FOR OMCIAL USE ONLY. Report No their official duties. Its contents may not otherwise be disclosed without World Bank authorization.

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Document of The World Bank FOR OMCIAL USE ONLY PROJECT COMPLETION REPORT HAITI URBAN DEVELOPMENT PROJECT (CREDITS 1338-HA & SFI-HA) DECEMBER 21, 1992 Infrastructure Division Country Department III Latin America and the Caribbean Regional Office Report No 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.

2 FISCAL YEAR October 1 to September 31 CURRENCY EQUIVALENTS Currency Unit Gourde (Gourdes) Exchange Rate Effective January 31, 1991 US$1.00 Gdes 5.00 US$0.20 Gdes 1.0 WEIGHTS AND MEASURES Metric System ABBREVIATIONS CDB EPPLS ERR GOH KFW MTPTC ONL SAGMP SAR UNDP Croix-des-Bossales Public Social Housing Promotion Company Economic Rate of Return Government of Haiti German Reconstruction Loan Corporation Ministry of Public Works, Transportation and Communications National Housing Office Metropolitan Market Management Agency Staff Appraisal Report United Nations Development Programme

3 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C U.S.A. Office of Director-General Operations Evaluation December 21, 1992 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on Haiti Urban Development Proiect (Credits 1338-HA and SFI-HA) Attached is a copy of the report entitled "Haiti - Urban Development Project (Credits 1338-HA and SFI-HA)" prepared by the Latin America and the Caribbean Regional Office. No contribution was received from the Borrower. Project implementation suffered from poor design and unfavorable country conditions. The closing date was extended by four years in total. Physical achievements were limited. Capacity building and policy development impacts were marginal. The project, typical of early generation urban projects, was too complex and demanding given the weakness of the institutional framework. The project outcome is rated as highly unsatisfactory. The Project Completion Report is of excellent quality. It provides, for each major component, an interesting discussion as to how implementation problems might have been avoided. No audit is planned. This document has a restrictedistribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without WorLd Bank authorization.

4

5 FOR OFFICIAL USE ONLY PROJECT COMPLETION REPORT HAITI URBAN DEVELOPMENT PROJECT (LNS HA & SFI-HA) TABLE OF CONTENTS Page No. Preface... Evaluation Summary... i ii PART I PROJECT REVIEW FROM BANK'S PERSPECTIVE 1. Project Identity Background Project Objectives and Description Project Design and Organization Project Implementation Project Results Project Sustainability Bank Performance and Lessons Learned Borrower's Performance Bank-Borrower Relationship Consulting Services Project Documentation and Data PART II BORROWER'S )MMENTS PART III STATISTICAL INFORMATION Related Bank Credits Project Timetable Credit Disbursements Project Implementation Status Project Costs & Financing Project Results: A. Direct Benefits B. Economic Impact C. Financial Impact D. Studies Status of Covenants Use of Bank Resources 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.

6

7 - i - PROJECT COMPLETION REPORT HAITI URBAN DEVELOPMENT PROJECT (CREDIT 1338-HA, SF1-HA) PREFACE 1. The following is the Project Completion Report on the Urban Development Project in Haiti under Credit 1338-HA and SF1-HA totalling SDR19.4 million (US$21.0 million equivalent). The Credits were signed on March 31, 1983, and the project was carried out from June 1983 through August 31, The original Closing Date of December 1987, was extended four times to December 31, 1991 when the Credits were to be closed. The Credits were not fully disbursed and SDR3.0 (US$3.9 million equivalent) will be canceled from the Credit Accounts as soon as the current suspension of disbursements is lifted. 2. This report was prepared by the Infrastructure Operations Division, Department III, of Latin America and the Caribbean Regional Office. It is based on the Staff Appraisal Report, the Credit and the Special Fund Agreements, information obtained from the project files. 3. Due to substantial personnel changes within the project executing agencies (EPPLS and MTPTC) that took place in the aftermath of the recent political upheavals, many of the key staff who were involved with the project are no longer available to prepare an analysis of key issues related to the project nor to make a retrospective assessment of the project implementation from the Borrower's perspective. Therefore, the Borrower did not prepare Part II of the report. The last IDA mission which collected data on the project visited Haiti in August 1991.

8

9 - ii - PROJECT COMPLETION REPORT HAITI URBAN DEVELOPMENT PROJECT (Credit 1338-HA and SF1-HA) EVALUATION SUMMARY Proiect Objectives i. The project's main objectives at appraisal were to finance the upgrading of the physical infrastructure and the institutional strengthening of the principal market in Port-au-Prince (Credit SF1-HA, SDR 12 million). The objectives were radically changed when their implementation proved impracticable in 1986/1987, due to project-related as well as political problems. Following the cancellation of the market component, Cr. SF1-HA was amended to finance a new component consisting of urban road improvements. The project's other component, aimed at providing low-cost housing (Credit 1338-HA, SDR 7.4 million), suffered fewer implementation problems. The components comprised (1) upgrading of the Croix-des-Bossales (CDB) market in Port-au-Prince, and provision of serviced sites for low-income families to be displaced from the CDB upgrading works site; (2) provision of 'sites and services' in Cap-Haitien and Les Cayes to improve living conditions of low-income groups (3) creation of SAGMAP, an agency to operate and promote the long-term development of markets in Port-au-Prince; and (4) institutional strengthening of EPPLS, the Public Social Housing Promotion Company (para and 3.02). ii. The market component was later replaced by an urban road component whose objectives were to (1) improve selected key roads in Port-au-Prince; (2) ameliorate maintenance of the city road network; and (3) to review the issues affecting transportation in Port-au-Prince (para. 3.03). Implementation Experience iii. Market component (para ). The Ministry of Public Works, Transport and Communications (MTPTC) was responsible for the implementation of this component; EPPLS was responsible to provide housing for those to be relocated from the market work site. Coordination between MTPTC and EPPLS was poor and IDA's supervision was limited. As a result, MTPTC delayed the contracting of the market upgrading works for two years (works were only to start in Jan. 1986). At that time, the construction of some 1,000 houses for the families living in the market had already been completed but the relocation was not initiated due to the lack of counterpart funds necessary to provide compensation payments to those to be displaced. In February 1986, Duvalier's regime fell, and, in the confused period that followed, part of the houses constructed for the market occupants were invaded by squatters; also, many market occupants no longer wished to leave the market because of the prevailing political instability. Under the circumstances, the contract for the market

