Colombia - Second Rural Roads Sector Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 16185 IMPLEMENTATION COMPLETION REPORT COLOMBIA SECOND RURAL ROADS SECTOR PROJECT (LOAN 3157-CO) December 20, 1996 Infrastructure Division Department In Latin America and the Caribbean 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 = Colombian Peso (Col $) (June each year) 1990 - 497,32 1991 - 623,67 1992 - 677,17 1993 - 784,49 1994 - 831,16 1995 - 874,52 1996 - 1,069.11 WEIGHTS AND MEASURES Metric System FISCAL YEAR OF BORROWER January 1-December 31 ABBREVIATIONS AND ACRONYMS CGR Office of the Comptroller General of the Republic CONPES National Council for Economic and Social Policy CORPES Regional Planning Councils DNP National Planning Department ERR Economic Rate of Return RCF Roads Cofinancing Fund (under FINDETER) FINDETER National Fund for Territorial Development FNCV National Rural Roads Fund GOC Govermnent of Colombia IBRD International Bank for Reconstruction and Development IDB Interamerican Development Bank ICR Implementation Completion Report INDERENA National Institute of Renewable Resources INVIAS National Institute of Highways MOPT Ministry of Public Works and Transport (up to 1993) MOT Ministry of Transport (from 1993 onward) NPV Net Present Value PNR National Rehabilitation Program SAR Staff Appraisal Report SPWs Secretariats of Public Works (Departmental) UDECOs Department's Cofinancing Offices Vice President: Shahid Javed Burki Director: Paul Isenman, LA3DR Division Chief: Peter Ludwig, LA31N Task Manager: Jose Luis Irigoyen, LA3IN  FOR OFFICIAL USE ONLY TABLE OF CONTENTS PREFACE ........................................... i EVALUATION SUMMARY ...........................................i PART I. PROJECT IMPLEMENTATION ASSESSMENT ...........................................1 A. STATEMENT/EVALUATION OF OBJECTIVES ...........................................1 B. ACHIEVEMENT OF OBJECTIVES ...........................................1 C. MAJOR FACTORS AFFECTING THE PROJECT ..........................................7 D. PROJECT SUSTAINABILITY .....9.....................................9 E. BANK PERFORMANCE .......................................... 10 F. BORROWER PERFORMANCE .......................................... 13 G. ASSESSMENT OF OUTCOME .......................................... 14 H. FUTURE OPERATIONS .......................................... 15 I. KEY LESSONS LEARNED .......................................... 17 PART II. STATISTICAL ANNEXES .......................................... 21 Table 1: Summary of Assessments .......................................... 21 Table 2: Related Bank Loans/Credits .......................................... 22 Table 3: Project Timetable .......................................... 23 Table 4: Loan Disbursement .......................................... 23 Table 5: Key Indicators for Project Implementation .......................................... 24 Table 6: Key Indicators for Project Operation .......................................... 25 Table 7: Studies Included in Project .......................................... 25 Table 8A: Project Costs (1) .......................................... 26 Table 8B: Project Financing(l) .......................................... 26 Table 9: Economic Cost and Benefits .......................................... 27 Table 10: Status of Legal Covenants .......................................... 28 Table 11: Compliance with Operational Statements .......................................... 31 Table 12: Bank Resources: Staff Inputs .......................................... 31 Table 13: Bank Resources Missions .......................................... 32 Appendix A: Completion Mission's Aide Memoire .......................................... 34 Appendix B: Borrower's Contribution to the ICR .......................................... 61 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not othenvise be disclosed wiLhout World Bank authorization.   IMPLEMENTATION COMPLETION REPORT COLOMBIA SECOND RURAL ROADS SECTOR PROJECT LOAN 3157-CO PREFACE This is the Implementation Completion Report (ICR) for the Second Rural Roads Sector Project in Colombia, for which Loan 3157-CO in the amount of US$55 million equivalent was approved on January 16, 1990 and made effective on June 6, 1990. The original closing date was December 31, 1995. The loan was 99.97% disbursed, and the last disbursement took place on May 9, 1996, without extension of the closing date. Undisbursed funds in the amount of US$13,865.71 equivalent were canceled on May 9, 1996. The ICR was prepared by Messrs. Jose Luis Irigoyen (LA31N) and German Ospina (Consultant), and reviewed by Messrs. Peter Ludwig (Division Chief, LA3IN) and Robert Crown (Project Adviser, LA3DR) of the Latin America and the Caribbean Region. Preparation of the ICR was begun during the Bank's final supervision mission in November 1995. It is based on material in the project file and discussions with the Borrower. The Borrower contributed to preparation of the ICR by conducting an ex-post evaluation of a sample of road subprojects financed under the loan, preparing an operational plan, and providing statistical data and a retrospective assessment of the project.  I  EVALUATION SUMMARY SECOND RURAL ROADS SECTOR PROJECT LOAN 3157-CO COLOMBIA Introduction The project followed two previous successful operations (Loans 1966-CO and 2668-CO) in the rural road subsector in Colombia. Project objectives The main goal of the project was to contribute towards achieving the country's economic and social goals (expand markets and reduce poverty in rural areas) by: (a) helping classify the road network and define functions and responsibilities of national, regional and local (municipal) authorities, and develop, strengthen and organize accordingly their institutional capacity for managing rural roads activities; (b) ensuring over a realistic time frame, satisfactory implementation of key policies in the roads sector; (c) consolidating positive institution-building advances in road management initiated by FNCV under the Rural Transport Sector Project (Loan 2668-CO) and during preparation of this project, especially in programming and supervision of civil works, strengthening of FNCV's Regional Offices and organizing the maintenance activities; and (d) contributing to financing a balanced four-year investment and maintenance programs (para. 1). Implementation experience and results The consolidation of the institutional advances within FNCV progressed satisfactorily until 1993, including the strengthning of its Regional Offices. However, this consolidation was later offset by the liquidation of FNCV mandated by the drastic decentralization agenda followed by the Government upon the approval of the 1991 Constitution. The liquidation of FNCV entailed retrenching about 1,600 staff and closing 15 out of the 24 Regional Offices. In spite of the disruptions created by the retrenchment program, FNCV substantially accomplished the objective of completing a four-year investment and maintenance program, with the exception of routine maintenance. The sector policy objectives, including a cofinancing system with municipalities and road funding strategies, were largely achieved, or are expected to be achieved once Government completes the implementation of its decentralization policies. The study on funding strategies made a valuable contribution highlighting the existing imbalances between responsibilities, capacity and financial resources, an aspect probably overlooked initially by the reform but that now is gaining support among the sector authorities. The road classification and inventory completed early in the project became a key instrument during implementation of the sector reforms, first defining responsibilities among the three levels of government, and later, supporting the transfer of secondary and rural roads to the departments. (paras 15 - 17).  - ii - With project funds, FNCV achieved the following physical targets: (a) construction of about 776 km of rural roads (or about 97% of the target planned at appraisal); (b) improvements on about 1053 km of roads (or about 81% of the target); (c) construction of about 1,134 meters of small bridges (or about 101% of the target); and (d) periodic maintenance on about 1,239 km of roads (or 55% of the initial target). The shortfall in periodic maintenance is compensated by a greater achievement under FNCV's 1988-1993 Investment and Maintenance Plan (which more than double the periodic maintenance target). The works, however, were completed one year behind the initial schedule (paras. 17 - 18). Several external factors had significant impact on project execution. The most critical included: (1) new economnic and social policies that shifted priorities from the past focus on rural development and poverty alleviation to the opening up of the economy and modernization of the state, (2) institutional reforms in the sector that brought major changes to the road sector and defined new goals for the decentralization agenda, (3) the liquidation of FNCV, as part of these reforms, that progressively weakened FNCV, (4) delays in implementing the new cofinancing system that added uncertainty to the devolution of roads and left part of the road network unattended, and (5) the introduction of new environmental policies and regulations that slowed down the execution of several road subprojects (paras. 21 - 28). Bank Performance Project preparation benefited from the previous experience with FNCV, as the same Bank team supervising the then on-going Rural Transport Sector Project (Ln. 2668-CO) was responsible for preparing and appraising the new project. A key contribution was to recognize that a major obstacle to developing an adequate framework for maintenance was the lack of definition of responsibilities between local, regional, and national administration levels. In view of the difficulties usually associated with the design and execution of institutional programs to build capacity at the local level, the appraisal should have elaborated on a more detailed implementation plan for this component (paras. 31 - 35). The major changes in the institutional set up and the further and swifter advancement of the decentralization agenda could not be foreseen at appraisal. When the modernization decrees were passed late in 1992, the Bank advised the Government about the risks associated with organizing the new units and cofinancing systems, and getting them to function smoothly, within the short transition periods established under the decrees. Though the Bank recognized that in the long run the reform would increase the efficiency in the provision of public services, it did not deem feasible to accomodate the project to the drastic decentralization process as this would have jeopardized the on-going works at a time when most institutional development components were completed. Bank supervision missions discussed the implications of the decentralization agenda for achieving the objectives of the project with high level officials at the Ministry of Transport, the National Fund for Territorial Development (FINDETER) and the National Planning Department, and drew the attention of the sector authorities on the issues that the agenda should address to ensure long-term sustainability (paras. 37 - 38). This dialogue bore fruit only towards the end of the project when, in 1995, the Government, acknowledging the difficulties of implementing the decentralization in the continuous absence of local technical, financial and administrative capacity, suspended the liquidation of FNCV, cancelled the transfer of rural roads, and established a cofinancing system that would take into account the capacity of the departnents.  - iii - Borrower Performance FNCV introduced between 1988-89 the cofinancing system ensuring that local government, particularly those in more need, had access to the system The screening and the economic and environmental evaluation of road construction and improvement subprojects were done using the methodologies designed and tested under the Rural Transport Sector Project (Ln. 2668-CO) (para. 41). The modemization decrees issued in 1992 were prepared at the last minute to comply with a Constitutional deadline, and by a few people in the Presidency and DNP, without extensive involvement of the institutions affected. This resulted in implementation problems during the transition period. Key elements of the reform needed further legislation. The reform process would have benefited from a more participative design and a well-structured implementation plan. However, political factors may sometimes require quick, radical changes while there is a favorable opportunity. This was the criteria that prevailed at the time of approving the plan for liquidation of FNCV. The progress in developing the new institutions and financial sources that were supposed to take over FNCV's responsibilities was slower than the phasing out of FNCV, to the extent that in June 1995 Congress suspended the liquidation of FNCV, which will now continue operating until the weakest departments and local governments demonstrate sufficient capacity to undertake this responsibility alone (paras. 