Madagascar - Antananarivo Urban Works Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No: 19160 IMPLEMENTATION COMPLETION REPORT REPUBLIC OF MADAGASCAR ANTANANARIVO URBAN WORKS PROJECT (Cr. 2591-MAG) April 19,1999 Water and Urban Unit Eastern and Southern Africa Africa Region 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.  CURRENCY EQUIVALENTS Currency unit = Malagasy Franc 1989: US$ 1,00 = FMG 1603 1994: US$ 1,00 = FMG 3067 1990: US$ 1,00 = FMG 1494 1995: US$ 1,00 = FMG 4266 1991 : US$ 1,00 = FMG 1835 1996: US$ 1,00 = FMG 4061 1992: US$ 1,00 = FMG 1864 1997: US$ 1,00 = FMG 5091 1993 US$ 1,00 = FMG 1914 1998: US$ 1,00 = FMG 5000 WEIGHTS AND MEASURES Metric System FISCAL YEAR January 1 - December 31 ABBREVIATIONS AND ACRONYMS (English/French Translation) AGETIP Agency for the Execution of Public Works Agence d'execution des travaux d'interet public AGETIPA Agency for the Execution of Infrastmcture Agence d'execution des travaux Works in Antananarivo d'infrastructures publiques d'Antananarivo AMEXTIPE Mauritanian Agency for the Execution of Agence mauritanienne d'ex6cution des Public Works and Employment travaux d'int6rft public et pour l'emploi AUWP Antananarivo Urban Works Project Projet de voiries urbaines d'Antananarivo CETE Technical Studies Center of the Ministry of Centre d'etude techniques du Ministere de Equipment 1' equipement CNFTP National Council for Technical and Centre national de formation technique et Professional Training professionnelle CTNDU National Technical Committee for Urban Comite technique national pour le Development developpement urbain FAC French Cooperation Fund Fonds d'aide et de cooperation FMG Malagasy Franc Franc malgache GNP Gross National Product Produit national brut (PNB) GOM Government of Madagascar Gouvemement de Madagascar IDA International Development Association Association intemationale de developpement JIRAMA (Jiro Sy Electricity and Water of Madagascar Eau et Electricite de Madagascar Rano Malagasy) MATV Ministry of Town and Country Planning Ministere de la Ville et de l'Amenagement du Territoire MTP NMinistry of Public Works Ministere des travaux publics PIP Public Investment Program Programme des investissements publics PPF Project Preparation Facilitation Avance de preparation de projet SMEs Small and Medium Enterprises Petites et Moyennes entreprises TELMA Telecommunications of Madagascar Tel6communications de Madagascar Vice President: Callisto Madavo, AFRVP Country Director: Michael Sarris, AFC08 Sector Manager: Jeffrey Racki, AFTU 1 Task Team Leader: Arnaud Guinard, AFTU1  TABLE OF CONTENTS FOR OFFICAL USE ONLY Page PREFACE EVALUATION SUMMARY PART I: PROJECT IMPLEMENTATION ASSESSMENT A. Project identity ...................................................................................................................I B. Introduction .................. . ....... . I C. Project Objectives . . . . ....l D. Project Description ......... . . . . .2 E. Project Preparation ........................................... .......... . 3 F. Project Implementation ............... . ......... .3 G. Achievement of Objectives ... . ..........................................................................................5 H. Major Factors Affecting the Project ................................... ........................6 I. Project Sustainability . . ...7 J. IDA Perfornance ............................................................8..................................................8 K. Borrower Performance .............. . . . ..............................................................................8......8 L. Project Outcome .......9 M. Key Lessons Leamed .............. . ........................................ .................9 PART II: STATISTICAL TABLES Table 1: Summary of Assessments Table 2: Related IDA Credits Table 3: Project Timetable Table 4: Credit Disbursements: Cumulative Estimated and Actual Table 5: Key Indicators for Project Performance Table 6: Studies included in Project Table 7A: Project Costs Table 7B: Project Financing Table 7C: Credit Allocations per Category Table 8: Status of Legal Covenants Table 9: Bank Resources: Staff Inputs in staff/weeks Table 10: Bank Resources: Missions Table 11: Contracts Distribution per Size of Enterprises Table 12: Estimated Wages According to Types of Works Annex 1: Borrower Comments on the ICR Annex 2: Completion Report by the Borrower Annex 3: ICR Mission Aide-Memoire Map IBRD No. 20035R1 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.  REPUBLIC OF MADAGASCAR ANTANANARIVO URBAN WORKS PROJECT CREDIT 2591-MAG IMPLEMENTATION COMPLETION REPORT PREFACE This is the Implementation Completion Report (ICR) of the Antananarivo Urban Works Project in Madagascar, for which Credit 2591-MAG in an amount of SDR 13,3 million (US $18,3 million equivalent) was approved on March 29, 1994, and made effective on August 1, 1994. The Credit closed on the original closing date of December 31, 1998. It was fully disbursed, and the last disbursement took place on August 31, 1998. Cofinancing for the Project was provided by the French Cooperation Fund (FAC) in an amount of US $170,000. A first draft of the ICR was prepared by Luc Cosyn, Consultant. The final report has been prepared by Arnaud Guinard, Task Team Leader, and reviewed by Jeffrey Racki, Sector Manager (AFTU1), Manorama Gotur, Operations Analyst (AFC08), and Michael Sanris, Country Director for Madagascar (AFC08). Preparation of this ICR was begun during the Bank's final completion mission in September 1998. It is based on materials in the project files, the Appraisal Report and the Development Credit Agreement. The Borrower contributed to the preparation of the ICR by providing its own completion report, as well as commenting on the draft ICR. -1i-  IMPLEMENTATION COMPLETION REPORT REPUBLIC OF MADAGASCAR ANTANANARIVO URBAN WORKS PROJECT: CREDIT 2591-MAG EVALUATION SUMMARY Introduction During the early 1990s the economy of Madagascar was mired in stagnation. Per capita GDP was clearly on the decline, hovering around US $230. Constraints affecting basic infrastructure in general, and roads in particular were identified as among the most serious impediments to Madagascar's economic development. This situation sparked