that the entire loan would only be disbursed for housing loans which incorporated loan terms designed to achieve full cost recovery. Accordingly, Bank management decided not to tranche the loan to achieve the elimination of interest rate subsidies. Full cost recovery was defined as lending at the CPP interest rate with unlimited interest capitalization, although full cost recovery would in fact require that housing loans be made at market interest rates (i.e. CPP plus a spread to cover the commercial banks risks, costs and investment return). The Bank also decided that the feasibility of allocating mortgage loans according to savings performance would be taken up in a savings mobilization study included as part of the project (see para. 4.06) Bank loan proceeds were to be provided to FOVI by BANOBRAS (Banco Nacional de Obras y Servicios Publicos, S.N.C., the borrower) in pesos at 1.5% below FOVI's relending rates to the commercial banks. The commercial banks would onlend the funds to project beneficiaries (private developers for construction loans and individual home buyers for longer term housing loans) at interest rates and amortization periods in accordance with Bank of Mexico's mandatory housing lending regulations which provided that the banks would repay FOVI on the same terms except that they would retain up to a 3 % spread. The terms on which loan funds were provided to FOVI were changed by the introduction of a new system for transferring foreign loan funds to project implementing agencies, as explained in para However, FOVI's relending rates were not affected by the new system. The financial spreads allowed under FOVI lending to the commercial banks were marginally attractive and were supplemented by commissions allowed by FOVI beginning in 1990 (para 5.12) There was a clear conceptual framework for the project with a primary focus on improving and extending the use of a new lending instrument to reduce subsidies within the 6% investment requirements, since without it banks might not have lent for housing. This was expected to pave the way for expanded commercial bank lending for housing on market terms. Other project objectives were closely related and were directed to increasing the mobilization of resources for housing and seeking less expensive housing

18 6 solutions to reach poorer segments of society. These were Government objectives as spelled out in its statement of housing sector policies provided to the Bank under the loan (para. 4.02). The project also included studies on savings mobilization and rental policy to assist in extending the dialogue on housing policy. An extension of the housing policy dialogue was being carried out in the framework not only of the Housing Finance Project but also of Bank lending through the government low-income housing agency, FONHAPO, through the Low Income Housing Project which was followed up by a second Low Income Housing Project in Project objectives were well understood and supported by all the parties including: SEDUE (Secretaria de Desarrollo Urbano y Ecologia) which was responsible for developing and administering overall housing policies in Mexico; the Secretaria de Hacienda y Credito Publico, which was responsible for implementing government public investment policy, of which housing investment was an important component because of its social and employment creating aspects; project beneficiaries, especially promoters, builders and prospective home owners who were interested in a more stable source of housing finance; and the Central Bank which was interested in perfecting FOVI's new housing finance instrument to make housing affordable and to reduce housing credit subsidies which were a burden on commercial banks and affected their profitability and financial condition negatively and were an obstacle to financial sector liberalization The project was well prepared by the government, principally by FOVI, a trust fund with some 25 years of experience in housing finance when the loan was made. The roles and responsibilities of the parties (the government, FOVI, promoters, commercial banks) responsible for each aspect of the project were clearly defined and well understood as were the risks which were mainly of a financial loss to promotors. The timing of the project was also fortuitous for commercial banks. World Bank funds became available to sustain housing finance just when new financing by the commercial banks was declining precipitously, because of concerns about providing long-term mortgage finance in a highly unstable inflationary environment. New commercial bank financing for housing was falling first as amounts of interest capitalization were increased (thereby increasing the nominal value of housing loans and fulfilling the mandatory requirement for lending for housing) and second when the mandatory requirements were eliminated entirely as a result of financial sector liberalization. 5. Project Implementation 5.01 The loan was approved by the Board on June 2, 1988 and signed on September 28, However, the loan was not made effective

