Ghana - National Feeder Roads Rehabilitation and Maintenance Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 9823-GH STAFF APPRAISAL REPORT REPUBLIC OF GHANA NATIONAL FEEDER ROADS REHABILITATION AND MAINTENANCE PROJECT NOVEMBER 19, 1991 Infrastructure Operations Division Western Africa Department Africa Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contb ' Vt rW" vw 7 '' " Wit ^ W ' D II ,, I. f,X  CURRENCY AND EQUIVALENT UNITS Currency Unit Cedis C USS1.0 a Cedis 375 1.0 Cedi * USS.0027 WEIGHTS AND MEASURES 1 meter (m) 3.28 feet (ft) 1 hectare (hs) 2.47 acres 1 kilometer (km) m 0.62 mile (mi) 1 metric ton (t) u 2,205 pounds Cib) ABBREVIATIONS AND ACRONYMS ADF - African Development Fund NTADS - Mediun Term Agricultural ADT - Average Daily Traffic Development Strategy AGSAP - AgriculturaL Sector MTC - Ministry of Transport and AdJustment Proaram Communications BHC - Bank for Housing and MTPU - Mechanical Training and Construction Production Unit CRP - Cocoa Rehabilitation Project NFRDP - National Feeder Roads CTC - Central Training Center Development Program DANIDA - Danish International NFRRMP - National Feeder Roads Developnent Agency Rehabilitation and DFR - Department of Feeder Roads Maintenance Project DUR - Department of Urban Roads NGO - Non-Governmental Organization EPC - Enviromnental Protection OPEC - Organization of Petroleun Council Exporting Countries ERP - Economic Recovery Program PANSCAD - Program of Actions to Mitigate ERR - Economic Rate of Return the Social Costs of Adjustment GDP - Gross Domestic Product in Development GHA - Ghana Highway Authority PCR - Project CompLetion Report GIMPA - Ghana Institute of Management PKU - Project Management Unit and Public Administration PNDC - Provisional National Defense GOG - Goverrnent of Ghana Council ICB - International Competitive PPF - Project Preparation Facility Bidding RRMP - Road Rehabilitation and IFAD - International Fund for Maintenance Program Agricultural Development RTPU - Road Training and Production IBRD - International Bank for Unit Reconstruction & Development SAR - Staff Appraisal Report ILO - International Labour SDR - Special Drawing Rights Organization SIAN - Spot Inprovement and INT - Intermediate Means of Maintenance Units Transport ShC - Single Man Contractors LCB - Local Competitive Bidding SOE - Statement of Expenditure LOI - Letter of Invitation TOR - Terms of Reference MA - Ministry of Agriculture TRP-1 - First Transport Rehabilitation MFEP - Ministry of Finance and Project Economic Planning TRP-2 - Second Transport MLG - Ministry of 'acal Goverrunent Rehabilitatirn Project MREMAT - Mobile Repair and UNDP - United Nations Development Maintenance Units Progra_ MPBS M Maintenance Performance and UNV - United Nations Volunteers Budgeting System USAID - United States Agency for NRH - Ministry of Roads and International Development Highways WFP - World Food Programme FISCAL YEAR January 1 - December 31  FOR OFFICIAL USE ONLY REPUBLIC OF GHANA NATIONAL FEEDER ROADS REHABILITATION AND MAINTENANCE PROJECT STAFF APPRAISAL REPORT Table of Contents Page No: CREDIT AND PROJECT SUMMARY ......... ....................... i I. INTRODUCTION . .................................... 1 II. BACKGROUND ....................................... 1 A. The Country and its Macroeconomic Setting .............. 1 B. The Agricultural Sector ........... ............... 2 C. The Transport Sector ............................ 3 D. Government Decentralization Policy and Strategy . ...... ... 5 E. Bank Group Assistance in Agriculture and Roads ... ....... 7 Ell. FEEDER ROAD CONSTRAINTS AND STRATEGY ............... 8 A. Feeder Road Organization. 8 B. Institutional Capacity. 9 C. District Routine Maintenance Organization .... ........... 11 D. Feeder Road Planning and Selection ................... 12 E. Design and Technology Choices ..................... 13 F. Feeder Road Development Strategy ............. ..... 14 IV. THE PROJECT ................................... 16 A. Pro ect Objectives ......... ... 16 B. Pro ect Description .............................. 16 C. Pro ect Cost and Financing .. : ..................... 21 D. Project Design and Implementation ...... ............. 24 E. Status of Project Preparation ....... ................ 27 F. Procurement ................................. 28 G. Disbursements ................................. 30 H. Accounting and Auditing ......... ................. 30 I. Rcporting and Monitoring ........ ................. 31 J. Environmental Assessment ........ ................ 31 V. ECONOMIC EVALUATION .................... 32 VI. AGREEMENTS AND RECOMMENDATION ..... ............. 35 The project was prepaped on the basis of a prm-appraisal mission in January/Fcbnary 1991 and an appraisal mission in uno /luly 1991 by Moss. A. Nickesen (Mission Leader, Transpott Specialist), S. Demissie (Highway Engineer), J. Gaviria (Transpott Specilist), K. Adarkwa (Transport Econonsti, Consultant), E. Connerley and IL Elison (Decentralization/Local Govemment SpeciaIsts, Conultants) and S. McCormick (Environmental Specialist, Consultant). Mr. T. Pankqj (Principal Transpoat Specialist) initiated the project and contributed the rural mobility comnponent. Mess. P. Taylor (Insitutional Specialist, Consultant) ad A. Coleman (ILO, Consultant) asseed the institutionl and labor-based construction components of the project, respectively. Editing was done by Mr. H. Young (Consultant). Mrs. A. Aruaza (Conaultant) provided computing asistance. Secreial work was done by Mrs. U. Raymond. Peer reviewers for the report were Mesrs. 1. Riverson (Transport Economist) and C. Mensckhoff (Principal Transport Specialist). Mesrs. J. Wright and E. Urm are the Division Chief and Director, respectively, of this opeation. 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.  ANNEXES 2-1 Ministty of Roads and Highways, Organization Chart 2-2 Administrative Structure of Decentralization 2-3 District Planning and Composite Budgeting Process 3-1 DFR Headquarters and Regional Offices, Organization Charts 3-2 Condition and Surface Types of Feeder Roads by Region (1991) 3-3 DFR Personnel in Post as of August 1991 3-4 The Road Committee System 3-5 Experience with Voluntary Community Labor 3-6 Road Area Prioritization and Feeder Road Pre-Selection 3-7 Foreign-Exchange Comparison Equipment/Labor-Intensive Works 3-8 Status of Domestic Construction Industry 3-9 DFR Road Work Accomplishments (1983-1990) 3-10 DFR Expenditures and Sources of Funding (1983-1990) 3-11 Feeder Road Development Program (1992-1999) 3-12 Total and Annual Funding Needs for Feeder Road Works (1992-1999) 3-13 Anticipated Financing Plan for Feeder Road Works (1992-1999) 4-1 DFR Equipment Status 4-2 Organization and Management Study 4-3 Maintenance Performance Budgeting System 4-4 Monitoring and Evaluation 4-5 Recommended DFR Organization 4-6 DFR Staffing Requirements 4-7 DFR Staff Training Program 4-8 Domestic Construction Industry Assessment and Development Study 4-9 Rural Mobility and Environment 4-10 Institutional Arrangements for District Routine Maintenance 4-11 District Resource Mobilization Potelitial 4-12 Summary and Detailed Project Cost Tables 4-13 Detailed Financing Plan 4-14 Schedule for Processing Procurement and Implementation 4-15 Project Supervision Plan 4-16 Monitoring Matrix 4-17 Detailed Time-Bound Action Program for DFR Institutional Strengthening 4-18 Detailed Procurement Table 4-19 Disbursement Schedule 4-20 Environmental Assessment 5-1 Economic Analysis 5-2 Listing of Selected Feeder Roads for Phases 1 and 2 6-1 Documents in the Project File IBRD 23004 Transport System, Project Zones, and Conservation Reserves AF4IN November 1991  i REPUBLIC -OF GHANA NATIONAL FEEDER ROADS REHABILITATION AND MAINTENANCE PROJECT CREDIT AND PMROJECT SUMMARY Mliorrower: The Republic of Ghana Beneficiaries: Ministry of Roads and Highways (MRH); Department of Feeder Roads (DFR) redit Amount: SDR 40.5 million (US$55.0 million equivalent) T&=ms: Standard IDA terms with 40 years maturity ProIect Description: The aim of the project is to support economic development in rural areas by removing physical constraints in feeder road infrastructure, reducing rural tramnmtl' costs for goods and passengers, and supporting farm production and marketing. The four year project includes the following main components: (a) a feeder road rehabilitation program to be executed by contract; (b) a feeder road periodic maintenance program to be executed by contract; and establishment of a sustainable maintenance management system; (c) an institutional strengthening and support program for feeder road planning, design, supervision and quality control; (d) a rural mobility and environmental program; and (e) support to DFR to assist in its decentralization. The project will support the goals of the Government's Economic Recovery Program (ERP) and the Medium Term Agricultural Development Strategy (MTADS) and will complement the trunk and cocoa road rehabilitation programs currently underway with IDA financing. PoQject Benefits and Risks: The main project benefit will be a more efficient and cost effective rural transport system that will stimulate agricultural production and marketing, aid rural employment. The principal quantifiable projict benefits are: (a) value added in additional agricultural production; and (b) reduced transport costs to road users. Key non-quantifiable benefits comprise: (a) increased direct employment in the rural economy; (b) improved access to social services through increased rural mobility; (c) improved quality of life especially for women; and (d) environmental improvements in terms of reduced soil erosion potential. There are no major risks associated with the implementation of this project. Timely