recognition of the need for a more effective infrastructure delivery system. The International Development Association (IDA) shared its experience with the Malagasy Government regarding the establishment of public works executing agencies. These discussions led the formulation and implementation of the Antananarivo Urban Works Project. Project Objectives The primary objective of the Antananarivo Urban Works Project was to demonstrate the benefits of the delegated contract management system as a tool for managing and maintaining infrastructure works. It was also intended to serve as model for subsequent broader application. The two central development objectives of the Project were: (i) to rapidly bring the infrastructure of Antananarivo to a satisfactory level, with a priority given to roads, and (ii) to develop formal domestic SMEs in the construction and engineering industry, strengthening their capacity to design, execute, and supervise a wide range of public works. The associated objectives were: (i) to strengthen the management capacity of the municipality, (ii) to create jobs by encouraging the use of labor-intensive technologies, whenever economically efficient, and (iii) to stimulate growth in the construction industry and in the economy as a whole. The project was based on the principle of community participation in order to increase the voice of users and stakeholders in maintenance planning and establish a closer link between the municipality's actions and the wishes of the public. Components The project resulted in (i) the creation of an executing agency,Agence d'Execution de Travaux d 'Infrastructure Publique d'Antananarivo (AGETIPA), (ii) the execution of sub-projects to improve the infrastructure in Antananarivo, (iii) the establishment of a training program for SMEs and public institutions, and (iv) the provision of services to inform the public about the approach of the project and the work program. Implementation Experience and Results Project implementation began very rapidly and proceeded throughout relatively smoothly despite difficult traffic conditions, inexperienced contractors and delays in mobilizing counterpart funds. Performance indicators for management and physical results mostly exceeded expectations throughout project execution. The total cost of the works and studies was US $20.5 million for a total of 874 signed contracts. Disbursements slightly exceeded appraisal estimates and the credit was fully disbursed on schedule. However, the speed with which the project was implemented and disbursements were made appears to have masked a number of problems during project execution. - Ui -  Technical audits complied only partially with the terms of reference set out in the Manual of Procedures; limited effort was made to examine the implications of choosing certain technical options and packaging of contracts on procurement arrangements, as well as to review compliance with IDA procurement procedures. Many sub-projects were subdivided into small contracts to an illogical and inordinate extent. The result was a total number of work contracts all out of proportion with the number estimated at preparation (732 instead of 310). Although small enterprises were supposed to have been the beneficiaries of such exercise, most of the contracts were executed by two large firms which built infrastructure costing FMG 48 billion, or 51,2 percent of the entire project. Compliance with procurement procedures was also deficient. In some instances, ICB should have been undertaken instead of LCB and requirements for IDA prior review were not met. Rehabilitation of urban works in Antananarivo. Under the project, 106 km of roads, 111 km of sidewalks, and 104 km of drains were rehabilitated, 8 km of stairs and alleyways and about 100 retaining walls were built or repaired, and five markets and various public places were developed. Sub-project viability was demonstrated by an internal rate of return of over 20 percent. A steady improvement in the quality of the works was observed as the project progressed. Job creation and poverty alleviation. Throughout, the project generated 133,000 man-months of employment in direct jobs. The resulting payroll amounted to some FMG 18 billion or about 20 percent of the total cost of the works. Many suppliers to whom contractors tumed for the works benefited in terms of employment and income. Together with this multiplier effect, increase mobility of the population also had a positive impact on job creation, particularly in outlying districts. Promotion of consulting firms and private enterprises and stimulation of growth in the public works sector. Overall, the number of failures among the SMEs which contracted with AGETIPA was low. Substantial training was provided and a considerable effort was undertaken to mobilize SMEs which enjoyed the benefits of technological progress, since most of them were involved in the project at its inception. A set of "Quality Standards," a list of technical specifications which they drew up together under AGETIP's guidance, also had a positive impact on the quality of works. SMEs in the civil works sector in Antananarivo also became better organized. An association was established during the project and entrepreneurs met periodically under its auspices to share their experience. Although the project expanded and stabilized existing SMEs in the sector, it led to the establishment of very few, if any, new firms. Strengthening the management capacity of the municipality. AGETIPA assisted the municipality in using its services as best as possible and improved the interaction of the municipality with its service users. This was achieved by involving district organizations on an individual basis through surveys, media campaigns and direct assistance to neighborhood committees during sub-project identification and execution. Borrower and IDA Performance The Borrower made a significant contribution during project formulation and preparation. It did its utmost to meet all the conditions for project effectiveness expeditiously. However during implementation, there were recurrent delays in mobilizing counterpart funds, and continuity in decision- making was somewhat hampered by the frequent tumover of