Mexico - Second Urban and Regional Development Project

Document of The World Bank FOR OMCIAL USE ONLY Report No. 12087 PERFORMANCE AUDIT REPORT MEICO SECOND URBAN AND REGIONAL DEVELOPMENT PROJECT (LOAN 1990-ME) JUNE 28, 1993 MICROGRAPHICS Report No: 12087 Type: PPAR 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. FOR OFFICIAL USE ONLY EXCHANGE RATES Currency Unit - MIP (Mexican Peso) 1981 24.5 1982 56.4 1983 120.1 1984 167.8 1985 256.9 1986 611.8 1987 1378.2 1988 2273.1 1989 2461.5 GLOSSARY OF ACRONYMS BANOBRAS National Development Bank for Public Works (Banco Nacional de Obras y Servicios PGblicos) CPAPA A State Water Company (Comisi6n Estatal do Ague Potable) FIDEIN Trust Fund Industrial Commercial Estate Development (Fideicomiso do Conjuntos, Parques, Ciudades y Centros Industriales) FIFAPA Investment Fund for Water Supply and Sewerage (Fondo de Inversiones Financieras pars Agua Potable y Alcantarillado) FOGAIN Small and Medium Industry Guarantee and Development Fund (Fondo de Garantia y Fomento a la Pequefna y Medians Industria) FOMUN Municipal Development Fund (Fondo do Fomento Municipal) FONHAPO National Fund for Low Cost Housing (Fondo Nacional de Habitaciones Populares) FORTAMUN First Municipal Strengthening Project (Primer Proyecto do Fortalecimiento Municipal) FOVI Housing Operational and Financing Fund (Fondo de Operaci6n y Financiamiento a la Vivienda) INVITAB Housing Institute INDECO - State of Tabasco (Instituto de Vivienda INDECO - Tabasco) MCMA Mexico City Metropolitan Area (Area Metropolitana de la Ciudad do M46xico) 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. GLOSSARY OF ACRONYMS (Cont.) PACDU Program of Concerted Action for Urban Development (Programa do Acci6n Concertads par& el Desarrollo Urbano) PEMER National Petroleum Company (Compala Nacional de Petr6leos do M6xico) PAR Performance Audit Report PCR Project Completion Report SAHOP Ministry of Human Settlements and Public Works (Secretaria de Asentamientos Humanos y Obras Pfblicae) SAPAET Water and Sewage Company - State of Tabasec (Servicios de Agua Potable y Alcantarillado del Estado do Tabasco) SAPAECH Water and Sewage Company - State of Chiapas (Servicios de Agua Potable y Alcantarillado del Estado de Chiapas) SAR Staff Appraisal Report SEDUE Ministry of Urban Development and Ecology (Secretaria do Ecologia y Desarrollo Urbano) SHCP Ministry of Planning and Budget (Secretaria de Hacienda y Cr6dito PGblico) SPP Programming and Budget Secretariat (Secretaria do Programaci6n y Presupuesto) FISCAL YEAR OF THE BORROWER January 1 - December 31 FOR OFFICIAL USE ONLY THE WORLD BANK Washington. D.C. 20433 U.S.A. Office of Dirctor*General Operations Evaluation June 28, 1993 EMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECTs Performance Audit Report on Mexico Second Urban and Regional Development Project (Loan 1990-ME) Attached is the report entitled "Performance Audit Report on Mexico - Second Urban and Regional Development Project (Loan 1990-ME)" prepared by the Operations Evaluation Department. The audited project was developed to meet the growing demand for services and to improve living conditions of the urban poor in Mexico's economically high potential and fast growing oil-producing region of Tehuantepec Isthmus. The project was formulated along the lines of an integrated project, typical at the time, offering a combination of new housing (serviced lots), upgrading and infrastructure for existing alums, city-wide infrastructure, social services, and credit for small businesses. The project was prepared at the time when Mexico's economic boom was slowing down. The project's programmatic approach with heavy reliance on a series of borrowing institutions failed to gain the necessary support and commitment from the government. The operation ended up as a series of marginal and uncoordinated investments in housing and water supply amounting to 20% of the originally expected program. The project did set in motion a search for viable municipal development lending mechanisms in Mexico. However, the Audit concludes that the project outcome was unsatisfactory. Institutional development is rated as negligible while sustainability of benefits is rated as unlikely. These audit ratings confirm the prior PCR assessments. This document has a restricted distribution and may te used by recipients onty in the perfo their officlat duties. its contents may not otherwise be disctosed without Wortd Bank authoriain PERFORMANCE AUDIT REPORT MEXICO SECOND URBAN AND REGIONAL DEVELOPMENT PROJECT (LOAN 1990-ME, TABLE OF CONTENTS Page No. preface................................ Basic Data Sheet ................................... 1. Evaluation Su=mary ........ ............. ....... iv I. BACKGROUND ..................................................... 1 II. TRE PROJECT .................................................... 3 A. Project Procesaing •••••••••••••••••••••3 l. Preparation ...................................... 3 2. Appraisal ....................................... 6 3. Negotiations ................................ 7 B. Project Description ........................................ 8 1. Objectives ...................... ..... 8 2. Institutional Arrangement8 ............................. 10 3. Project Area ......................... 10 4. Components, Costø, and Financing ....................... 10 III. IMPLEMENTATION, ACRIEVEMENTS, AND IMPACT ....................... 12 A. Time Delays, Cost Overrune, and Cancellations .............. 12 8. Achievements ............................................... 14 1. Shelter Components ..................................... 16 2. Citywide Infrastructure and Upgrading ................... 16 C. Policy Impact ............................. 17 D. Institutional Impact ................................... 18 E. Poverty Impact ............................................. 18 IV. POINTS OP INTEREST ................................ 18 A. coat Recovery ........................ 18 I. Sustainability ......... 19 C. Follow-on Projects ........................................ 19 D. Integrated Projects versus Sectoral Programa ............... 19 E. Deconcentration and Decentralization ....................... 20 V. TRE ROLE OF TME BANK . 20 VI. CONCLUSIONS AND LESSONS LEARNED ................................ 21 A. Conclusions ............. ............ 21 B. Lessons Learned ......................................... 21 Comente from the Borrover ..4................................. 24 I-ERFORMANCE AUDIT REPORT MEXICO SECOND URBAN AND REGIjNAL DEVELO?JENT PROJECT (LOAN 1990-ME) PREFACE 1. This is the Performance Audit Report (PAR) of the Mexico - Second Urban and Regional Development Project. Approved on May 12, 1981, the project was supported by Loan 1990-HE for US$164.0 million. Of this loan amount, US$81.8 million was reallocaced to support the Mexico City earthquake reconstruction effort in 1985. An uncommitted portion of the loan, US$22.40 million, was eventually canceled. The loan closed on December 31, 1989. 2. The PAR consists of an Evaluation Summary and a Performance Audit Report prepared by the Operations Evaluation Department (OED). A Project Completion Report (PCR) prepared by the Infrastructure and Energy Operations Division of Country Department II of the Latin America and Caribbean Region (LA2IE) was previously submitted to the Board of Executive Directors as Report No. 11548, dated January 12, 1993. 3. The PCR provides an account of the project exsrience, its achievements and failures. In order to give a wider perspective, the Audit provides additional information about the history of the project, its outcome, the issues raised, implementation problems, and the principal lessons learned. The PAR is based on the Staff Appraisal Report (SAR), the President's Report (PR), the Loan Agreement, and projuct files, as well as the PCR, a completion report prepared by the Borrower, BANOBRAS, and other relevant material. It also draws on discussions vith Bank staff in Washington and interviews with representatives of BANOBRAS and other Mexican agencies during an audit mission in August-September 1992. 4. While the Audit agrees with the broad conclusions of the PCR, the use of a broader data base has led to differences in quantitative data and emphases. The PCR appears to have been largely based on the project files and the SAR, while the Audit obtained access to additional Mexican Government documents which were not in the files. The Audit also made greater use of the Loan Agreement and expands upon the lessons learned from the operation. It highlights the importance of the Borrower's "ownership" of the proposed project and acceptance and understanding of all its objectives. 5. Following standard OED procedures, the draft PAR was sent to the Borrower for comments on March 25, 1993. Comments received from Secretariat of Finance and Public Credit, BANOBRAS, and SAPAET are reflected in and attached to the report as Annex 1. ii PERFORMANCE AUDIT REPORT MEXICO SECOND URBAN AND REGIONAL DEVELOPMENT PJCT L 1990a- ) BASIC DA 031E REY r4ROJLCT DATA Actual or Actual as 2 Appraisal Current of Appraisal item Expectation zs;imate Eatkmate Total Project Cost (US$ m) 468.3 95.6 2C.4 Loan Amount (US$ a) 164.0 59.81 36.5 Economic Rate of Return 1&.02 N/A CUMULATIVE ESTIMTED AND ACTUAL DISBURSEMENTS Appraisal Estimate (US$H) 164.0 Actual Disbursed (US$M) 59.8 Reallocated (US$M) 81.8 Cancelled (US$M) 22.4 Actual as 2 of Appraisal (1) 36.5 Date of Final Disbursements September, 1989 PROJECT DATES Orixinal Actual Identification 04/78 04/78 Preparation 06/78 06/78 Pre-Appraisal 10/78 10/78 Appraisal Mission 02/79 05/79 Post-Appraisal 08/79 02/80 Negotiations 05/80 03/06/81 Board Approval 06/80 05/12/81 Loan Signature 08/13/81 Loan Effectiveness 10/80 01/12/82 Project Completion 06/30/86 12/31/89 Loan Closing 12/31/86 12/31/89 I US$81.8 was reallocated to support another project, Loan 2665-Mg. No re-estimate made as project was not fully completed. STAFF INPUTS (staff-weeks) FY?8 FY79 fYSS FY81 fL fl8 flM L1 FY87 flM E82 12 In19 Preappraisai 33.2 85.6 .5 - - * * - - - - - 119.3 AppraisaL * 52.3 92.3 6.2 * - * * * - 150.8 Negotiation * * - 23.0 * - - * - * 23.0 Supervidon - - * 4.5 24.9 16.7 18.5 11.9 11.2 6.4 6.6 .8 .6 102.1 Other - 6.9 15.4 16.4 - - * * * - * * 38.7 MISSION DATA Date No. of No. of MonthfYear Days Persons Identification 04/78 6 3 Preparation 06178 10 6 Pre-appraisal 10/78 20 7 Appraisal 05/80 33 8 Post-appraival 02/81 10 4 Project Launch 05/81 3 3 2-pervision 1 11/82 5 2 Supervision 2 03/83 10 3 Supervision 3 12/83 5 2 Supervision 4 05/84 12 2 Supervision 5 04/85 10 2 Supervision 6 04/86 5 3 Supervision 7 11/87 9 2 Supervision 8 04/88 7 2 Supervision 9 10 2 Supervision 10 5 2 Supervision 11 7 1 Supervision 12 10 2 Supervision 13 5 1 Supervision 14 14 1 Total 126 OTHER PROJECT DATA Borrowers Banco Nacional de Obras y Servicios PGblicos (BANOBRAS) Executing Agencies: Banco Nacional de Obras y Servicios Pfiblicos (BANOBRAS). iv PERFORMANCE AUDIT REPORT MEXICO SECOND IMR3AN AND REGIONAL DEVELOPMENT PROJECT (Loan 1990-ME) EVALUATION SUMMARY Introductiou 1. The Bank' s Second Urban and and infrastructure for existing Regional Development Project i. slums, social services, and credit to Mexico (hereafter Urban II) supported small businesses. Given the dynamic multiple investments in selected nature of the Isthmus region, the towns and cities along the Caribbean Bank attempted to develop a broader side of the Isthmus of Tehuantepec in project which would include the states of Veracruz, Chiapas, and investment in highways, bridges, and Tabasco. From preparation through river control works (land implementation, the project straddled reclamation) in addition to the a complicated era in whi;h Mexico's standard urban menus in the context economic boom turned to crisis, its of a line of credit. Presidents changed, and with them, policies concerning urban and 3. Project preparation involved regional development. At the same working with state governors on new time, the Bank's own policy focus institutional arrangements thus evolved and caused redefinition of bypassing regular channels of its urban lending concept. With communication. This became %concern regionalization of the Bank's urban to some Mexican functions-ies, and portfolio, projects included more generated tensions with other "country-specific solutions" and divisions in the Bank. The Mexican began to follow the sector-wide line Government requested the Bank to end of credit approach. In the case of project preparation and leave the Urban II, the desire to develop such country on two occasions and the a sector program relying on a series project's continuation twice required of local borrowing institutions, was intervention from the Regional Vice foremost in Bank thinking but not President's office. During this shared by the Mexicans (para 15). difficult process which lasted four years, the project underwent two Objectives and Description appraisals and two rounds of negotiations, which dragged on for 2. Both Bank and Mexican studies two and a half years. By this time, had identified the oil-producing local support had not been built up. region of the Tehuantepec Isthmus as Although the participating Mexican fast-growing with high potential for agencies stayed with the project, few economic development. At the time, a in Mexico could claim to "oW or typical Bank urban project followed a even fully understand the project. general formula which aimed to improve the living standards of the 4. The project included financing poor by offering a combination of new for shelter (serviced lots, urban housing (serviced lots), upgrading upgrading, and home construction v loans), citywide water and arrangements, the process of making sanitation, municipal facilities, sublana to municipalities and local productive credits, and additional entities becam extremely