execution of the physical components of the project will depend largely on early Government approval of procurement of works and selection of design and supervision consultants. Risks concern difficulties in developing a sustainable feeder road maintenance  ii system; constraints in DFR institutional capacity; and increases in cost of feeder road rehabilitation works. These risks will be addressed by obtaining Government assurance for adequate and timely funding of routine and recurrent maintenance; by establishment of a maintenance performance budgeting system to be operated by DFR; by DFR staff training, hiring of sflicient middle level professional and technical staff as well as technical assistance to fill critical positions; and by adoption of cost effective methods of feeder road design and prompt payment to contractors for completed wo,ks. Estimated Proiect Costs Local Foreign Total ---US$ Millions-- A. Feeder Road Proram, 1. Feeder Road Rehabilitation 21.9 17.5 39.4 2. Feeder Road Regravelling 9.8 10.0 19.8 3. Culverts/Spot Improvements 3.6 1.1 4.7 4. Maintenance/Workshop Equipment 0.3 3.5 3.8 5. Design and Supervision 1.3 1.4 2.7 Subtotal 36.9 33.5 70.4 B. Institutional Sunport Program 1. Management Line Positions 0.2 1.0 1.2 2. Technical Line Positions 0.2 1.4 1.6 3. United Nations Volunteers 0.1 0.1 0.2 4. Studies 0.3 0.6 0.9 5. Training 0.3 1.0 1.3 6. Construction Industry Support 0.5 4.1 4.6 7. Rural Mobility and Environment 0.8 1.1 1.9 8. DFR Decentralization Support Q, 6 1.4 Subtotal 3.2 9.9 13.1 Total Project Base Cost 40.1 43.4 83.5 Physical Contingencies 3.8 3.4 7.2 Price Contingencies 5 1 m Lk Total Project Cost 49.0 50.3 99.3 Taxes & Duties , =f6: Total Project Cost (including Taxes) N5106  tii Financing Plan LpS21 Foreign Total -----US$ Millions-- IDA 27.3 27.7 55.0 Government of Ghana (Road Fund and Budget) 8.4 0.0 8.4 USAID 20.0 1.0 21.0 DANIDA 0.0 15.0 15.0 OPEC 0.0 5.0 5.0 Japanese Grant Fund 0,0 li 16 Total SSl Estimated IDA Disbursements, IDA FiscalYear EY92 E FY94 FY5 FY97 ---US$ Millions - Annual 0.5 11.5 14.9 14.3 10.4 3.4 Cumulative 0.5 12.0 26.9 41.2 51.6 55.0 Economic Rate of Return: 44%, based on quantifiable benefits from project components covering 76% of total project costs.  REPUBLIC OF GHANA NATONAL FEEDER ROADS REHABILITATION AND MAINTENANCE PROJECI 1. INTRODUCTION 1.01 Inadequacy of feeder roads poses a serious constraint on agricultural production and marketing in Ghana. In many rural areas, transport costs account for up to 70 percent of marketing costs as a result of the poor state of the feeder road network. Such costs would decrease substantially if the rural road system wtfu improved. Many fertile farming areas have lost access to markets due to poor roads which hinder mobility and often make head-porterage the only alternative for transporting agri.kultural produce and inputs. However, headloading is costly and slow, it constrains labor supply at harvest time, and it contributes to rural poverty. Thus, feeder road bottlenecks arising from prolonged neglect of maintenance constrain Ghana's farm production and marketing, factor mobility, exports and most non-farm activities vital for economic recovery and growth. 1.02 Feeder road rehabilitation and maintenance is a crucial part of Ghana's agricultural development and its economic recovery and growth strategies, to alleviate poverty and ensure food security. A start was made under earlier IDA trunk road and transport projects to upgrade the maintenance of feeder roads and to strengthen the Department of Feeder Roads (DFR), the agency responsible for them. While progress has been made in rehabilitating feeder roads in selected areas and in building up the DFR, the main tasks still lie ahead. The restoration of feeder roads and further strengthening of the DFR will take about a decade of concerted action. The Government of Ghana (GOG) has sought IDA's support in designing and implementing such a program, and the proposed project would finance the first phase of the program; complementing the ongoing transport sector operations under the Transport Rehabilitation (TRP-1 and TRP-2) Projects whose main emphasis is on rehabilitation and periodic maintenance of trunk roads. These projects are needed to rebuild Ghana's road network, and are part of a phased restoration of the transport sector. 11. BACKGROUN12 A. The Country and its Macroeconomic Setting 2.01 Resource Endowment. Ghana is a medium-sized country (239,000 sq km) with a widely scattered population estimated at about 14 million in mid-1991. More than 70 percent live in rural areas. The country's economic potential is great: it has vast timber resources, its agricultural possibilities are considerable and it was once the world's leading cocoa producer. Mineral resources include manganese, bauxite, gold, and diamonds, and petroleum exploration is under way. In the first decade after Independence (1956), Ghana had one of the highest per capita incomes in Africa, and it had one of the best levels of physical infrastructure and human resources in Sub-Saharan Africa. 2.02 Past PeRformanc. In spite of the country's potential, a continuous decline in per capita income after the early 1960s increased absolute poverty and worsened the income distribution. For two decades prior to 1983 the Ghanaian economy was characterized by (i) a bias in favor of industrialization; (ii) a large but inefficient public sector; and (iii) a neglect of agriculture with a decline in agricultural output, even though it had been the country's most important foreign exchange  - 2 - earner. The resulting macroeconomic imbalance led to a widespread system of price and import controls. By 1983, per capita income was half that in 1970, most economic activity was in the parallel market, physical inftastructure had deteriorated substantially, and the bHlance of payment situation was so weakened that the country accumulated large external debt and payments arrears. 2.03 Present Situation. It was against this background that the Government introduced the Economic Recovery Program (ERP) in April 1983 in order to (i) re-establish fiscal and monetary discipline as a means of stabilizing the economy; (ii) restore relative prices in order to improve producer incentives and encourage private sector participation in economic development; and (iii, rebuild the co-itry's social, economic, and physical infrastructure. The ERP was the most extensive of those attempted in Sub-Saharan Africa. Policies in the first three years focused on stabilizing the economy by reducing fiscal deficits, reducing demand pressure through reduced credit, and stabilizing prices. The reforms of the exchange rate system improved incentives for exporters and helped to improve export performance. From 1986, the program moved into the structural adjustment phase which focused on incentive policies, domestic resource mobilization, and public resource management. The Government has also undertaken reforms in specific sectors, particularly agriculture. 2.04 As a result of the fundamental macroeconomic reforms implemented since the launching of the ERP, the economy expanded. The ERP measures led to an annual growth rate in GDP of 5.3 percent between 1984 and 1988, 6 percent in 1989 and 2.7 percent in 1990 in real terms. This is in sharp contrast to average declines of 0.9 per cera per annum between 1975 and 1982, and of 4.6 percent in 1983. The reform program has laid the foundation for long-term growth. The civil service has been rationalized, and the financial sector has been strengthened by the revision of banking rules. Significant resources are being and will continue to be invested in transport infrastructure rehabilitation and maintenance, including trunk and feeder roads, ports, and railways. B. TheAfricuitural Sector 2.05 Sector Characteristics. Agriculture is the mainstay of the Ghanaian ecinomy, contributing more than 75 percent of merchandise exports and about 50 percent of GDP. It employs about 60 percent of the labor force. The sector is predominantly smallholder, traditional and rain-fed. Farming systems have developed over time as adaptations to the major agro-ecological zones in the country. Mixed cropping to minimize risks, and the widespread integration of livestock into the farming systems are frequent. In the forest zone, important tree crops are cocoa, oil palm, coffee and rubber, while inter-cropped mixtures of maize, plantain, cocoyam and cassava are the principal food crops. The transition zone is characterized by mixed or sole cropping of maize, legunes, cocoyam or yam, with tobacco and cotton as the predominant cash crops. In the savannah zone, main food crops are sorghum, maize, cowpeas and yam in the wetter areas, whilst sorghum, millet and cowpeas prevail in the drier north. Cotton and tobacco are also important cash crops in this zone. Rice is grown in seasonally flooded valley bottoms. In general, average yields are low throughout the country. 2.06 Sector Performance. As with other sectors, agriculture was severely hindered by the economic policies of the 1970s and early 1980s. While agricultural growth between 1975 and 1982 declined steadily at about 1.0 percent per annum, and in 1983 decreased even further by 4.6 percent due to a severe drought, the decline was reversed in 1984 as a result of improved incentives brought about by the ERP. The agricultural sector is now adjusting to fundamental changes in macro- economic policy and the elimination of a policy bias against agriculture. The growth rate of agricultural production between 1984 and 1988 averaged about 1.9 percent per annum; in 1989 it was  about 5.9 percent and in 1990 it was 1.5 percent. The recovery was brought about mainly by the performance of the cocoa and forestry sub-sectors. The slow growth in agriculture was largely due to continuing structural impediments such as (i) poor condition of the road infrastructure; (ii) weak marketing systems; and (iii) the poor state of agricultural support services. 