44 - 45). Summary of Findings The project's outcome can be considered marginally satisfactory. Under difficult institutional circumstances, the works component was largely executed. The project contributed to the definition of the reforms mandated by the decentralization agenda and key sector policies advanced by the Transport Law. The road classification completed under the project was an instrument in the definition of responsibilities among the three levels of government prior to the reforms, and the study on road-user charges and funding strategies highlighted the imbalances between responsibilities and financial resources. The Transport Law set key principles to increase efficiency and effectiveness in road sector management, but most of them are not yet fully implemented (para. 47). The results of the long term efforts to build capacity at the local level and mobilize municipios to assume their responsibilities for road maintenance were mixed. As the experience of FNCV shows, it will take time until local governments fully assume such responsibility for road maintenance and the principle that maintenance should be provided on a continuous basis; but that experience also recognizes that municipios take action when they perceive the roads need repair. In retrospect, the project's outcome would have been better, had the reform mandated a more gradual liquidation of FNCV or focused on a new role for FNCV, consistent with the new decentralization policies. The sustainability of the project investments remains at risk since in the long run it relies on (1) developing capacity for maintenance at the level of the departments and local governments, and (2) implementing appropriate financing mechanisms for the departments and local governments to secure a stable flow of funds for this purpose. Though there are issues yet to be addressed, the prospects for improved sustainability are reasonably good since Government has shown its commitment to furthering the decentralization of the road sector and make it work (para. 48). Future Operations Government is taking actions in three main areas that will eventually contribute to improving project sustainability, namely: (a) strengthening the role of MOT to coordinate and monitor the implementation of  - iv - technical assistance programs for subnational governments and assisting the departments in the preparation of their road plans; (b) strengthening road sector funding by (1) increasing investment levels to clear the existing road infrastructure backlog, (2) enhancing the structure of road user charges, to redirect the use of matching grants under the cofinancing system, and (3) making operational the cofinancing system to finance maintenance and rehabilitation of secondary and tertiary roads; and (c) reincorporating the reorganized FNCV into the national cofinancing system to assist the weakest departments and municipios in expanding, improving and maintaining their rural road networks (para. 51 - 56). Lessons Learned The experience gained so far in implementing the Colombian institutional reforms confirms that reform processes do not immediately produce all the intended results. However, they open the dialogue to continue the reform and allow permeation of "cultural" change, assuming that commitment to the reform is steadfastly sustained. It is not enough to pass legislation, it is also necessary to internalize strategies and concepts and this entails time. Political considerations may force quick, radical changes, even without a sound implementation strategy in place but the risks of delays, and even reversals increase substantially, and may jeopardize the achievement of the objectives of the reform. The dismantling of FNCV started without having in place adequate financing mechanisms and the institutions that would take over its responsibilities (para. 58). The transfer of roads from the central government to the departments turned into protracted negotiations over the control of the resources associated with it, instead of focusing on building capacity through some time-bound programs. The financial mechanisms to support the transfer should be clearly defined in advance to avoid this risk. The experience is demonstrating the need to develop financing mechanisms based on user-charges --following the tariff concept established by the Transport Law-- to avoid the current situation in which departments heavily depend on inter-governmental transfers to finance their programs (para. 59). Designing effective institutional programs aimed at building local capacity poses probably one of the greatest challenges for development projects. The quality and sustainability of these programs become even more critical when dealing with rural communities which are institutionally weaker and more difficult to reach. The effectiveness and the prospects for sustainability of these programs can be increased by: (a) promoting the dissemination of good practices adopted by municipios and the use of strategic partnerships. The former carries strong motivation, thus becoming an effective way of speeding up the development of local capacity. The latter may help address issues in training and mobilizing resources for maintenance; (b) designing flexible programs, tailored to local or regional demands and provided in a decentralized manner; and (c) ensuring sustained commitment while seeking progressive results. Though it is generally accepted that building institutional capacity is a long term effort, this is often neglected in practice when the objectives of such programs are set in a manner that is not consistent with the time needed to produce the desired changes. More attention should be devoted to defining an appropriate sequence for the interventions needed (which may go beyond the scope of one project) and setting progressive benchmarks. The use of these benchmarks will help focus the programs on a few prioritized goals, facilitate the assessment of progress and outcome, and increase motivation by showing early results (para. 62). The experience of FNCV shows how difficult it is to move municipalities towards a "continuous" maintenance culture, in spite of the increased "ownership" than can be expected from the cofinancing system. The project suggests some elements that may help local governments to move to a culture of  maintenance: (a) some previous investment is needed to improve the condition of the roads to a condition in which they are maintainable; (b) the effectiveness of routine maintenance becomes a demonstrated reality to both road users and local authorities only after having delivered the services for a reasonable period of time; (c) an affordable mechanism of road user charges secures a stable flow of funds for this purpose. While the use of community-based organizations or microenterprises is an affordable solution for the municipios, their sustainability requires prompt payment; (d) those with a vested interest in good roads evaluate the results and exercise their voice before the local authorities. Local participation in screening and identification of subprojects, favored by FNCV, should also be expanded to include decisions regarding maintenance (para. 64).   IMPLEMENTATION COMPLETION REPORT COLOMBIA SECOND RURAL ROADS PROJECT (LOAN 3157-CO) PART I. PROJECT IMPLEMENTATION ASSESSMENT A. STATEMENT/EVALUATION OF OBJECTIVES 1. The main objective of the project was to contribute towards achieving the country's economic and social goals by: (a) helping classify the road network and define functions and responsibilities of national, regional and local (municipal) authorities, and develop, strengthen and organize accordingly their institutional capacity for managing rural roads activities; (b) ensuring over a realistic time frame, satisfactory implementation of key policies in the roads sector, including those related to the administrative decentralization process, road maintenance and investment programming, coordination and resource mobilization between agencies at different administrative levels; (c) consolidating positive institution- building advances in road management initiated by FNCV under the Rural Transport Sector Project (Loan 2668-CO) and during preparation of this project, especially in progranuming and supervision of civil works, strengthening of its Regional Offices and organizing the maintenance activities; and (d) contributing to financing a balanced four-year investment and maintenance program. 2. These objectives were consistent with the Government's overall policies of expanding markets, improving living conditions in rural areas and creating a favorable environment for private investment, employment growth, and access to social resources. They were also in line with the Bank's country and sector strategies that emphasized the importance of effectively maintaining and improving the access of rural communities to social services and markets. While the objectives remained unchanged, project implementation witnessed major changes in economic policy, institutional reform and decentralization. The reforms implied slating FNCV for liquidation by end of 1995. 3. The project design was simple. Its execution, however, was somewhat complex because it entailed implementing (1) a large number of small subprojects scattered mostly throughout the poorest municipalities, (2) a cofinancing system developed in 1989, which required mobilizing local governments in the selection and financing of subprojects, and (3) more stringent environmental guidelines for the execution of road works. B. ACHIEVEMENT OF OBJECTIVES 4. Macroeconomic and social goals. Farm access to markets in Colombia --and the competitiveness of farm produce-- is conditioned by the extent and quality of rural access roads. As reported in an ex-post evaluation of a sample of 18 subprojects, the project made a positive contribution in raising the living standards of the rural population in the project areas by reducing transportation costs from producing  - 2 - centers to markets, and increasing integration arnong rural communities and their local govenmments. This, in tum, has helped bring stability to areas of social unrest. 5. Classifying the road network and defining responsibilities. At this stage, these objectives have been only partially achieved because the decentralization agenda --reshaped in 1993 to set new goals far beyond those initially contemplated under the project-- is still under implementation. Though several issues remain to be addressed, the new agenda sets the right direction by putting functional responsibilities, resources and accountability closer to the administrative level that is better positioned to know the needs and respond to them. 6. The road classification established in October 1990 was a first step towards clarifying responsibilities for road sector management'. FNCV, MOPT and the Secretariats of Public Works in each department, under the guidance of MOPT, carried out an inventory of all public roads and assigned functional responsibilities over each road. Later, these inventories became key tools at the time of planning the transfer of roads to subnational governments. However, the classification fell short in addressing the financing issues associated with these responsibilities (see para. 14). The classification assigned to local governments the responsibility to maintain the local roads and even FNCV's network; the cofinancing system introduced in 1988-89 made this clear, and required for all cofinanced maintenance subprojects that local governments contribute with the labor needed to carry out the non-mechanized works. 