19 7 for more than a year until November 16, This long delay occurred because the government used this project to introduce a new system for transferring foreign loan funds to project implementing agencies (spelled out in the "Normatividad") aimed at establishing a transparent system of transferring resources to trust funds and development banks and controlling their administrative expenses and credit subsidies. The Housing Finance Project was the first Bank loan to be affected by these new arrangements and it took some months for the Bank to agree on their use and to reflect them in the transfer of funds agreement required under the loan. Moreover, the new arrangement for transferring funds through BANOBRAS, the borrower, to the Central Bank, trustee for FOVI, the project implementing agency, required a change in the terms for onlending resources under the loan. Instead of onlending the proceeds of the loan to FOVI at 1.5 % below FOVI's relending rate to the commercial banks as originally specified, BANOBRAS would relend the funds to FOVI at the government's Treasury bill rate (CETES). FOVI's relending rates to commercial banks were unaffected by the new procedures. The government agreed to provide FOVI with resources equivalent to any subsidy it provided and FOVI's administrative costs were budgeted, monitored and adjusted if required by the government. It took some months before the new arrangements could be satisfactorily reflected in a subsidiary loan agreement between BANOBRAS and the Central Bank. Moreover, the Loan, Guarantee and Project Agreements had to be formally amended and, as the amendments were intended to serve as a model for modifying the legal documents of other previously approved Bank loans, Board approval was sought. Approval was obtained in September Despite the more than one year delay in making it effective, the loan was fully disbursed by the end of April 1992, some 26 months faster than estimated at appraisal. Loan funds financed housing construction carried out during the years 1989 through 1991 rather than the longer 1988 through 1993 period forecast during project appraisal. To a large extent, as discussed below (para 5.03), FOVI made up for the significant drop in commercial bank lending for lower cost housing after the 6% mandatory requirement was removed in the context of the Financial Adjustment Loan (FSAL, Loan ME) approved in The significant increase in the average selling price of houses financed also increased the rate of loan disbursements. It was estimated at appraisal that 300,000 Type A houses would be financed between 1988 and 1993, involving total financing of US$1,419 million, of which about 146,000, involving financing of US$691 million would be financed during the 1989 through 1991 period. Average financing was expected to amount to US$4,730 per housing unit. However, prior to the initiation of disbursements, and during the project implementation period, there was a steady increase in housing costs in real terms (see para 5.06) and the average value of the mortgages financed by FOVI in the period 1989-

20 was US$10,043, more than double the expected amount. FOVI financed 78,815 mortgage loans during the period mainly for Type A housing with a total value of US$791.6 million. Some 31,411 of these mortgages were financed with Bank resources, with an average mortgage value of US$9,550 (Type A houses alone had a mortgage value of US$9,470). Bank funds were estimated to have benefitted some 170,000 people (see Part II) The commercial banks, which were expected to finance with their own resources about 59,000 of the 146,000 Type A houses during the period, virtually withdrew from this segment of the market, preferring to lend for higher cost housing. This was the direct result of government action in March 1989, to restructure and liberalize the financial system (and with the support of the FSAL), to eliminate interest rate controls and mandatory commercial bank lending requirements. In the years , commercial banks made 167,530 mortgage loans with their own resources totalling some US$4,717.5 million representing an average mortgage loan of US$28,160 equivalent. While FOVI provided financing with Bank loan proceeds and with other resources provided to it, mainly by Central Bank loans and from mortgage loan repayments, the number of Type A houses financed fell well short of estimates made at appraisal due to the steady increase of housing costs. While FOVI responded by eliminating the minimum size of houses financed and by financing less expensive housing solutions such as improved lots, partial housing solutions and home improvements, the challenge still remains to find lower cost housing solutions (see para 5.09 on the results of the pilot program for marketing lower-cost houses) and ways to reduce the cost of housing finance and increase affordability. Despite the smaller number of Type A houses financed, overall housing finance provided increased dramatically during the period as the commercial banks increased the amount of housing finance provided with their own resources from an average of US$880 million a year in the years 1984 through 1987 to an average of US$1,570 million in the years , and reaching some US$5.2 billion in In March 1988, the Central Bank had eliminated any limitation on the capitalization of interest to assure full cost recovery on housing credits. However, there was still an obstacle at the local level as the cost of mortgage registration was very high as it was related to the total registered amount of the credit guarantee including any capitalization of interest. This problem was eventually resolved throughout the country by defining housing credits in multiples of the minimum of wage, thereby limiting registration costs to the current value of the housing loan. However, this required agreements amongst the Federal Government, State governments, and associations of notaries which were somewhat more difficult and time