2.07 Government Obiectives. GOG's long term objectives for the sector include food self sufficiency and security, higher employment and incomes in rural areas, regional balanced growth and a larger contribution of the sector to GDP, foreign exchange earnings and government revenues. Agriculture will contintue to be the mainstay of the economy for the foreseeable future. The country has the potential to use its agriculture as the basis for sustained economic growth and development, given its immense agricultural resource base. The key to realizing this potential lies in providing an enabling environment for private producers by: (i) continued liberalization of prices; (ii) increasing access for the private sector to markets and agriculturtl inputs; (iii) raising agricultural productivity through a vigorous and responsive agricultural services system; (iv) encouraging more efficient processing, storage and marketing systems; and (v) improving feeder road and other transport infrastructure. 2.08 Sector Strategy. To achieve these objectives, GOG and the Bank have prepared the Medium Term Agricultural Development Strategy (MTADS). This strategy is based on the premise that Ghanaian farmers will respond to suitable incentives to increase production of both food and cash crops. The strategy uses simultaneous implementation of export promotion (tree crops and fisheries) and import substitution (cotton, maize, rice and meat). It is intended to: (i) improve the agricu!tural policy and marketing framework allowing prices to be market based, introducing greater marker efficiency and increasing investment in marketing infrastructure (including feeder roads and drying and storage facilities); (ii) strengthen agricultural sector coordination; (iii) establish a framework for improved allocation of public sector resources; and (iv) improve the focus on poverty alleviation and the environment. In order to overcome the sector specific policy and structural impediments, GOG is presently preparing an Agricultural Sector Adjustment Program (AGSAP) with IDA support. This program is in line with the MTADS and will support it. C. The Tranpoll Sct 2.09 Sector Overview. Key elements of Ghana's transport system are: (i) a network of about 14,400 km of trunk and urban arterial roads and about 21,300 km of feeder roads; (ii) an over- aged road fleet of some 130,000 vehicles; (iii) a 950 km railway system linking Accra, Tema, Kumasi and Takoradi; (iv) deep-water ports at Tema and Takoradi; (v) an inland water transport system on the Volta Lake; (vi) an international airport at Accra and domestic airports at Kumasi and Tamale; and (vii) a national airline providing international and domestic services. Road transport is the dominant mode in the system, and about 90 percent of it is operated by the private sector, with easy market entry and little government regulation of its operations. Ports, lake transport, railways and civil aviation are dominated by the public sector. Though it intervened on a large scale in the transport sector in the past, GOG now clearly favors st.engthening private sector operations, streamlining and reorganizing public transport enterpri.;>, and privatizing many of them. 2.10 Str Qbjectives. Government's primary sector objectives emphasize (i) rehabilitation and maintenance of existing transport infrastructure in a phased program, with a minimal amount of new construction, and (ii) strengthening institutions in the sector. The network of trunk and rural roads is by far the most important part of the transport system. Though recent projects are gradually improving road conditions, particularly for the trunk roads, the overall road  - 4 - network is still in a generally poor sate. It therefore needs the largest share of resources and the longest period to regain capacity.1/ 2.11 Road Administration. The Ministry of Roads and Highways (MRH) and its three agencies, Ghana Highway Authority (GHA), DFR and Department of Urban Roads (DUR), are the organizations which plan and implement road construction, rehabilitation and maintenance. MRH was set up in 1982, as a separate ministry in charge of roads and highways, with a nucleus of technical staff that reports to the Secretary of Roads and Highways. GHA was established in 1974 as an autonomous body with its own Board of Directors appointed by the Government, but the Board has recently been replaced by an Interim Management Committee. GHA has a staff of some 6,500 spread over the central office in Accra, 10 regional offices and 32 road area offices. It also has a central workshop and a number of smal! regional workshops. GHA has 70 engineers, 60 technician engineers as well as some 670 technical and supervising staff. DFR and DUR were established in 1981 and 1983, respectively, and function as civil service agencies under MRH. DFR staff includes 36 engineers, some 40 other professional staff and 330 technical personnel. It maintains small offices in all regions and in 10 road areas. DUR has a staff of about 200, including 10 engineers and offices in the main cities, where it works with th- city councils. Annex 2-1 shows the organization chart of MRH. 2.12 Institutional Achievements and Requirements. The most important achievements in recent transport and road projects include: (i) implementing a phased 10-year program (1988-1997) to rehabilitate and maintain the trunk road network, and start of work to organize feeder roads planning, rehabilitation and maintenance; (ii) strengthening MRH, GHA, DFR and DUR through technical assistance in key areas and through training; (iii) reducing the share of force account works carried out by direct departmental labor; (iv) setting up a Road Fund to ensure the regular flow of funds to the agencies and contractors; (v) starting pilot projects for introduc.ing labor-intensive rural road rehabilitation; and (vi) improving contractor capacity through sub-loans for equipment and spare parts, training, competitive bidding for all works, and ensuring prompt payments to contractors. The main institutional support needs which remain in MRH, GHA, DFR, and DUR are: (i) technical assistance in specialized areas where local engineers are not available; (ii) an enlarged inter-agency training program to improve existing staff skills and to train newly graduated engineers; (iii) incentives such as housing for key officials and food aid through the World Food Program (WFP), which has helped to improve productivity; and (iv) training and advice to domestic contractors and consultants to fuither improve their technical and managerial efficiency. The ongoing First and Second Transport Rehabilitation Projects and the Second Urban Project provide such support to MRH, GHA, DFR and DUR; additional support for DFR will be provided through this project. 2.13 DIomestic Contracting and Consulting Industries. With the increased road rehabilitation and maintenance works financed under ongoing road and transport projects, the domestic contracting and consulting industries are developing well. There are some 145 local road contractors classified by the Government in categories Al to A4 according tc capacity, some 310 small bridge and strictures contractors (classifications Bi to B4), and some 47 labor-intensive local road contractors classified as class C. Previous projects have included substantial provisions for lines 1/ More detailed information on Ghana's transport sector was given in SAR-6921 GH for the First Transport Rehabilitation Project (TRP-1), presented to the Board on May 27, 1988, updated in SAR-8945-GH for the Second Transport Rehabilitation Project (TRP-2), presented to the Board on December 13, 1990. The following presentation concentrates on road sub- sector issues which are of relevance to feeder roads.  - 5 - of credit and technical assistance to strengthen domestic contractors. The economic reforms implemented in the country including foreign exchange 1 .eralization, have enabled contractors to have easy access to foreign exchange for spare parts and equipment. They have also encouraged the return of Ghanaian engineers to Ghana, particularly to consultancy firms which, often with foreign cour. -rparts, are undertaking a substantial share of trunk and feeder road engineering design and construction supervision. There are at present some 12 local consulting firms working on road projects. In addition, a large number of foreign contractors and consultants are active in Ghana's road sub-sector. 2.14 R-oad Sub-Sector Strategy. The Government strategy for road rehabilitation and maintenance recently evolved into two separate branches, one for trunk roads and another for feede roads. A trunk road network stabilization plan supported by the TransDort Rehabilitation Priects (r1RP) was developed in 1987, with the aim of clearing the backlog of rehabilitation and periodic maintenance requirements on paved and gravel roads during the 1988-1997 period. The purpose of the plan is to stabilize the trunk road network condition at an acceptable 70% good and 30% fair or poor ratio, thus almost reversing the ratio prevailing in 1987. On the basis of the stabilization plan, the 1988-1997 requirements for trunk roads total about 11,300 km of rehabilitation and periodic maintenance, and some 1,500 km of reconstruction and upgrading works. 