10 - iii - upgrading had to be canceled, and the remaining US$12 million under Cr. SF1-HA were reallocated in 1987, to finance the urban road component. iv. Housing component (para ). Construction of about 1,000 shelters of acceptable quality was completed in Port-au-Prince by mid Provision of serviced sites in Cap-Haitien (883 units) and Les Cayes (254 units) was completed, with delays, by mid-1987, due to EPPLS's limited capacity and political problems. In 1987, IDA agreed to finance additional housing units in Cap Haitien and Les Cayes. However, due to problems with squatters, only the program in Les Cayes was implemented. Overall, more than 25% of the units have been illegally occupied by squatters. A new extension of the housing program was agreed in 1989, but no work was done due to constant changes among EPPLS high level staff and political uncertainties. The quality of the works done in Cap Haitien and Les Cayes was acceptable due to on-site visits by IDA supervision missions. EPPLS's cost recovery performance was poor throughout the life of the project (35% on average), mainly because of the difficulty to collect rents in areas where houses were occupied by squatters, and due to EPPLS's unwillingness to contract out rent collection in line with IDA's recommendations. v. Urban road component (para ). This component experienced major delays, cost overruns, expropriation problems and misprocurement resulting in the cancellation of a significant part of it. As a result, at the end of 1989, some US$11 million remained undisbursed. These funds were allocated to a new road maintenance project, including the paving of 7 km of Port-au-Prince secondary streets and periodic maintenance of about 330 km of interurban roads. Implementation of this new component was acceptable and was completed around mid (para. 5.20). vi. A fourth extension of the closing date until December 1991 was granted in April 1991 to finance a labor-intensive program requested by the Government. A political coup in September 1991 disrupted implementation of the program and prompted IDA to suspend disbursements. Some US$3.9 million still remained undisbursed under Credits 1338-HA and SF1-HA when the credits closed in December 1991 (para. 5.21). Proiect Results vii. Overall, the results of the project are hardly satisfactory. Its main component, namely, to upgrade the CDB market and create SAGMAP, had to be canceled. The only positive results of this project were the physical achievements under the housing and the urban road components (para ). viii. The achievement under the housing component was the construction of 2,661 low-cost housing units of acceptable quality, that benefitted some 10,000 people. However, the overall results of this component are unsatisfactory because (1) the basic services were not provided in a timely manner, or at all; (2) the project houses were often not occupied by the target beneficiaries; (3) the institutional objectives of the housing component were hardly met: average cost recovery never

11 - iv - exceeded 35 percent, and EPPLS failed to develop any sound housing sector policies. ix. The results of the urban road component are mixed. The initial component designed in 1987 suffered from major reductions of its scope. The works, mostly carried out by force account, were generally of poor quality. Concerning the revised road maintenance and labor-intensive components described in the June 1989 and April 1991 amendments, most of their objectives were met, including the improvement of important urban and interurban roads, and the initiation of the labor-intensive program. Those components represented a successful pilot experience and an opportunity to improve MTPTC's procedures for the management of road programs, in line with the reorganization of the Ministry to be implemented under the proposed Road Maintenance Project. Proiect Sustainability x. Whether the benefits derived from the construction of sites and services and the 2,661 housing units will be sustained will depend on the commitment of the beneficiaries to preserve their properties through regular maintenance. Given the difficult economic situation and the limited interest in maintenance works in Haiti, it is possible that these benefits will not be long lasting. The benefits from the rehabilitation and paving of a limited number of roads may not be sustainable in the future mainly because of lack of timely maintenance caused by delays in launching the proposed Road Maintenance Project. Findings and Lessons Learned xi. The project is a typical example of the early generation of urban projects: complex components with weak implementing agencies and poor coordination arrangements. The generally unsatisfactory results of the project strongly indicate the need (1) to pay more attention to implementation arrangements to ensure that sufficient capacity exists within project executing agencies and that coordination is well defined and agreed upon; (2) to encourage private sector participation in all "external" activities such as rent collection (EPPLS) and road maintenance (MTPTC); and (3) to put due emphasis on supervision missions whenever a complex project experiences major problems (para ). xii. More specifically, MTPTC should be encouraged to contract out most of the works and to improve its project management skills. Regarding the housing sector, emphasis should be given to the provision of basic infrastructure, and the review of the issues affecting the sector. The present EPPLS's role, which mainly amounts to providing houses, should be reviewed in this context, in view of the limited number of housing units constructed and managed by EPPLS compared to the overall country-wide demand.

12

13 PROJECT COMPLETION REPORT HAITI URBAN DEVELOPMENT PROJECT (Credit 1338-HA and SF1-HA) 1. Proiect identity I. PROJECT REVIEW FROM BANK'S PERSPECTIVE Project Name: Credit Number: RVP Unit: Country: Sector: Subsector: Urban Development Project HA and SF1-HA. Latin America and Caribbean Region Country Department III Haiti. Urban Development. Housing & Market. 2. Background 2.01 IDA's approach to the urban sector was based on the findings of an urban sector study conducted in 1977, which concluded that in view of Haiti's limited resources, policies of spatial decentralization should be approached cautiously and progressively, and that, in the short- and medium-term, the major priority should be the provision of urban services in Port-au-Prince. Simultaneously, institutional capacity and basic infrastructure were to be developed in selected provincial cities. This approach was closely linked with macro policy objectives to improve and expand basic infrastructure, and to strengthen local institutions The project was designed in line with the above sectoral and macro policy, to (i) address the most pressing needs in Port-au-Prince, Cap-Haitian and Les Cayes, the cities with the highest urban growth rates and most visible urbanization problems; and (ii) develop policies and institutions to deal with medium and long term urban problems. 3. Proiect Obiectives and Description 3.01 Proiect Obiectives. The main objectives of the project were to: (a) upgrade and rehabilitate the Croix-des-Bossales (CDB) market in Port-au- Prince, and to provide serviced sites for the low-income families who would be displaced as a result of the works necessary to upgrade CDB market services; (b) improve the living conditions of low-income groups in Cap-Haitian and Les Cayes by financing pilot "sites and services" projects; (c) help establish a metropolitan market agency, SAGMAP, that would operate the markets in Port-au- Prince and promote their long-term development; and (d) support institutional development of the Public Social Housing Promotion Company, EPPLS, to assist the Government in its efforts to develop a sound institutional and policy framework for low-cost housing in urban areas.