Ministers. While the municipality received adequate assistance to identify sub-projects and establish priorities, its ability to maintain rehabilitated works remains inadequate because of lack of financial resources. Overall, AGETIPA performed well under difficult conditions and achieved the major objectives assigned to it. Its performance in complying with procurement procedures was, however, deficient. - i11 -  IDA played a key role in project identification and was successful in winning the support of the Malagasy authorities for the delegated contract management system. The appraisal mission went very smoothly, paving the way for rapid implementation. IDA performance during these phases of project processing was highly satisfactory. During project execution, IDA supervised the project by sending two to three missions annually, It also relied on the work undertaken by technical assistance missions financed by the French Cooperation Fund (FAC). All these missions focussed primarily on the progress and quality of works, and on training aspects. However, important issues which emerged during implementation concerning the coverage of technical audits, choices of technical options, packaging of contacts and compliance with IDA procurement procedures were overlooked. IDA performance in addressing these issues and taking corrective measures during supervision is deemed unsatisfactory. Project Outcome The project has achieved the objectives intended at appraisal and its results are considered satisfactory. By demonstrating the benefits of using contracted out services, AGETIPA has proven it is an effective instrument for assisting municipalities in infrastructure rehabilitation, promoting SMEs in the public works sector and creating jobs. After initial reluctance from Govemment officials, the delegated contract management system appears to have won the support of many local officials. IDA has financed a second project covering five secondary cities (Credit 2968-MAG) whose implementation is now picking up after an initial slow start. While physical results achieved during the project are substantial, the sustainability of rehabilitated works remains uncertain in the absence of sufficient resources for maintenance. The longer term sustainability of the delegated contract management system and of AGETIPA greatly depends on mobilizing additional resources not only from extemal donors other than IDA but also from domestic sources, in particular through the establishment of the Road Maintenance Fund. Key Lessons Learned Key lessons learned from this project can be divided in two main categories: those specific to this operation and AGETIPA itself, and those of a general nature which can apply to other similar AGETIP operations. Project related lessons * On the institutional front; it is necessary to expand the General Assembly of AGETIPA to reflect the composition envisaged in the appraisal report (including, for example, mayors, major utilities, and industry representatives). Steps have now been taken under the urban infrastructure project (Credit No. 2968-MAG) to ensure a broader representation at the Board. * With regard to financial matters, and in the interest of ensuring the sustainability of AGETIPA, it is essential that it accesses additional financial resources needed to carry out its activities from other donors and local sources (e.g. the Road Maintenance Fund). AGETIPA should also set aside sufficient resources to meet cash requirements from administrative expenses arising from temporary but recurrent declines in its level of activities. - iv -  * Regarding the packaging of contracts, there should be a gradual shift towards larger contracts. This would reduce administrative costs and have a positive impact on the development of small enterprises. * With respect to procurement, IDA shall ensure that technical audits systematically include a review of contracts and an analysis of compliance with procurement procedures set forth in the Manual of Procedures and the Credit Agreement. General lessons * During the sub-project identification phase and preparation of work plans of AGETIP projects, greater attention should be paid to the city's overall infrastructure rehabilitation needs, and a multiyear program and budget should be developed for approval by the municipality. Besides ensuring that municipalities meet their counterpart obligations, this approach would encourage elected officials to conduct a complete inventory of, and then classify, the city's entire urban infrastructure on the basis of actual rehabilitation requirements. Estimates of rates of return might then prove more accurate if a group of projects could be examined together in the context of a coherent strategy. * Satisfactory implementation of AGETIP projects require close supervision, especially during the initial phases of execution, supplemented by strong emphasis on the quality and adequate coverage of financial and technical audits. * Conceming the sustainability of project works municipalities should be required to draw up plans and budget for maintaining repaired infrastructure. Consideration should be given to require municipalities to contribute up to three years of maintenance funds up front as a criteria for project eligibility. * With regard to choice of technical options for repairing roads, job creation and maintenance requirements should receive increased attention. Giving preference to technical solutions based on paving techniques would be an important step in this direction.  REPUBLIC OF MADAGASCAR ANTANANARIVO URBAN WORKS PROJECT: CREDIT 2591-MAG IMPLEMENTATION COMPLETION REPORT PART I: PROJECT IMPLEMENTATION ASSESSMENT A. PROJECT IDENTIFICATION Name: Antananarivo Urban Works Project Credit Number: 2591-MAG Credit Amount: SDR 13.3 million (USD 18.3 million equivalent) Vice-Presidency: Africa Region Country: Madagascar Sector: Infrastructure Sub-Sector: Urban Development B. INTRODUCTION I. During the early 1990s the economy of Madagascar was mired in stagnation. As studies unanimously indicated, economic growth and investments were inadequate to bring about the economic recovery the country wished to achieve. Per capita GDP was clearly on the decline, hovering around US $230. hi response to this situation, major efforts to spur economic growth were launched focusing on greater openness, transparency, unfettered competition, private sector development, and public sector performance. 