2.15 Feeder road rehabilitation and maintenance became relatively well organized around 1986/87 after DFR became a stronger organization. So far, the feeder road program has concentrated on: (i) rehabilitation of cocoa roads (trunk and feeder roads in cocoa growing areas) under the ongoing Cocoa Rehabilitation Project; (ii) rehabilitation of non-cocoa feeder roads on a selective basis; and (iii) introduction of a successful labor-intensive road rehabilitation program based on light equipment and contractor training and support. IDA has assisted DFR in these activities and in preparing a long-term rural road strategy. To support its MTADS, the Government, with IDA and other donor support, has recently prepared a 7 year National Feeder Roads Development Program (NFRDP) for the mid-1992 to mid-1999 period. This project will be the first 4-year slice of this program, starting in early 1992 and concentrating on institutional strengthening, feeder road rehabilitation and maintenance, and establishment of a sustainable maintenance performance budgeting system involving local governments in the districts and local wmnunities for routine maintenance. D. Government Decentralization Policy and Strategy 2.16 Despite a long and complex history of attempts to decentralize government in Ghana, the Government remains highly centralized in Accra. Decentralization has been an announced, high priority policy objective of the Provisional National Defense Council (PNDC) since 1983. Previous Governments have also sensed the need to improve government performance through a variety of decentralization schemes. In large part, these efforts had failed to achieve their objectives in the past. 2.17 The principal recent accomplishment of the PNDC Government with respect to its intent to decentralize has been the enactment of PNDC Law 207, the 'Local Government Law of 1988." Law 207 deals with virtually every aspect of local government such as Regional Coordinating Councils, District Assemblies and Metropolitan Authorities. However, its primary focus is on district governments, whose powers are greatly enhanced thereby. 2.18 PNDC Law 207 defines some of the key mechanisms by which the Government intends to decentralize its operations. The primary strategy is to devolve the central government's administrative and political authority to local levels (i.e. regions, districts, towns and villages), with  - 6 - the District Assembly being the centerpiece of administrative and political authority. Among the principal functions of the 110 district governments are the following: - Supervision of the field operations of twenty-two central government departments; - Responsibility for overall development of the district; - Preparation and submission of annual development plans and composite budgets to the PNDC (via Regional Coordinating Councils); - Formulation of programs and strategies for mobilizing and utilizing district human and financial resources; - Promotion and support of productive activity and social development; and - Initiation of programs for the development of basic infrastructure, municipal works and services. 2.19 PNDC Law 207 mandates that District Assemblies shall have administrative and financial control of the district-level operations of twenty-two departments of the central government, including the DFR. For all practical purposes the staff of the District Assembly would therefore include not only core officers of district administration, but also all civil service employees of the twenty-two departments. The administrative structure of decentralization contemplated by PNDC Law 207 is described in Annex 2-2. Annex 2-3 contains a summary of the district planning and composite budgeting process. 2.20 To date, no department has transferred budget administration to direct district control. Some departments have begun to place staff under district supervision, particularly those that were already involved in local operations prior to Law 207. For instance, the Department of Community Development has officers in each District Assembly headquarters who are supervised by District Secretaries. However, this department has not yet decentralized its budget. 2.21 Numerous delays in implementing key aspects of Law 207 should not be interpreted as a lack of interest. Implementation is slow precisely because it involves very significant shifts of authority and resources, primarily from the ministries and departments to the districts and District Assemblies. Nevertheless, many officials believe that the commitment to decentralization is irreversible. 2.22 The principal objective sought by the PNDC through decentralization is to place political and professional supervision as close to implementation as possible. The PNDC believes that there is currently little linkage between political leadership and implementation activities. In the process of decentralization it is hoped that the current vertical linkages, which run on departmental and ministerial lines, will be broken down and replaced by a combination of direct vertical linkages to the PNDC and better horizontal linkages across departments at the district level. It is hoped that this will speed up decision-making, while leaving the ministries with policy formulation, project monitoring and evaluation.  - 7 - E. Bank Group Assistance in Agriculture and Roads 2.23 In addition to the support for the ERP through the MTADS and its AGSAP, a number of projects have been financed within the agricultural sector to overcome particular constraints. These include the Second Oil Palm Project (Credit 1498-GH in 1984) and the Cocoa Rehabilitation Proiejt (Credit 1854-GH in 1988, comprising a large cocoa roads component), to provide a broad spectrum of support to the oil palm and cocoa sub-sectors; the Agricultural Services Rehabilitation Project (Credit 1801-GH in 1987), to develop a range of agricultural services; the Rural Finance Proe (Credit 2040-GH) to improve the efficiency of financial institut;ons; and the Agricultural Diversification Prgiect (Credit 2180-GH in 1990), to promote the production of tree and horticultural crops, and gradually reduce Ghana's over-dependence on cocoa. Support to the forestry sub-sector is being provided through the Forest Resource Management Project (Credit 1976-GH in 1989). A National Agricultural Research Project has recently been approved. The key element in all these projects is the development of the role of the private sector in agricultural production and input supply. 2.24 In the roads sub-sector, apart from an engineering credit for US$1.5 m in 1969 for the preparation of road projects, the Bank Group has provided four highway credits (one credit/loan combined), two transport credits, and two urban credits with road components. The First Highway Uroiec (Credit 438-GH for US$13.0 m in 1974) was completed in early 1981 while the Socond Highway Project (Credit 594-GH for US$10.0 m and Loan 1182-GH for US$18.0 m in 1975) was completed in 1983. The Third Higihway Project (Credit 1029-GH for US$25.0 m in 1980) was fully disbursed by late 1986. These three projects were completed five, four, and three years, respectively, behind the SARs' estimated completion dates. The projects focused on trunk road rehabilitation and maintenance; institution building of GHA; and development of domestic construction and private trading industries through provision of funds for the purchase of equipment and spare parts on an on- lending arrangement through the Bank for Housing and Construction (BHC). 2.25 The main conclusions from experience with these early road projects covering the period from 1974 to 1986 (based on audit completion reports) were that: (a) the objectives were sound but the targets for almost all components were unrealistically high because negative aspects of the economic environment were either discounted or overlooked; (b) targets for institutional building were only partially met mainly because they were over-optimistic compared to the prevailing conditions and Government commitment; (c) the domestic construction industry benefitted through carrying out small civil engineering projects, though some contractors did poorly because works awarded to them were beyond their technical and financial capacities; (d) the difficult economic conditions of the country and implementation deficiencies slowed down progress; and (e) all projects have suffered from inadequate local funding and from shortage of foreign exchange needed to enable the Government to fulfill local financial obligations. 2.26 Since the first three highway projects, the Bank Group has provided funds for the Road RehabilitatiQn and Maintenance Proiect (RRMP) (CR 1601-GH for SDR 40.4 m, with additional CR A-1GH for SDR 10.1 m from the Bank's Facility for Sub-Saharan Africa, and cofinanced by Japanese Grant 1601-IGH for Yen 100.0 m, African Development Bank, UNDP and MFEP in 1985); the First Transport Rehabilitation Proiect (TRP-1) (CR 1858-GH for SDR 46.9 m in 1988); and the Second Transport Rehabilitation Project i RP-2) (CR 2192-GH for SDR 69.0 m in 1991). RRMP was closed on June 30, 1991; TRP-1 funds are 90% committed as of June, 1991; and TRP-2 became effective in May, 1991. These three projects vary in their composition of components as follows: RRMIP consisted of civil works for trunk and feeder roads; training; support to local contractors; and consultant and TA assistance; TRP-1 consists of civil works for trunk and feeder roads; bridges;  -8- railway rehabilitation; support to transport sector institutions; pilot programs for road transport and rural transport; and support for infrastructure planning; and TRP-2 consists of civil works for trunk roads and bridges; support to the MRH; railway rehabilitation; support to transport sector institutions; and a Northern Region pilot scheme. 