14 3.02 Proiect Description. The project included the following component: (a) (b) (c) (d) (e) Rehabilitation and upgrading of the CDB central market facilities within an area of about 12 hectares in the center of Port-au- Prince, comprising: (i) a covered retail market area for the sale of agricultural products, manufactured goods, meat and fish, and other services; (ii) an open-air retail market; (iii) a wholesale market (designed for future conversion into a retail facility); (iv) a water supply network for the market; (v) an interior road system interconnecting various sub-markets; and (vi) an administrative building to house SAGMAP, the Metropolitan Market Agency; a dormitory for about 180 caretakers employed by SAGMAP, with a separate building for storage of maintenance equipment. The Ministry of Public Works, MTPTC, was responsible for procurement and the supervision of the required works. Sites and services project for 1,036 housing units in Cite Soleil, to be occupied by the relocated families whose dwellings on the CDB market site would be demolished to make room for the construction of the proposed CDB market improvements. Provision of 1,093 housing units, a small market and a community center in Cap-Haitian, and of 278 housing units, a community center and a health post in Les Cayes. In these two provincial towns, as well as in Port-au Prince, the standards of the serviced sites were kept to a minimum but included communal water taps and sanitary facilities, surfaced main roads, footpaths, storm drainage, and public lighting along the major streets. Establishment and strengthening of SAGMAP, which would be responsible for the operation and maintenance of all metropolitan markets (starting with the CDB market). SAGMAP was also expected to generate sufficient revenues to cover its operating costs and at least 60% of depreciation on its assets. This institutional development was supported by the provision of limited financial assistance for studies, staff training, vehicles and equipment, and for 14 man-months of international technical assistance. Institutional development of EPPLS, which was created in November 1982 to assume the functions of the National Housing Office, ONL, that was responsible for housing projects in Port-au-Prince. EPPLS was given the nation-wide mandate to (i) take the lead in formulating housing policy, and (ii) make housing available on a cost-recovery basis, so that the income generated from sales and rentals would be sufficient to cover investment and recurrent costs and allow replication of the housing programs. The institutional development program to support EPPLS included preparation of studies; acquisition of equipment; and the provision of 78 man-months of technical assistance services by international experts for training EPPLS staff in the management

15 of housing projects, to be implemented by the Ministry of Public Works (MTPTC) IDA financing for the project totalling US$21 million equivalent was provided under Special Fund 001-HA (US$ 13 million) for the market component, and under Credit 1338-HA (US$ 8 million) for the housing component. Due to the cancellation of the entire market component in 1986, as will be explained below, the Government proposed to IDA that the funds available in SF1-HA account (about US$12 million) be used to finance urban roads projects in Port-au-Prince which the Government considered necessary, the most important of which was the Carrefour Road project. Carrefour Road is the only road link connecting the Northern and the Southern parts of the country. Its capacity to carry the traffic was completely inadequate, and its reconstruction and widening was of high priority. In addition to improving some key road links, the revised objectives were to improve the maintenance of the Port-au-Prince road network and to review Port-au-Prince's transport system. The proposed major change of the project's scope was approved by the Board in September 1987, and the project description was amended to include: (a) (b) (c) implementation of a five-year maintenance program for Port-au- Prince road network, including purchase of equipment for MTPTC's maintenance brigades, upgrading of MTPTC's asphalt plant, and rehabilitation and maintenance works; reconstruction (including widening) of about 7 km of the Carrefour road, and of the Canap6-Vert road (2.6 km); and technical assistance to help MTPTC (i) prepare and implement the above works program, and (ii) prepare a diagnostic and recommendations concerning the organization and funding issues needed to resolve the transportation problems in Port-au-Prince Finally, at the request of the new elected Government, both Credits were amended in April 1991 to enable all remaining funds (about US$3.5 million) to be reallocated to a labor-intensive road program. 4. Prolect Design and Organization 4.01 Conceptual Foundation. The project was decided to address problems of extremely poor sanitary facilities and inefficient working conditions inside the CDB market, which affected thousands of people whose activities were directly or indirectly connected with the market. The project was designed to upgrade the CDB market and to create an autonomous agency responsible for its operation. The shortage of housing in Port-au-Prince, and the need to relocate about 1,000 families living inside the market, required preparation of a housing program in Port-au-Prince. Additional housing programs in two provincial towns were later added on to the housing component of the project, as a pilot operation for IDA's first involvement in the housing sector in Haiti.

16 Institutional building was an important component of the project design and included the creation and strengthening of autonomous agencies fully responsible for (i) the design and implementation of policies and programs for Port-au-Prince markets (SAGMAP), and the housing sector of the entire country (EPPLS); and (ii) the management of the corresponding infrastructure based on a full recovery of costs Innovative concepts. The proposed design differed from previous Haitian practices under which: (i) the responsibility for Port-au-Prince markets was shared among its four poorly-staffed and financially-weak municipalities; (ii) ONL (EPPLS' predecessor) was not involved in the design and implementation of the housing projects that it managed; and (iii) housing programs and markets were not operating under a full cost recovery scheme Prolect scope. In retrospect, the overall scope of this first IDA-financed urban project appears too ambitious considering that its implementation was entrusted to weak institutions, MTPTC and EPPLS. MTPTC was to manage all the works related to the market upgrading, but it had at the time limited satisfactory experience in managing large and complex contracts; ONL, EPPLS's predecessor, had completed only its first large-scale housing project (879 housing units financed by UNDP), between 1979 and 1982, yet EPPLS was expected to implement at the same time three new significant housing projects both in and outside Port-au-Prince totalling 4,400 housing units (750 UNDP-financed and 1,250 KfW-financed housing units, in addition to the 2,400 units to be financed by IDA) Prolect preparation. The technical preparation of the project was generally adequate. By appraisal, most detailed engineering studies had been completed and all the land necessary for the project acquired by the Government. With regards to the market component, it is now evident that arrangements agreed with the Government during project preparation did not adequately take into account the need for close collaboration between the Ministry of Public Works and EPPLS (which reported to the Ministry of Social Affairs, MAS) to implement the market upgrading. The project design made it imperative that the completion of the house construction program in Linthau II, under EPPLS's responsibility, coincide with the initiation of the works on the market site, under MTPTC's responsibility. The need for a strong coordinating agency to oversee and synchronize all CDB market-related activities was all the more important because of (i) MTPTC's weak project management capacity for the complexity of the market component; and (ii) the fact that EPPLS, had no voice concerning the timing of the market upgrading activities to be arranged by the Ministry of Public Works With regards to the housing component, it appears that the proposed full-cost recovery scheme promoted under the project may have been premature because (i) ONL's organization inherited by EPPLS was weak; and (ii) the cost recovery target under current ONL's housing programs never exceeded 50 percent. Furthermore, EPPLS, a parastatal company subsidized by the Government, lacked the incentive to press its clients for the payments due to EPPLS.