7. In December 1993, Congress approved the Transport Law which provides the legal framework for the redistribution of responsibilities over transport infrastructure among central government, departments and municipalities. The law sets key principles for increasing efficiency in the sector. Basically, it: (a) mandates central government to transfer to the departments responsibility over about 13,000 km of secondary roads under MOPT and all rural roads under FNCV, in total about 32% of the road system, together with the resources needed to maintain or improve those roads. To this end, the Nation had to agree with each department on a plan to gradually transfer such responsibilities and resources within a three-year period, starting in January 1994; (b) requires the setting of road user-charges to fully recover infrastructure maintenance, operation and development costs, and gives ample powers to the departments and municipalities to charge tariffs, tolls, betterment levies and even surcharges on gasoline taxes exclusively for transport infrastructure financing. To support the devolution of roads to the departments, the law also establishes under FINDETER (a national financial agency)2 a new Road Cofinancing Fund (RCF) to pass matching grants, exclusively for roads; (c) allows the departments to restructure their road agencies, including the possibility of establishing autonomous private/public entities with capacity to raise their own financing (ie., bonds issues, loans); and (d) provides the basic framework for franchising infrastructure facilities to the private sector. ' / Under such classification, MOPT was responsible for about 25,600 km of national roads (primary and secondary roads), the departments for about 39,264 km of secondary or regional roads, local governments for about 26,300 km of rural roads, while FNCV for about 23,000 km of mostly rural roads (those which had been built by FNCV in the past). 2/ Under the institutional set-up created by the law, departments prepare together with local governments the road investment and maintenance programs proposed for cofinancing through the RCF (central goverrunent matching grants). Specialized units (UDECOs, Unidades Departamentales Especiales de Cofinanciacion) have been set in each department, with representation of local governments, to coordinate these efforts and approve the programs proposed for cofinancing in the various sectors. FINDETER approves the matching grants and makes payments under the RCF. Other funds under the cofinancing system created in 1992 include the Social Investment Fund (health and education), the Urban Infrastructure Fund, and the Integrated Rural Development Fund.  -3 - Box 1: An Overlook of Colombia's Decentralization and Cofinancing Policies 1983.86 Law 14 of 1983 grants municipalities greater revenue raising powers. Law 12 of 1986 increases the municipalities' share of the national sales tax (VAT) with a formula that explicitly rewards fiscal effort and favors small municipalities with low fiscal capacity 1987-91 Decree 77 of 1987 assigns to local governments responsibility over tertiary or local roads. FNCV is transformed into a cofinancing agency which must secure the contribution of local and regional governments in the form of budgetary transfers, labor, equipment/materials, or scrvices (cngineering dcsigns, supervision). FNCV cstablishcs a cofinancing formula in 1988 and modifies the investment selection procedures to include local/regional pre-screening and evaluation of candidate projects (final approval remains within FNCV). .992-95 The new Constitution of 1991 mandates the modemization of the state and further decentralization. Decree 2171 of 1992 restructurcs MOPT into a Ministry of Transport (MOT), creates the National Institute of Highways (INVIAS) and slates FNCV for liquidation by the end of 1995. The Transport Law of 1993 passes on to the departments responsibility over a substantial part of the national road network, including the road network under FNCV. It also establishes a new Road Cofinancing Fund (RCF) under FINDETER to substitute FNCV's cofinancing activities. Decree 1820 of 1993 defines the plan for liquidation of FNCV and closure of its Regional Offices in a three-year pcriod. By mid-1995 (1) FNCV's staff totals about 160 officials from about 1,750 in 1992, (2) only ninc regional offices remain in operation, and (3) about 16,000 km of roads still remain under FNCV, while about 7,000 km are transferred to the departments. ................................................................ .................I..... . . ....................... .... . ................................................................................................................................. 19-to Law I88 of 1995 temporarily suspends the liquidation of FNCV until departments and local date govcrnmcnts devclop capacity to manage thcir road nctworks. The decentralization and corinancing systems are kept, but the transfer of roads from FNCV is halted. Decree 2128 and 2222 "reorganizc" FNCV to cofinance construction and maintenance activities on rural roads not yet transferred to the departments and evaluate the administrative, financial and technical capacity of the departments beforc passing responsibility ovcr the rural roads. 8. The decision to pass FNCV's road network on to the departments was taken on account that most of the municipalities were too weak to receive them and too many (more than one thousand) to efficiently manage the transfer. Though this decision has simplified the transfer, it may also result in an unclear definition of jurisdiction, with local governments not acknowledging their day-to-day responsibility for rural road maintenance. To date, the transfer of all secondary national roads and of about 7,000 km of rural roads has been completed. While endorsing the decentralization process mandated by the Colombian Constitution of 1991, a subsequent law in 1995 halted the liquidation of FNCV, which will continue operating until subnational governments demonstrate they have the capacity to assume full responsibility for the rural road system. 9. Implementing key policies in the road sector. The project was very successful in putting in place the first cofinancing system with the municipios and the administrative decentralization within FNCV. The cofinancing mechanism introduced in 1988-89 substantially increased FNCV's responsiveness to local needs and built "ownership" on the part of the municipios, which had to contribute in cash or kind to defray the cost of the investments3. In particular, FNCV was very successful in ensuring that the poorest 3/ The cofinancing formulae used by FNCV between 1990-95 established that municipalities with population over 500,000 inhabitants should contribute with 30% of the total cost, municipalities with population between 100,000  municipios had access to the cofinancing system, and in managing through its 24 Regional Offices the work programs agreed with each municipio. Throughout project implementation, FNCV together with the benefited municipios selected about 135 road subprojects (55 for road construction and 80 for road improvement) located within the boundaries of about 300 municipalities. These subprojects were chosen following the guidelines set under the national plan for eradication of extreme poverty and the requests submitted by the municipios, and giving priority to completing road subprojects started under the previous Rural Roads Transport Project (Loan 2668-CO). FNCV's involvement contributed to emphasizing the priority of road improvements and maintenance over new construction. 10. The cofinancing process also changed the arrangements for selecting the projects, placing the initiative squarely on the local governments. Many local governments required FNCV to undertake additional works on roads other than those strictly under its mandate in order to agree on the investments to be cofinanced. While this reflected the priorities assigned by local governments to the road networks in their area (independently of jurisdiction) it forced FNCV to disperse its resources over an increasingly expanded road network (from about 23,000 in 1989 to about 37,000 km in 1993), and weakened the institutional framework for road maintenance planning. This was compounded by the fact that, in line with the decentralization agenda, FNCV was an agency neither designed to routinely maintain the entire rural road network nor allowed to grow in order to achieve such goal. 11. Long-term efforts to build up institutional capacity at the local level proved to be more difficult, in part because FNCV's work programs concentrated on those municipios with most prevailing needs, low institutional capacity and, in many cases, conflicting working environments (presence of violence or social unrest). In the very small municipios (typically with populations below 10,000) the mayor becomes a one- man band being in charge of most activities that require certain degree of qualification, imposing certain fragility for the sustainability of the capacity development process. With the assistance from the Advance School for Public Administration (ESAP), FNCV conducted training courses on road maintenance management, cofinancing and economic-environmental screening of road investments, for relevant staff in its Regional Offices and municipal administrations involved in the project. In 1993 when the decision to phase out FNCV was announced, FNCV organized workshops in each region to transfer its know-how (planning, engineering, contracting and supervising projects) to the municipal environment. Three additional elements were brought to the agenda: (1) the updating of the road inventory to support the transfer of roads, (2) the costing of typical road activities (including a fair assessment of the costs of owning and operating equipment fleets, and (3) the dissemination of good practices, taken from initiatives successfully implemented by other comparable municipios (such as arrangements to share equipment among members of an association of municipalities, the expanded use of contract maintenance, and the setting of rural road committees to manage maintenance programs). 12. The transfer of roads under FNCV to the departments mandated by the Transport Law in December 1993, also implied shifting the priorities for the institutional strengthening program, as it required accommodating a new cofinancing system which assigned a (new) critical role to the departments (with whom FNCV had almost no past experience). The new decentralization framework has required, in different ways, that the municipios work together with their departments. Departments have a coordinating, monitoring and assisting role to play. However, with the exception of a few departments, and 500,000 inhabitants should contribute 20%, and municipalities with less than 100,000 inhabitants should contribute 1 to 15% (the actual percentage based on the amount of value-added taxes being transferred by central government); for maintenance works, the minimum cofinancing requirement was 5%.  there is still a significant gap between this proposed role and the observed reality. Furthermore, many municipios have either confrontational or distant relationships with the departmental government. The sustainability of the new cofinancing system requires addressing the institutional weakness of most of the departments and strengthening the relationship with their municipios, in order to position them adequately to play the role of partners and enablers of local efforts. 13. In line with the new functional responsibilities in the sector and the weakened capacity of FNCV to continue with these efforts, after 1994 the Ministry of Transport and FINDETER took over the responsibility for institutional building and for coordinating resource mobilization between agencies at the subnational level. The series of seminars undertaken in 25 of the 32 departments focused on (1) the decentralization of the road network and the national cofinancing system, (2) identification, preparation, evaluation and approval of road subprojects, (3) preparation of transport plans at the departmental level, and (4) preparation and supervision of environmental management plans for road subprojects. Long-term institutional building programs will be furthered in the context of a Departmental Roads Project, which GOC is presently preparing with IDB. Implementing the decentralization of the road sector is taking longer than initially anticipated. 