cumbersome. studies. A US$164.0 million Bank Bank missions likuwise turned into loan was expected to finance 35 large and unwieldy affairs. percent of total project costs, estimated at US$468.0 million (para 6. The investment profile 30). The number of components was contained in the SAR and the substantially pared down from the President's Report aggregates the Bank's original proposal and the components differently than in the project area reduced considerably Mexican proposal presented during from what the Mexicans had projet preparation (para 29). This anticipated. Urban II contained most took the form of an increase of s of the standard elements of US$50 Qillion in the water component "integrated" urban development and a corTGS,-)udiug reduction in the projects and was focused on low- shelter component. The project income families located in ten underwent another adjustment when the municipalities in the Gulf of Mexico decline in oil prices led to a oil producing region. The reduction virtual collapse in local government in the project area prior to approval resources in late 1982. The Bank increased project-related investments offered to pay 70 percent of project per family to about US$5,000, costs (with the same loan amount), equivalent to about 7.7 years of while the need for counterpart funds median family income, making it fell to 30 percent. According to extremely difficult for the calculations made by the Audit, total participating towns to absorb the project costs should then have full loan amount. Project goals were dropped to approximately US$270 exceedingly ambitious; they were million, or 581 of the original shared with the Mexicans too late in amount. But again no readjustment of the process (para 26). project targets appears in the files. Implementation Experience 7. Implementation proceeded slowly. At the original closing 5. Implementation was extremely date, less than 40 percent of the complex, involving entities from loan had been disbursed and the three levels of government in project appeared to be losing rather addition to BANOBRAS, the official than gaining momentum. About half of executing agency, which operated both the loan, US$81.8 million, was in Mexico City and the field. The reallocated for use in the project relied on the creation of six reconstruction effort after the new institutions and a multitude of Mexico City earthquake of 1985. "special units" within existing Another US$22.4 million was canceled agencies. These were to work in a in 1987. Total investment was multi-tiered fashion (under equivalent to US$95.6 million by guidelines and norms) with the closing, or just 35 percent of the approval of the line ministries from initial project target. While Mexico City, coordinated by the state investments made in shelter and urban planning commissions and operated and upgrading, particularly in Tabasco, maintained by local authorities or by were not iusignificant, they were trust funds administered by well short of the project's ambitious representatives of all three levels original scope and objectives (para. of government. Under these 45). Annual disbursements under the vi operation averaged US$12 million, project preparation relied on suggesting that the project (even projections of need that after revision of the disbursement excluded any analysis of costs, ratio) would have taken another benefits, capacity to invest, fifteen years to complete. or creditworthiness by participating local 8. There were numerous reasons for governments the project's deterioration. In addition to those identified by the Results PCR, the Audit highlights the following$ 9. Productive activities and studies components were canceled, (i) Local governments and corresponding to about 10 percent of institutions did not understand the original total cost, and the what was expected of them under municipal facilities component was the project. More seriously, dropped after investment, of just the fact that the Borrower was US$2 million out of an anticipated not made well aware of the US$40 million. Even reduced to material in the SAR suggests housing and infrastructure, that the Mexican functionaries investments remained far below had little idea of what the targets: less than 7 percent and 50 Bank envisaged as their role; percent respectively. Only small fractions of the targeted families (ii) After four years of preparation actually benefitted from the and an additional year prior to project's shelter components: 28 loan effectiveness, the project percent for serviced lots; 11 percent lost whatever momentum it had for materials loan.; and 19 percent mustered at the local level; for upgrading. Even though infrastructure was the more dynamic (iii) As oil prices dropped, revenues component, with investments totalling fell in wealthy Tabasco and almost US$80 million over six years, Veracruz states, as did the none of the contracts was carried to enthusiasm for large completion. Furthermore, with the investments. Even though the dissolution of one of the water disbursement percentage was companies (SAPAECH of Chiapas) and raised and BANOBRAS set up a the reorganization of another system for loaning counterpart (CRAPA), it has become impossibl, to funds to the municipalities, confirm which items were actually these changes took time to have built under the project, or at all. an impact; 10. The project's impact was (iv) With complex implementation extremely limited. A number of arrangements, processing serviced lots were built and slum subloans became a time- areas upgraded. No information consuming and cumbersome exists about beneficiary income process; and levels, so its poverty impact cannot be determined. The project's policy (v) The capacity of ten (or later objectives were so extensive (par". twelve) small towns to absorb 23 & 48), they probably could not the amount of money programmed have been met under the best of was never analyzed. In its circumstances. With respect to its PCR, BANOBRAS has shown that primary aim -- to make regional vii planning an integral part of the eroding effects of future governn-t policy -- the project inflatiot. According to the Bank's failed to the extent that it own projections at appraisal, undersc=.ed the comparative ease and inflation would have been higher than success of sectoral rather than the interest rate charged until the comprehensive approaches. Other last year of project implementation. policy concepts such as cost At any rates the concept of cost recovery, low-income housing recovery died with the executing solutions, and local participation agencies. The water supply company appear to have become lost. of Tabasco reports that its service charges cover only 10 percent of its Sustainabilitl operating and investment costs at the completion of the project. 11. The establishment of a self- Repayments to BANOBRAS clearly came sustaining mechanism for from federal government subsidies not comprehensive regional development from users. was not achieved under the project. The housing and infrastructure 13. Inteirated V3. Sector Avproach. investments have been continued and Urban 11 was the Bank's last sustained by virtue of on-going "integrated" urban development sectoral programs which have also project in Mexico. After 1982, both received Bank support (FIFAPA, the Mexican Government and the Bank followed by the water sector loans began to favor sector programs that FONEAPO and FOVI programs in the case separated housing, water supply and of housing.) Perhaps more municipal facilities and financed significant was the role played by time slices through lines of credit the project in pointing out the need rather than specific projects. The for greater institutional development project can be considered something of the municipalities, an issue of a transitions it confirmed the addressed by the FORTAMUN project difficulty to cover various sectors (para. 49). In this sense, the real in a single lending operation and led sustaining feature of the project has the way to projects involving single been the fact that search for subsectors, such as housing, urban municipal development and for a transport, and solid waste viable mechanism for sector lending management. was set in motion under Urban II. The Role of the Ban Points of Interest 14. Bank performance was poor. 12. Cost recovery. The subloans Preparation was long, and did not were contracted at fixed interest muster either local support for, nor rates, well below current levels of understanding of, the project. inflation. The Bank was setting Supervision, at six month intervals, subsidized, fixed interest on-lending was insufficient. Project records rates at a time when it calculated were badly kept. More serious loan amounts based on increasing perhaps Bank management appears to costs and charged adjustable rates have ignored the operation when it itself. The interest rates used in needed most help. connection with the housing component were in fact 6 points higher than those proposed by the Mexican Government,t a low margin in view of viii Conclusions and Lessons 18. The main lessons illustrated by the Audit include the following: 15. Urban II achieved few tangible results, limited to some city-wide (i) The hirtory of water supply and infrastructure improvements and housing projects in Mexico marginal progress in provision of suggests that the learning shelter. The Audit rates the process can be long and drawn Project's overall performance as out, but also assures that unsatisfactory; its institutional failures can result in changes achievements as negligible. With for the better (para. 61); poor performance in cost recovery, the sustainability of project (ii) Despite their theoretical benefits is unlikely. The PCR-based appeal, "integrated" urban ratings were the same. projects have rarely lived up to their promise. As borne out 16. The Bank's second urban project by subsequent experience in in Mexico attempted a very ambitious Mexicos lines of credit are approach to regional development by better managed when applied to simultaneously offering several lines individual separate subsectors of investment credit in different (para. 62); sectors. With no system of accountability and no real incentive (iii) Projects must rely on strong for institutional action, the project support within the client was unable to sustain much interest. country. Unlike Urban 11, the Disbursements and investments lagged. success stories in Mexico are What began as a multi-sectoral vision projects which have originated ended up as a series of limited and in and relied on a dedicated largely uncoordinated investments in group of professionals to carry water supply and housing. Neither them out (pares. 63 & 64); and the spacing of investments nor the number of beneficiaries of these two (1v) Projections of economic growth components appear to have differed or inflation have proven from the normal course of events. extremely unreliable and should not be used for setting f ixed 17. Thus, with its introduction of interest rates (para. 65). In a complicated administrative and addition, any line of credit to conceptual apparatus, the project municipalities should be based actually accomplished very little. on a prior analysis of their It certainly never "provided an financial and administrative attractive project model for capacity to borrow and repay replication project in other regions the loans. Preparatiu of any of Mexico." Furthermore, it fell out such program should also of phase with evolving national include a full review of priorities, did not establish a available sources of funding general credit mechanism for local and real demand for loans participation in development because (para. 66). it was spread over too many sectors, and failed to advance Bank-client 19. Finally, the project confirms dialogue on urban issues because it the importance of careful supervision lacked "owners" in either Mexico or and decisiveness in cutting losses the Bank. (para 67). Urban obtained the reputation of a "loser" early on, but ix the Bank was unable to make the changes needed to improve it or take action to terminate it, despite pleas from the project officer. While Bank management played a key role in pushing the project through preparation and negotiations, it vas notably absent when strong intervention was required during implementation. PERFORMANCE AUDIT REPORT MEXICO SECOND URBAN AND REGIONAL DEVELOPMENT PROJECT (LOAN 1990-MF) I. BACKGROUND 1. The Second Urban and Regional Development Project in Mexico (hereafter Urban II) was to support a multitude of investments in a selected number of towns and cities along the Caribbean side of the Isthmus of Tehuantepec (see map) in the states of Veracruz, Chiapas, and Tabasco. This report presents, from both Bank and Borrower perspectives, information which the Audit has attempted to clarify, complement, and on occasion, correct, but not repeat from the PCR. Thus, the Audit emphasizes themes not dealt with in other reviewe, specifically Bank-Mexico background, project preparation, and selected crucial aspects of project implementation, especially complexities due to project design and adjustments made after project approval. 