2. Shortcomings affecting basic infrastructure in general, and urban infrastructure in particular, were identified as among the most serious impediments to Madagascar's economic development. The constraints in the public infrastructure sector, especially those affecting roads, were familiar ones: insufficient financing, sorely inadequate implementation capacity, and cumbersome administrative procedures. 3. These shortcomings caused cumulative delays in the maintenance and rehabilitation of infrastructure required to sustain a market economy. This situation sparked recognition of the need for a more effective delivery system. The International Development Association (IDA) shared its experience with the Malagasy Government regarding the establishment of public works executing agencies. The creation of such agencies made it possible to increase the absorptive capacity for civil engineering works in all sectors. This approach was also consistent with the Government's structural adjustment efforts and its efforts to streamline its own role. C. PROJECT OTBJECTWVES 4. The primary objective of the Antananarivo Urban Works Project was to demonstrate the benefits of the delegated contract management system as tool for managing and maintaining infrastructure works. It was also intended to serve as a model for subsequent broader application. - 1 -  5. The project had two fundamental development objectives: (i) to restore the infrastructure of Antananarivo - and particularly its roads - to satisfactory condition rapidly and (ii) to promote the development of domestic small and medium enterprises (SMEs) in the construction and civil engineering industry by strengthening their capacity to design, execute, and supervise a wide range of public works undertakings. 6. The project's related objectives were to strengthen the management capacity of the municipality, create jobs by encouraging the use of labor-intensive technologies where economically efficient, and stimulate growth in the construction industry and in the economy as a whole. 7. The project was based on the principle of community participation in order to increase the voice of users and stakeholders in maintenance planning and establish a closer link between the municipality's actions and the wishes of the public. D. PROJECT DESCRIPTION 8. The project resulted in (i) the creation of an executing agency, Agence d'Ex6cution de Travaux d 'Infrastructure Publique d 'Antananarivo (AGETIPA); (ii) the execution of sub-projects to improve the infrastructure in Antananarivo; (iii) the establishment of a training program for SMEs and public institutions; and (iv) the provision of services to inform the public about the approach of the project and the work program. 9. The sub-projects were identified by the municipality and other public agencies and were reviewed and approved by the executing agency according to clear eligibility criteria. The sub-projects were then packaged into an annual work program, and focused primarily on urban road repair and rehabilitation of public areas. 10. AGETIPA was expected to oversee the contracting of about 25 sub-projects a year costing approximately US $200,000 each on average. The Agency was to be headed by a General Manager with broad powers who was recruited from the private sector and whose responsibilities were clearly defined in the statutes of the Agency. He was to report to a General Assembly comprising representatives of the water and electricity company, associations of domestic contractors and consultants, road transporters, and labor unions. User and stakeholder ownership was to be promoted through a participatory approach. 11. The Government and AGETIPA were to sign a convention under which AGETIPA assumed full responsibility for (i) reviewing and approving sub-projects proposed for execution; (ii) inviting tenders from domestic SMEs; (iii) evaluating tenders and awarding contracts (for engineering studies, works, and supervision); (iv) administering contracts efficiently and in a timely manner; and (v) providing technical assistance to SMEs and consulting with public institutions: AGETIPA was to perform these operations in accordance with a detailed Manual of Procedures, acceptable to IDA, including sample bidding documents. Its activities were monitored on the basis of monthly progress reports and technical and financial audits conducted by extemal auditors - acceptable to IDA - quarterly during the initial months and biannually thereafter. 12. The training program for SMEs covered execution of sub-projects and basic business management. This program was coordinated with the newly established National Center for Technical and Vocational Training (CNFTP) and the ongoing Financial Sector and Private Enterprises Development Credit project (APEX, Credit No. 21040-MAG). Promising consulting firms were trained to become future executing agencies in their own right with similar responsibilities. The provision of information services consisted mainly of using the mass media. -2 -  E. PROJECT PREPARATION 13. The idea of a delegated contract management system was first discussed in the early 1990s during the preparation of the seventh road project (Cr. 1905-MAG). Various Bank experts explained the Bank's experience in Africa in recent years with the establishment of public works executing agencies to the Malagasy Government.. IDA proposed to the Government that it introduce and test the system in an urban works project in Antananarivo and the Government agreed in December 1992. The Bank agreed to a PPF request to finance project preparation. Project appraisal took place in June 1993. The General Manager of AGETIPA, deemed acceptable to IDA, was appointed on July 25, 1993. The statutes of AGETIPA were registered and it was declared a public-interest entity on August 20, 1993. 