21 consuming than expected to achieve. These changes were largely achieved through FOVI's promotional efforts at the state level In May 1989, the Central Bank introduced an auction system for assigning FOVI funds to make the allocation of its funds more equitable and transparent by allocating its long-term financing to the highest bidder. The auctions have allowed the government to capture in the bid prices received by FOVI the premium that home builders and borrowers are willing to pay for long-term mortgage finance. The auctions also reward the most efficient home builders, who can bid more for FOVI funds (para. 6.03) Two waivers were granted to loan covenants during project implementation. On June 15, 1990, the Bank authorized on an exceptional basis financing of all "Type A" houses that participated successfully in FOVI's auctions during 1989 and January 1990 which had a final selling price of up to 130 minimum monthly salaries and were acquired by beneficiaries with a household income up to 4.5 minimum wages. The request was made on the basis of the requirement that within a year of loan effectiveness the Central Bank would review with the Bank the selling price of eligible housing taking into consideration changes in market conditions. The cost of the houses FOVI financed had increased principally due to: (i) FOVI's decision no longer to finance construction subloans which were made by commercial banks at market rates which were higher than FOVI rates; (ii) the additional costs of the auctioned funds; (iii) the large increase in housing lots; and (iv) contractor's costs of entering new markets as housing finance for the three major metropolitan markets (Mexico City, Monterrey and Guadalajara) was reduced On October 30, 1991, the Bank approved another waiver to allow the use of "valores sustitutos" as the basis for establishing the maximum sale price of FOVI "Type A" houses eligible for Bank financing for which contractors made successful bids on or after November 1, Under the concept of "valores sustitutos", FOVI was able to increase gradually throughout the year the maximum sales price of the Type A houses it finances (in terms of multiples of the minimum wage), to compensate for the lag between inflation and the adjustment in the minimum wage, which after 1990 was being increased only once annually. On December 4, 1991, the waiver was extended to make it retroactive to May 1, During project implementation, at the request of the Mexican authorities, the Bank agreed to postpone the deadline for completing the savings mobilization study from March 31, 1989 to March 31, 1990 and the test marketing program for lower-cost houses from March 31, 1989 to June 30, Initially the savings mobilization study was to be carried out by a

22 10 task force of the Mexican Association of Banks (as indicated in the SAR). However, there were several delays in starting the study and the Bank of Mexico undertook responsibility for it in February 1990, completed it and submitted it to the Bank in early April The rental housing study was completed and submitted to the Bank close to the target date of December 31, The quality and coverage of both studies was fully satisfactory, and the studies recommendations have been or are being implemented (see paras 5.10 and 5.11) The policy dialogue was advanced during project implementation by discussions between the government and the Bank on the rental and savings mobilization studies and on the findings of the pilot program for marketing lower-cost houses. As a result of the efforts to reform mortgage registration at the local level, FOVI learned more about the high costs of such local regulations generally. While FOVI concluded that the pilot program for lower-cost housing was not generally replicable because of the special arrangements required between key parties such as the federal government, municipalities and home builders to simplify technical and administrative procedures, reduce financial costs and achieve economies of scale, it did demonstrate that there was a market for progressive housing solutions and, more importantly, that to reduce housing costs generally, efforts would need to be made at the local level to simplify building codes and approval procedures, reduce mortgage registration costs and the costs of introducing water, electricity and drainage services and to provide an adequate supply of land. Since 1990, FOVI has been active in identifying key regulatory bottlenecks at the state and local level. To encourage states to reform standards and regulations, FOVI began signing agreements with states which set out specific targets for reasonable building standards, time limits for approving permits and licenses and a schedule to lower overall bureaucratic costs. A follow on Bank-finance project with FOVI, the Housing Market Development Project, is supporting these types of regulatory reforms in a well focussed and systematic way The savings mobilization study led to a discussion of ways to mobilize resources to meet the growing needs of the lowest income groups, to reduce the over-reliance on standard housing solutions and the need to extend financing beyond new construction to purchases of existing homes. The government has taken steps to increase financial savings by removing interest rate controls on commercial banks, which led to an increase in deposits held by banks of some 35% in real terms between April 1989 and the end of In February 1992, the government set up a new retirement saving system that will substantially increase the volume of long-term finance available to banks for mortgage and other lending and will bring low-income savers in contact with the financial system. Under this system, private