2. From preparation to scheduled date of completion, the project straddled two presidential administrations and several changeovers in local government. In the case of Urban II, the intervening years also marked the end of Mexico's impressive thirty year period of sustained economic growth, the beginning of the debt crisis and a cycle of uncontrolled inflation, and the consequent refocusing of government's role. During the 1960s and 1970s, the public sector was looked at "to carry out direct investment in infrastructure and key industries such as power, steel, and petroleum, while creating a stable regulatory and institutional framework."' This involved building massive and highly visible construction projects, a tendency described as the "Aztec complex" by the succeeding generation of planners, who would abandon state-directed development in favor of modernization and deregulation in the 1980s. 3. When Urban II was conceived, both the Bank and the Mexicans thought of urban development in terms of physical planning and the creation of new cities and towns. In the second half of the 1970s, the Bank carried out a series of studies to support Mexico's spatial distribution policies (PCR, para. 2.01) and to identify the regions with greatest investment potential. These efforts were reflected in the following Bank reports: The Isthmus Development Study of 1976; Urban Development in Mexico, 1977; Spatial Structure and Urban Development, also 1977; and the Impact of Oil Income on Tabasco Regional Development in 1980. The first three documents specifically identified the Tehuantepec Isthmus as an area of high growth potential, while the last highlighted its future as an oil producing region. All stressed the importance of investing in regions where economic growth was most dynamic and in specific projects or sectors that could help bolster that growth. PCR of the L&zaro CArdenas Conurbation Development Project (Loan 1554-ME), Report No. 7006, November 1987. 2 4. The creation of SAROP in 1977 (PCR, para. 2.02) paralleled the spread of "planning fever" in Mexico.2 SAHOP was charged with producing the first National Urban Development Plan, which implied the production of urban development plans for all the municipalities of more than 100,000 population, as well as the 242 municipalities in the priority region of the Isthmus of Tehuantepec (BANOBRAS PCR, para. 7). At this stage, planning entailed an assessment of projected needs, but applied little financial or institutional analysis. The National Urban Development Plan followed contemporary planning concepts: it set target populations for urban areas and defined strategies to keep growth from exceeding these limits (mostly by making alternatives attractive); located industrial parks in key areas; emphasized an Atlantic- Pacific link; and proposed rationalized physical planning, which meant increasing densities in residential areas and lowering densities in central business districts. SAHOP's regional planning in the 1970s was not particularly concerned with low-income housing.S Although SAHOP's functions included construction, the agency did not move far beyond producing master plans before being reorganized in 1982 into a strictly normative Ministry, SEDUE,' of broader scope but with fewer powers. 5. In addition to administrative and policy changes in Mexico, the project straddled an important restructuring within the Bank; a revision of the Bank's conception of urban projects, and a redefinition of Bank's urban lending to focus more on an integrated approach and Borrower participation. At the end of the 1970s, the original Urban Projects Department was divided into regional divisions. 6. Throughout the 1970s, the Bank supported the Mexican Government's approach to development and financed both construction and a series of studies complementing the notion that government should orchestrate urban (and other) development through large-scale projects, together with zoning practices and fiscal incentives. By the early 1980s, with the debt crisis affecting the Bank as much as its borrowers, the focus shifted to financial issues and fiscal discipline. Finally, the same crisis and a succession of problematic projects led Bank officials to emphasize dialogue, collaboration, and more of a supporting role in the 1980s, compared with the 1970s when it used to take the lead in identifying and preparing projects. These elements combined make projects prepared in the 1970s somewhat out of fashion during the 1980s. 7. Design and development of Urban II reflected these changes at the Bank. The original independent department oversaw the promotion of the "integrated" 2 Interviews with former SAROP employees, September, 1992. (Translated from the Spanish "planitis".) 3 Interviews with the former Vice-Ministers of Housing and Urban Development of SAHOP, September, 1992. 4 Secretaria de Ecologia y Desarrollo Urbano - Ministry of Ecology and Urban Development. 3 approach, combining components from diverse sectors into single projects applying them with little variation in client countries around the world. With regionalization, project officers attempted to develop more "country-specific solutions" and, later, beginning to follow the "sector line of credit approach." Problems arose because the traditional definition of the "urban sector" included elements borrowed from several others. In the case of Urban II, the Bank's desire to develop a sector approach, and its associated line of credit for a series of local borrowing institutions, was not shared by the Mexicans and led to the problems of a project with little, if any, local support (paras. 9, 15, 18, 22, 34, 35 and 53). II. THE PROJECT A. Proiect Processing 1. Preparation 8. Bank and Mexican Government studies identified the oil-producing region of the Tehuantepec Isthmus (the states of Veracruz, Chiapas, Oaxaca, Tabasco, Yucut&n, and Quintana Roo) as fast growing with a high potential for economic development (PCR, para. 2.02; BANOBEAS PCR, paras. 7-8). During the 1970s, the Bank's Urban Project Department offered two general areas of credit to Mexico; one for tourism projects, soon to be discontinued' and another aimed at improving the living standard of the urban poor. The Tehuantepec Isthmus offered an attractive opportunity to attempt a regional development project (para. 1.03). By locating the first integrated regional operation in an oil-rich region which commanded full national attention as a new "growth pole," Bank urban staff hoped that Urban II would combine the best of the project-specific approach to a sector line of credit approach with a regional focus. Named PACDU in Spanish, the project promised a harmonious injection of resources to spur development of a key region. The Mexican vision was expressed in a 1979 proposal to develop potential growth poles on both coasts so as "to create or expand port and industrial infrastructure facilities." 9. Project preparation took four years, from 1977 to 1981. While the intentions of the Urban Projects Department and the task manager appeared honorable enough, the preparation strategy, which involved bypassing SAROP and direct liaising with state governors and one high-ranking official in the Ministry of Finance, alienated some groups both in the Bank and the Mexican Government. For a line of credit project, the preparation seriously lacked: (i) a strategy for promoting subprojects and coordinating their implementation; (ii) financing for subproject preparation; (iii) selection and evaluation criteria for After financing three projects in Mexico which assisted in the development of Ixtapa, Zihuatanejo and Los Cabos in Baja California. 2 Programa de Acci6n Concertada para el Desarrollo Urbano - Integrated Urban Development Project. 4 subprojects; and (iv) any analysis of the debt service and implementation capacity of target municipalities and institutions. 10. Tersions with Mexicans officials arose particularly over institutional arrangements, since the Bank asked for clarification regarding project management and felt that its counterparts were not completely responsive. An internal memorandum indicates some of the main problems encountered during preparation as: "the delay of almost 2 years in getting institutional arrangements defined ... the complete lack of coordination within SAEOP; ... failure to coordinate with state governors; failure to establish contacts with PEMEX; etc."3 The problem was compounded by the Bank's desire to replicate a formula recently applied to an integrated rural development project, "PIDER." This implied shifting responsibility for the project from SAHOP, the Ministry charged with urban and regional planning and development and official channel for sector dialogue with the Bank, and the Ministry of Finance to the (rival) Ministry of Planning and Budget (SPP). From the Bank's perspective, SAHOP's basic failure was its refusal to accept the institutional arrangements as defined by the Bank. 11. Seen from the Mexican point of view, Urban II and the Bank's approach could have appeared threatening for several reasons: general uncertainty about what the Bank was doing; vagueness of project content; and finally the institutional arrangements pushed by the Bank. In this case, the issue of control over the project was not adequately clarified. In particular, Bank missions appeared to be bypassing SAHOP, whose performance had proven disappointing in the faltering L&zaro CArdenas project (PCR, para. 5.04) but nevertheless remained the official Ministry for the urban sector, in favor of the small but powerful Secretariat for Planning and Budget (SPP). The issue of local participation, likewise proposed by the Bank, generated additional malaise. By 1979, the Bank's Country Program. Division was also involved in the fray, warning that the project had little support in Mexico, and that the urban missions were earning it a bad name. A typical memo from the Programs Division affirmed: During Mr. MclNamara's visit, the Mexican officials ... complained about the urban development missions ... More specifically, they complained of excessive and not focused urban missions, missions making excessive demands on the Government, failure to keep appointments organized by the Hacienda (i.e., Finance] officials, trying to play one Mexican agency against another and of inadequate briefing prior to leaving Mexico. Officials of SHCP, SPP, SAROP, and BANOBRAS voiced ... complaints ... that the missions were deliberately leaving them in the dark about its dealings with SPP, which is interested in controlling the project at the detriment of SAROP, to the point of avoiding to meet SAHOP staff ... [T]he impression is that the Bank wants to take away from SAROP and BANOBRAS designated responsibilities of the respective institutions and rely on SPP to deliver the program through State Governors. Throughout the week of July 16, 1 could not find a single taker for the project at the federal level, not even in SPP ... I am disturbed Internal memorandum, July 17, 1979. 5 at the continuing deterioration in relations. We should learn to maintain correct official relations, even if we do not always find it possible to respect their professional expertise.4 12. The deterioration in relations was reflected in the way the Bank's project officer tended to view preparation as a competitive rather than collaborative effort. A memo in the files states: "having worked now for some years in Mexico, I have learned to live with the common practice of the (very able and shrewd) Mexican officials to fish for opportunities and to score points wherever and whenever they can."5 It seems here that the Bank (project officer) was attempting to beat the Mexicans rather than help them -- an unfortunate attitude in the view of the Audit and one which almost certainly bodes as badly for a project as for Bank-client relations generally. In hindsight, the same project officer has reflected on the rather heady atmosphere in the Urban Projects Department at the time of project preparation: "we had the complete support of Bank management and we really thought we could do no wrong."' 13. Project preparation sparked its share of tensions within Bank as well, the most significant encounters occurring between the Projects and Programs Divisions. Most commonly, Programs Division pushed Projects Division to develop operations faster. In Urban II, however, the reverse occurred. Programs had numerous concerns about the operation, from its complicated design to its problematic effect on Bank-Mexican relations (para. 11). The regional Vice President had to intervene twice to keep the project alive.' Programs also defended the viewpoint of other divisions which objected to the multitude of proposed project components. 14. Urban II initially proposed to open a general line of credit to states and municipalities for investments in highways, bridges, flood control -- actually regional irrigation and drainage systems -- and pollution reduction all of which were considered the terrain of other sectors. While the regional development approach necessarily implied investments like these, which stretched beyond both the traditional municipal and urban sector limits, their combination in a single operation under the urban masthead proved unworkable early on. Other Bank sectors had developed independent strategies, on-going loans, financing terms, and project criteria which were not entirely consistent with those in the proposed isthmus development scheme. As preparation advanced, these "non-urban" components gradually dropped out or were reclaimed by their original sectors, leaving the project very much as a traditional urban poverty operation with components for housing, slum upgrading, small business credit and training, and some municipal works, presumably aimed at generating economic production. Despite the considerable reduction in scope, the project retained its initial 4 Internal memorandum, August 16, 1979. Internal memorandum, August 1, 1980. Interview with former task manager, September 14, 1992. Interview with former Country Programs officer, September 1992. 6 approach, identifying certain specific projects and relying on an open line of credit to finance further investments. 