14. The municipality of Antananarivo and the Govenmnent identified eight sub-projects to be executed before project implementation. They were reviewed and approved during a post-appraisal mission in September 1993. AGETIPA and the Government signed an initial convention, and AGETIPA and the municipality signed conventions for the contracting of the eight sub-projects. The relevant works were financed with an advance of US $300,000 (PPF No. 800.0) provided on July 20, 1993. Works were executed between November 1993 and January 1994. As a result of the satisfactory outcome of these sub-projects, and given the reconstruction needs in the aftermath of cyclone Geralda, an increase of US $975,000 to the PPF advance was made for seven additional sub-projects pending credit effectiveness. While these sub-projects were under way, various technical studies were also undertaken during the preparation phase (see Table 6). 15. Negotiations were held in Washington from January 31 -February 5, 1994. The credit of SDR 13.3 million was approved by the Executive Board of the World Bank in March 1994 and signed on April 7, 1994. Credit effectiveness, initially expected in July 1994, was declared on August 1, 1994. F. PROJECT IMPLEMENTATION 16. Project implementation began very rapidly in 1994 and proceeded relatively smoothly. Despite difficult traffic conditions, inexperienced contractors and delays in mobilizing counterpart funds, AGETIPA's performance indicators for management and physical results mostly exceeded expectations throughout project execution (see Table 5). This success was due in part to the speed of implementation of sub-projects; AGETIPA issued contracts in 7 days on average and paid work orders in a maximum of 10 days. The Agency's overhead costs amounted to only about 5 percent of works executed. AGETIPA had only 6 professional staff (out of a total of 12 staff) to manage a program costing US $20.5 million for a total of 874 signed contracts. The total cost of the works was FMG 92 billion (see Table 7A). Disbursements slightly exceeded appraisal estimates and the credit was fully disbursed on schedule (see Table 4). However, the speed with which the project was implemented and disbursements were made appears to have masked a number of problems, particularly with respect to choices of technical options, packaging of contracts, and compliance with IDA procurement guidelines: 17. The technical audits complied only partially with the terms of reference set out in the Manual of Procedures, since these audits focussed essentially on the quality and technical execution of the works. Limited effort was made to examine the implications of choosing certain technical options and packaging work contracts on procurement arrangements, as well as to review compliance with IDA procurement procedures. This led to a number of deficiencies as explained below. 18. Choice of technical options: Many of the sub-projects focused on resurfacing bitumen-covered roads with bituminous concrete. This technique was not included in the list of eligible works contained in the Manual of Procedures. Moreover, it has two drawbacks: it is not labor-intensive and only large firms - 3 -  can undertake such works. Because of political pressure and since the surface of the major road network had deteriorated so badly and no alternative options existed, it was inevitable that this technique would have to be used. This could have been anticipated during project preparation and reflected in the project design. Moreover, during project execution and supervision, this situation should have been officially acknowledged and dealt with as necessary, including revising the Manual of Procedures. 19. Packaging of contracts: Many sub-projects were subdivided into small contracts to an illogical and inordinate extent. The result was a number of contracts all out of proportion to the number estimated at preparation (732 instead of 310). The reasons given by AGETIPA were: (i) the need to adapt to the capacity of the construction enterprises; (ii) the desire to have the maximum possible enterprises working simultaneously, mainly in order to minimize execution delays (to prevent traffic disruption), but also (iii) to dispense with IDA procurement prior review "in order to save time." This splitting up of contracts certainly generated additional administrative costs, and also very likely raised the cost of the works themselves. Although small enterprises were supposed to have been the beneficiaries of such exercise, it should be noted that most of the contracts were executed by two large companies with an annual income in excess of FMG 14 billion. These two firms in fact built works costing FMG 48 billion, or 51.2 percent of the entire project (see Table 12). 20. Procurement procedures and Award of contracts: Compliance with procurement procedures was deficient. The Nanisana road project can be cited as an example. The total cost of this sub-project was FMG 3.6 billion (US $700,000), for which ICB should have been undertaken instead of LCB. Only two bids were submitted. According to the Manual of Procedures, since fewer than three bids were received, re-tendering should have been carried out. This was not done and the contract was awarded nevertheless to the winning firm. However, the contract was then subdivided into 15 small separate contracts, each of these smaller contracts averaged FMG 240 million to remain below IDA prior review threshold. In addition, the Manual of Procedures calls for an automatic prior review by IDA where a tendering process results in fewer than five bids, or where the successful bidder has not submitted the lowest bid. A review of the project files suggests that this condition was not met. 21. The Manual of Procedures provides for eliminating lowest bids on the basis of objective criteria. A review of the tendering process for the project as a whole shows that 28 different criteria were used to eliminate the lowest bids. This number is neither realistic nor manageable. Moreover, as the following table indicates, the number of lowest bids eliminated as the project progressed kept rising: Year % of lowest bids rejected 1993 5% 1994 6% 1995 18% 1996 40% 1997 53% 22. This pattem is surprising, to say the least, since one would expect that over time the enterprises should have grown more familiar with AGETIPA's procedures. The argument that finns were saturated by an excessive workload defies credibility. In fact, the number of lowest bids rejected between 1996 and 1997 was much higher, despite a noticeable decline in the volume of works during this period. - 4 -  G. ACHIEVEMENT OF OBJECTIVES 23. Rehabilitation of urban works in Antananarivo. Under the project, 106 km of roads, 111 km of sidewalks, and 104 km of drains were rehabilitated, 8 km of stairs and alleyways and about 100 retaining walls were built, and five markets and various public spaces were developed. Sub-project viability was demonstrated by an internal rate of return of over 20 percent. This assessment was borne out in user surveys. Besides these substantial physical achievements, a steady improvement in the quality of the works was observed as the project progressed despite inexperienced SMEs. However, traffic conditions in Antananarivo still remain difficult because of the increase in motorization. 