17 4.07 Roles and responsibilities. The project was prepared by IDA jointly with EPPLS for the housing component, and with an inter-ministerial work group for the CDB market component. The roles and responsibilities of the various implementing agencies were defined during project preparation and clearly spelled out in the SAR: EPPLS was responsible for the housing component, MTPTC for the upgrading of CDB market, and SAGMAP for the management of the market. SAGMAP was to become operational only after the completion of the market upgrading. The inter-ministerial work group was only consultative and lacked executive authority. The lack of a unified command for the implementation of the project became critical during the market upgrading phase and constituted the main flaw in the project design. Coordination of EPPLS's and MTPTC's project related activities during the difficult initial phase (see above para. 4.05) should have been the main role of SAGMAP Coordination between EPPLS and MTPTC was instead entrusted to the EPPLS Director, who he had no decision-making authority within MTPTC, especially in matters of project procurement, whose delays affected implementation of the project. Neither was anyone put in charge to prevent newcomers from moving into the market area, after the Government announced (around 1980) that the upgrading of the market had top priority. As a result, the number of families living in the market increased from 1,000 in 1980 up to about 1,900 in 1983, when the project actually started, delaying further the relocation process that was so important to begin the upgrading works. The relocation problem was a serious threat to the viability of the market component even before the 1986 political upheaval. 5. Prolect Implementation 5.01 The Project was appraised by IDA in July 1982, negotiated with the Government in February and approved by Board in March It became effective on June 30, 1983; its original Completion Date of June 1987 and the Closing Date of December 1987, were extended each year through December Implementation of the project suffered from managerial shortcomings, including inadequate project coordination, and from socio-political upheavals after As a result, the project required five amendments and four extensions, and was completed in 8.5 years, compared to the 4.5 years estimated at appraisal. The 1987 amendment comprised a radical change of the project scope and objectives after the cancellation of the market component. As of the final closing date (December 31, 1991), about SDR 3 million (US$3.9 million equivalent) remained undisbursed In the case of the CDB market component, MTPTC's lack of experience in procurement caused major delays in contracting works for the market upgrading program. More than two years elapsed between Credit effectiveness (June 1983) and the signature of the contract for the market upgrading works (September 1985). This was due to MTPTC's weak capacity to follow IDA's procurement guidelines, and to a change of MTPTC's staff in 1984, following which MTPTC insisted on reviewing the bid evaluation prepared by the previous administration. The contractor was finally ready to start in January 1986, the works which at appraisal were scheduled for completion in 1985.

18 Although the Linthau II housing program was completed by mid-1985, relocation of the families living in the market had not yet started in January These unexpected relocation problems were the result of Government's inability to provide the necessary counterpart funds to finance the compensation payments to the CDB market occupants. This was mainly caused by an unexpected increase from US$0.8 million to US$ 1.6 million in the amount of compensation payments, following the increase of the number of families living in the market from about 1,000 in 1980, up to 1,900 in The increase was only noted in mid-1983, when the 1980 survey of the market occupants was updated. By April 1985, the Government had still not provided the necessary counterpart funds, and IDA had threatened to reconsider its financing of the entire market component The delays and the lack of coordination of the Linthau II house construction and the market upgrading programs had become critical as 1985 drew on, and jeopardized the viability of the entire market component. Close synchronization between the two programs was essential for two reasons: (i) the market occupants (mostly merchants) were not willing to move out until construction to upgrade the market was to begin, for fear that they might lose their place of business to squatters, if they had moved out and the upgrading works had not been started; and (ii) completion of the houses before they could be occupied would cause EPPLS to lose revenues and would increase the risk of invasion and vandalism of the houses. Thus, EPPLS had slowed down its construction program since end Despite this late coordination effort, the Linthau II houses had been completed by mid In August 1985, IDA had agreed with EPPLS that the Linthau II site had to be occupied by January 1986 and that the priority access of the market occupants to the Linthau II houses would, therefore, expire in December This prompted the Government to schedule the beginning of the relocation for December 1985, and to budget US$1 million for compensation payments. Unfortunately, further deterioration of the socio-political situation disrupted all project activities by the end of 1985, and led to the fall of Duvalier's regime in February After Duvalier's fall, part of the houses constructed for the market occupants were illegally invaded by squatters. As a result of the political instability, many market occupants no longer wished to leave the market, their workplace. Since the contractor was charging for mobilization (up to US$60,000/month) while the relocation of the market occupants was no longer feasible, the contract was canceled in May 1986, at a cost of US$1.5 million. This amount was paid for mobilization and penalties for cancellation of the contract Possible avoidance of the cancellation of the CDB market component. The early 1986 political and social upheaval was unforeseeable, and, during that period, nothing could have prevented the Linthau II site from being invaded. In fact, most EPPLS sites suffered from the same invasion and the EPPLS Director, responsible for the surveillance of the housing sites, was personally threatened and had to leave the country.

19 By mid-1985, there were concerns that implementation was likely to suffer from further delays, cost overruns and possibly reduction of its scope, even in a stable political situation. This was mainly due to the inadequate coordination between EPPLS and MTPTC and to limited support provided by the Government to the project, that was mentioned by IDA supervision missions. If a strong coordinating agency, solely responsible for the oversight of project implementation, had been put in place from the very beginning, it could have been in a position to (i) control the settlement of new families in the CDB market after the upgrading project was agreed to by the Government; (ii) prompt MTPTC to carry out the market upgrading in a timely manner; and (iii) coordinate MTPTC's and EPPLS's actions, including putting the Linthau II construction program on hold. Such a coordination effort was particularly needed because (i) MTPTC either did not understand the complexity of or was not interested in the project, as was noted by the January 1984 supervision mission; (ii) EPPLS had commercial interests in selling its houses and, therefore, did not have any incentives to stop its construction program; (iii) timely implementation of the relocation program was essential, because the market occupants could not sustain their motivation to relocate over any considerable length of time, as was concluded by the 1983 sociologic survey of the CDB market The appraisal report (SAR, para. 7.10) mentions the weak implementation capacity of the existing executing agencies and the lack of Government support as the most important project risks. In the case of the market component, this risk was deemed mitigated by the existence of a Project Implementing Unit that would be set up within MTPTC. However, this executing unit was to become active only after the works had begun, with the result that no assistance in project management was provided to MTPTC during the early stages of the project. Given the Government's limited support for the project, MTPTC's weakness and EPPLS's inability to act as a coordinating agency for the market upgrading, IDA supervision might have insisted that the Government (i) give SAGMAP (which had been created) the mandate to oversee all market-related activities; and (ii) set up a strong executing unit within MTPTC to speed up the market upgrading activities. Had such arrangements been implemented, the project might have been less vulnerable to the circumstances that led to the cancellation of its main component Provision of serviced sites. In Port-au-Prince (Linthau II area), construction of about 1,000 shelters started only in mid-1984, one year after Credit effectiveness, due to procurement delays. After the slow-down of the housing construction to synchronize its completion with the start of works at the CDB market (see para. 5.04), the Linthau II housing program was completed around mid In general, the quality of works on this highly visible housing component was satisfactory. After Duvalier's fall, squatters took advantage of the inter-regnum situation to invade about 250 houses. By end 1986, the status of these squatters as authorized occupants was approved by EPPLS Provision of the serviced sites was delayed in Cap-Haitian and Les Cayes more than in Port-au-Prince, and the quality of works was inferior. Because EPPLS was focusing on Port-au-Prince, works in Cap Haitian started only at the end of 1984 and in Les Cayes in mid After Duvalier's fall,