15. Slum upgrading implied a hefty investment in water and sanitation, not only for specific low-income neighborhoods, but also for entire low-income towns. In this case, the water and sanitation division accepted the idea of proceeding with aa urban project, albeit with some misgivings (para. 18), because of the poverty alleviation objective. Neither the poverty objective nor the development of a line of credit mechanism, however, were of great interest to the Mexicans, who had a different kind of project in mind (para. 8) and who saw both objectives as unnecessary in one of the country's wealthiest regions. Apparently to avoid problems, Bank missions simply never mentioned either aspect during preparation (para. 62). This led to misunderstandings which hindered implementation in the long run and complicated negotiations in the short run. 2. Appraisal 16. Project appraisal was first scheduled for 1978. However, the Mexican Government, uncomfortable with the line of credit approach and associated institutional arruagements, as well as the Bank's working method, requested a halt in missions until it could present its own proposal for institutional management. Eight months passed with no word from the Mexican Government and the programs division intervened. At their behest, the Mexican Government issued an invitation to appraise but ruled out the Bank's proposals for SPP's control and the "sector approach," indicating that the Bank should not count on further support from the high-ranking official in the Ministry of Finance (para. 9). Despite this, the Bank continued to insist on its preferred institutional arrangements during the appraisal mission in July 1979. The results were not satisfactory to the Mexicans. According to the programs officer, "[SHCP and SAHOP] were not clear as to what project the mission had appraised. During my meetings with officials of BANOBRAS and SPP they expressed considerable uneasiness with the appraisal mission and skepticism regarding the project."* Again the Bank was asked to suspend preparation until further notice and again no such notice was forthcoming. The project was finally rescued by higher level Bank management, who met with the high-ranking official in the Ministry of Finance during the annual Bank-Fund meetings. Together, they agreed on a new appraisal mission, but only on condition that SAHOP be left in command. 17. In the meantime, the Appraisal Report was generating extensive commentary within the Bank. The greatest preoccupation was with the large number of new institutions and subsidiary agreements proposed as conditions for negotiations (para. 27), although concern was also expressed about: (i) state participation in a project about which they still knew nothing; (ii) land acquisition -- 370 hectares were needed but only 36 hectares were well in hand; (iii) local financing -- 65 percent was to come from local sources, which was considered reasonable for states with high oil income (i.e., Veracruz and Tabasco) but difficult to defend for other states which were among the poorest in the federation (Chiapas, Oaxaca.) As a compromise, the project region was limited to 10 citien and towns in three states, all directly affected by the oil boom a Memo from Programs Officer to Files, August 16, 1979. 7 Veracruz (Coatzacoalcos, Minatitl&n, and Cosoleacaque), Tabasco (Villahermosa, C&rdenas, Cunduacan, Comalcalco, Paralso, and Macuspana), and Chiapas (Reforma). Unfortunately, the oil boom was waning just as appraisal was taking place. 18. During the original appraisal, the project was estimated to cost US$276 million and Bank financing would have covered 40 to 50 percent. A review from the Water Supply Division, however, questioned the institutional arrangements established to take care of the water and sanitation components, as well as their costs and recommended a second appraisal. The Water Division's desire was reinforced by the Mexicans, who were neither well-informed about, nor particularly interested in the low-income housing component, but were interested in water supply (although not perhaps in the context of slum upgrading). The project underwent a second appraisal, which may have led to the changes in the Loan Agreement investment profile (discussed in para. 29 below). A year later it emerged with a still more complicated institutional set-up and around 60 percent more costly. 3. Negotiations 19. Negotiations began in October 1980 but broke down over the issue of interest rates for the housing component. In fact, with inflation at 26 percent, it would have made little difference which rate was selected, given that both options were fixedt the Bank's proposal of a fixed 15 percent or the Mexican proposal of a fixed 9.5 percent over eight years. Each side held to its own projections for a coming downturn in the inflation rate (para. 64). Issues were finally settled during a second round of negotiations in March 1981, nearly two years after the first appraisal mission. At this time, the Mexican Government proposed a new solution for project management, BANOBRAS. This settled the internal dispute among the participating agencies (i.e., SAHOP and SPP) by naming a totally new one with no experience in implementing urban programs, but, as it turned out, with a sincere willingness to learn over time (para. 48). 20. On other points, a compromise interest rate for housing (10 percent) was accepted, but the housing and credit components were reduced (and subsequent actions left certain doubt as to the initial intentions of fulfilling them at all) to the benefit of the water supply component. Curiously, no mention is made of these decisions in the notes in the project files on the negotiations. The lack of information in the files and the inability of the persons involved to recall events leaves the Audit with contradictory documents (para. 39) and with no satisfactory explanation for what actually happened. 21. The project which finally emerged suffered serious design inconsistencies, largely because of a large number of fundamental changes made during preparation, appraisal, and negotiation which were never really assimilated. The operation remained a visionary regional development and multi-sectoral integrated program in terms of its goals and institutional arrangements. But in terms of components, it was simply an integrated shelter project, typical of the 1970s. In terms of budget, the project could have financed the multi-sectoral development needs of a region, but, in actual area, it reached only ten towns with a total population of half a million. Despite the changes in project costs and investment profile, local affordability was never reassessed, nor were 8 project areas or physical targets, even though more than two and a half years went by between preparation of the SAR and finalization of the Loan Agreement. 22. More troubling, the Bank's insistence on its own approach and institutional arrangements eroded any support for a project which had never generated much entlusiasm in Mexico to begin with. This should have been obvious by the Mexican Government's attempts to stall the appraisal missions and its evident reluctance to negotiate. The persisting preparation delays were certainly commented on by the project officer (para. 10) and the programs officer noted the Mexican's gneral disenchantment with the whole exercise. Given these messages one can ask in hindsight why the Bank continued to push the project. In answer, the programs officer at the time resa1ls that Bank management was pressing to conclude a large loan with Mexico and that SHCP needed to fill out its annual pipeline. It would also appear that the eagerness of the task manager to process the loan led to a misreading of all the signals which implied a polite latin refusal, and to the decision to forge ahead with a project that had virtually no real support in Mexico and very little within the Bank. B. Project Description 1. Obiectives 23. The PCR (para. 3.01) lists only the first three objectives mentioned in the SAR (para. 1.25). The SAR, however, actually went on to fill two more pages and eight paragraphs with assorted additional goals, includings' (i) to address the global problems of a region in a concerted manner; (ii) to help establish an institutional mechanism through which state and local authorities could effectively participate in programs which affect them; (iii) to establish a financial mechanism that will permit the simultaneous channeling of funds to a number of interrelated activities ... one of the necessary conditions for the efficient implementation of comprehensive urban and regional development programs; (iv) to resolve both critical housing and general infrastructure deficits and to provide complementary social and economic investments ... by broadening the application of cost recovery procedures already tested under the L&zaro C&rdenas project; and, (v) to provide an attractive model for additional such projects in other regions of Mexico.10 9 SAR, paras. 1.25-1.31. 1o SAR, paras. 1.25-1.31. 9 In addition of being unrealistic in the given situation, these objectives were not adequately reflected in the project design and its implementation plan. 24. In memoranda written to the Bank's Loan Comittee, the project officer and the division chief for country programs expounded furthers The primary objectives of this project would be to: (a) support the development of a high-potential region which has received the highest priority in the Government's national urban plan; (b) help the Government in its efforts to decentralize planning, implementing and operating responsibilities from the present highly centralized federal bureaucracy to state, municipal, and community levels; (c) establish the institutional capacity to carry out this and similar projects in the future; and (d) strengthen the Government's goal to assist low-income urban groups with self-sustained programs that will enable these groups to become productive partners in Mexico's development. 25. After reading the SAR, the senior country economist spelled out the problem which the multitude of project goals represented: "Overkill. I suspect that we are overloading the appraisal document with objectives the loan would never be able to meet. Thus, the project objectives are such a tall order that the Mexican Government would need about a decade and a whole army of civil servants to implement them. If we compare the project's objectives with the project's components, we can easily see that there is but a tenuous link between the two."1 26. In retrospect, the Bank's agenda appears to have been not only overly ambitious, but also overly secretive. Aide memoirs and other documents shared with the Mexicans make note of the Government's regional development and decentralization programs, but do not mention institutional reform, planning and implementation capabilities, or poverty alleviation among the project's goals. Any reference to project objectives is notably absent from the President's Report. Moreover, at that time, the Bank did not share SARs with the Mexicans as a matter of general policy.12 The Mexican project document, in turn, limits objectives to "establishing a program to issue credits for the realization of works of urban development as well as promoting the productive activities program." The significant difference between the Government's and the Bank's understanding of Urban II should have been a clear warning sign of serious problems during implementation (para. 62). 11 Memorandum, April 10, 1980, para. i. 12 Interview with BANOBRAS official, September, 3, 1992. 10 2. Institutional Arrangemento 27. As finally approved, BANOBRAS was to act as the regional development bank, on-lending project funds to a variety of institutions, agencies, and trust funds at the state and municipal levels, where the different subprojects were to be implemented. The state planning commissions were to set priorities according to which the agencies would propose new projects, under guidelines, norms, and with the approval of the line ministries from Mexico City. All was subject to final review by the Bank in Washington. The idea was to include the local authorities in the planning and implementation process, and these arrangements certainly did just that. But as they included state and federal governments, together with one international agency, they made the project development process extremely long and complicated while diluting responsibility for project execution. 3. Proiect Area 28. As Part II of the PCR explains, the original project area spread over the entire Isthmus of Tehuantepec, including all towns in four statess Veracruz, Tabasco, Chiapas, and Oaxaca. The latter, with its 240 municipalities, justified the project's initial scale. Whether Oaxaca was eventually dropped because it did not fit the urban project profile, as the project officer claims -- it is a very poor state with no oil income which happens to border on the oil producing states -- or whether it was excluded because the Governor declared himself against the operation at the outset, as BANOBRAS engineers claim, is not a matter of record. As the regional development components were progressively eliminated during preparation, the project area was pared down accordingly, from a four- state region to a beltway running through three states and covering ten towns (para. 17). The reduction in project area meant the total population affected was on the order of half a million when the project was appraised, which means that the shelter component alone would have covered about half the population. As a result, project cost came to about US$5,000 per family, or about 7.7 years of median family income. During the last years of the project, when it came to a standstill, the Bank tried to help BANOBRAS expand the project area to include more towns and states. Two more towns, Estaci6n Juarez and Pichucalco in Chiapas, were added in 1983 but local absorptive capacity remained a serious constraint. 