23 11 employers are required to deposit 2 percent of a worker's wages into pension funds administered by commercial banks. At start up in late 1992, employers deposited more than US$1 billion equivalent in new individual workers' accounts opened in commercial banks. Also in 1992 the government passed legislation allowing the creation of savings banks (Cajas de Ahorro) which are expected to finance housing and support phased housing solutions (progressive construction and starter homes) and improvements to existing homes. In response to the recommendations of the savings mobilization study, FOVI started to finance improved lots and partial housing solutions and is now financing housing resales through lines of credit to commercial banks The rental study identified the need to eliminate rent controls and to simplify legal procedures for evicting non-paying tenants. Subsequently, in the housing policy statement formulated as part of the Housing Market Development Project, the government reconfirmed its intention to address these issues. In mid-1993, public officials in both Mexico City's Federal District and the State of Nuevo Leon, whose capital is Monterrey, the country's third largest city, passed legislation that will gradually abolish all rent controls and facilitate evictions of non-paying occupants In general, project objectives were achieved and the policy dialogue steadily advanced. There were some minor delays in carrying out project studies but the studies proved to be useful vehicles for continuing the policy dialogue on housing and housing finance. The SAR indicates that there were no major project risks other than the risk of slow disbursements due to delays in reforming registration procedures at the state level. However, the states moved quickly to allow mortgage liens permitting the full capitalization of interest. The major risks turned out to be the reduction in the minimum wage in real terms and the increase in housing costs in real terms making housing financed under the loan more costly and less affordable to lower income groups. It would have been difficult to foresee the further reduction in minimum wages in real terms. Real wages had already fallen significantly and steadily between 1982 and 1988 and were expected to recover. At the same time, debt service was made easier by the sharp drop in nominal and real interest rates. Nevertheless, FOVI was forced to increase its eligibility limits to meet changing market conditions and to seek ways of increasing affordability by reducing monthly payments, increasing the maximum repayment period of loans and reducing income requirements. Also the commercial banks completely withdrew, for a while, from financing low cost housing with their own resources after mandatory investments in housing was abolished. While they resumed lending for housing with their own resources in 1990 it was exclusively for higher priced

24 12 housing. Moreover, FOVI permitted banks to charge commissions on FOVI financed housing loans to increase their yield to market levels thereby increasing the overall cost of housing. This basically provided for full cost recovery on housing loans made with FOVI resources. 6. Project Results 6.01 Although the amounts of housing finance going to the lowest housing solutions financed by FOVI declined from an average of about US$450 million in the mid-1980s when the commercial banks were forced to lend for this kind of housing, to US$320 million average in 1990/1991, overall housing finance increased dramatically as subsidies were eliminated, banks adopted financing mechanisms similar to FOVI's and charged market rates on such lending with their own resources. As a result, commercial bank lending for housing with their own resources increased from US$450 million in 1989 to US$ 5.2 billion in 1992, compared with an average of about US$800 million between 1983 and The reduction of housing subsidies was an important step in getting commercial banks to lend voluntarily for housing. The elimination of mandatory requirements and the removal of interest rate controls and financial sector reform was eventually followed by the government's 1990 decision to reprivatize the commercial banks. Interest rate liberalization led to a significant increase in deposits in real terms and an increase in resources available for housing finance which was an attractive vehicle for bank lending FOVI's introduction of a public funds auction in 1989, to replace the direct placement of financing to developers and eventually to home buyers, has had several advantages. The auctions allow the government to capture in the bid prices received by FOVI, the premium that home builders and borrowers are willing to pay for long-term mortgage finance. They also reward the most efficient home builders who can bid more for FOVI funds. In addition, the FOVI auctions which awards funds to the highest bidders tend to reward those states and municipalities with the lowest regulatory burdens by allowing developers and mortgage lenders to compete better for FOVI funds to finance and build housing The project has resulted in a much stronger FOVI as it enhanced its reputation by taking a leading role in promoting the agreements with state governments to simplify regulatory procedures and reduce the administrative burden and costs of housing. These agreements have resulted in significantly lower additional costs to housing in those states which have pushed reform. In the face of a fall in affordability due to higher real