24. Job creation and poverty alleviation. Throughout, the project generated 133,000 man-months of employment in direct jobs. The resulting payroll represented approximately 20 percent of the cost of the works, or some FMG 18 billion (see Table 12). Many suppliers (including quarry and sandpit operators) to whom enterprises turned for the works benefited in terms of employment and income. Together with this multiplier effect, increased mobility of the population also had a positive impact on job creation, particularly in outlying districts. 25. Payroll accounted for 20 percent of total costs on average, although the figure varied considerably depending on the nature of the works and choice of technique. For example: (a) Resurfacing with bituminous concrete absorbed 12 percent of the credit but generated only 0.6 percent of the project payroll (labor accounts for only 1 percent of the cost of such works). (b) Street pavin, including supplied paving stones, accounted for 8 percent of project credits but generated 25 percent of the payroll (in this case labor accounts for 62 per cent of the cost). 26. Promotion of consulting firms and private enterprises and stimulation of growth in the public works sector. The enterprises and consulting firms with which AGETIPA dealt initially were weak. They lacked professionalism, were notoriously under-equipped and had limited access to banking services. Training was provided and a considerable effort was undertaken to mobilize SMEs, which had for so long found it necessary to diversify their activities excessively across various sectors owning to a paucity of work (prior to 1994). AGETIPA took various steps to enhance their performance, such as providing targeted training and on-the-job training, encouraging the profession to organize itself, and fostering the preparation of "Quality Standards" and a work-site manual. Forty training modules were developed. Some sixty training sessions were organized, attracting 1,111 participants. Nine on-the-job training modules were also used to provide instruction to 101 professional staff and 57 laborers. 27. SMEs working for AGETIPA enjoyed the benefits of technological progress, as most of them became involved in the project at its inception. The CVs of their permanent staff, their references, and the equipment they owned show that they were active mainly in the construction sub-sector. A set of "Quality Standards," a list of technical specifications which they drew up together under AGETIPA's guidance, also had a positive effect. In addition to developing greater technological skills, participating firms ended up engaging in a natural self-selection process, with the result that the enterprises keenest to boost their professionalism stood out. Overall, the number of failures among the SMEs which contracted with AGETIPA was low. 28. SMEs in the civil works sector in Antananarivo also became better organized. An association was established during the project and entrepreneurs met periodically under its auspices to share their experience. There were constructive sessions, as evidenced by the proposals which they forwarded to -5-  AGETIPA for improving relations among the delegated contract manager, general contractors, and SMEs. Such initiatives had been umheard of earlier. 29. With respect to stimulating the growth of the public works sector, the following table shows the total of SMEs' contracts with AGETIPA as a proportion of their overall business was 20 percent on average and evolved as follows: 1994 1995 1 1996 1997 7% 23% 27% 18% 30. With regard to promoting consulting firms and enterprises, the following figures indicate that: * AGETIPA worked with 387 enterprises; * Bids were invited for 171 studies and works supervision projects; *. Eighty consulting finns participated in the bidding and 39 were awarded contracts; * Bids were invited for 700 works contracts; * One hundred seventy-two firms participated and 105 were awarded contracts. 31. Overall, many enterprises derived considerable benefit from the AGETIPA project, which was deemed a sound program that provide them with business opportunities beyond the short term. The project also enabled SMEs to demonstrate their ability to seize these opportunities and adapt to working with several contract managers, whether delegated or not. It should be noted that although the project expanded and stabilized existing SMEs in the sector, it led to the establishment of very few, if any, new firms. 32. Strengthening the management capacity of the municipality. Activities were mostly focused on assisting the municipality in using the services of a contracted out agency as best as possible and geared at improving the interaction of the municipality with its service users. During project startup, operations were initially identified by involving district organizations on an individual basis (spontaneously or otherwise) through surveys, media, campaigns, and direct assistance to neighborhood committees. AGETIPA then pre-qualified the projects and submitted them to the municipality for review. Final approval was made by the General Assembly of AGETIPA. The advantage of this method - borne out by surveys - was that it reflected the situation in the field. There were drawbacks, however, in that coherence of sub-projects and a global strategic vision proved at times to be elusive goals. 