14. The Road Funding Strategies and User Charges Study was completed early in 1995. Though late to contribute to shaping the reforms, nevertheless it made an important contribution by bringing attention to the need to balance responsibilities, institutional capacity and financial resources at subnational levels, a key element initially overlooked by the reforms. The study was included in the project to identify funding alternatives for FNCV and the rural roads subsector, but its scope was extended to review the structure of road user charges for the entire road system to accommodate the institutional changes being introduced by the 1992-1993 reforms. Among other conclusions, the study showed that: (a) cost recovery at the national level exceeded the 1994 expenditures, but would need to be complemented in order to support the expansion and modernization of the network as planned under the National Development Plan for 1995- 1998; (b) while actual expenditures were not known, cost recovery at the subnational levels is about 15% of the projected needs; (c) the structure of road user charges should be revised to better reflect marginal cost pricing (underestimated in the case of heavy vehicles) and a framework for its implementation conveniently delineated in line with the provisions of the Transport Law; and (d) as important (if not more) than the structure of road user charges or the extent of cost recovery, is the interplay between the responsibilities that different levels of governments have regarding the road network, their financing resources and their implementation capabilities. The study highlighted that as a result of the existing imbalances between responsibilities, institutional capacity and financial resources at the departmental and local levels, the RCF will neither address the existing financial gaps at subnational levels nor put in place an adequate framework for road maintenance, in the absence of other explicit counterpart sources managed by the departments as part of a well-established road-user charges system. A second stage of this study is presently underway (funded by Loan 3453-CO) to define a coherent roads funding strategy at each government level and recommend mechanisms to ensure a stable flow of funds for road maintenance. 15. Consolidating institutional advances within FNCV. The progress in consolidating institutional advances within FNCV was remarkable till 1993. FNCV strengthened its Regional Offices with sufficient delegation of authority, qualified technical staff, and resources to assist the local governments involved in the programs --a key objective of the strategy set to foster a fruitful dialogue with local governments. However, the substantial efficiencies consolidated between 1988-1992 were later offset by the mandated liquidation of FNCV. This process entailed retrenching about 1,600 staff, and closing 15 out of the 24 Regional Offices. Though most of FNCV's know-how and procedures were passed on to FINDETER's  RCF in 1993 to promote their application by subnational governments, valuable information was lost during the liquidation process. All in all, the positive experience gained through the deconcentrated FNCV became a model for departments to work together with their own municipalities, but will take time and considerable institutional building efforts until the departments develop that capacity. 16. One major objective in the SAR was to strengthen investment programming and maintenance managernent capability with particular emphasis on training staff of the regional offices and the Maintenance Division, which was established by project approval. The Maintenance Division implemented a comprehensive maintenance management system to program and control maintenance operations, especially those carried out by force account. In spite of productivity gains, FNCV's maintenance capacity never reached the targets envisioned under the project in part due to the constraints imposed by an aging fleet that could not be renewed and by a cofinancing system that gave the municipio the right to choose which roads should be attended. From 1993 and on, the progressive closure of Regional Offices, the retrenchment of equipment operators, and the departure of qualified personnel at both central and regional level, further reduced such capacity and ultimately led to phasing out the force account program. The Maintenance Division and the Environmental Unit, both established under the project, were disbanded by end- 1994 and mid-1995, respectively. 17. Financing a balanced four-year investment and maintenance program. In spite of the disruptions created by the retrenchment program, FNCV substantially accomplished this objective, with the exception of routine maintenance for the reasons explained in para. 18 below. The works, however, were completed one year behind the initial schedule. With project funds, FNCV achieved the following physical targets: (a) construction of about 776 kmn of rural roads (or about 97% of the target planned at appraisal); (b) improvements on about 1053 km of roads (or about 81 % of the target); (c) construction of about 1,134 meters of small bridges (or about 101% of the target); and (d) periodic maintenance on about 1,239 km of roads (or 55% of the initial target). The shortfall in periodic maintenance is compensated by a greater achievement under FNCV's 1988-1993 Investment and Maintenance Plan (which more than doubled the periodic maintenance target). This suggests that local governments favored cofinancing maintenance when done by force account (which required contributions in kind and/or labor) rather than by contract (which required financial contributions). FNCV cofinanced routine maintenance on about 17,890 km of rural roads (or about 69% of the appraisal target) in four years. On average this represents an annual coverage of about 20% of the entire network. 18. Though the Investment and Maintenance Program gave priority to periodic maintenance, several reasons prevented FNCV from achieving the targets planned for routine maintenance: (a) the deteriorated condition of the rural road network called for more periodic maintenance, and local mayors favored periodic maintenance over routine maintenance when cofinancing with FNCV; (b) many municipalities lacked the resources (financial, institutional and human) to carry out routine maintenance on a continuos basis; and (c) the targets under the force account program were progressively reduced in line with the liquidation of FNCV, while delays in setting the RCF prevented other agencies from stepping in and complementing FNCV's force account program. Not only the force account targets were revised down from 1994 and on, but compliance with those targets became increasingly poor (around 40 to 50% of the annual targets) reflecting the deteriorating working environments in most of the Regional Offices. 19. Economic Results. FNCV conducted an ex-post evaluation of a representative sample of 18 road sub-projects, to assess whether the projects achieved the Net Present Values (NPV) and Benefit/Cost Ratios (B/C) estimated at the time they were selected and appraised. The sample comprised construction  and improvement subprojects, located in different regions of the country. The analysis entailed field surveys undertaken between May and July 1996. The outcome of the analysis is shown in Table 9. 20. The economic indicators are satisfactory, though generally lower than those anticipated at the respective appraisals (although 6 subprojects out of 18 show higher returns). The initial investment targets were reduced in some cases due to poor initial estimates, in others because the scope of the works was reduced to match the availability of funds in the approved budget. The results of the retrospective evaluation carried out by FNCV and, in particular, the analysis of the results obtained in the ex-post evaluation, highlight the following conclusions: a) most of the construction subprojects showed increases in the areas cultivated and minor improvements in productivity. The benefits notably exceeded appraisal projections in two subprojects where industrialized fishery substituted traditional crops. Field surveys proved that local peasant's access to new markets and social services also improved; b) other factors accounted for reductions in subproject benefits, namely (1) a drop in local agriculture prices during project implementation due to reductions in Colombian trade protections and in international prices, (2) a severe drought that affected most of the country, and (3) increased rural insecurity in the region. In the sample, the presence of social unrest and violence in the area was acknowledged in 11 of the 18 subprojects (or about 61% of the sample); and c) local government's sense of ownership --as expressed by commitment to cofinance the works and assume future maintenance responsibilities-- varied depending on the purpose served by the road. Local governments favored cofinancing those roads providing access to the municipal capital or connecting the municipal capital with main corridors, with less interest in rural roads reaching small communities and almost none in rural roads serving communities within the municipal boundaries but with no commercial interaction with the municipal capital. Most of the roads in the sample showed that although some maintenance had been done after the construction or improvement works, maintenance was generally neglected. However even those roads in critical condition were being used by local traffic with positive benefits for the living conditions of local peasants. C. MAJOR FACTORS AFFECTING THE PROJECT 21. Several external factors had significant negative impact on project execution. The most critical included: (1) new economic and social policies that shifted priorities from the past focus on rural development and poverty alleviation to the opening up of the economy and modernization of the state, (2) institutional reforms in the sector that brought major changes to the road sector and defined new goals for the decentralization agenda, (3) the attempt to liquidate FNCV, as part of these reforms, that progressively weakened FNCV, (4) delays in implementing the new cofinancing system that added uncertainty to the devolution of roads and left part of the road network unattended, and (5) the introduction of new environmental policies and regulations that increased the number of institutions involved in issuing environmental permits and, in turn, slowed down the execution of several road subprojects. The fact that FNCV could complete the project within this adverse environment demonstrates the high degree of professionalism it had achieved by 1992.  - 8 - 22. New economic and social policies. The new administration that took office in August 1990 adopted new economic policies aimed at opening up the economy to international competition and lowering protection for national industries and services. Government efforts shifted to substantiating the major reforms needed to modernize the state and support that policy, giving less priority than the preceding administration to rural development and poverty eradication. For the rural road subsector, this translated into tighter fiscal policies and substantial cuts in public investments. Delays in approving counterpart, funds significantly slowed down project execution in 1990 and early 1991. 23. Major institutional reforms. To support the policy of opening up the economy, Govermment embarked on a series of reforms to redefine the role of the state, eliminate its presence in areas considered unnecessary or unproductive, and reinforce decentralization by shifting responsibility for public investments to the departments and municipalities. Article 20 of the July 1991 Constitution gave the Executive Branch of the government transitional powers, for a period of 18 months, to modernize the organization of the state and ensure consistency between the organization of public entities and the constitutional reforms. Just three days before the transitional period expired, Government issued a total of 61 decrees that eliminated, restructured, fused, or created various public sector entities. 24. The modernization decrees were written keeping in mind the need to minimize transitional implementation problems and unnecessary changes in projects being executed with external funding. Decree 2171, dated December 30, 1992, (1) transformed MOPT into a Ministry of Transport (MOT) responsible only for planning and policy making, not its execution, (2) eliminated the National Highway Fund and entrusted to a new autonomous agency --the Instituto Nacional de Vias, INVIAS-- management of a well-defined network of import/export corridors and trunk roads, (3) eliminated by end-1995, FNCV and MOPT's field districts (the former in support of the expanded agenda for decentralization and the latter in support of a policy of contracting out all works, including maintenance). The three-year time frame established for liquidation of FNCV was just enough to complete the execution of the project as originally planned; several programs under FNCV not directly funded by the project were severely curtailed (among these was the force account maintenance program). As the actual experience showed, this time frame proved too short to develop the new cofinancing system that was intended to replace FNCV. 25. Implementing the Decree implied transferring the responsibility for FNCV's road network. Because of limitations in the powers entrusted by the Constitution, the Decree could not address this issue. After protracted discussions, Congress passed on December 30, 1993 --one year later-- the Transport Law that redistributed responsibilities in the sector. As noted in para. 7, the responsibility over FNCV's network was passed on to the departments, instead of the municipalities with whom FNCV had been working with. The institutional strengthening programs that aimed at strengthening local governments had to be revised to include the departments, at a stage where FNCV was too weak to react. 