2.27 RRMP was closed after a one-year extension and the credits were fully disbursed. Physical targets were 70% achieved; policy and institutional objectives have been achieved; the contracting system was reformed; full competitive bidding was established; domestic contractors and consultants' capacity have increased; and the feeder roads labor-intensive pilot program proved to be successful. Implementation of TRP-1, after a slow start, is now progressing well. Accomplishment of trunk road rehabilitation and maintenance works will be less than the SAR target because of substantial increases in cost per km due to further deterioration in the condition of the roads. The increasing backlog in physical works will entail increased deterioration in road conditions requiring more rehabilitation work rather than the periodic resealing and resurfacing works stipulated under GPA's trunk road network stabilization program. Start-up of TRP-2 was about six months behind the agreed implementation schedule, mainly due to the time taken by the Government in approving procurement. The linkages between the lessons learnt from completed and ongoing road projects and the design of the feeder roads project are discussed in paras. 4.30 and 4.31. 2.28 Measures for simplifying the procurement approval process were announced at the Akuse II meeting in May 1991 and confirmed at the infrastructure sector implementation review held in Accra in July 1991. Once these measures are implemented, progress of the TRP projects is expected to improve. The Accra District Rehabilitation Project (Urban-I) (Credit 1564-GH for SDR 22.5 m in 1985) included, inter alia, new construction of the ring road in Accra which has been completed. The Second Urban Project (Urban-2) (Credit 2157-GH for SDR 53.7 m in 1990) became effective in May 1991. This project includes, inter alia, reconstruction and upgrading of selected urban road sections in Accra. HI. FEEDER ROAD CONSTRAINTS AND STRATEGY A. Feeder Road Organization 3.01 Feeder Road Administration. Prior to the establishment of DFR in 1981, feeder roads were the responsibility of various agencies including GHA. At present, feeder road planning, construction, rehabilitation and maintenance are the sole responsibilities of DFR. Annex 3-1 shows the existing organization of DFR headquarters and its regional offices. Since its creation, DFR has been actively supported as the focal point for feeder road development. This support has been through (i) technical assistance and training; (ii) setting up a Road Fund shared with the trunk road program to improve resource availability; (iii) training and supporting labor-based contractors for feeder road rehabilitation; (iv) executing most feeder road rehabilitation works through local contractors; and (v) expanding the role of local consultants for engineering design and works supervision. IDA has supported these efforts through feeder road components in three recent transport and road projects. 3.02 Network Condition and Surface Types. Although existing policies have already resulted in an enhancement of the DFR role, the overall condition of the feeder road network remains poor. At present, only 16 percent (some 3,300 km) of feeder roads are in good condition, constituting DFR's maintainable feeder road network. Roads in this category were rehabilitated  - 9 - between 1983 and August 1991 and are being maintained to the extent present funding levels allow. Of the remainder, 24 percent (5,100 km) of the network is in fair condition, and 60 percent (12,900 kin) is in poor or very poor condition. Only 39 percent (8,300 km) of the network is gravel-surfaced while 60 percent (12,600 km) consists of earth roads. The amount of bitumen-surfaced feeder roads (300 km or I percent) is negligible. Annex 3-2 shows the condition and surface types of Ghana's feeder road network by region as of August 1991. This condition mix and DFR's limited absorptive capacity indicate the need for a concerted effort to step-up physical rehabilitation and establish a sustainable feeder road maintenance performance budgeting svstem. 3.03 Feeder Road Issues. With Bank and other donor support, and with active GOG backing, DFR has become an important agency, but its organization is not yet fully developed. Any large expansion of DFR's activities mandates a close examination of its present capacity and constraints. Feeder road rehabilitation and maintenance on a larger scale will require: (i) building institutional capacity; (ii) establishing a sustainable mechanism for district routine maintenance organizations; (iii) expanding DFR's feeder road planning and selection capacity to integrate future rehabilitation and maintenance activities, in agreement with national, regional and district government policies; (iv) reaching a consensus on design and technology choices; and (v) developing a medium-term feeder road development program including its funding. These issues are dealt with in the following. B. Institutional Capacity 3.04 Institutional Structure and Staffing. Country-wide feeder road works require substantial preparation and supervision. DFR is a young organization which operates at four levels in: (i) the head office in Accra, (ii) 10 regional offices in the regional capitals, (iii) 10 out of 32 road area offices, and (iv) some 30 out of 110 districts. Improvement in some critical areas is essential. While staffing levels for DFR are generally appropriate, out of 67 engineering posts, only 34 (51 percent) are filled. Experienced middle level engineers are in particularly short supply. Only 33 percent of principal engineers' positions are occupied, and 6 percent of senior engineers' posts are filled. In the regional organization, only 42 percent of positions for engineers are occupied. The need for engineers must be met if DFR is to fulfill its role as the focal point for feeder road development in Ghana. Pay conditions do not compare favorably with GHA or the private sector, and civil service policies provide inadequate salary differentials. There is also a lack of trained accountants, and career opportunities for technician engineers and quantity surveyors are poor. Annex 3-3 gives details of DFR's 1,430 personnel at post as of August 1991. 3.05 Organizational Issues. DFR top managers are qualified, experienced, and well motivated. Their enthusiasm does much to engender the good level of morale in the organization. The distribution of headquarters management responsibilities for different operations is sound and efficient. The current division of responsibilities between the maintenance and development divisions appears appropriate. Work loads are allocated flexibly and distributed evenly, and no changes are needed. Recently, more emphasis has already been placed on feeder road planning and selection, and integrating feeder road rehabilitation and maintenance into trunk road and agricultural priorities. A DFR organizational study which was conducted for the preparation of this project recommended the creation of a separate planning division. Also, a maintenance performance budgeting system will be developed under the project to ensure that every feeder road once it has been rehabilitated is followed up by systematic maintenance. The management structure of the three-level regional organization (regional, road area, and district offices) was considered adequate by the organizational study. However, DFR needs to achieve staffing levels, especially in terms of engineering capacity, that will  - 10 - enable it to design and supervise all of its labor-based works while it will continue to rely on consultants for the design and supervision of all other contract works. 3.06 Regional Operation. At the regional level, DFR engineers serve as chief advisors to the Regional Secretaries, and coordinate the distribution of resources among road area offices. A regional engineer is backed by one additional road engineer, and five regions already have an additional engineer to oversee labor-intensive operations, including Brong Ahafo, Ashanti, Western, Eastern and Northern Regions. The regional engineer is also in charge of distributing the services of the DFR-supported Mobile Repair and Maintenance (MOREMAT) units among the road area offices in the region. MOREMAT units are force account operations which execute emergency repair and heavy blading works. The programming of MOREMAT operations requires trained engineers as well1 as an equipment maintenance program. Most communities contribute to the operation of MOREMAT units in the form of diesel for graders and trucks, and food and accommodation for equipment operators. 3.07 Field Offices. Each road area office serves between 3 to 4 districts. At present, only 3 area offices are well established, and 7 are partially established. DFR endorses the need for a total of 32 road area offices to coordinate work in the 110 districts, and to support GOG's decentralization policy. Each area office will be in charge of an average of 650 kilometers of feeder roads. Each of the already established area offices has one technician engineer who represents DFR at the district level and is expected to become an advisor to the District Secretary and the District Assembly for prioritizing and costing feeder road recurrent and routine maintenance activities. Once the MOREMAT units are transferred from the regional to the area offices, each technician engineer will require training to program the activities of MOREMAT in coordination with the regional engineer. Only three office buildings out of the existing 10 area offices are considered adequate. Three area offices have a vehicle each, and 7 others are expected to receive a vehicle each shortly. A foreman has been assigned to each of 30 district offices to implement routine maintenance. Presently, DFR- supported force account Spot Improvement and Maintenance (SIAM) units are partially operational in 35 to 40 districts only. Road committees have been established in some 36 districts. However, the coordination of these road committees has proven difficult. Annex 34 describes the road committee system while Annex 3-5 summarizes DFR's unsatisfactory experience with voluntary community labor. 