33. Subsequently, once elected municipal authorities took office, AGETIPA adapted the sub-project identification process to this new situation. From that point on, the municipality provided AGETIPA directly with a list of its priorities. AGETIPA then determined the eligibility of each sub-project. H. MAJOR FACTORS AFFECTING THE PROJECT 34. The major factors affecting project implementation are set out below: Factors outside the control of the Government and the municipality: * The depreciation of Madagascar's currency by half during project implementation hampered the ability of the Government and the municipality to pay their share of sub-project financing. This also translated into a quasi absence of funding for maintenance. * Antananarivo experienced a very rapid increase in traffic congestion, which seriously hindered works construction and limited bypass options. -6 -  Factors within the control of the Government and the municipality: * The transformation of the transaction tax to a value-added-tax (VAT) and rate variations during the initial stage of the project complicated financial transactions; * Cumbersome administrative procedures for raising counterpart funds owed by the Government and the municipality resulted in delays in approving work programs; * The lack of funds to maintain rehabilitated works resulted in poor maintenance and still raises serious concerns as to the sustainability of investments; * The limited capacity of the municipality to coordinate and supervise works executed by major utilities such as JIRAMA and TELMA resulted in implementation delays of some sub-projects and in some instances degradation of works already completed. The absence of representatives of utilities on AGETIPA's General Assembly may have compounded this problem. Factors within the control of AGETIPA: * Very few representatives of civil society served on the Board of Directors, resulting in a smaller Board than had been expected at project preparation. The Board had only four members, one of whom was the General Manager of AGETIPA. Little effort was made to broaden representation at the Board because of the perceived risk of political interference in AGETIPA's operations, at least initially. Stronger efforts should nevertheless have been made to include at least representatives of major utilities and civil society. * Because the city's infrastructure had deteriorated severely and was in need of urgent repair, it was difficult to carry out very detailed engineering studies. As a result, many changes were made to the sub-projects in the course of project execution. Managing this situation was complicated by the fact that the enterprises and consulting firms were not very experienced. It thus became necessary for AGETIPA officials to get directly involved in project execution, a situation that placed the Agency in the difficult position of having to serve as both contract manager and contractor. I. PROJECT SUSTAiNABILrrY 35. Sustainability of the works. This will depend primarily on the municipality's capacity to keep up with maintenance requirements. In the absence of adequate financial resources, this is uncertain at present. Improved resource mobilization on the part of the municipality and the establishment of a Road Maintenance Fund under a proposed IDA financed Transport APL in FYOO remain critical. If this is achieved, AGETIPA definitely should play an important role as the municipality's delegated contract manager for maintenance works. 36. Sustainability of the delegated contract management system. After initial reluctance, a number of Malagasy institutions have now recognized the benefits and usefulness of the system. Several ministerial departments (Public Works, Regional and Urban Development) and AGETIPA are involved in preparing legislation concerning delegated contract management. A draft of the proposed legislation is expected to be submitted to the National Assembly in 1999. 37. Sustainability of AGETIPA. AGETIPA demonstrated its usefulness and efficiency. This success was not lost on the elected mayors of Madagascar's large cities. Their requests for government assistance - 7 -  have prompted the Government to seek IDA support to extend the system to the country's major cities. In early 1997, an amendment to Credit No. 2591 -MAG was made authorizing the financing of sub-projects in five other cities. At the same time, a new IDA credit (Cr. No. 2968-MAG), which was prepared in 1996 and became effective in 1997, calls for the extension of the delegated contract management system under AGETIPA's auspices to these five cities. 38. Thus far, IDA is the sole donor financing AGETIPA operations. If AGETIPA is to prove sustainable, it is crucial that it secures additional funding not only from other donors but also from local sources (e.g. Road Maintenance Fund). To this end, AGETIPA needs to project an image other than a provider of infrastructure and publicize the social and economic benefits derived from its activities. It should also ensure broader representativity on its Board. As an autonomous institution, AGETIPA also needs to set aside sufficient resources to meet cash requirements from administrative expenses arising from temporary but recurrent declines in its level of activities. J. IDA PERFORMANCE 39. IDA played a key role during project identification and preparation and was successful in winning the support of the Malagasy authorities for the delegated contract management system. The appraisal mission went very smoothly, paving the way for rapid implementation. IDA performance during these phases of project processing was highly satisfactory. During project execution, IDA supervised the project by sending two to three missions annually (see Table 10). It also relied on the work undertaken by technical assistance missions financed by FAC. All these missions focused primarily on the progress and quality of works, and on training aspects. However, as mentioned in paras. 