20 -8- works were interrupted for several months. They resumed in September 1986, and were completed around mid-1987, including 883 houses in Cap Haitian and 254 in Les Cayes. At that time, the housing component was proceeding satisfactorily, but with delays of up to two years. IDA agreed, therefore, to the Government's request to use the remaining funds under Cr HA (extended till December 1988) as well as SDR300,000 from Cr. SF1-HA, to finance construction of 1,240 additional housing units in Cap-Haitian and 480 units in Les Cayes. Due to problems with squatters, the Cap Haitian component was eventually abandoned and EPPLS focused on the program in Les Cayes, where housing was completed in 1989, after a second extension of Cr HA had been granted through December IDA also agreed that the remaining funds under Cr HA (estimated at US$1.8 million) be used to finance (i) another phase of the housing program in Linthau II, Cap-Haitian and Les Cayes, (ii) equipment for garbage collection and cleaning of latrines, and (iii) technical consultants to assist EPPLS in reviewing its overall strategy and to prepare a new pluriannual housing program. Because of political instability, little progress was made in 1989, and IDA agreed once again to extend the closing date, until December However, due to continuous changes in EPPLS' high level staff, EPPLS' newly hired inexperienced staff was unable to comply with IDA's procurement requirements for the contracting of works or the purchase of equipment. The remaining funds under Cr HA (when finally extended through December 1991) were reallocated to a labor-intensive program, in line with the priorities of the newly elected Government The quality of the works carried out under the Cap Haitian and Les Cayes housing programs was generally adequate, thanks to the intervention of the few field visits made by IDA supervision missions. The September 1987 supervision mission, which made the first technical field visit to the housing program outside Port-au-Prince, found that many completed units had major construction defects including improper roofs and sanitary pits, as a result of deficient design, and poor construction and supervision. Eventually, most of these construction deficiencies were corrected. EPPLS's management of the completed sites was not satisfactory: basic utilities were not provided on time, and garbage collection and sanitary pit cleaning programs have never materialized EPPLS's cost recovery performance (35% on average) was also poor throughout the life of the project, and particularly low in 1986 and 1987 (see Part III.6 of the PCR). The main reason was the difficulty to collect rents in areas where houses were occupied by squatters. Another reason for a low cost recovery was poor delivery of basic services. Finally, there was the problem of non-uniform rents. Some housing projects were located in low-lying areas that were subject to flooding. These areas had to be filled prior to construction of the housing units. The cost of such units was higher than that of the houses built on lots where the fill was not necessary. This was reflected in the monthly payments made by the purchasers of the housing units constructed under IDA projects, and made it difficult for EPPLS to collect payments from house owners who objected to paying more than the purchasers of seemingly equivalent houses constructed under projects financed by other agencies. At one time, EPPLS explored the possibility to reduce the amount of

21 - 9 - periodic payments by extending the total repayment period, but this effort was not successful Institutional strengthening of EPPLS. The project was to finance technical assistance to strengthen EPPLS in internal administration, operation of the serviced sites, rent collection, and the development of sound housing policies. In fact, due to the operational weakness of EPPLS, its staff and consultants had to focus exclusively on the supervision of the works, accounting, financial controls, and computerization of EPPLS operations and had no opportunity to work on the formulation of housing policies Possible avoidance of Variances within the above housing component could have been avoided if the following action had been taken: (a) (b) (c) The poor technical quality of the serviced sites could have been reduced by more frequent technical supervision by IDA. Field visits made since September 1987 proved very efficient. Also, more works should have been contracted out because EPPLS's attempts to do works by force account produced poor results. Occupation of the project houses by squatters was mainly possible because (i) construction of the units was completed before they were sold to the intended beneficiaries, (ii) the housing sites were always located close to existing poor neighborhoods where the housing demand was tremendous and the temptation for illegal occupation high, and (iii) the climate of social unrest since 1986 made it impossible to oust squatters once the sites were invaded. Nevertheless, EPPLS observed in 1988, that community participation in the works and an early take over of the housing sites by the beneficiaries, precluded illegal invasion. If such a scheme had been adopted from the beginning, much of the invasion problem might have been avoided. Low cost-recovery, poor operation and lack of maintenance of the sites were mainly a consequence of an unrealistic assessment by IDA that EPPLS would act as a private enterprise (see SAR 2.12). Responsibility for "external activities", such as maintenance of the sites and rent collection, should have been transferred to a private institution under EPPLS' supervision. In 1987, as a condition to extend Cr HA, EPPLS was to propose such a privatization scheme to be implemented in 1988, at least for the collection of payments due from purchasers and renters, which might have helped resolve EPPLS financial problems. In the absence of EPPLS's interest, such a privatization scheme has not been implemented. In retrospect, it seems that IDA should have been more persistent to resolve these important issues, especially since KfW, the other main donor participating in the housing sector, had similar problems and fully agreed on the concept of privatization Urban Road component. In 1987, following the cancellation of the market component, IDA agreed that, in spite of MTPTC's poor performance, the

22 Ministry could use the US$12 million available under Credit SF1-HA to finance an urban road component. This was motivated by (i) the priority of the road works to be financed (see para and (ii) IDA's willingness to support the new Government which had replaced Duvalier's dictatorial regime. The new component, which was approved by Board in September 1987, included the reconstruction of the Carrefour and the Canape Vert roads, as well as road maintenance works The Carrefour road preliminary design prepared by local consultants was presented as technically and financially feasible by MTPTC to IDA during the January 1987 supervision mission. Its cost was estimated at US$8.5 million, including US$3 million for expropriation. It was on this basis that the 1987 amendment was agreed. However, in 1988, a detailed technical and economic evaluation of the Carrefour road project, including expropriation costs, were carried out by a consulting firm. The estimated cost of the projects of US$24 million (including US$7 million for expropriation), was based on final engineering. The cost exceeded the Ministry preliminary estimates and the available external financing. Furthermore, the Government did not have the financial capacity nor the political will to complete expropriation within a reasonable period of time and MTPTC requested that this component be withdrawn from the project The Canap6-Vert road project was initiated early in 1987 and completed by mid The initial cost estimate was US$1.6 million (of which IDA accepted to finance US$1.2 million). The final cost to complete the works amounted to US$2.8 million, or more than US$1 million per km. The works, which were carried out partly by force account (for about US$1 million) and partly by contract, were of poor quality. By the end of 1988, some sections were already badly eroded, threatened the stability of nearby houses, and required costly repairs. Many contracts had not been awarded in accordance with IDA's procurement guidelines, including some that had even been awarded to unqualified contractors. As a result, IDA decided, in May 1988, not to finance the works carried out by contract, and financed only US$0.6 million of the total cost (US$2.8 million) Road Maintenance Component. Agreement was reached with IDA in April 1988 that this component would include: equipment for MTPTC maintenance brigades, rehabilitation of M,TPTC's asphalt plant and road maintenance works by force account (total cost of US$615,000); road maintenance works by contract (US$380,000); and installation of traffic lights at the main Port-au- Prince intersections (US$590,000). To carry out the works, Credit SF1-HA was extended until December Maintenance works by force account started in May 1988, and were limited to filling potholes (for a total cost of US$0.62 million), for lack of maintenance equipment and MTPTC's failure to repair its asphalt plant. Similarly, contracts for maintenance works approved by IDA to improve streets connecting the Canape-Vert road with the center of Port-au- Prince, could not start because of expropriation problems. Finally, MTPTC could not satisfy IDA procurement requirements for the traffic lights and the component was abandoned by MTPTC.