4. Components, Costs, and Financing 29. The PCR (para. 3.02) lists the items included in the project as they appeared in the SAR and in the President's Report. They are shown here in Table 1, together with their estimated costs both in dollars and as a percentage of the project total. Table 1 shows the cost profile and compares it to the proposal made by the Mexican Government during project preparation. The most important differences between these two profiles occur in the percentages assigned to housing and water supply. 11 Table 1 Mexico Urban II - Resource Allocation (US$ millions) SAR Mexican Request (Annex B)" Component Amount Percent Amount Percent Shelter-related 269.4 57.6 215.0 47.0 Citywide Infrast. 119.0 25.4 164.0 35.9 Municipal Facil. 41.1 8.8 42.0 9.2 Productive Activ. 33.9 7.2 31.4 6.9 Studies & Support 4.6 1.0 4.6 1.0 Unallocated - - - - Total 468.0 100.0 457.0 100.0 30. The original project costs were estimated at US$468 million (SAR, pg. 16), of which the Bank loan would cover US$164 million," with Bank financing accounting for approximately 35 percent of the total. However, in 1982, when the drop in oil prices led to a serious decline in local government resources, the Mexican Government requested a revision in the project financing terms. In mid- 1983, the Bank agreed to increase its share of project funding to 70 percent against a counterpart of 30 percent. Since the initial loan amount was maintained, the Audit calculates that, as a reflection of this change in funding arrangements, project costs were reduced to approximately US$270 million equivalent, or 58 percent of the original total.'? Again, it would be logical to assume that the project's goals would have been reduced accordingly (para. 56), but no readjustment appears to have taken place. 13 Contingencies allocated for various components. 14 Loan Agreement, Article II, para. 2.01. The Audit assumes that para. 11 of the Borrower's PCR contains a typographical error. is The Audit calculated investments on a year by year basis using data in the BANOBRAS PCR. 12 III. Implementation. Achievements, and Impact A. Time Delays, Cost Overruns, and Cancellations 31. The project moved extremely slowly. As the PCR points out, less than 40 percent of the loan had been disbursed by its presumed closing date and the project appeared to be losing rather than gaining momentum. Half of the loan US$81.8 million was reallocated for use in the reconstruction effort in hexico City after the 1985 earthquake. Another US$22.4 million was canceled in 1987.16 The project was finally closed in December 1989 with a total investment of just US$95.6 million, about 35 percent of the reduced (estimated) project target, or roughly 20 percent of the original target. Despite a marked pick-up in the last two years -- half of total disbursements came in 1985-86 -- annual investments averaged US $12 million, which, at this rate of expenditures, suggests that even the revised project would have required another fifteen years to complete. 32. Table 2 shows how the investment profile was also altered as the production and studies components were canceled (corresponding tG about 10 percent of the total). Finally, the municipal facilities component, originally expected to absorb some US$40 million, faded after generating investments of only US$2 million for a central wholesale market and food distribution center in Villahermosa (and a memorial structure for CArdenas, Tabasco). The original budget, which programmed 90 percent of the investment for Veracruz and Tabasco, with over 70 percent going to Veracruz for massive housing and upgrading programs, was subsequently overhauled. In the final analysis, Veracruz received over 40 percent of project funds, but the housing component was canceled after only 911 serviced lots were built and 157 building materials loans financed: As a result, Veracruz accounts for only 17 percent of project housing beneficiaries. 33. Another reversal affected expenditure patterns for housing and infrastructure components, which together account for 98 percent of the final project cost. While the SAR and the Loan Agreement both anticipated a significantly greater investment in shelter than in infrastructure, at project completion, infrastructure investments were some five and a half times those in housing. Problems in assessing the results of these investments are discussed in para. 40. What follows is a discussion of the reasons for the significant implementation delays, the ultimate cancellations, and the generally unpredictable course of events. 34. The PCR affirms that project implementation slowed down due to a lack of new proposals once the civil works identified at appraisal had been completed (para. 56). It also indicates that local governments and institutions did not understand what was expected of them under the project (para. 53). In support of this argument, the Audit found that the BANOBRAS engineers and local officers had no idea that the project was supposed to be receiving, evaluating, and 16 As the files do not provide complete information, the Audit's source is a BANOBRAS official, interviewed on September 4, 1992. 13 financing additional construction proposals.17 Furthermore, the Borrower's PCR, an extremely rigorous document of some 250 pages, makes no mention of the plan to generate new projects. Since the Mexicans were not made well aware of the material in the SAR until the project was already being implemented, they really could not have known how the Bank envisaged their role. The earlier predictions of BANOBRAS and the warnings of the Bank programs officer proved true (pareas. 16 and 26). 35. Why was there so little Mexican interest in the project? After such a long preparation period and another year of becoming effective, the project lost whatever momentum and energy it had mustered at the local level. National politics aside, the local governments and their economic situation both changed during project implementation. Oil prices fell and the Government instituted a new policy whereby PEMEX (the national petroleum company) revenues were appropriated into a general fund by the central government, rather than remaining under PEMEX's control to return to the oil producing states. Thus, state revenues fell in wealthy Tabasco and Veracruz, dragging with them the enthusiasm for large investments. The project's financing terms, which obliged municipalities to match every Bank dollar with two of their own, put the line of credit out of the reach of the project townsl. When the financing terms were reversed in 1983 (para. 30), it was too late to incorporate new projects into short-term municipal planning. As the participating towns had difficulty coming up with even the reduced counterpart funding, BANOBRAS opened a special credit line to finance counterpart resources, but this, in turn, required the establishment of a special account for World Bank financing, all of which took time. 36. Why were not the municipalities interested in borrowing? Another assertion of the Bank's PCR, that the municipalities had immense fiscal revenues or had access to other, cheaper credit, is disputed by BANOBRAS. Actually, the interest rate for on-lending was still extremely attractive, with inflation running at over 20 percent a years a fixed 10 percent for housing (SAR, para. 4.02), was preferable to the indexed competition -- another Bank loan through FONHAPO, and 17 Interviews with BANOBRAS officials, September 1992. The accepted explanation for the unspent project funds is that Oaxaca, with over 50 municipalities in need of basic services, was dropped at the last minute but the financing, which had been calculated to include Oaxaca, was never reduced. The BANOBRAS PCR cites local functionaries as follows: "in interviews, the municipal functionaries claimed not to know that the various water supply, sanitary, and other municipalworks systems could be financed through PACDU. This could well result from the changes which take place every three years in the municipal authorities which, combined with the inadequate supervision of the municipal works, hindered the new authorities from having an adequate notion of the same." (pp.158-159.) is Interview with BANOBRAS official, September 2, 1992. 14 9-15 percent for water supply"' (not the 18 percent mentioned in the PCR). Against the advantageous financial terms, municipalities had to consider the costs of PACDU's cumbersome procedures. The project's administrative structure meant that a subproject needed the approval of several different levels of government, and a positive review from Washington, a process that could easily outlast any governor's or mayor's term in office. Although tie SAR and the supervision reports mention several proposals for municipal facilities, all but two were financed directly by municipal revenues in the end, largely in order to avoid the alternative red tape. 37. The most serious problem, however, was the limited capacity of ten, or later even twelve, small towns to absorb the amount of money programmed (PCR 2.22). BANOBRAS' project team has shown that project preparation depended on projections of need which excluded any analysis of costs, benefits, capacity to invest, or debt capacity, as far as the local governments were concerned. Mayors and governors who did not rush to increase the indebtedness of their localities under the project probably had good reasons. 38. Why didn't the Bank push harder (a question raised in the PCR)? In fact, the Bank made several efforts to salvage the project, beginning with the change in the project financing in 1983 (para. 30) and the attempt to establish a special account to help with counterpart funding. The second project officer made valiant efforts to expand the project area either by including new states or adding further towns in the original states, but Bank management moved extremely slowly. The revised project financing took over a year and a half to be approved, the special fund two years, and the second project officer's memoranda for project expansion, like her recommendations for cancellation, went unanswered for over three years. 39. If project preparation could be characterized as "Bank-dominated," then implementation would have to be labeled "Bank-abandoned." After carrying out several preparation missions per year during the five years prior to negotiations, the Bank held a three-day seminar in Mexico to "launch" the project and then largely dropped out of sight (para. 56), implying that it could only go so far and, if the project was to fail, it would be a Mexican failure. Given the antagonisms the project had generated both in Mexico and the Bank, it did not take much for the project to be declared a failure, sotto voce. From the beginning, it lacked a group of promoters and believers in Mexico. And when the original task manager left the project, it lost much support in the Bank as well. B. Achievements 40. It has proven very difficult to confirm exactly what the project funded, although it is clear that the results were considerably less than those aspired to at appraisal. Only two sectors, shelter and infrastructure, enjoyed any significant investments under the project, but each fell far short of the expectations at project initiation. Less than 7 percent of the original target to The first loan to CEAPA of Veracruz was made at 9 percent. Thereafter the loans made to CEAPA, SAPAET, and SAPAECH were at 15 percent (BANOBRAS PCR, pp. 101-106). 15 for housing was attained, while the corresponding figure for infrastructure was less than 50 percent. Although the latter appears to have been the more dynamic sector (para. 33), a problem of definition makes even this uncertain because the infrastructure contracts included works for upgrading which was technically part of the shelter component. Even if there were a way to break down the contracts signed for infrastructure works to determine which part financed what items, an even more intractable problem emerges. None of the infrastructure contracts were carried to completion and, with the dissolution of one of the water companies (SAPAECH of Chiapas) and the reorganization of another (CEAPA), it has become impossible to confirm which items were actually built under the project's terms of reference, or were built at all.20 41. The main sources for determining what was accomplished under the project are Bank supervision reports, which tend to be contradictory, cheerfully reporting the construction and allocation of thousands of serviced lots, which subsequently drop out of discussion, by BANOBRAS and the state housing and water agencies. In any case, the matter is complicated because BANOBRAS evidently mixed funds and the state agencies are not clear on the source of their financing (PACDU, FONRAPO, FIFAPA, or a mixture). In Table 2, the Audit has compiled a comparative chart of project investment and beneficiaries based on information from BANOBRAS confirmed by local housing and water supply agencies. The results shown there are discussed by component below. Tatle 2 Mexico Urban II - Investments and Beneficiaries Amount Invested No. of Items (US $ millions) SAR Ln. Agrmt Adjusted* Completion SAR Completion Component Shelter 255 217 130 14.3 Serviced Lots 19,000 5,389 Materials Loans 10,000 1,131 Upgrading 14,000 2,700** Citywide Infras. 110 161 97 79.4 not included 38,000 (No. of connections) Productive Activ. 53 33 19 0 Individual Loans 1,800 0 Business Loans 600 0 Municipal Facil. 