26. Liquidation of FNCV. The liquidation of FNCV posed, no doubt, the greatest challenge for completing the project investments on time. The three-year retrenchment program started in 1993 and progressively weakened FNCV. FNCV's staff was regarded as one of the most charismatic groups in the sector, but following the decision to eliminate the agency, qualified and experienced staff departed as better opportunities became available for them in the public sector (new institutions in the road subsector) and the private sector (existing firms and new small firms formed by former MOPT and FNCV staff). From a total of about 1,750 staff in 1992 before the liquidation, the staff was reduced to about 1,250 by December 1993, about 660 by June 1994, about 380 by December 1994, and about 160 by May 1995. Similarly, three Regional Offices were closed as early as end-1993, 12 remained in operation by end-1994, and 9 by  - 9 - June 1996, when the liquidation was reversed. The initial plan had called for a much more gradual approach, with the Regional Offices kept in operation till the end to retain capacity for key services (such as maintenance, liaison with local governments) during the transition to the new set up. With the disbanding of several units within FNCV and its Regional Offices, the supervision arrangements became a bottleneck and its quality deteriorated from 1994 and on. Political instability also shook FNCV and compounded this adverse environment. The turnover in top management posts was very high, with five Directors/Liquidators appointed during the last four years. Also, early in 1994, a top official was removed from his post accused of mismanaging work contracts. 27. Delays in implementing the new cofinancing system. Implementing the new cofinancing system took much longer than anticipated, creating even more uncertainty and, ultimately, leaving a large portion of the rural road network unattended and in rapid deterioration. Both, the cofinancing matrix for apportioning funds among departments and investment programs under the RCF and the necessary budgetary provisions to support them, were not defined till late 1994. The departments were concerned that they would receive the roads but not the funds needed to keep them up, in view of the little progress made in implementing the RCF, and the uncertainties regarding the cofinancing matrix and the level of funding that would become available through the RCF. Similarly, it took more time than planned to make operational the specialized units (UDECOs) created in each department to coordinate preparation of the investment programs proposed for cofinancing. As a result, the devolution of roads turned into time consuming negotiations with the departments, demanding substantial attention and resources from an FNCV already overwhelmed by the difficulties inherent to its liquidation. 28. New environmental policies and procedures. The reforms included the creation of the Ministry of the Environment to replace INDERENA in December 1993. The introduction of this new set up brought more stringent regulations; the Corporaciones Autonomas Regionales (CARs) became responsible for granting environmental clearances, but most of them were too weak to fulfill their duties in a timely and efficient manner. This stalled the execution of several road subprojects and led to delays in obtaining pending environmental permits from the CARs. FNCV's Environmental Unit was overstretched monitoring the implementation of increasingly complex environmental plans. Some road construction subprojects along sensitive areas (high mountain ranges) were reduced in length to mitigate negative environmental impacts or to free funds to improve drainage and slope conditions on existing contiguous sections. The Environmental Unit, composed mostly by consultants, was disbanded in 1995. D. PROJECT SUSTAINABILITY 29. The sustainability of the project investments remains critical since in the long run it relies on (I) developing capacity for maintenance at the level of the departments and local governments, and (2) implementing appropriate financing mechanisms for the departments and local governments to secure a stable flow of funds for this purpose. Though there are issues yet to be addressed, the prospects for improved sustainability are reasonably good since Government has shown its commitment to furthering the decentralization of the road sector and make it work. Paragraphs 51 - 56 briefly describe key actions currently in progress that demonstrate such commitment. 30. In furthering the agenda for decentralization, the following elements are key to the sustainability of the project investrnents:  - 10- a) further clarification of responsibilities between FNCV and RCF, and between departments and local governments. Though now functional responsibilities over national, departmental, and local roads are more clearly assigned, and departments and local governments acknowledge their responsibilities for road maintenance, the transitional arrangements under the cofinancing system may in practice dilute such responsibilities. In the short term, from a budgetary view point, it has been established that FNCV will continue cofinancing works on those roads not yet transferred to the departments, while the RCF will cofinance investments on the roads under direct control of departments and local governments. In the long run, the transfer of roads to the departments must be completed and the departments and their municipios must define together a workable framework to split day-to-day responsibilities, with routine maintenance of all tertiary and rural roads in the hands of local governments. Departments should not be exposed to the kind of "negotiations" that FNCV suffered when cofinancing maintenance with the municipios; b) improved financial mechanisms. The first priority is to implement the tariff concept established under the Transport Law, based on road user-charges managed at the national, departmental and local levels in order to reduce the existing dependence on intergovernmental transfers. This will address some of the pending issues under the cofmancing system (ie., definition of funding levels among the various programs, percentages for the cofinancing matrixes), and redirect the use of matching grants to that of passing back to subnational governments that portion of the road user charges collected at the national level, and of supporting national government priorities.; c) more access to technical and institutional support. With some notable exceptions, most of the departments and local governments not only need to build up capacity for road maintenance but also to ingrain a culture of "continuous" maintenance. As the experience of FNCV shows, many mayors believe that maintenance resources should be invested in other priorities as long as the road is passable. The interventions needed to increase support for maintenance and produce a behavior change must go beyond technical assistance and work, indirectly, by promoting responsible leadership and community participation at the local level and demonstrational projects; and d) an adequate information system for increased accountability. Road construction and maintenance is an area in which municipios have been very active since the beginning of the decentralization process. However, no systematic information is available regarding the coverage of these services or the outcome of such initiatives. The information systems developed by FNCV were not directly applied to monitor municipal programs, in part because they were too structured for the municipio's capacity. At the local level, an effective decentralization starts with creating a "public" to whom become accountable to. This public must know the results accomplished by the municipal authorities, and to judge them appropriately, must be able to compare them with those in other municipios. At the national level, MOT also needs to monitor the progress of those programs executed through departmental and municipal governments to assess the outcome of national transport policies. The national evaluation system mandated by the Constitution provides an excellent opportunity to put the information and monitoring systems to work. E. BANK PERFORMANCE 31. Project identification and preparation. The project followed two previous successful operations (Loans 1966-CO and 2668-CO) in the rural road subsector in Colombia. Project preparation benefited  - 11 - from the previous experience with FNCV, as the same Bank team supervising the then on-going Rural Transport Sector Project (Ln. 2668-CO) was responsible for preparing and appraising the new project. 32. The Bank's involvement through the two past loans was instrumental in sustaining the expansion of the rural road network and building FNCV's project implementation capacity. While continuing its support to FNCV's investment program, this third operation aimed to further develop FNCV's institutional strength, especially for maintenance, which in the past did not receive sufficient attention. Dunng the 80s, FNCV had expanded its rural road network from about 16,300 km in 1979 to about 23,000 in 1989. So the emphasis placed on maintenance under the project became very relevant and helped FNCV shift its attention from expanding the network to building capacity for maintenance and balancing its investment program. A key contribution was to recognize that a major obstacle to developing an adequate framework for maintenance was the lack of definition of responsibilities between local, regional, and national administration levels. FNCV was regarded as the only agency capable of adequately managing rural roads, implementing government policies in the sector, and assisting local authorities in the decentralization scheme introduced in 1988-89. 33. Project appraisal. The appraisal correctly identified the main risks, namely the ability and readiness of local governments to organize and develop technical capability to maintain the local road networks effectively, and the level of funding needed at both local and central levels for this purpose. Both continue to be major issues even today. It was foreseen that FNCV's Regional Offices, strengthened under the project, would be an important vehicle to channel the required technical assistance to the local governments. No consideration or function at all was given to the departments because a possible Departmental Highways Maintenance Project was to follow later to address issues and investment requirements of the secondary network. The radical reforms introduced by the modernzation decrees of 1992 or the Transport Law of 1993 could not be envisioned at the time of appraisal. Environmental considerations were also correctly assessed, but again they were exceeded by the changes that follow the creation of the Ministry of the Environment and the Environmental Law. 34. Ihe targets set for FNCV's force account maintenance program turned out to be overly optimistic. They assumed that FNCV's capacity for routine maintenance would almost double dunrng project execution. However, in reality FNCV was an agency not allowed to grow because of the decentralization trend and inhibited by an aging equipment fleet, strict budgetary constraints, and a cofinancing system that implied mobilizing local governments to contribute to the cost and provide the labor needed for maintenance. 35. In view of the difficulties usually associated with the design and execution of institutional programs to build capacity at the local level, the appraisal should have elaborated on a more detailed implementation plan for this component, including the definition of strategic objectives and key milestones to be achieved in both the Regional Offices and the municipios involved in project implementation. Also, the percentages of Bank financing in the various loan disbursement categories were set in a way that favored civil works (60%) over consultant services and training (45% and 50% respectively). Most of the training activities were funded directly by FNCV with its own funds (later with the reform, other funding sources became available through the RCF and MOT). 