3.08 Institutional Strategy. The Bank, in coordination with other donors, needs to continue its support of DFR as the focal point for feeder road policies and operations in Ghana. Interagency coordination with the Ministries of Agriculture (MA) and Local Government (MLG), GHA, the Environmental Protection Council (EPC) and district authorities will also be improved. Expansion of DFR will emphasize a build-up of maintenance capacity. At the Head Office, a planning unit was created two years ago and will be upgraded to a division as a condition of credit effectiveness. The appointment of a deputy director of planning, not later than December 31, 1992 is a requirement for this project. Some 20 new road area offices will be established, starting July 1, 1992, as maintenance activities are expanded. These issues were agreed during negotiations. 3.09 All 110 districts require arrangements for routine maintenance. DFR capacity expansion at that level will follow Government's decentralization objectives by expanding its technical support to the districts and, starting from July 1, 1992, as needed train about 80 additional foremen for the supervision of district feeder road maintenance activities. Ihis was agreed during negotiations. Even though DFR has made impressive progress in training labor-based contractors and its own staff, overall training capacity of DFR needs to be strengthened. Training is needed at all levels, including technician engineers, equipment operators, foremen and labor-based contractors.  - 11 - The provision of permanent facilities and staff at the training school at Koforidua will allow it to perform this function. This school will be strengthened under the proposed project to become the West-African equivalent of the East-African labor-based training school in Kisii, Kenya. C. District Routine Maintenance Organization 3.10 DFR has historically relied on unpaid community labor to accomplish feeder road routine maintenance. It has supplied tools and technical supervision to volunteer laborers mobilized by the multi-tiered road committee system. Technical supervision is provided by DFR foremen based at the districts. Where it has proven difficult to mobilize sufficient and reliable voluntary labor, food aid has sometimes been given to laborers. However, a fully implemented community routine maintenance system exists in no more than ten percent of the districts and has generally failed to provide sustained, timely routine maintenance. There are not enough foremen available and communities in these districts are inconsistent in volunteering the required labor. 3.11 Community supported routine maintenance is, in theory, backed up by DFR force account SIAM units based at the road area level. These units include mobile culvert gangs, backlog routine maintenance teams and towed grader teams. However, SIAM units are in place in only twelve of the thirty-two road areas and are often the primary providers of routine maintenance rather than providing backup to community efforts. Recurrent maintenance will continue to be provided by DFR force account MOREMAT units, with the support of towed grader teams from the SIAM units. Graders for a nation-wide MOREMAT system have been procured, but staffing and training of grader operators lags behind the available equipment. This will be rectified under the proposed project. 3.12 Apart from a number of problems described above, the intended routine and recurrent maintenance system, with the exception of reliance on voluntary community labor, seems to be reasonably well designed and attainable over the intermediate term (5 years). Progress in resolving many of the identified problems is evident. 3.13 The most prominent gaps in the intended routine and recurrent maintenance system have occurred at the district level. These gaps are the result of both a failure to effectively implement the intended system in many districts, and failures of the system to perform successfully where implemented. Therefore, the early stages of this project will include efforts to develop and test alternative institutional arrangements for district-level organization and finance of routine maintenance. 3.14 Because of the apparent variety in district circumstances and the lack of experience with successful routine feeder road maintenance in Ghana, no single "blueprint" for the organization of routine maintenance can be offered. Each district must be taken as somewhat unique until the circumstances of the districts and the requirements of successful routine maintenance are better known. When the circumstances and requirements are better known, it is unlikely that a single, standard model of routine maintenance organization will be adequate. 3.15 Appropriate systems for the organization and financing of sustainable routine feeder road maintenance will be developed as follows: (a) District Assemblies will be responsible for routine and recurrent feeder road maintenance in accordance with the Government's expressed decentralization policy and PNDC Law 207. This project will provide technical assistance to support DFR's  - 12 - decentralization and the strengthening of routine maintenance planning, contracting, supervision and financial management capacities of the involved districts; (b) Central Government will ensure financial sustainability of routine and recurrent maintenance by granting sufficient funds in a timely manner to the 110 districts for a four year period until end 1996, starting early 1993. Grants will cover 100 percent of estimated costs of routine and recurrent maintenance. DFR has already prepared detailed cost estimates comprising all maintainabl feeder roads in each of the 110 districts. Only after a thorough assessment of the resource mobilization potential in each district would Central Government be able to decide to what extent individual districts would be expected to contribute to future feeder road maintenance funding. Until such time full routine and recurrent maintenance funding is expected to come from Central Government budget resources; (c) Specific arrangements for the organization and eventual financing of sustained routine maintenance have been identified as (i) combined rehabilitation and maintenance contracts, (ii) single man contractors, (iii) labor-based contractors, and (iv) contracting with intermediary groups. The relative effectiveness of each of these arrangements will be evaluated in the first two and one-half years of the project in an experimental program in up to eight "pilot' districts. D. Feeder Road Planning and Selection 3.16 Feeder Road Planning. The proportion of feeder road expenditures to total road expenditures was around 17 percent in recent years. This compares favorably with other countries in the region. However, there is no firm mechanism to split expenditures between rehabilitation and maintenance for feeder roads, since planning so far has largely been limited to rehabilitation. In past years, planning for feeder roads was limited to the preparation of annual work programs by DFR, based on urgent requests from field offices and local communities. Planning capacity has been bolstered as part of TRP-1 with the creation of a planning unit, which has focused initially on developing planning methodologies and criteria for feeder road selection and economic analysis. Selection and design methods for labor-based road works have also been introduced. Good cooperation with the planning unit in the Ministry of Agriculture (MA) has been achieved. 3.17 Regional and District Planning. At the regional level, DFR has developed a prioritization mechanism for road areas in close coordination with MA, and has agreed to focus future investments in road areas with high to medium agricultural potential. Within each region, network based planning is beginning to be coordinated with rehabilitation works in the trunk road network to ensure linkage to feeder roads, coordination that until recently was only coincidental. At the district level, DFR has developed and implemented a methodology to pre-select and evaluate teeder roads on a corridor basis based on economic criteria. The methodology which consists of pre-selection and final economic evaluation includes substantial local participation. Local constituencies are involved at the pre-selection stage. DFR has started operationalizing this methodology in 16 road areas which have been selected for this project. Annex 3& describes the road area prioritization and feeder road pre-selection methodologies. 3.18 PlanninggStIa1g. DFR will adopt a multi-tiered planning and programming system allowing the participation of local constituencies in the districts. The existing planning capacity, limited for the most part to route selection, will evolve into a system that allocates resources for feeder road rehabilitation and maintenance at the national, regional, road area and district levels.  - 13 - DFR management will develop targets for an adequate split of resources between rehabilitation and maintenance, ensuring linkage to trunk road rehabilitation and maintenance activities. Given the present condition of the feeder road network, most expenditures are concentrated on rehabilitation. However, in the future, DFR is expected to substantially increase allocations for maintenance. Regional distribution of resources will continue to be linked to the priority areas set by the MTADS. The prioritization of road areas will continue to be reviewed by the planning units in MA and DFR in the context of national priorities. In the short term, DFR is expected to concentrate rehabilitation investments in districts with high agricultural potential. In the long term, DFR's rehabilitation activities will expand to cover other districts with medium agricultural potential. 3.19 Local planning requires involvement of local constituencies for the selection of feeder roads and for increased accountability and community participation in routine maintenance activities. The road selection process will formalize the methodology started with the preparation of this project, including pre-selection and final economic evaluation. A maintenance performance budgeting system to program maintenance activities will be developed and implemented as part of project activities. A consolidated framework for network based planning and budgeting is needed so that maintenance requirements are considered along with those of rehabilitation. Planning for feeder roads will evolve into a system using these methods and criteria as well as involving key constituencies at the center (GHA and MA), and at the regional and district levels (District Assemblies). E. D-esign and Technology Choices 3.20 Design Standard. In the past, the emphasis on the concept of "design speed" and overly wide roadways resulted in unnecessarily high standards and excessive construction and maintenance costs. Serviceability and durability are now receiving greater emphasis in design. The former emphasizes all-weather access as against speed and riding comfort. DFR has already established three feeder road standards with formation width of: (i) 5 m for Average Daily Traffic (ADT) of 10 to 20; (ii) 6 m for ADT of 20 to 50; and (iii) 7 m for ADT of 50 to 100. Gravel layer thickness ranging from 10 to 15 cm for sub-base and surface courses is considered sufficient for feeder roads. The thickness variation from 10 to 15 centimeters will depend on subgrade materials, traffic density and axle loads. Geometric design standards have also been specified for horizontal and vertical aligments and maximum gradient, taking into account the need to minimize earthworks by following as much as practical the existing alignments. Durability is being introduced as a design concept during the training of domestic contractors, thus giving more attention to compaction and road drainage requirements. 3.21 Choice of Work Methods. In the past, the use of capital-intensive methods for rehabilitation and maintenance of roads has sometimes been plagued with frequent equipment breakdowns, delays in procurement of equipment and spare parts, and foreign exchange constraints. In response, GOG and DFR, with assistance from IDA and technical assistance from the United Nations Development Programme (UNDP) and the International Labour Organization (ILO), have successfully developed a labor-based contractor pool. After a pilot project and four subsequent training courses, 47 contractors will be trained by the end of 1991 to conduct labor-based works with a potential output in feeder road rehabilitation of some 1,000 to 1,200 km per year. These labor- based works have proven technicallv and economically viable. In addition, there appears to be sufficient labor available in most rural areas in Ghana. 3.22 The technology has key advantages such as: (i) costs are between 15% to 20% lower than the capital-intensive method; (ii) about 10 times more employment is generated; (iii) foreign exchange costs are reduced; (iv) multiplier effects help the rural economy; and (v) the community is  - 14 - involved. A small contractor using appropriate low-cost equipment for labor-intensive works employs between 150 and 200 workers daily and rehabilitates about 2 km of feeder road per month or 25 km per year. 3.23 Annex 3-7 contains a detailed foreign exchange comparison between the capital- intensive and labor-intensive work methods for a standardized km of feeder road. The analysis was carried out on a work item basis, separately for the four principal feeder road operations DFR undertakes: (i) rehabilitation; (ii) regravelling; (iii) spot improvement and reshaping; and (iv) culverting. The key result of the analysis is that on average, the labor-based technology requires less than 50 percent foreign exchange compared to the capital-intensive work method. 3.24 The continued development of private domestic contractors for road works of acceptable quality requires further assistance to the industry which will be provided under this project. Support and training is needed to increase the number of technically and financially viable domestic contractors. In addition, the performance and sustainability of all types of domestic contractors need to be monitored properly and regularly. Some labor-based contractors may lean towards more equipment-intensive operations while capital-intensive contractors may start to compete for labor- based works. Annex 3-8 contains a summary of the status of Ghana's domestic construction industry including its performance problems. 3.25 Force Account versus Contract Qperations. DFR's objective is to continue executing rehabilitation and periodic maintenance works by contract. Routine maintenance operations on feeder roads by force account are severely limited and are presently executed by SIAM units only. Emergency repairs and heavy blading to restore basic access will be continued using the MOREMAT units which are being expanded with the arrival of 70 new graders. Communities have been encouraged to contribute labor to the SIAM units with mixed results. Mechanisms for involving communities in feeder road routine maintenance need to be further strengthened, which will be done under this project. F. Feeder Road Development Strategy 3.26 Past Physical Achievements. Between 1983 and 1990, some 2,700 km of feeder roads were fully rehabilitated. DFR relied mainly on domestic contractors for these works. Some 1,413 culverts were constructed, reconstructed or extended; some 7,070 km of roads reshaped; and some 2,310 km regravelled. While regravelling was mostly by contract, reshaping and culvert construction were executed by force account (MOREMAT units). Substantial increases in feeder road rehabilitation works have taken place over the last 2 to 3 years. An additional 900 km of feeder road rehabilitation is ongoing and will be completed by the end of 1991. This will bring the total of fully rehabilitated feeder roads, defined as the maintainable network, to 3,600 km by end-1991. Annex 3-2 gives details on DFR's road work accomplishments for the period 1983-1990. 3.27 Past Exnenditure. DFR's sources of funding for the period 1983-1990 are shown in Annex 3-10. Domestic funding sources are the GOG Central Budget and the Road and Cocoa Funds. Central Budget contributions reached on average Cedis 450 million per annum between 1983 and 1990 and were spent on recurrent (reshaping) and capital (rehabilitation, culverts, regravelling) works. Road and Cocoa Fund contributions have increased rapidly and reached about Cedis 800 million and Cedis 2,500 million in 1990, respectively. Regular external financing for feeder roads started in 1989 and is being provided by the United States Agency for International Development (USAID), the International Fund for Agricultural Development (IFAD) and IDA through its Program of Actions to Mitigate the Social Costs of Adjustment in Development (PAMSCAD), and its TRP-1  - 15 - and CRP projects. While total support for feeder road works between 1983 and 1987 varied between US$6.0 and 8.0 million annually (with the exception of 1985 when virtually no funds were available), overall funding increased to US$15.0 million in 1988, decreased to US$13.0 million in 1989, and reached US$19.0 million in 1990. The 1991 funding level (in terms of both domestic and external resources) is estimated at US$20.0 million. 3.28 Propose Development Program. As noted in para. 2.15, DFR, with IDA assistance, has developed a 7 year feeder road development program (mid-1992 to mid-1999). Details of the proposed program are shown in Annex 3-11. Physical targets to be achieved between 1992 and 1999 are: (i) full rehabilitation of 8,400 km of high priority feeder roads at an average 1,200 km per annum; (ii) regravelling of 8,850 km of fully rehabilitated high priority feeder roads; (iii) spot improvement and culverting of 1,500 km of additional lower priority feeder roads; and (iv) keeping the oc rehabilitated and spot improved feeder roads under permanent recurrent (blading) and routine maintenance. 3.29 During the mid-1992 to mid-1996 period (implementation period of this project), a total of 4,800 km of full feeder road rehabilitation will be financed by (i) PAMSCAD (250 km); (ii) IDA, the African Development Fund (ADF), and the Cocoa Fund and GOG Capital Budget under the ongoing Cocoa Rehabilitation Project (2,050 km); and (iii) IDA, USAID, the Danish International Development Agency (DANIDA), the Organization of Petroleum Exporting Countries (OPEC Fund), and GOG Road Fund and Capital Budget under this feeder roads project (2,500 km). During the same period DFR will also carry out under the feeder roads project: (i) regravelling of some 2,850 km of feeder roads which were flXly rehabilitated between 1983 and 1990 and need urgent periodic maintenance, and (ii) spot improvement of some 720 km of lower priority roads. Under the feeder roads development program, adequate recurrent and routine maintenance on all rehabilitated and spot improved feeder roads u.ndr maintenanc (9,120 km by 1996) will be provided through Government funds. 