46 40 24 1.9 Total 468 470 270 95.6 * Audit estimate after change in financing ratio. ** Refers to projects in Chiapas (see para. 45) 20 BANOBRAS PCR. 16 1. Shelter Components 42. As Table 2 shows, only a small percentage of the target number of families benefitted from the project's shelter component: in serviced lots, 28 percent; in materials loans, 11 percent; and, in upgrading 19 percent. Even when the targets are reduced according to the adjusted project size and investment profile, the number of serviced lots comes to only 54 percent of the goal. It should be stressed that the idea of serviced lots was still not easily accepted by local authorities or state governors, which explains in part the withdrawal of the Veracruz housing agency from the program (para. 32). The housing agencies, arguing that they could build core houses for the same cost as serviced lots and make them affordable to the same income groups, continued to largely reject the idea of serviced lots until they were promoted under FONHAPO's national housing program. Another hindrance stemmed from the lack of available land. This issue, which has stymied many Bank housing projects, is more complex in Mexico because of the "eJido" communal rural land ownership system which greatly restricts the availability of land for urban expansion. 43. In all, five subdivisions with serviced lots were financed in Tabasco and one in Veracruz. Cost per beneficiary proved to be considerably less than the Bank estimate. Serviced lots averaged less than US$3,000, despite problems of land acquisition and infill, as opposed to the US$7,000 estimated at appraisal, which explains how the Mexicans could meet close to 20 percent of the proposed number of beneficiaries with less than 7 percent of the programmed investment. The discrepancy implies either that the SAR cost estimates were unsound (done, perhaps, without much local collaboration) or that the Mexicans made considerable effort to reduce housing costs during implementation. It has been observed in connection with a simultaneous project (FONAPO I) that the Mexicans did indeed work to reduce housing costs (FONHAPO, PCR, para. 5.04), but Bank estimates of housing costs were also higher for Urban II than FONHAPO I. 44. Consolidation of project sites has been uneven. La Parilla II in Tabasco, for example, consists of 783 lots, of which only about 15 percent have actually been built on and occupied today. In contrast, the other four subdivisions in Tabasco have grown into model communities and have stimulated recent private sector development on their outskirts. While serviced lots vent on to become a standard housing solution under subsequent low-income housing programs in Mexico, the materials loans program, which floundered under PACDU, was discarded on a national scale. 2. City-wide Infrastructure and Upgrading 45. While specific urban upgrading projects reached only 2,700 families in three cities in Chiapas, citywide infrastructure projects benefitted families in Veracruz and in Tabasco in much the same way. Water and sanitation were introduced into five towns in Tabasco, (Paraiso, Marcuspana, C&rdenas, Cunduac&n, and Comalcalco), serving 11,000 connections. Local authorities credit PACDUwith "a positive transformation of five towns which were total slums and would never 17 have been attended otherwise."1 They also claim that it inspired the subsequent upgrading of 2,000 irregular properties in Villahermosa (Miguel Hidalgo and La Selva). The city-wide infrastructure program was estimated to benefit another 18,000 families in Veracruz. Difficulties in evaluating such projects arise from the fact that a large percentage of financing was retroactive (33 million Mexican pesos), the new contracts were never formally terminated, and the terms of reference are not always specific. The Tabasco project did include the purchase of some 11,000 water meters in its terms of reference, although implementation was never confirmed. 46. Authorities in Tabasco credit the water and sanitation components with improving the population's health, but they also admit that technically, the solutions were not the most appropriate. Alternative solutions with lower O&H costs might have been more economic in the long run. There were also quality problems as in the case of three water tanks financed for the city of Villahermosa which are functioning only at one-third capacity because they were not built to adequate performance specifications. Altogether, the experience suggests to the Audit that the format of the integrated urban project does not lend itself to successful supervision of water and sanitation subprojects and offers a lesson regarding multi-sector and multi-specialty loans under the general umbrella of urban development. 47. As the PCR indicates, the component for productive economic activities disintegrated and the line of credit for municipal facilities financed a wholesale market in Villahermosa. The projected 2,400 loans to individuals and small businesses were never made. Given the frequent failure of these relatively small components in the Bank's urban projects of the 1970's, neither the PCR or the Audit dwell on this issue (one other municipal project was also financed, a memorial structure in the city plaza of C&rdenas, Tabasco). The SAR's vision of US$25 million in markets, slaughterhouses, refuse collection, and new industrial estates never materialized. C. Policy Impact 48. The project had such extensive objectives (para. 23), that no amount of policy change could meet its aspirations. As far as its primary aim -- to make integral regional planning part of government policy -- the project had the reverse effect to the extent that it underlined the comparative ease and success of sectoral rather than comprehensive approaches. However, the project served, albeit indirectly, as a forerunner of subsequent municipal development operations, but in itself led to no policy changes. Other policy concepts such as cost recovery, low-income housing solutions, and local participation cannot be said to have advanced. They were lost in the complexities of project management. 21 Interview, Manager of INVITAB, September 7, 1992. 18 D. Institutional Impact 49. The project fostered the creation of three state water agencies,22 two special trusts,23 a special unit for productive credits within BANOBRAS,a2 and a special committee within BANOBRAS.2 While the units within BANOBRAS are now long gone, the two special trusts together with two water supply companies (SAPAECH in Tabasco and CEAPA in Veracruz) have continued for better or worse, while SAPAECH (from Chiapas) was dissolved in 1984. The project's most important institutional impact appears to have occurred, however, in BANOBRAS. Urban II gave BANOBRAS a certain impetus to set up regional offices and to begin, albeit timidly, to exercise an important role as a vehicle for on-lending World Bank funds to local implementing agencies. Subsequent projects have developed BANOBRAS' role further and in recent years the agency has managed several very successful projects for the Bank. E. Poverty Impact 50. As designed, the project sought to have a significant poverty impact despite its location in the high-income part (Veracruz and Tabasco) of the Isthmus. No information, however, is available concerning the income levels of final project beneficiaries although it is known that serviced lots were affordable to families earning well below the median national income. IV. POINTS OF INTEREST A. Cost Recovery 51. The BANOBRAS PCR (November 1990) includes a long discussion of cost recovery by the on-lending institution (BANOBRAS), a topic rarely addressed in PCR9,26 especially in borrower PCRs. As of November, 1990, 47 percent of the subloans had been repaid to BANOBRAS, and, as of the time of the Audit, almost two years later, the loans had been fully paid off.27 However, as the PCR 22 SAR para. 3.06. 23 SAR, para. 3.09. SAR. para. 3.08. 25 SAR, para. 3.27. 26 Most PCR's shortly follow the end of loan disbursements, when recovery has barely had a chance to begin. The implementation of Urban II, on the other hand, dragged on for so long, and was extended to cover a good part of the earthquake construction. It thus offers an interesting example of a project partially paid for at its formal completion date. 27 Interviews with BANOBRAS officials, also confirmed by interviews with officials of INVITAB and SAPAET, September 1992. 19 points out (paras. 8.01 and 8.02), this is far from a sterling performance via a via cost recovery. The information available concerns the repayments by borrower-institutions, and not from final beneficiaries. All the subloans were contracted at fixed interest rates well below the level of inflation. The water company of Tabasco, the only one which still maintains information on the project, reports that service charges cover 10 percent of the company's operating coste (current and investment). SAPAET is presently undertaking a study of its pricing and charging practices and was expecting to reach an equilibrium point by the end of 1992. The single example of SAPAET suggests that the concept of cost recovery, as promoted by the project, was never clearly grasped by its participating institutions. B. Sustainability 52. The objective of establishing a self-sustaining mechanism for comprehensive regional development was not achieved under the project. As regards specific components, the housing and infrastructure investments have been continued and sustained by virtue of on-going sectoral programs which have also received Bank support (FIPAPA, followed by the water sector loans, FONHAPO, and FOVI). Perhaps more significant was the role that the project played in pointing out the need for greater institutional development of municipalities, an issue addressed by the FORTAMUN project (para. 53). In this sense, the real sustaining feature of the project has been the fact that search for municipal development and for a viable mechanism for sector lending, was set in motion under Urban II. C. Follow-on Projects 53. In its attempt to combine integrated urban development with a sector loan and in its component for "municipal facilities," Urban II was a forerunner of the First Municipal Strengthening Project in Mexico (FORTAMUN), approved by the Bank in 1986. Unfortunately, FORTAMUN came too soon after PACDU to absorb its lessons of the need for administrative simplification and the importance of dedicated national supporters. 'While considerably smaller and more focussed than its predecessor, FORTAMUN faced the same disbursement problems as PACDU. After six slow years of implementation with only about 30 percent of the loan funds disbursed, project resources are being reallocated to a disaster reconstruction operation in Guadalajara. Efforts continue to develop a municipal development loan in Mexico. So far, however, local support has not materialized and the Mexicans do not appear interested in rallying around a project concept brought from Washington (para. 62). D. Intexrated Projects versus Sectoral Programs 54. PACDU was the last attempt by the Bank to do an "integrated" urban development project in Mexico. After 1982, changes in approach led both Mexico and the Bank to favor sector programs which separated housing from water supply and from municipal facilities and financed time slices of on-going lines of credit rather than specifically identified projects. The first loan exclusively for housing was made in 1985 (FONHAPO I). On its successful completion, a second loan for FORAPO was approved, as well as two other housing loans which focus on the financial markets (FOVI I and FOVI II). In water supply and sanitation, after three FIFAPA loans, BANOBRAS finally considers that it has perfected the 20 program design in this sector. While municipal facilities have yet to produce a viable sector program, the Mexican Government and the Bank are honing solid waste and urban transport projects. In retrospect, Urban II, although as a failure itself, can be considered something of a "bridge" from the project to the program approach. As an "integrated regional sector development program," it demonstrated the difficulty of covering various sectors in a single lending operation and led the way for the development of the single sector loan. E. Deconcentration and Decentralization 55. The project was intended to deconcentrate investments and decentralize administrative and executive capacities. While it left local governments in much the same state as it had found them, the project did help decentralize BANOBRAS. As the executing agency, BANOBRAS found it could not manage the project from Mexico City, particularly as it began to take on other credit programs to local governments (FORTAMUN, FIFAPA, IDB-financed projects). In 1987, BANOBRAS put a new policy of administrative decentralization into practice and established permanent offices in each of Mexico's 32 states. Despite efforts by the Bank to strengthen municipal and local governments, decentralization in Mexico has been measurably successful only in terms of "deconcentration" of national agencies. While they still respond to their national headquarters in Mexico City, they have acquired greater capabilities than the local governments. V. THE ROLE OF THE BANK 56. Beginning with the four years required for project preparation and continuing on with inadequate project supervision and deficient documentation and record-keeping, Bank performance in Urban 11 was poor. While souring Bank- Mexican relations, the preparation team appears to have sold the project to the Bank without ever fully explaining it to the Mexicans (paras. 10, 26, and 63). The PCR (para. 9.01) notes that supervision at six month intervals was insufficient, and it could be -dded that missions actually came on an average of only once every nine months during the three crucial years prior to the earthquake. 