25 13 housing costs and lower real wages, smaller housing units and partial housing solutions were promoted and eligibility requirements were lowered by reducing income requirements and lengthening the repayment periods of mortgages The achievements of FOVI in designing the new financing instrument for housing and refining it to the point where it produced positive market returns were not only important for financing by commercial banks but aspects of the FOVI financing instrument were adopted by the other agencies and payroll contribution funds important in housing finance in Mexico including FONHAPO and INFONAVIT The studies carried out under the loan identified measures needed to mobilize savings and to promote investment in rental housing. Steps have been taken in both these areas. FOVI concluded that the pilot program for lower cost housing is not replicable because it was a laboratory type experiment requiring special arrangements with financial institutions and local government authorities to carry out successfully. However, it was instrumental in identifying needed reforms in local regulatory rules for housing, land requirements, etc. These findings were important in motivating FOVI to push for local regulatory forms through its agreements with the states Experience under the loan was also instrumental in identifying the need for finding another more suitable index for adjusting mortgage payments and balances which better reflect changes in real earnings to replace the minimum wage. The minimum wage is an administered wage, set by government decree, and not determined by labor market developments. Currently, it is adjusted once a year and because it does not reflect changes in real wages, it is being applied in fewer and fewer labor contracts. FOVI, together with other governient agencies, is developing a suitable replacement, in accordance with a commitment made under the Housing Market Development Project The loan was important for advancing the sectoral dialogue and in identifying the policies needed to be addressed in the follow-up loan with FOVI, the Housing Market Development Project. In addition to increasing the focus on reducing overregulation in housing, this loan focuses on the need for dealing with the shortages of urbanized land, dealing with the segmentation and inefficiencies of housing finance programs, improving FOVI's mortgage instrument and its onlending mechanism so as to stimulate commercial bank interest in low-cost housing and improving information on the housing market. Many of these issues were identified as important

26 constraints to future sector development during the implementation of the Housing Finance Project. 7. Project Sustainability Two Bank loans have been made subsequent to the Housing Finance Loan, the Second Low-Income Housing Loan to FONHAPO in 1989 and a Housing Market Development Loan to FOVI in Both projects have extended and cemented the reforms made under the Housing Finance Loan In October 1992, President Salinas convened a meeting of all 31 state governors, the mayor of the Federal District and representatives of FOVI, and virtually all state institutions dealing with housing, to sign agreements that deregulate the housing sector nationwide. In these agreements the states have pledged to reduce to less than 10% of the cost of a house total charges for building permits, service rights, title registration and taxes (surveys indicated that regulatory costs have added as much as 25 % to the cost of a house). Also the National Association of Notaries agreed to limit notary fees for title registration to I % of home value (they had in the past increased house costs by up to 10%) The government is closely monitoring these deregulation efforts. Moreover, FOVI completed its third survey of regulatory costs in early 1993 and published them in the Mexican Homebuilder's Association trade magazine. In order to reduce housing costs, in November 1992 the government exempted the commissions earned by commercial banks (initially for the period 1992 to 1994) on housing finance from value added tax and reduced the transfer tax on property sales from 6% to 4 % in January The government, on the basis of a constitutional amendment passed primarily to modernize the agricultural sector, has initiated land certification and titling programs in each state. These reforms would free landholders, known as "ejidatarios", to associate with builders or sell their land to investors for housing development. As many of Mexico's cities are currently surrounded by "ejido" land, this constitutional reform is expected eventually to relieve one of the major constraints to housing development, a shortage of reasonably priced, well-located land An important objective of the project was to widen and supplement the housing policy dialogue in the areas of housing program coordination and housing finance harmonization. In October 1992, INFONAVIT introduced new policies and mechanisms to improve the efficiency, financial self-sufficiency and transparency of its operations. Under these new policies, INFONAVIT allows borrowers to decide on the

27 15 type of house they want to acquire (i.e. new, used or self-built) and awards mortgage loans, which are wage-indexed, and charge a positive real rate of interest on the basis of a borrower's age, salary, contributions to the fund and savings level. INFONAVIT has also begun to assign housing construction contracts through public bidding. In January 1993, FOVISSSTE, the smaller public sector employees housing fund, implemented similar reforms. Finally, the government is contemplating a reorganization of FONHAPO to address better the housing needs of the poor The government is also carrying out a study to develop a secondary market for mortgages to broaden the housing finance market and make it more liquid. It is also searching for a mechanism to replace the minimum wage to adjust mortgage payments which better reflects changes in earnings. The overall government commitment to strengthening housing finance and extend the benefits of housing continues to be quite strong as reflected in government actions and policy reform measures in the last few years. 8. Bank Performance 8.01 The Bank acted with imagination and foresight in identifying and supporting a key government program which had important implications for housing and financial sector reform. It engaged in a prolonged and fruitful dialogue with government to define and strengthen sector objectives. It supported a key institution, FOVI, in housing finance and its efforts to improve an important new financial instrument which it had developed to sustain housing finance in a period of high inflation. The comparatively small Bank team, the continuity of task managers (two from project identification to completion), and their close cooperation with the FOVI team, helped to resolve project issues quickly and to heighten the Bank's effectiveness. The relatively small number of Bank supervision missions (Part III, Table 5) is somewhat misleading, as supervision of the project took place during the preparation of the Housing Market Development Project (Loan 3497-ME). It also reflects the project's relatively short disbursement period Bank support to FOVI was important to assist in sustaining finance for housing during a period of considerable economic uncertainty, marked by high and volatile inflation, and while substantial financial sector reform was being defined and carried out. Bank participation in the project led to the clarification and extension of housing sector policies during project preparation and as a result of studies included in the project to resolve sector issues identified in the dialogue between the government and the Bank.