23 Due to the cancellation of the Carrefour road project, and the poor management by MTPTC of the Canape-Vert road and the urban road maintenance components, almost US$11 million equivalent remained undiabursed under Cr. SF1-HA in mid-1989, prompting IDA to approve the Government's request to amend the Credit. In view of the state of deterioration of the entire road network, the amendment provided financing for: (i) periodic maintenance of 330 km of interurban roads (US$6.5 million); (ii) the rehabilitation of 7.5 km of secondary Port-au-Prince streets (US$2.5 million); and (iii) an institutional study of MTPTC and preparation activities for a proposed Road Maintenance Project. In October 1989, the Closing Date of Cr. SF1-HA was extended until December 1990, to enable implementation of the program. More delays occurred in preparing contract documents and making arrangements for the supervision of the project activities. Works started in September 1990, and were finally completed around mid-1991, after protracted delays caused by social unrest in early A fourth extension of the closing date till December 1991 was granted in April 1991, to (i) enable the completion of the above program, which was concluded around mid-1991; and (ii) use the remaining funds under Credit SF1-HA and 1338-HA (about US$3.9 million equivalent) to finance a labor-intensive program requested by the Government. The political coup of September 1991 disrupted implementation of the program and prompted IDA to suspend disbursements Possible avoidance of variances within the urban road component: Some of the problems could possibly have been avoided by the following actions: (i) Improvement of project management by MTPTC. Poor management markedly delayed implementation of the urban road component. MTPTC needed to strengthen its capacity in engineering, cost control, procurement procedures and management of contracted works. The urban road component would have been better managed by a specially created executing unit staffed with competent local consultants, assisted by expatriate experts in key fields such as engineering and procurement. Finally, payments should have been made directly to all contractors and consultants to avoid repeated misuse of the project's special account, which eventually led to IDA's decision, in 1990, not to open new special accounts; and (ii) The contracting of most of the project works to avoid uneven quality of the works carried out by force account, and to have better financial controls. In view of the unsatisfactory quality of the force account works carried out under the Canape-Vert and the urban road maintenance components, IDA required upon amending Cr. SF1-HA in June 1989, that the additional works to be executed should be contracted The issue of expropriation was underestimated by MTPTC. The Ministry never succeeded in expropriating the necessary properties for the Carrefour road nor for the extension of the Canap6-Vert road. The main problems were (i) the difficulty to assess property values, (ii) social and

24 political issues involved, and (iii) the lack of Government funds for compensation payments (IDA does not finances expropriation costs). In retrospect, road projects which require major expropriation should not be considered for IDA financing unless (i) firm and detailed expropriation costs are available, (ii) the corresponding amount of Government funds is set aside for expropriation payments in a specific account, and (iii) a formal arrangement has been made with the occupants to be expropriated. 6. Proiect Results 6.01 Overall, the results from the project are hardly satisfactory. Its main objective, to upgrade the CDB market and create SAGMAP, had to be canceled. The positive results included achievements under the housing and the urban road components. Institutional strengthening of the responsible agencies, EPPLS and MTPTC, was insignificant CDB market upgrading was canceled in 1986, and SAGMAP was never set up. As a result of the cancellation, US$1.5 million were paid to the contractor Housing component. The only achievement was the construction of 2661 low-cost housing units of acceptable quality (at an average cost of US$2,450, ie. 4% less than appraisal estimates). It is estimated that some 10,000 people benefitted from the project. However, the result of the housing program may be considered unsatisfactory for the following reasons: (a) (b) (c) (d) The basic services (ie. water supply, electricity, garbage collection) were not provided in a timely manner or at all; The project houses were often not occupied by the target beneficiaries. Many units were taken over illegally by squatters, or purchased by private entrepreneurs as rental houses. This makes it difficult to assess the number of beneficiaries who actually were within the targeted income category; The institutional objectives of the housing component were hardly met. The averaged cost recovery never exceeded 35 percent. EPPLS failed to develop as a financially autonomous agency, and was unable to reimburse the loan contracted with the Government and reinvest its remaining profit in new housing programs. Instead, as stated in the 1990 audit of EPPLS, a significant share of the scarce resources coming from rent collection served, under political pressure, to finance an increasing number of EPPLS permanent staff, which increased up to 420 in the late 1980's, before it was eventually reduced to 160. The full cost recovery target (including almost all investment costs) was probably overambitious, given ONL past experience with housing programs which never achieved a cost recovery of more than 50 percent; Finally, EPPLS failed to develop any sound housing sector policies to address, on a significant scale, the problems confronted by the urban poor in the housing sector.

25 The results of the urban road component are not satisfactory, because most objectives were not met, namely: (a) (b) (c) The reconstruction of the Carrefour road was not possible for lack of funds. A traffic master plan, including a review of the necessary key road investments, was finally scheduled to be financed under the proposed Road Maintenance Project. Nonetheless, this important arterial, the only link between the southern and northern part of the country, will remain badly congested until the mid-1990s, at a great cost to the Haitian economy. The results of the reconstruction of the Canap6-Vert road are also unsatisfactory. The cost of reconstruction of the 2.6 km road section was excessively high (more than US$1 million per km), for a project of standard design. The quality of works was, at best, fair. The actual reduction of traffic congestion in the area was smaller than expected because MTPTC failed to complete the widening of an intersection at the down-town end of Canap6-Vert area, that was necessary to relieve traffic congestion on the Canape-Vert road. Finally, a significant share of the maintenance works that were included in this component were not implemented. Works actually carried out were limited to filling potholes, because MTPTC failed to acquire the minimum of necessary equipment and to upgrade its asphalt plant, for which IDA financing was available Concerning the revised road maintenance component described in the June 1989 and April 1991 amendments, most of the objectives were met: (a) (b) (c) (d) The paving of 7.5 km of secondary streets in Port-au-Prince was completed by mid-91. These works were all contracted out, were generally of good quality and succeeded in promoting paving with concrete blocks, a labor-intensive and easy-to-maintain paving technique which should be extensively used under the proposed Road Maintenance Project; Periodic maintenance on about 330 km of interurban roads was completed in early The quality of works was acceptable, despite poor supervision. An institutional study to reorganize MTPTC and other preparatory activities were financed under Cr. SF1-HA and were instrumental in the preparation of the proposed Road Maintenance Project; Under the labor-intensive program, it was possible to initiate, before the September 1991 coup, projects designed to generate about 100,000 man-days of employment. This program also served as successful pilot experience to implement MTPTC's new procedures