57. The Audit found the project files incomplete. Memoranda and reports are missing, as is the Back-to-Office Report for the first supervision mission in February 1982. The PCR (Table 8-B), relying on the project files, fails to mention this first supervision mission, which is, in fact, alluded to in the report of the following mission and in terms of reference. The PCR also skips over the second supervision mission in July 1982, although the relevant report appears in the files. Finally, the PCR prepared by BANOBRAS (para. 1.01) was not found in the files. The Audit received another copy from BANOBRAS itself, but has found that the information contained therein does not always agree with that presented in the PCR. 58. The Bank appears to have had a bright moment under its second project officer, and the files include a highly complementary letter written by the Mexican authorities on the occasion of her retirement from the project. However, 21 her efforts to restructure or cut off the project were squelched either by Bank management's slowness to react or confusion in the Mexican bureaucracy. VI. CONCLUSIONS AND LESSONS LEARNED A. Conclusions 59. The Bank's second urban project in Mexico attempted a very ambitious approach to regional development by offering several lines of investmeat credit for different sectors simultaneously. The complicated institutional arrangements which were designed to include a vast array of agencies and ministries resulted in a very cumbersome and prolonged process for approving particular investments and, at the same time, left the project without any real support base at either the local or national level. With no system of accountability and no real incentive for institutional action, the project was unable to sustain much interest. Disbursements and investments lagged. Whatever was achieved resulted from the existing plans, capacity, and interest of different and disparate authorities, not from the inspiration of a new coordinated regional development effort. What began as a multi-sectoral vision ended up as a series of limited and largely uncoordinated investments mainly in water supply and, to a limited extent, in housing. Neither the sequence of investment nor the number of beneficiaries of the water and housing projects seem to have varied at all from what would probably have occurred in the absence of the project. 60. Thus, with its introduction of a complex administrative and conceptual apparatus, the project actually accomplished very little. It certainly never "provided an attractive model for (similar] projects for replication in other regions of Mexico." The project failed in its regional development objectives largely because it fell out of phase with national priorities. It failed in its attempt to establish a general credit mechanism for local participation in development because it was extended over too many sectors. It failed to generate or to advance Bank-client dialogue on urban issues because it lacked clear "owners" either in Mexico or the Bank. In summary, the Audit rates the project overall performance as unsatisfactory and its institutional achievements as negligible. With poor performance in cost recovery, the sustainability of project benefits is unlikely. This agrees with the ratings given in the PCR. B. Lessons Learned 61. Although extremely successful sector programs in housing and water supply have recently been developed, a workable program for urban development or urban management continues to elude the Bank in Mexico. The history of water, housing, and BANOBRAS in Mexico suggests that the learning process can be long and drawn out, but also assures that failures can help to force changes for the better. The current Director of International Operations at BANOBRAS has remarked that "PACDU showed us what mistakes to avoid in the future, even if we are slow in 22 learning."" He implied that when the urban sector is finally tamed in Mexico, the true usefulness of Urban II will become apparent. 62. Among the other lessons drawn by BANOBRAS were the followings (i) the danger of working in a limited area where any political (para. 31) or economic shift can throw a project off course; and (ii) that, despite their theoretical appeal, integrated development projects have rarely lived up to their promise. A parallel can be drawn with the behavior of a commercial bank. Few banks will give individuals or companies single loans for a variety of purposes. Even though the guarantor is the same person in the final analysis, the guarantees, loan terms, and payment levels usually differ for home mortgage loans, car loans, education loans, and business loans. 63. One lesson, to date learned better, perhaps, by the Mexicans than the Bank, involves the importance of project support within the client country. A corollary would be that support is best mustered by sharing information and intentions. Neither worked well in Urban II (paras. 10, and 26). The success stories in the urban sector in Mexico are projects which originate with and rely on a dedicated group of persons to carry them through. Urban II lacked supporters in the country who believed in -- or for that matter understood -- its Bank-intended objectives and approach. Nonetheless, as the history of project preparation demonstrates, the Mexicans accepted the Bank's proposal, albeit after prolonged discussion. In recent years, the Mexican Government has shown itself more resistent to ideas or proposals which do not enjoy the full support of the relevant Government institutions. Projects have been stopped during preparation at Government instigation, an important change from the 1970s, when Government occasionally allowed discussion to drag on until the Bank tired of project preparation or resulted in operations such as Urban II. 64. Ownership, in short, comes up as a significant issue. A review of project preparation leaves the distinct impression that the project was the brainchild of the Bank with very limited support in Mexico mainly by a few individuals. But a distinction should be made between individuals and institutions. And the project had, in fact, no institutional base, and no group of persons who actively "claimed" it for their own. Project preparation, and appraisal must take into account the need for clear institutional ownership and support during implementation. This is to enable changes in that institution to be sustained whether there is staff continuity or not. 65. Project preparation suffered from various analytic flaws. For the most part, these have been corrected in subsequent sector operations, but they are still worth repeating. Among them, cost recovery and interest rates are delicate issues which have now been settled in Mexico by overall agreements with the Government. However, this project and several others which followed clearly cast doubt on the wisdom of relying on projections of economic growth or inflation. The projections used invariably showed income increasing steadily and inflation decreasing from the moment of the analysis. Reality has rarely followed such a course. 28 Interview, September 2, 1992. 23 66. Another lesson involves the importance of fully assessing the administrative and financial (including repayment) capacity of municipalities before designing on-lending programs. In addition, preparation of any such program should include a full review of available sources of financing and real demand for subloans. 67. Finally, the project teaches the need for supervision, fast reactions and fast cutting of losses, especially in difficult projects. Urban II obtained a reputation as a "loser" early on, yet no one in the Bank was able to make the necessary changes to improve it or take action to cut it off, despite the mounting pleas in unanswered memoranda from the second project officer, who followed in the tradition set by her predecessor of reporting the project as experiencing only "moderate" problems, notwithstanding attempts to improve disbursements or finally to cancel it. This experience points out that Bank management must be willing to make decisions and take actions in difficult cases. Bank management played a hefty role in pushing Urban II through preparation, appraisal, and negotiations (para. 39), but appears to have been largely absent when needed during implementation. 24 Annex 1 Page 1 of 10 Comments from the Borrower (Translation from Spanish original) International Financing Agencies Department Public Credit Division Secretariat of Finance and Public Credit United Mexican States Ref. No.: 393.111.1.155 Subject: Comments on the Performance Audit Relport for the Second Urban and Regional Development Project (Loan 1990-MEl Mexico City, May 12, 1993 Mr. Yves Albouy Director, Evaluations Operations Department Infrastructure and Energy Division The World Bank Dear Mr. Albouy: I am writing in reply to your letter of March 25 seeking our comments on the Performance Audit Report prepared by your Department for the Second Urban and Regional Development Project. Below I summarize our comments, in particular the conclusions we feel must be taken into account in the preparation and execution of other projects: Problems arise with projects designed by the central government authorities and not enjoying local support, since if the state and municipal authorities are not involved, the needs to be addressed by the project cannot be accurately assessed. Moreover, multisectoral regional projects require greater capacity on the part of the state and municipal authorities, which may be lacking. Sectoral programs that do not involve too many people in their implementation tend to be more successful because monitoring is much easier. The Second Urban and Regional Development Project failed because, in addition to being very ambitious in its objectives, these were viewed differently by the Bank and the Government. Additionally, the Mexican portion of the project was very paternalistic, involving considerable red tape, which slowed down the pace of funding and ultimately sapped interest in the project. 25 Annex 1 Page 2 of 10 With regard to the municipalities and final beneficiaries, creditworthiness was not examined before the works were carried out. Consequently, there was no interest in monitoring cost recovery or works maintenance, as there was no local accountability residing with either the municipality or even the final beneficiary. As regards the contents of the report, in general we agree with the data and information given. However, we would like to see several statements eliminated or clarified: Para. 3 of the Evaluation Summary. "The Bank's Urban Projects Division may have misread signals which should have been interpreted as a polite latin refusal, when it forget ahead with the operation." This should be clarified, mentioning the signals in question, or otherwise eliminated. Para. 12 of the Evaluation Summary. The report states that service charges cover only 10% of Tabasco's operating and investment costs. It should be clarified that this was true when the Second Development Project ended, but that at present there is an efficient water supply company. This suggestion is based on the fact that progress has been made in sectors such as housing and transportation thanks to subsequent Bank loans, but no mention is made that there is also a water program. (See comment on para. 51 of the report.) Concerning the body of the report, we feel certain changes are in order: Para. 2. Eliminate "Given Mexico's highly centralized political system, presidential changes imply policy changes (PCR, para. 4.03)." While this was true in the past, recent administrations have had continuity in their policies and priorities. Para. 11. The statement "Institutions within the Mexican Government maintain a very delicate equilibrium of functions and political weight, so that small shifts can cause major political imbalances. In this case, the issue of control over the project produced just such a shift." suggests a weak government and should therefore be stricken. Alternatively, the text could say that it was not clear as to what federal entity was to monitor the project. Para. 35. "...that national priorities shifted between the time when the project was prepared and when it was finally implemented." Same comment as for para. 2. Para 51. There is no mention that the water companies were consolidated under the current First Water Sector Project, financed in part by the Bank (Loan 3271- ME). However, other sections of the report refer to FONHAPO housing loans, explaining the subsequent attainment of certain objectives. 26 Annex 1 Pace 3 of 10 I hope that these coments can be reflected in the final version of the report. Very truly yours, Lorena Gabilondo Alp1zar for the Director, International Financing Agencies Deputy Director, Services Sector Projects - 27 - Annex I Paee 4 of 10 Directorate of Development Financing Subdirectorate of Urban Infrastructure Financing Division of Water Supply and Sanitation Financing MEMORANDUM Subject: Comments on the Performance Audit Report Second Urban and Regional Development Project (Urban II) The Performance Audit Report on the Second Urban and Regional Development Project (Integrated Action Program for Urban Development of the Gulf Coast Region and Tehuantepec Isthmus - Urban II) was prepared by World Bank representatives in the fourth quarter of 1992 and the first quarter of 1993, to evaluate the Urban II project and identify the weaknesses that prevented it from meeting its goals. Since an Audit Report has to adopt both an analytical approach and a broad-based perspective, this document was based both on documents and on personal interviews. Following are certain points of the Audit on which we should like to comment: L Background In the OBackground' section, the Audit Report refers to the studies carried out in support of Urban II. However, from preparation to implementation, the project passed through a difficult era in which the oil *boom* was followed by a severe crisis, with the resultant changes in government policy in relation to urban and regional development programs. This was also the time when the World Bank was redefining its lending policies and developing 'country-specific solutions.