17-22, there were some important issues which emerged during implementation concerning the coverage of technical audits, choices of technical options, packaging of contracts and compliance with procurement procedures. IDA performance in addressing these issues and taking corrective measures during supervision is deemed unsatisfactory. K. BORROWER PERFORMANCE 40. Government performance. The Government made a significant contribution during project formulation and the final phase. It did its utmost to meet all the conditions for project effectiveness expeditiously. However, there were recurrent delays in raising its financial contribution, and continuity in decision-making was also somewhat hampered by the frequent turnover of Ministers. 41. Performance of the municipality. The municipality required and received assistance to identify sub-projects and establish priorities. Its ability to maintain rehabilitated works was and remains inadequate mainly because of lack of financial resources. An effort is currently under way to streamline and improve the resource mobilization process, including revised procedures for assessing the tax base and collecting taxes under the Antananarivo Urban Works Project (Credit No. 2117-MAG) and with technical support to the municipality from the French Cooperation Fund (FAC). Further efforts and assistance are nevertheless still needed to improve urban infrastructure management at the municipal level. 42. Performance ofAGETIPA. Overall, AGETIPA performed well and achieved the major objectives assigned to it. * The level of infrastructure service provided in the Antananarivo Urban Area was substantially improved through the rehabilitation of a significant segment of the infrastructure (including - 8 -  roads, sidewalks, sanitation, alleyways, and stairways). However, because of the substantial infrastructure backlog which accumulated for the last twenty years and the rapid increase in motorization during project execution, traffic conditions still remain difficult despite these achievements. * Support for SMEs (enterprises and consultants) is now well-established. Their professionalism was gradually enhanced by providing many opportunities for staff training and by actively involving AGETIPA staff in the preparation of studies and execution of works. * A communication system is in place to keep users informed on a regular basis. The project's socioeconomic and enviromnental impact was assessed periodically through user surveys, leading to improvements in project execution with a view to meeting the public's expectations more effectively. * After initial reluctance from Government officials, the delegated contract management system has been put to the test and has now won the support of many local officials. * Sub-project design and the quality of works have gradually been improved through periodic technical assistance missions. * Adequate financial management systems were put in place and financial audits were prepared and submitted to IDA on time. IDA had, however, to inform the auditors that there were several issues which were not sufficiently covered in their reports such as internal control and adequacy of expenses in relation to contracts awarded. L. PROJECT OUTCOME 43. The project has achieved the objectives intended at appraisal and its results are considered satisfactory. By demonstrating the benefits of using contracted out services, AGETIPA has proven it is an effective instrument for assisting municipalities in infrastructure rehabilitation, promoting SMEs in the public works sector and creating jobs. After initial reluctance from Government officials, the delegated contract management system has now won the support of many local officials. IDA has financed a second project covering five secondary cities (Credit 2968-MAG) whose implementation is now picking up after an initial slow start. 44. While physical results achieved during the project are substantial, the sustainability of rehabilitated works is uncertain in the absence of sufficient resources for maintenance. The longer term sustainability of the delegated contract management system and of AGETIPA greatly depends on mobilizing additional resources not only from external donors other than IDA but also from domestic sources, in particular through the establishment of the Road Maintenance Fund. M. KEY LESSONS LEARNED 45. Key lessons learned from this project can be divided in two main categories: those specific to this operation and AGETIPA itself, and those of a general nature which can apply to other similar AGETIP operations. Project related lessons * On the institutional front, it is necessary to expand the General Assembly of AGETIPA to reflect the composition envisaged in the appraisal report (including, for example, mayors, major utilities, and industry representatives). Steps have now been taken under the urban -9-  infrastructure project (Credit No. 2968-MAG) to ensure a broader representation at the Board. * With regard to financial matters, and in the interest of ensuring the sustainability of AGETIPA, it is essential that it accesses additional financial resources needed to carry out its activities from other donors and local sources (e.g. the Road Maintenance Fund). AGETIPA should also set aside sufficient resources to meet cash requirements from administrative expenses arising from temporary but recurrent declines in its level of activities. * Regarding the packaging of contracts, there should be a gradual shift towards larger contracts. This would reduce administrative costs and have a positive impact on the development of small enterprises. * With respect to procurement, IDA shall ensure that technical audits systematically include a review of contracts and an analysis of compliance with procurement procedures set forth in the Manual of Procedures and the Credit Agreement. General lessons * During the sub-project identification phase andpreparation of work plans of AGETIP projects, greater attention should be paid to the city's overall infrastructure rehabilitation needs, and a multiyear program and budget should be developed for approval by the municipality. Besides ensuring that municipalities meet their counterpart obligations, this approach would encourage elected officials to conduct a complete inventory of, and then classify, the city's entire urban infrastructure on the basis of actual rehabilitation requirements. Estimates of rates of return might then prove more accurate if a group of projects could be examined together in the context of a coherent strategy.