36. Supervision. Project supervision utilized about 63 staff weeks throughout the implementation period. Eleven supervision missions visited FNCV. The Back-to-Office reports provide for an easy follow up of the project implementation and the developments under the decentralization agenda that had an  - 12 - impact on the achievement of the project development objectives. Bank missions drew the attention of FNCV, FINDETER and the new Ministry of Transport on key steps to address the long-term sustainability of the reform (paras. 29 - 30 of this report reflect recommendations given by a Bank mission in September 1994). Though acknowledging the difficulties surrounding the implementation of the project, the monitoring indicators for project implementation and development objectives were rated satisfactory from 1990 to mid-1995. In November 1995, in view of the uncertainties regarding the reorganization of FNCV and the future of the decentralization agenda. which rendered critical the sustainability of the project investments, those ratings were downgraded to unsatisfactory. The final ratings were raised to marginally satisfactory in 1996, based on the signals given by GOC that implementation of the decentralization agenda will continue with greater support of technical assistance programs aimed at developing capacity at both departmental and local levels. 37. In spite of having an on-going dialogue on main policies for the modermization of the state, the Bank had little involvement in delineating the modernization decrees issued in December 1992. The decrees affected a large number of projects in the portfolio in several ways (some required changing the borrower or implementing agency, re-appraising the capacity of a new implementing agency, or restructuring the projects). The Bank advised the Government about the risks associated with organizing the new units and cofinancing systems, and getting them to function smoothly within the short transition periods allowed by the decrees. But overall the Bank recognized that in the long run the reform would increase the efficiency in the provision of public services through the key strategies (redefining/reducing the role of the state, outsourcing, decentralization) being pursued. 38. For the rural road sector the reform meant disbanding an efficient organization, FNCV, in a three- year period that was perhaps sufficient to complete the project (if appropriate measures were undertaken to minimize the disruptive impact of the reforms), but too short to develop the capacity of those who were going to replace FNCV. The Bank recommended that FNCV adjust project implementation by: (a) reviewing FNCV's Rural Road Investment Program, giving priority to completing on-going subprojects and not approving new subprojects beyond 1994; (b) monitoring closely project execution through increased reliance on consultants; (c) carrying out a gradual liquidation of FNCV under timetables tailored to the capacity of the departments. This gradual approach envisaged that FNCV should start disbanding its equipment fleet and force account crews, while retaining till the end other technical services such as contract administration, supervision, and asssistance to local governments, provided by the Regional Offices. This was not possible because GOC wanted a more abrupt liquidation to prevent any possibility of reversing the reforms; (d) carrying out the targets for road maintenance (force account) defined under the project through an appropriate arrangement consistent with the phasing out of FNCV. It was envisaged that FNCV would still be able to accomplish these targets in 1993, and about 50% of them in 1994, so this would require contracting out the remaining works to private contractors and/or increasing the participation of local governments in delivering routine maintenance; (e) expanding the scope of the study of the Borrower's financing structure to take into account the new institutional framework introduced by the reform; and (f) a parallel effort to build the department and local governments' capacity to prepare and execute road projects, and establish a sustainable system for road maintenance through a combination of contracted maintenance and increased participation of local governments using labor-intensive methods. 39. The major changes in the institutional set up and the further and swifter advancement of the decentralization agenda could not be foreseen at appraisal. When the modernization decrees were passed late in 1992, the Bank advised the Government about the risks associated with organizing the new units and cofinancing systems, and getting them to function smoothly, within the short transition periods established  - 13- under the decrees. At the moment, the Bank did not deem feasible to restructure the project and accomodate it to the drastic decentralization process as this would have jeopardized the on-going works at a time when most institutional development components were completed. Bank supervision missions discussed the implications of the decentralization agenda for achieving the objectives of the project with high level officials at the Ministry of Transport, the National Fund for Territorial Development (FINDETER) and the National Planning Department, and drew the attention of the sector authorities on the issues that the agenda should address to ensure long-term sustainability. This dialogue bore fruit only towards the end of the project when, in 1995, the Government, acknowledging the difficulties of implementing the decentralization in the continuous absence of local technical, financial and administrative capacity, suspended the liquidation of FNCV, cancelled the transfer of rural roads, and established a cofinancing system that would take into account the capacity of the departments. F. BORROWER PERFORMANCE 40. Preparation The Government's National Rehabilitation and Poverty Alleviation Program had emphasized development and maintenance of rural roads in an effort, consistent with its overall priorities, of expanding exports, improving living conditions in rural areas and creating incentives for employment creation and access to social resources. FNCV's 1988-1993 Road Investment and Maintenance Program was at the center of this strategy, when the project was identified. 41. FNCV carried out the project preparation activities in a timely and efficient manner. Between 1988-89 FNCV introduced the cofinancing system ensuring that local government, particularly those in more need, had access to the system The screening and the economic and environmental evaluation of road construction and improvement subprojects were done using the methodologies previously designed and tested under the Rural Transport Sector Project (Ln. 2668-CO). During project preparation, FNCV correctly stated that some of the project objectives were outside the control of the agency; in line with this, the loan documents assigned to MOPT responsibility for classifying the road network and defining functional responsibilities among the central, regional and local levels. 42. Project implementation. The loan agreement was declared effective six months later than initially planned, once Government complied with the Bank general conditions for effectiveness. Implementation of the project works also suffered delays in 1990 due to shortages in counterpart funds. However, counterpart funds were increased later in 1991 and the pace of project implementation accelerated. 43. FNCV's use of small local contractors had proven cost-effective. While the experience under the project confirms that strategy, the rate of problem contracts arose after 1993 and many road subprojects experienced delays. Several factors contributed to this, most notably: (a) contractors' poor performance, generally derived from the use of inadequate equipment, or the inability to cope with adverse weather conditions. Many contractors were concerned about deploying new equipment units in areas where violence could infringe damage to them or later hinder their redeployment outside the job sites; (b) enviromnental considerations, which in some subprojects led to redesigning the works; and (c) local govemments' late compliance with the cofinancing agreements, especially with regard to making the financial contributions (an issue observed by Contraloria General de la Repiiblica, CGR). Though FNCV' management of these situations was reasonably satisfactory, there was scope for improvement. The quality of work supervision declined from 1994 and on, with the retrenchment of qualified staff and the  - 14 - closure of Regional Offices. The impact of FNCV's liquidation on the execution of the project is briefly discussed in paragraph 26. 44. Though the reforms had been under discussion over 1991-92, it appears that many of the modernization decrees issued in 1992 were prepared at the last minute to comply with the Constitutional deadline, and by a few people in the Presidency and DNP, without extensive involvement of the institutions affected. This resulted in implementation problems during the transition period. Key elements of the reform needed further legislation, because the Constitution had not empowered the Executive Branch to redistribute competencies among the central, regional and local levels of government. From a technical view point, it can be concluded that the reform process would have benefited from a more participative design and a well-structured implementation plan. However, political factors may sometimes require quick, radical changes while there is a favorable opportunity. 45. The progress in developing the new institutions and financial sources that were supposed to take over FNCV's responsibilities was slower than the phasing out of FNCV, to the extent that some congressmen started questioning late in 1994 whether the liquidation of FNCV should proceed as initially planned. Finally in June 1995, Congress approved the National Development and Investment Plan 1995- 1998 (Law 188-95), which suspended the liquidation of FNCV, but it was not until December 1995 that the decree reorganizing FNCV was passed. As the legislation states it, the reorganization of FNCV does not mean that the decentralization policy is being reversed, but rather that FNCV becomes an additional option to provide continuity in the rural roads programs until the weakest departments and local governments demnonstrate sufficient capacity to undertake this responsibility alone. 46. Compliance with loan covenants. Compliance with loan covenants was generally satisfactory, with the exception of the financial and audit covenants. In its audits reports CGR, pointed out several deficiencies in FNCV's accounts and internal control procedures which led to adverse opinions on the financial statements of the agency and in various occasions on the project accounts. The annual audit reports were always late, four to eight months later than the covenant date, giving little chance for FNCV to discuss and clear the deficiencies observed by CGR. The administrative decentralization to the Regional Offices compounded the problem since many of them were lacking staff with appropriate accounting skills. In 1994, FNCV made remarkable efforts to clear those deficiencies and reinforced the internal control and accounting functions in both headquarters and the Regional Offices still in operation, an effort that was acknowledged by CGR. Nevertheless, FNCV's financial statements of December 31, 1995 received again a negative opinion from CGR. G. ASSESSMENT OF OUTCOME 47. Overall rating. The project's outcome can be considered marginally satisfactory. Given the implementation scenario, overall performance was better than could be expected. The majority of the sector policy objectives were achieved, or are expected to be achieved once Government completes the implementation of its decentralization policies. The road classification and inventory completed early in the project became a key instrument during implementation of the reforms, first defining responsibilities among the three levels of government prior to the reform, and later, supporting the transfer of secondary and rural roads to the departments. With the transfer of secondary roads, the national road network is now confined to a well-defined network of import/export corridors and trunk roads, which removes the pressures that in the past led to spreading thin national resources to satisfy regional balances. The Transport Law sets key principles to increase efficiency and effectiveness in road sector management. The study on road-user  - 15 - charges and funding strategies highlighted the existing imbalances between responsibilities, capacity and financial resources, an aspect probably