28 16 9. FOVI Performance 9.01 FOVI, with a staff of fewer than 100, is a well managed institution with some 25 years of experience in housing finance when the loan was made. It, together with senior staff of the Bank of Mexico, FOVI's trustee, who provided important intellectual leadership in the development of the project and in policy design, were sophisticated interlocutors during project preparation and responded quickly and effectively to problems during project execution. It was effective too as a trust fund in the Central Bank in obtaining government support for the project, in assuring that project objectives were clearly defined and supported and were consistent with government sector policies. When the savings mobilization study was delayed due to changing priorities of the banking association, FOVI and the Bank of Mexico responded effectively to make sound alternative arrangements to carry it out. 10. Project Relationship The relationship between the Bank and the Mexican housing sector authorities remained strong throughout project preparation and implementation and the sector dialogue, initiated under the Low Income Housing Project (Loan 2612-ME), approved in 1985, and extended under the Earthquake Reconstruction Loan, approved in 1986, was expanded further under the Housing Finance Loan. Working with the highly experienced and well managed project implementation agency FOVI was probably crucial, not only in resolving project issues, but also in extending the policy dialogue, especially in the mortgage finance and subsidy rationalization areas. 11. Consulting Services Outside consultants played no role in project preparation or implementation. FOVI staff, with Bank support, was responsible for project development and execution. Studies included in the project were also carried out by government agencies--the Central Bank (savings mobilization), FOVI (test marketing) and SEDESOL (rental housing). 12. Project Documentation and Data The legal agreements for the project were adequate. Covenants were met on a timely basis, except for the timetable for studies which needed to be extended briefly. Required amendments were made without any problems or delays after a brief review of the issues between FOVI and Bank staff.

29 The SAR provided a useful framework for the Bank during project execution. Data required for preparation of the PCR were readily available from project files or from the study of the project prepared by FOVI, although reliable data on housing demand, housing stocks and changes in affordability (i.e. real income data) are not available. However, FOVI is working closely with the commercial banks to get data on housing and housing finance and with the states to get reliable data on housing regulatory costs. SEDESOL is also working with state governments to get data on housing registrations, etc. and plans to strengthen the overall data base on housing. 13. Findings and Lessons Learned As spelled out in Part 1', the objectives of the project were achieved and exceeded in several areas. The project was well designed, timely and supported by a sound and comprehensive set of sectoral policy objectives. Despite a significant delay in the making the loan effective, once done it was rapidly and successfully implemented. When problems emerged, creative solutions were developed, for example, to adjust financing terms to make housing more affordable as housing costs increased in real terms relative to the minimum wage. Also, the FOVI auction system became a useful device to assist promoters to plan their home building projects, to channel resources to the more efficient builders, and to provide state and local governments with an incentive to simplify and make more efficient their regulatory frameworks for housing and to reduce housing costs, as home builders could construct and sell houses at a lower cost in those states. These achievements can be attributed in large part to entrusting project implementation to a highly competent trust fund which had a long record of achievement and which had the full support of its trustee, the Bank of Mexico, which provided it with intellectual leadership and support in achieving project objectives Another important factor in project success was that the project was developed and designed in Mexico, based on studies carried out by the government and reflected sectoral policies fully supported by the government. Bank staff played a supportive role in extending and defining more clearly the policy framework for the loan, including the need for further studies to develop less expensive housing solutions, mobilize savings, and review the need for rental housing reforms, and widen and supplement the housing policy dialogue in the areas of housing program coordination and housing finance harmonization. These were fully in accordance with government policy for the sector and contributed to project success and the achievement of policy reform in housing finance and the housing sector such as the