26 - 14 _ for road program management, in line with the proposed reorganization of the Ministry. 7. Prolect Sustainabilitv 7.01 The sustainability of the project's benefits is problematical. No benefits have been derived from the CDB market project due to its cancellation. Whether the benefits derived from the construction of sites and services and the 2,661 housing units will be sustained will depend on the commitment of the beneficiaries to preserve their properties through regular maintenance. Given the difficult economic situation and limited interest in maintenance works in Haiti, it is possible that these benefits will not be long lasting The benefits from the rehabilitation and paving of a limited number of roads may not be sustainable in the future because of (i) defective construction (eg. Canape-Vert road); and (ii) the lack of timely maintenance due to delays in starting the proposed Road Maintenance Project, in the wake of the September 1991 political crisis. This situation could affect the condition of 100 km of regravelled roads that need extensive maintenance after rainy seasons. Streets paved with concrete blocks should produce the expected benefits to road users, as this type of pavement requires very limited maintenance. Finally, sustainability of the rather limited institutional strengthening of MTPTC is doubtful, given that the Ministry has no permanent civil service cadre because of the tradition that key personnel are changed as new ministers assume office. 8. Bank Performance and Lessons Learned 8.01 Preparation and Appraisal. During project preparation, IDA made a thorough evaluation of the technical aspects of the project and ensured that most detailed engineering studies were completed by appraisal. However, IDA did not realistically assess the complexity of the market component and the subsequent need for a local agency with decision-making authority to be responsible for that component from the very beginning. In addition, IDA apparently overestimated the capacity of EPPLS to properly operate its serviced sites and to meet full cost recovery targets; and overestimated MTPTC's capacity and commitment to proceed with the implementation of the market component. The lack of commitment was evidenced by protracted delays in matters of procurement and the lack of coordination with EPPLS The very concept of building houses to address the housing needs of the urban poor, promoted by IDA during preparation of the project, should be reviewed. The number of housing units financed by the project was 2,661; in total, EPPLS program during the 1980's included less than 10,000 units. This is insignificant compared to the overall demand from low-income families, which was estimated, in early 1980, at 115,000 units in Port-au-Prince only, with 6-8% annual growth. In retrospect, it seems that the impact of the housing project might have been more significant if (i) EPPLS had primarily focused on the provision of basic infrastructure (sites and services including

27 construction of primary roads and drainage) and let the private sector or the beneficiary communities be in charge of house construction (which would also have reduced the risk of illegal invasion); and (ii) EPPLS had carried out a review of issues affecting the housing sector (eg land management) Supervision. IDA allocated no more than 10 staff-week per year for project supervision (including field and office work). Two field supervision missions, on average, per year were insufficient in light of the unusually high supervision requirements for a project whose performance was rated (3) (ie major problems) for more than four years, and (2) (ie moderate problems) for about three years, over a total implementation period of 8.5 years. The main weakness of IDA's performance during supervision are as follows: (a) (b) (c) (d) Limited supervision by IDA and relatively high turnover of IDA task managers (change of task managers after 1.5 years of supervision) during the early stage of the project, whose implementation was experiencing serious delays. When the lack of coordination of the market component, the inadequate Government support and the limited interest of MTPTC in the project had become obvious in 1984, IDA should have been more active in assisting in the implementation of this complex component, and should have required the setting up of (i) a strong coordinating agency for the market component; and (ii) an executing unit within MTPTC to speed up the market upgrading works. EPPLS's inadequate operation and maintenance of the serviced sites and weak rent collection capacity should have prompted IDA to require that these activities be contracted out, before any extension of the housing program was granted. This idea to contract out "external activities" was shared by KfW, the other EPPLS's main donor, and a strong common position could have been adopted. Concerning the quality of the works on both EPPLS' housing program and MTPTC's road program, the few IDA field visits made proved useful and prompted EPPLS and MTPTC to correct defects noted in the already completed works. Such technical field visits should have been made more frequently. In view of the MTPTC's insufficient capacity to implement the first phase of the urban road component (see para to 5.19), IDA should have insisted that a strong executing unit be set up or, at least, that competent local consultants be recruited, instead of relying on long-term expatriates whose performance was inconclusive The main lessons learned from the above are that (i) IDA should pay due attention to implementation arrangements to ensure that sufficient capacity exists within the project executing agencies and that coordination is real. In the absence of a strong executing agency, recruitment of competent local consultants should be encouraged, in preference to long-term expatriates; (ii) private sector participation should be encouraged or, at

28 least, tested whenever in-house execution has no clear comparative advantage; (iii) more emphasis should be put on supervision missions in general and on field visits by IDA engineers, in particular, whenever a construction project experiences major problems To assist in project implementation, the Credit agreement for the housing component (Cr HA) could have spelled out more clearly key indicators to monitor EPPLS' progress necessary to achieve the agreed objectives of financial autonomy. Similarly, dated covenants to prompt the Government to clear the market site of the existing housing units in a timely manner and to initiate the market upgrading works, could have given IDA more leverage to persuade the Government to act promptly during the two-year period when these actions were unnecessarily delayed. 9. Borrower Performance and Lessons Learned 9.01 Borrower performance during project implementation was not satisfactory. The main project objectives were not met due to the social and political instability that prevailed in Haiti throughout the life of the project, and to inadequate Government support for the project and the weakness of the Borrower's executing agencies, namely: (a) (b) (c) lack of Government support for (i) the market component, evidenced by its failure to set up a strong coordinating agency and to prompt MTPTC to speed up the market upgrading; and for (ii) EPPLS's effort to prevent illegal occupancy of the housing units. In addition, from 1986 on, the Government staffed EPPLS's key positions with political appointees rather than with competent professionals; lack of interest of EPPLS and MTPTC in improving their performance by contracting out "external activities" (eg housing site maintenance, rent collection; road maintenance); and poor capacity of EPPLS and MTPTC to manage the project, including: poor engineering and cost estimates; lack of experience in procurement and contracting; inadequate monitoring of both the contractors and supervisory consultants; and poor accounting and mismanagement of the project's special account Lessons learned. In view of the limited impact of the Government's housing programs and the poor performance of EPPLS, its housing agency, the Government should reformulate its housing strategies, possibly in line with recent Bank experience in housing, which tends to focus more on (i) facilitating the working of the private market for housing, particularly through financial policies; (ii) targeting of subsidies directly to intended beneficiaries rather than to suppliers of housing: and (iii) providing basic infrastructure instead of building houses To correct its weaknesses, MTPTC should (i) contract out most of the works; (ii) improve management of works contracts by, inter alia, streamlining payment procedures and selecting contractors on the basis of