overlooked initially by the reform but that now is gaining support among the sector authorities. Unfortunately the study came too late to implement its recommendations within the project's time frame, but this effort is being pursued with support from the on-going Loan 3453- Co. 48. The results of the long term efforts to build capacity at the local level and mobilize municipios to assume their responsibilities for road maintenance were mixed. As the experience of FNCV shows, it will take time until local governments fully assume their responsibility for road maintenance and the principle that maintenance should be provided on a continuous basis; but that experience also recognizes that municipios take action when they perceive the roads need repair. A study performed by the Bank4on a sample of 16 muncipios concludes that the experience of those municipios suggests that municipal governments can be more reliable service providers than it is sometimes assumed. The study confirms that roads is an area in which municipios have been very active since the beginning of the decentralization process. Furthermore, the evidence collected shows an expansion in the local road networks, and notable emphasis on road maintenance in ten of the sixteen municipios. Though there is scope for improvement, the project made an important contribution in this area. 49. The achievement of the physical targets planned in the SAR was substantial, contributing to expanding and improving the condition of the rural road network. The economic analysis performed upon project completion on a sample of 18 subprojects shows that the investments had positive impact in reducing transportation costs and improving access to markets. The economic indicators are satisfactory, though generally lower than those anticipated at the respective appraisals. Agriculture production in the rural areas suffered with the reduction in trade barriers that came with the opening up of the economy. However, this has increased farmer's awareness about transport cost constraints, which is likely to lead to a better exploitation of the cost savings to be made at the farm-to-local market level. 50. In retrospect, the project's outcome would have been better, had the reform mandated a more gradual liquidation of FNCV or focused on a new role for FNCV, consistent with the new decentralization policies. H. FUTURE OPERATIONS 51. Government is taking actions in three main areas that will eventually contribute to improving project sustainability, namely (1) strengthening the role of MOT, (2) strengthening road sector funding, and (3) re-incorporating the reorganized FNCV into the national cofinancing system. 52. Strengthening the role of the Ministry of Transport. MOT, through its Directorate of Rural Roads, has to play a leading role to coordinate and monitor the implementation of technical assistance programs for subnational governments. MOT will assess whether a department has achieved the technical, administrative and financial capacity to receive the road network still under FNCV, using a criteria that is currently under discussion. The evaluation will be based on the department's performance in managing the secondary roads that were transferred from the national level. With funding provided under the RCF, 4 J"Colombia Local Government Capacity: Beyond Technical Assistance", Country Operations Division I, Country Department HII. Report No. 14085-CO. July 7, 1995  - 16 - MOT is assisting the departments in the preparation of their road plans. GOC is preparing with IDB a new operation aimed at building the institutional capacity of the departments. 53. Strengthening road sector funding. This is being pursued by increasing investment levels, enhancing the structure of road user charges, and making operational the cofinancing system to finance maintenance and rehabilitation of secondary and tertiary roads (with special emphasis on those roads transferred from the national level). Government has substantially increased investments in the road sector, from about 1% of GDP during 1991-94 to about 2.3% for 1995-98, in order to clear the existing road infrastructure backlog and accommodate the transport system to present needs. The Development Plan entails investing about US$7.2 billion, out of which US$1.2 billion would be apportioned to support the devolution of roads to departments. With funding provided under the on-going Loan 3453-CO, Government is undertaking a second-phase of the study on Road User Charges and Funding Strategies to, among other objectives, put in place a better structure, consistent with the principles of the Transport Law and more adequate for road maintenance financing. This would allow redirecting the matching grants under the cofinancing system to target poverty and externalities between departments, provide incentives for reforms or increased outsourcing, and passing back to subnational governments that portion of the road user charges that are collected at the national level. Government is also streamlining the procedures to access the cofmancing system. 54. Reincorporating FNCV into the National Cofinancing System. The reorganized FNCV will continue expanding the rural road network, improving the condition of those rural roads that have not yet been transferred to subnational governments, and building up capacity for road maintenance at the local level in line with the decentralization policy adopted by the Colombian government. The new role emphasizes the latter objective over the executing functions. FNCV will coexist with the departmental cofinancing units (UDECOs); the former will be charged with implementing those subprojects specifically requested by local governments and identified in the central government annual budget, while the latter will be responsible for putting together the departmental road programs that will be submitted for cofinancing to the RCF. 55. The reorganized FNCV will adopt the same cofinancing procedures established under the RCF. Hence, FNCV will contract out all works to private firms, community-based and local organizations, regional entities, phasing out its force account program. It is envisaged that by phasing out the force account program, the Regional Offices will be in a better position to channel technical assistance to departments and municipios and focus on monitoring departments and local governments' compliance with the targets agreed under the rural roads maintenance programs. To support this process, FNCV plans to develop micro-enterprises for routine maintenance, expanding the system so successfully used by MOPT and now INVIAS on national roads. These cooperatives will be charged with maintaining an average of 30 km of roads in the local area, under the supervision of FNCV's Regional Offices. Periodic maintenance activities will be undertaken by private contractors or associations of municipios. 56. The operational plan prepared by FNCV in October 1996 is shown in Appendix B. The plan follows the structure of a logical framework. Key targets sought by the plan are highlighted below: Objective Performnance Indicator Upgrade the condition of the In two years time... rural road network => the condition of about 2,000 km of rural roads is improved; => about 20,000 km of rural roads are annually maintained, with periodic  - 17 - maintenance activities over 4,000 kin; Develop institutional capacity, In three months time... complete the transfer of roads = FNCV is reorganized (including headquarters and Regional Offices) to the departments,. and = the distribution of responsibilities between FNCV, RCF, departments and local implement at the local level the governments is clarified and understood by subnational governments; maintenance program based on In six months time... microenterprises => the institutional building plan for departmental and local governments is agreed with MOT and its implementation begins with funding through FNCV and RCF's budgets; the first 50 pilot microenterprises are established to routinely maintain rural roads networks; => the mechanisms for the transfer of roads to the departments are established; In one year time... > about 200 microenterprises formed to routinely maintain rural roads networks; > about 300 municipios and the departments are involved in the technical assistance program; In two years time... about 900 microenterprises formed to maintain rural roads networks under the supervision of local governments and FNCV's Regional Offices I. KEY LESSONS LEARNED 57. Several lessons can be derived from the implementation of this project and of the major institutional reform that took place between 1992-1993. For their possible application to the design of other projects in Colombia, and perhaps in other countries as well, the lessons discussed below refer to the implementation of institutional reform, the design of institutional building programs to reach local governments, and the establishment of an enabling environment for road maintenance. 58. The experience gained so far in implementing the Colombian institutional reforms confirms that reform processes do not immediately produce all the intended results. However, they open the dialogue to continue the reform and allow permeation of "cultural" change, assuming that commitment to the reform is steadfastly sustained. It is not enough to pass legislation, it is also necessary to internalize strategies and concepts, which entails time, and to lauch a program of law regulations which will evolve in light of the experience gained. The Transport Law states key concepts (tariff for services, public-private partnerships for Departments) but most of them have not been fully implemented because of the weak institutional base, and insufficient efforts to overcome it through institutional building programs. Last minute approval of legislation and the lack of an integral strategy to support the changes --in a country with notable capacity for innovation but with a weaker record in implementation-- leads to few results. Political considerations may force quick, radical changes, even without a sound implementation strategy in place, but the risks of delays and setbacks increase substantially, and may jeopardize the achievement of the objectives of the reform. The dismantling of FNCV started without having in place adequate financing mechanisms and the institutions that would take over its responsibilities. Seen retrospectively, the retrenchment of FNCV's highly committed and experience staff was a loss. It could have been avoided, had the reform provided for a longer transitional period, linked with the progress made in establishing the new arrangements. 59. The transfer of roads from the central government to the departments turned into protracted negotiations over the control of the resources associated with it, instead of focusing on building capacity through some time-bound programs. Many departmental governments expressed concern that the level of  - 18 - funding that would become available through the RCF would not match the expanded responsibilities. The mechanisms, financial and otherwise, to support the transfer should be clearly defined in advance to avoid this risk The financial incentive added later (one time financial contribution per km of road transferred) was an attractive means to promote the transfer but does little to ensure the sustainability of the process. The RCF has to give more support to the weakest departments when assigning resources. There is still a risk that the roads transferred to the weakest departments may reverse to national agencies if these departments do not promptly develop enough capacity to maintain their networks. In the case of rural roads, this risk is more evident due to the higher likelihood of road closures, particularly in mountainous regions. The experience is demonstrating the need to develop financing mechanisms based on user-charges --following the tariff concept established by the Transport Law-- to avoid the current situation in which departments heavily depend on inter-governmental transfers (through the RCF) and support from municipalities to finance their programs. 