* It was this approach that led it to promote Urban II, regardle- of the fact that the Mexican Government was not fully convinced about participating in projects of this type. In this section, we would add that the agreement on Loan 1990-ME was reached basically because of the following two factors: A. At the time, the World Bank did not have any loan requests from other members that would enable it to introduce its new lending program, and Mexico, with its sound economy, was a viable candidate for a project of this size, for which reason it decided to offer this urban and regional development loan to the Mexican Government; B. During the presidency of Jos6 L6pez Portillo, certain SAHOP officials expressed interest in carrying out a development program in the then Priority Region of the Gulf and Tehuantepec Isthmus. Subsequently, with the policy changes introduced during the presidency of Miguel de la Madrid, urban development programs were relegated to second place, with the result that the programs proposed by SEDUE in that area were not adequately supported, a situation that weakened the significance of and interest in Urban II. - 28 - Page 5 of 10 II. The project In this chapter, tLe report comments on the opportunities offered by Urban H within the project area, and notes that project preparation took four years, owing to excessive red tape resulting from political confrontation with Mexican officials. The Audit states that in order to expedite preparation of Urban H, Bank officials involved certain ministers in participating under new institutional arrangements, bypassing the normal channels of communication. This was confusing to some of the Mexican officials, and created tensions with other Bank Divisions. As a result, the Mexican Government asked the Bank to end project preparation and, on two occasions, to leave the country. The Bank's Regional Vice President had to intervene to persuade the Mexican officials to continue with the project. In order to complete preparation of Urban II and come to an agreement with Mexican government officials, it was necessary to pare down the project goals, which would have involved more government agencies, leaving only the program of assistance to slum dwellers [clases marginadas]. However, although government agencies took part in its preparation, it may be said that there never was any interest on the part of the government officials in implementing an urban development project of this scale. Preparation of the Staff Appraisal Report fell seriously behind schedule owing to: (i) lack of interest on the part of the Mexican authorities in participating actively in its preparation; (ii) shortage of land for the shelter programs; and (iii) absence of agreements on counterpart contributions and interest rates. Similarly, the Audit notes that during the Loan 1990-ME negotiations, differences arose because of lack of agreement concerning the interest rates to be charged. The *Project Descriptiona section comments on the project objectives and components. It notes that despite the ambitious nature of the program, and the reduction in project area, project cost came fo about US$5,000 per family, or about 7.7 years of income, which put a strain upon the capacity of local governments to absorb the total cost of the loan. IIL Implementation The Audit Report notes that project implementation was extremely complex, owing to the fact that it involved the three levels of government in addition to BANOBRAS. The program promoted the creation of six new institutions and a large number of *special units* within existing agencies. By the original closing date, less than 40% of the loan had been disbursed, and the project seemed to be losing momentum. Total investment in the project was US$95.6 million, or 35% of the proposed goal. With an average annual disbursement of US$12 million, it would have taken 15 more years to complete disbursement of the loan. In addition to the issues set forth above, various other reasons also contributed to the project's deterioration, including the following: - 29 - Annex l Pane 6 of 10 (a) Officials at the three levels of government did not understand what was expected of them under the project. The Audit comments that a serious factor was that BANOBRAS did not apply the suggestions offered in the Staff Appraisal Report concerning the level of understanding of the program on the part of the Mexican officials; (b) After four years of preparation and the additional year preceding loan effectiveness, the project lost its original momentum and thrust; (c) With complex implementation arrangements, the onlending processes became extremely long drawn-out and difficult. Combined with this, the enormous fiscal deficits of the local authorities [ayaamlentos] forced them to seek financing with more concessional interest rates and with more flexible contracting procedures than under Urban 1; (d) The capacity of 10 (and later 12) small towns to absorb the money programmed was never analyzed. In the final report, BANOBRAS showed that project preparation excluded any analysis of costs, benefits, capacity to invest, or debt capacity, as far as the local governments were concerned. IV. Achievements In this chapter, the Audit Report examines the achievements under each of the project components, concluding that project impact in the area was limited. A small number of serviced lots were constructed and slum areas received basic services, but this might perhaps have come about more rapidly without Urban II. V. Points of Interest Cost recovery In this section, the Audit makes the point that subloans were contracted at compound interest rates below prevailing inflation levels. The Bank set subsidized compound interest rates for the project subloans at a time when it was itself charging variable interest rates throughout the world. Integrated Projects versus Sectoral Programs This section notes that Urban II was the last Integrated Urban Development Project in Mexico. After 1982, both the Mexican Government and the World Bank started to favor sector programs which separated housing from water supply and from municipal facilities, preferring to finance lines of credit rather than integrated projects. - 30 - Ann.1 Page.7 of 10 Deconcentration and Decentralization The report comments that one of the project's achievements was its support for the decentralization and deconcentration of public functions into offices located in the different States, as in the case of BANOBRAS, which opened 31 branch offices in 1987. VI. The Role of the Bank The Bank's performance is rated poor, owing to the following facts: (i) preparation of the project took an extremely long time, and did not eqjoy local support; (ii) project supervision was inadequate; and (iii) the project records were badly kept, the Staff Appraisal Report containing discrepancies, errors and omissions in relation to the President's Report and the Loan Agreement, and the project reports in turn containing contradictory information. What is most serious is that Bank management apparently ignored the project even when its help was most urgently needed. VII. Conclusions and Lessons Learned (a) The Urban II project had few tangible results, other than certain improvements in the area of infrastructure and marginal progress in provision of shelter. The Audit rates the project's overall performance as unsatisfactory, with poor performance in cost recovery and little likelihood of sustainability of project benefits. (b) The project attempted an ambitious approach to regional development by offering several lines of credit in different sectors, but, lacking accounting systems and any real incentives for institutional action, the project was incapable of awakening much interest. (c) With the introduction of a complex administrative and conceptual apparatus, the project actually accomplished very little. It never tried to be an attractive model for replication in other regions of Mexico, and even when it attempted to observe national priorities, it did not establish general credit mechanisms for local participation in development because it was extended over too many sectors. It failed to generate adequate dialogue on urban issues between the Bank and BANOBRAS, and there was no clear "ownership' of the project, either in Mexico or in the Bank. (d) The Audit includes the following among the lessons learned: (i) The history of water and housing projects in Mexico suggests that the learning process can be long and drawn out, but also assures that failures can help to force changes for the better. (ii) Despite their theoretical appearance, 'integrated' urban projects rarely live up to their initial objectives. As has been shown in Mexico, credit lines are better managed when applied on the basis of sectoral criteria. (iii) Projects need true technical and political support within the client country. - 31 - An..1 Page 8 of 10 (iv) Projections of economic growth or inflation have proven to lack reliability, and should not be used to set compound interest rates. In addition, any credit line to a municipality should be based on prior assessment of its financial and administrative capacity to request and repay loans. Likewise, preparation of any program should include a full review of available sources of financing and real demand for loans. Finally, the project confirms the need for careful supervision and decisiveness to avoid losses to the borrower, resulting from attempts on the part of the Bank to keep a project alive when it is not receiving the requisite local aupport. Observations and Comments The Audit Report is a useful document, although it is viewed as a delayed effort on the Bank's part to determine the reasons for the failure of the Urban II project. Following are some of the factors omitted by the Audit Report that would have been crucial to the project's success: 1. Lack of acreage for housing, resulting from the scarcity and high price of private lands, and SEDUE's tardiness in expediting the formalities required for the transfer of federal lands earmarked to become part of the State and municipal territorial reserves. 2. Land purchases in regions unsuitable for housing construction, which, once urbanized, proved too costly in relation to their sale price. 3. Changes in policy occurring during the presidency of Miguel de la Madrid relegated urban development programs to second place, with the result that the programs proposed by SEDUE were not adequately supported, a situation that weakened the significance of and interest in Urban II. 4. The rigid financial framework of Urban H and the lack of support from Bank management made it impossible to shift any of the financing to other sectors or agencies to promote and consolidate programs of social interest or projects to create new industries and generate employment. 5. The Bank's supervision missions were understaffed and lacked decisionmaking authority, showing little interest or understanding of the Urban 11 program or in the methods of operation and administration at the three levels of government, and lacking the support of specialized technical personnel. This absence of decisionmaking authority was patently obvious during meetings with Mexican officials, as was the lack of clearcut guidelines for implementation and development of Urban II, which over time led to an attitude of mistrust and rejection toward the project, the impression being that the Bank was imposing its desires and goals rather than aiming to achieve real benefits for the region at the federal, State and municipal levels. June 7, 1993 - 32 - Annex 1 Pante 9 of 10 Servicios do Agua Potable y Alcantarillado del Estado do Tabasco - SAPART (Water and Sewerage Company - State of Tabasco) Villahermosa Tabasco Mexico May 6, 1993 Mr. Yves Albouy Chief, Infrastructure and Energy Division Operations Evaluation Department World Bank Washington, D.C. Subiects Comments on World Bank Loan Dear Mr. Albouys I am writing in reply to your letter dated March 25, 1993, received on April 19, 1993, in which you asked us for our comments on the Performance Audit Report on the Second Urban and Regional Development Project (Loan 1990-ME). I have the following comments to makes During the seventies the State of Tabasco experienced the benefits and problems inherint in the development of oil production, which impacted the immediate production areas directly but also affected the state as a whole. Faced with this situation, the government authorities in office at the time embarked on the task of creating the social infrastructure needed in order to meet the ever-growing demand for services. The shortfall in these services was very high, however, particularly where water supply and sewerage services were concerned. The magnitude of funds needed was such that it proved necessary to negotiate a loan from the World Bank, which was used to establish the Water Supply and Sewerage Program for the Development of the Urban Areas of the Tehuantepec Coast and Isthmus, known by the acronym PACDU, which continued until 1987. - 33 - Annex 1 Paae 10 of 10 Thanks to this loan it was possible to fully rehabilitate and expand the water supply and sewerage systems of five municipal seats and to construct sewage treatment plants. The Loan Agreement was signed in 1984 for more than Mex$5.1 billion, benefitting five cities, viz. CArdenas, Comalcalco, Cunduac&n, Macuspana and Paraiso, and improving the living conditions of a total population that numbered 106,000 at the time. It was thanks to this financial assistance that the infrastructure created in the above-mentioned municipal seats became fully operational, in accordance with the corresponding projects, which meant that the funds granted by the World Bank had a far-reaching impact on the region. I hope that you find the foregoing comments useful and remain, Very truly yours, /s/ Andr6a Granier Helo Director General, SAPART Villahermosa, Tabasco