29 their past experience and capacity; (iii) strengthen its administrative capacity to manage projects, especially in the fields of procurement and accounting by using IDA standard bidding procedures and documents, and by relying on specialized consultants for the bidding of large or complex contracts, and (iv) pay more attention to expropriation problems. 10. Bank-Borrower Relationship The Bank-Borrower relationship was correct but not intense. Field supervision missions to monitor project implementation were not frequent enough. The repeated assurances by EPPLS and MTPTC to take adequate corrective actions, the social aspects of some project components in a time of economic crisis, and the prospect of the establishment of democracy after Duvalier's fall, all induced IDA to be understanding of the project's problems. (During the later years of the project IDA was virtually the only source of external financing available to the Government after key bilateral donors pulled out). Although a more rigid position might have given more incentives to the Government to take timely corrective actions, IDA tried to make best use of the funds that remained uncommitted during different stages of the project's implementation, by designing and preparing totally new projects that met current priorities (eg. the road maintenance activities and the laborintensive program) Following the 1990 elections, IDA established a fruitful dialogue with the new Government regarding preparation of a comprehensive Road Maintenance Project. MTPTC accepted, as a condition for the launching of the labor-intensive program, that its most important deficiencies (as described in para. 9.01) be corrected, including contracting out all works, setting-up an adequate executing unit, and recruiting external accountants. The dialogue with EPPLS continued to deteriorate. EPPLS appeared to have little interest in carrying out a review of its housing strategies. In 1991, Cr HA's remaining funds (US$1.2 million) were transferred to the labor-intensive program and IDA did not encourage additional housing programs. IDA's decision was appropriate considering that EPPLS failed to show the required administrative or technical capabilities for leadership in this field. This was the first IDA-financed housing operation in a complex sector which required more than one project to establish sound sector policies and to strengthen local agencies, and the dialogue with the Government concerning the housing sector will have to be renewed IDA denied the Government's request for the fifth extension of the Credit 1338-HA and SF1-HA, both of which are in the process of being canceled. The total of undisbursed credit funds is SDR 3.0 million (US$3.9 million equivalent). The project was implemented during a difficult post-duvalier period during which the performance of newly formed governments that followed and their agencies responsible for the implementation of the projects, was poor, often having a negative effect on the implementation of the project. There were cases of non-compliance with the Credit covenants which gave IDA ample legal grounds to consider cancellation of the Credit Agreement, which IDA discussed with the client. On one occasion (para 5.18) IDA declined a request from MTPTC to reimburse about US$1 million equivalent which the

30 Government had paid to contractors who were not contracted in accordance with Bank guidelines The principal reason which mitigated against cancellation of the Credit funds during those turbulent years, was an optimistic expectation that these funds could be used effectively to prevent complete deterioration of the country's major asset, its road network, whose replacement value was estimated at about US$400 million. The labor-intensive character of the then proposed maintenance works provided an additional important economic benefit to the country through generating employment. These programs were launched and successfully implemented by communities until problems arose with payments for the works. The problem was that such payments had to be made through the Ministry of Works rather than directly to the communities, causing unacceptable delays and demoralizing the workers. A solution that would permit making such payments to the communities directly was being worked out but could not be finalized due to the eruption of the current political turmoil and the suspension of all project activities. The country arguments for continuing disbursements took on even more weight after the 1987 elections were aborted and the key bilateral donors, USA and France, increasingly canceled their financial support through non-government agencies and private voluntary agencies. IDA, one of Haiti's few remaining official sources of external financing, felt it had a special responsibility to continue supporting the road maintenance program, provided that the Government gave evidence of moving towards better economic policies and more economic and efficient use of public funds. IDA's continued involvement with the project assumed even greater importance as the benefits of the expanding laborintensive road maintenance programs became apparent. 11. Consulting Services The technical assistance aimed at improving accounting and financial control at EPPLS was fair, at best, and did not significantly improve either EPPLS' cost recovery performance or its accounting or filing system (see EPPLS 1990 audit report). Technical assistance for MTPTC road maintenance program by foreign consultants did little to improve MTPTC's capacity in project management. During their long-term assignment, the consultants neglected the project's training objectives and were actually assimilated as "permanent" staff of the Ministry. In the future, qualified local consultants should be encouraged to provide long-term assistance Consultants' services for detailed engineering studies for the housing program, the Carrefour road and Canap6-Vert road projects, were of uneven quality, characterized by design deficiencies and poor cost estimates Civil works supervision by local consultants was satisfactory in Port-au-Prince (eg. Linthau II housing program, secondary street paving), but inferior in the case of less visible projects such as the housing programs in Les Cayes and Cap Haitien, or the maintenance works on inter-urban roads With few exceptions, civil works contractors showed inadequate technical skills and performed poorly. In many cases, their performance was

31 also hampered by insufficient financial capacity to confront major delays of payments for their services, especially during the execution of the road maintenance program. 12. Prolect Documentation and Data Data needed for the preparation of the PCR was generally available in Bank files. The supervision reports and corresponding aide memoirs were particularly helpful. It was difficult, however, to collect the data necessary to carry out a comprehensive ex-post evaluation of (i) the final cost of the housing construction programs, including the Government contribution; (ii) the evolution of EPPLS' cost recovery performance from 1984 to 1991; and (iii) the data on the actual beneficiaries of the project. This difficulty was mostly due to frequent staff changes within EPPLS, inadequate administrative procedures to keep an inventory of assets; poor accounting practices; and the lack of political will to take corrective measures that might have resulted in increasing social tensions, that were characteristic during the long implementation period of the project. Preparation of the quarterly progress reports was lax. It is necessary that, in the future, IDA insist on receiving well documented quarterly or semi-annual reports in which Clients review all administrative and technical achievements and/or problems associated with IDA projects' implementation. The lay-out of such reports should be standardized following, in principle, the format used for the preparation of the PCRs, and modified as needed to better reflect specific characteristics of particular projects. Non-compliance with this requirement should be seriously considered as grounds for suspending disbursements.