60. Designing effective institutional programs aimed at building local capacity poses probably one of the greatest challenges for development projects. The quality and sustainability of these programs become even more critical when dealing with rural communities which are institutionally weaker and more difficult to reach. FNCV was very successful in reaching the poorest municipios when it introduced the first cofinancing system in 1989 (before 1989 FNCV had financed the works without requiring any contribution from the beneficiaries). Some key elements of that success include: (a) a clear purpose (ie., training municipios in the procedures to access the cofinancing system) well understood by the participating municipalities. The municipios immediately connected the relevance of the future investments with the benefits that those investments were going to bring about; (b) continuity. The program was designed centrally but carried out in a decentralized manner. FNCV's Regional Offices remained there close to the municipios to work together in identifying candidate subprojects, adopting the cofinancing arrangements, and preparing the necessary studies (the latter a task most generally undertaken by FNCV through consultants); (c) trust. FNCV was an agency with recognized prestige and a proven record for working together with the poorest nuniclpios; and (d) demand-driven. The municipios showed great interest in gaining access to a cofinancing system, which was flexible enough to accommodate local conditions. The requests exceeded FNCV capacity to undertake projects. The program gave the opportunity for FNCV and the municipios involved to learn by doing while motivating participation by showing results. 61. Achieving similar results in long term institutional building programs is more difficult, particularly because some of the above elements may not be evident. Sustainable development of capacity at the local level is possible when there is effective demand by local administrations. The new national cofinancing system may play a role in mobilizing such demand, but is not as flexible to respond to the enornous diversity of regional and local conditions in Colombia, as the previous more limited efforts were. 62. However, the effectiveness and the prospects for sustainability of these programs can be increased by: (a) promoting the dissemination of good practices and use of strategic partnerships. Examples of "best practices" and of solutions adopted by municipios, which are clearly successful but yet perceived as at reach by other municipios, carry strong motivation, thus becoming an effective way of speeding up the development of local capacity. Several municipios of varying size (eg., Manizales, Libano, La Mesa, Ipiales) increased the coverage and efficiency in delivering road maintenance services by creating associations of municipios to share road maintenance equipment, contract out the services or establish twinning arrangements; (b) designing flexible programs, tailored to local or regional demands and provided in a decentralized manner. INVIAS and MOT are launching an institutional program that rather than  - 19 - following a rigid agenda, would finance organization and management consultancies in those departments that request the assistance, with the agenda defined mostly by the benefited department. Though it is too early to assess the outcome of this initiative, it is interesting to note that the strategy was conceived after realizing the little impact of more traditional approaches; and (c) ensuring sustained commitment while seeking progressive results. Though it is generally accepted that building institutional capacity is a long term effort, this is often neglected in practice when the objectives of such programs are set in a manner that is not consistent with the time needed to produce the desired changes. More attention should be devoted to defining an appropriate sequence for the interventions needed (which may go beyond the scope of one project) and setting progressive benchmarks. The use of these benchmarks will help focus the programs on a few prioritized goals, facilitate the assessment of progress and outcome, and increase motivation by showing early results. However, the difficulty of tailoring both, sequence and benchmarks, to the variety of local conditions prevailing in the country (or even a region of the country) cannot be ignored; in particular, the cultural, social, and political factors that usually remain outside the control of a project but have significant impact, call for flexibility in the design. This is an area where more research is needed. 63. The cofinancing arrangements used by FNCV increased ownership and responsiveness to local needs. Through this arrangements FNCV also stressed that the responsibility for road maintenance relied on the local governments. The agreements made this very clear, and specifically those for routine maintenance required that local movements provide the labor needed. On the other hand, the cofinancing system to a certain extent implied accepting the priorities defined by the local governments. With so many roads impassable and severe budget constraints, it is understandable that mayors would give preference to periodic maintenance and rehabilitation of those roads (even if not part of FNCV's road network) over routine maintenance. They will likely step in and undertake corrective works when the trafficability is seen at risk. Under these circumstances, it is difficult to conclude whether such behavior reflects lack of local capacity or rather response to local priorities. 64. The experience of FNCV shows how difficult it is to move municipalities towards a "continuous" maintenance culture, in spite of the increased "ownership" that could be expected from the cofinancing system. The project suggests some elements that may help local governments to move to a culture of maintenance: (a) some previous investment is needed to improve the condition of the roads to a condition in which they are maintainable; (b) the effectiveness of routine maintenance becomes a demonstrated reality to both road users and local authorities only after having delivered the services for a reasonable period of time. Using locally acceptable schemes ensures the program will not be discontinued (the formation of microenterprises and road maintenance cooperatives at the local level, now supported by FNCV, is an effective solution on both technical and social grounds with its potential for building a local lobby for continued maintenance); (c) an affordable mechanism of road user charges secures a stable flow of funds for this purpose. While the use of community-based organizations or microenterprises is an affordable solution for the municipios, their sustainability requires prompt payment; (d) those with a vested interest in good roads evaluate the results and exercise their voice before the local authorities. Local participation in screening and identification of subprojects, favored by FNCV, should also be expanded to include decisions regarding maintenance. In many of the municipios that worked with FNCV, PNR (through the Juntas de Accion Comunal) played an important positive role in fostering participation in rural areas and among poor communities, organizing formerly disenfranchised groups and helping them voice demands and acquire a space in the local political debate. FNCV implemented several years ago the peon caminero or lengthmen program to maintain road stretches about 4 km long and give local employment. The program failed because local authorities became excessively clientelistic at the time of awarding the petty contracts, thus creating rivalries and distrust. Former FNCV staff believe that the Juntas can play a key role in developing  - 20 - these local initiatives for the most simple works in the remote "veredas" where the sustainability of microenterprises may be more difficult. 65. Aiming at eventually achieving full-fledged transfer of road maintenance responsibilities to local governments, FNCV together with the local administrations involved in the project were required to draw up action programs with time-bound targets. In practice, these action programs were necessarily limited because FNCV had no control over other key elements of the decentralization agenda, in particular the financing mechanisms. Since the primary function of the Regional Offices was to keep the work program, going, particularly the force account program, institutional building was frequently a second priority. This may improve now, with the elimination of the force account program, the emphasis on the provision of technical assistance services under the reorganized FNCV, and a better understanding of the financing mechanisms in the decentralization agenda.  - 21 - PART II. STATISTICAL ANNEXES Table 1: Summary of Assessments A. Achievement of Objectives Substantial Partial Negligible Not Applicable Macroeconomic policies 1' Sector policies l Financial objectives l Institutional development Physical objectives l Poverty reduction l Gender concerns I Other social objectives l Environmental objectives I Public sector management Private sector development I C. Bank Performance Highly satisfactory Satisfactory Deficient Identification Preparation assistance Appraisal I Supervision l D. Borrower performance Highly satisfactory Satisfactory Deficient Preparation l Implementation Covenant compliance Operation (if applicable) E. Assessment of outcome Highly Marginally Highly SaSatisfacry Satisfacto Unsatisfactory Unsatisfactory I  - 22 - Table 2: Related Bank Loans/Credits Loan/Credit Title Purpose Year of Status approal Preceeding Operations Rural Roads Sector l.First Rural Roads Support FNCV's 1981-84 Program for rural March 1981 Completed Project, Ln. 1966-CO roads construction and rehabilitate and improve June 1986 FNCV's managerial performance 2.Rural Transport Sector Support 1982-92 Investment and Maintenance July 1982 Completed Project, Ln. 2668-CO Program and improve FNCV's institutional Dec 1992 efficiency Agriculture Sector 3. Second Integrated Included financing for construction of about June 1982 Completed Rural Development 700 km of rural roads and improvement of 400 June 1990 Project, Ln. 2174-CO km. FNCV carried out the rural roads component for the Integrated Rural Development Fund (DRI) Highways Sector 4.First Highway Sector Upgrade the transport system, adjust sector April 1982 Completed Project, Ln. 2121-CO policies, strengthen MOPT's highway Dec 1988 management capacity 5.Second National Improve MOPT's internal efficiency, implement October 1987 Completed Highways Sector policy reforms in the transport sector and Dec 1993 Project, Ln. 2829-CO upgrade the condition of the highway network Following Operations Highways Sector 6.Third National Roads Reduce cost of transport, improve project March 1992 In progress Sector Project, Ln. preparation capacity within MOPT, and 3453-CO implement a rolling five year investment program Agriculture Sector 7.Rural Development Includes financing for construction of about 800 July 1990 In progress Investment Program, km of roads, improvement of about 516 km, Ln. 3250-CO under the road component for the Integrated Rural Development Fund (DRI) Includes projects in the same sector/subsector as this project. A limit of 10 years is observed when listing preceeding operations  - 23 - Table 3: Project Timetable Steps in project cycle Date planned Date actual Identification January 1988 January1988 ................ . ..... .. .................................... ............ ......................... ....................... ................ ..... Pre paration (preappraisal) May 1988 June 1988 Appraisal .... ,November 1988 Febmary 1989 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.....PP ,,.... ... ........................................... .............. . ................................. Negotiations ... . April 1989 October 1989 Letter of development polcy (if applicable) n.a. n.a P oApr d pscation, June 1989 January 16, 1990 [Sgig....... .... ........ ...... .............. ......................... ..................... .... ............ ~ .1.1 ,,. ,,,,., .,,MachX 990 Effectiveness Jan ay 1990 June 6, 1 990 ffidterm review~~~~~~~~~~~~~~~~~~~~~~. .................. b!4,,,,...........................,,,,,,,,,,,, n . ....................................................... F ~irst '.. rn"c..h"e ...r"ele'..asec . ("if ..a"p ..p"l i c"able") ...................... n.a. n.a Midterm review (ifapplicabl) n.a. n.a Second (and third) tranche release (if applicable) n.a. n.a ......................................... ......... ....... ..... ........ ..................... ........................................... ..........................................-1 Projectcompletion December 31, 1994 December 31, 1995 ~~~~~~~~~~~~.............. ....... .................... .............. .......... ........................................ ......... .... ...................................................... Loan closing December 31, 1995 December 31, 1995 Table 4: Loan Disbursement FY 90 91 92 93 94 95 96 Appraisal Estimate 9,1 16,6 24,3 32,0 39,7 47,4 55,0 Appraisal Revised Estimate 9,1 16,6 24,3 32,00 39,8 54,0 55,0 Actual 0 6,5 13,8 25,8 41,8 50,6 55,0 Actual as % of Estimate 0