3.30 During the 1996 - 1999 period some additional 3,600 km of feeder roads will be fully rehabilitated, bringing the high priority feeder road network under maintenance to a total of 12,000 km, which is considered as an adequate feeder road network for the country. Also, regravelling on some 6,000 km of high priority feeder roads is intended, together with 780 km of spot improving and culverting, and recurrent and routine maintenance on a total of 13,500 km on feeder roads by end- 1999. 3.31 Development Program Costing. To arrive at DFR's feeder road development and follow-up maintenance funding requirements, a preliminary maintenance performance budgeting system was developed during project appraisal. It was based on DFR and mission experience, and by adapting GHA's maintenance management system to feeder roads. The total and annual funding needs for civil works for the 1992-1999 NFRDP are shown in Annex 3-12. Total costs are estimated at US$311.0 million with a foreign exchange component of US$127.0 million (41 percent). DFR will need substantial external funding for rehabilitation and periodic maintenance; consultancy and technical assistance services for engineering design and works supervision, and institutional development; and funding provision for purchase of materials and equipment for routine and recurrent maintenance. 3.32 The required funding level for recurrent and routine maintenance operations between 1993 and 1999 has been estimated at a total of US$31.5 million. While recurrent maintenance will cost US$19.1 million, roui.:Le maintenance will amount to US$12.4 million. US$12.7 million will be needed for both routine and recurrent maintenance for the 1993 to 1996 period, and US$18.8 million  - 16 - for the 1997 to 1999 period. These cost estimates include 10% physical and 17% price contingencies. DFR has prepared district maps showing maintainable feeder roads, thereby enabling accurate annual estimates of required routine and recurrent maintenance funding needs pga istrict. Routine and recurrent feeder road maintenance funds will have to be provided by Central Government budget allocations and, increasingly in future years by contributions towards sustainable feeder road maintenance from the districts' own financial resources. Road Fund resources are mainly used as counterpart funding for rehabilitation and periodic maintenance operations. 3.33 Funding Stragy. Annex 3-13 contains an anticipated financing plan for the 1992- 1999 civil works program. The lUrgest financial uncertainty relates to the availability of sufficient funds for recurrent and routine maintenance. DFR has until now expected to complement the GOG budget for recurrent expenditures with voluntary contributions from the communities, with unsatisfactory results, since voluntary contributions for recurrent and routine maintenance have been very small in recent years. Sustainable financing sources for recurrent and routine maintenance have to be secured from Central Government and from the individual districts and matched to the proposed rehabilitation and regravelling (periodic maintenance) programs. The recent surge in external financing for capital investments such as rehabilitation and periodic maintenance requires proper and timely coordination among present and future donors. IDA will assist the Government in this effort. IV. THE PROJECT A. Projact Objectives 4.01 The project is consistent with the IDA-supported MTADS and TRP projects and will complement and reinforce the objectives of the ERP. The aim of the project is to assist Ghana's economic recovery by: (a) providing improved feedez r 3ad access to transport agricultural produce and agricultural inputs to and from farms or villages and nearby markets and thereby increasing food and cash crop production; (b) improving mobility and economic opportunity for the rural poor; and (c) improving the institutional capacity of DFR to sustain the feeder road program over time and to ensure the maintenance of the rehabilitated road network. B. Project Description 4.02 The main thrust of the 1992-99 national feeder roads development program is to selectively improve and expand the feeder road network, and to improve maintenance management. An initial four-year time slice has been identified, but may be amended from time to time to ensure that all sub-projects meet agreed selection criteria. Proposed project components are: (a) full rehabilitation of 2,500 km of feeder roads in 16 selected road areas with high and medium agricultural potential;  - 17 - (h) regravelling of 2,850 km of feeder roads which were rehabilitated since 1983 and are now under maintenance; (c) construction of up to 5,000 culverts including spot improvements on some 720 km of selected feeder roads to provide minimum access; (d) road maintenance and workshop equipment and tools and spare parts; (e) consultant services for engineering design and contract supervision for (a), (b) and (c) above and for an organization and management study; a maintenance performance budgeting system development; and a socio-economic impact study; (f) technical assistance to support DFR in contract management, road planning, road maintenance and equipment maintenance/repair operations; (g) overseas training for DFR staff and support for the Koforidua labor-based training school; (h) support to the local contracting industry including assessment of industry performance; provision of training, and light equipment for labor-based contractors; (i) rural mobility and environmental improvements and assistance to NGOs and women; and (3) DFR decentralization support to develop and test a district-based routine maintenance system on feeder roads involving local communities. 4.03 Rehabilitation of Feeder Roads (2,500 kwm. The project will cover the first four year phase (mid-1992 to mid-1996) of the 7 year NFRDP (mid-1992 to mid-1999) and will provide funds for full rehabilitation of 2,500 km of feeder roads, mainly in the 16 selected high and medium agricultural potential road areas (see map). The rehabilitation works will be carried out using capital- intensive contractors for 1,500 km through IDA, OPEC and COG funding, and labor-intensive contractors for 1,000 km through USAID and DANIDA financing, applying ICB and LCB procurement procedures. Out of the 2,500 km of feeder roads, 560 km were selected earlier and are being designed under TRP-1 funding. Selection of these roads took place prior to the adoption of the selection criteria; therefore, some of the road links fall outside the 16 selected high priority road areas. Since all roads proposed for full rehabilitation under the project have ERRs above 10% 2/, those outside the priority road areas are packaged as the phase-i roads of the project. The selection of additional 1,140 km, constituting phase-2 feeder roads, has been completed and engineering design has started under PPF and a Japanese grant and is expected to be completed in the first quarter of 1992. Feeder road selection for the remaining 800 km (phase-3 roads) is presently underway. The Bank will review and approve the engineering designs of the feeder roads. 2/ Investments with ERRs estimated as low as 10% based on quantifiable benefits have been included in the project since large, unquantifiable benefits are typically associated with feeder road development and are likely to raise estimated ERRs substantially.  - 18 - 4.04 Feeder roads in high rainfall areas such as the Western Region exhibit very fast rates of deterioration making frequent and costly rehabilitation and heavy maintenance a necessity. During negotiations, agreement was reached on the tarring of selected feeder roads in high rainfall areas on a pilot basis, provided the roads selected are economically viable. This will enable comparison of performance, durability and maintenance needs of bitumen-surfaced feeder roads with the traditional gravel-surfaced feeder roads. The project contains sufficient funds for tarring some feeder roads on an experimental basis. 4.05 hRegr ling of Feeder Roads Rehabilitated since 1983 (2,85k. A total of 2,700 km of feeder roads has been rehabilitated between 1983 and the end of 1990. This total will increase to 3,600 km by the end of 1991 when on-going rehabilitation works are completed. Rehabilitated feeder roads require regravelling on average every 6 years. In the past, regravelling requirements were not fully met due to deficient maintenance management and financial constraints. As the network of rehabilitated feeder roads expands, the backlog of regravelling needs will increase, leading to more deterioration of rehabilitated roads. Th brefore, an expansion in rehabilitation can be justified only if the follow-up regravelling requirements can be met. 4.06 To reverse the present trend, the project will provide funds from IDA, OPEC, USAID, DANIDA and GOG for regravelling of about 2,850 km of feeder roads during the 1992-96 project implementation period. All regravelling works will be carried out through ICB and LCB contracts. The project will also provide funds from IDA for routine road maintenance equipment, while GOG will provide funds for the execution of routine and recurrent maintenance from its budget. Routine maintenance activities such as grass cutting, ditch and culvert cleaning will mostly be carried out using single man contractors (SMC) for 3 to 5 km road stretches (length-man system of contracting). 4.07 Culverting ana Spot Improvement. The 7 year NFRDP aims at rehabilitating some 8,400 km of priority feeder roads out of the total 21,300 km of feeder roads in Ghana. This would normally imply that low priority areas could not be considered for full rehabilitation of feeder roads. However, to enable increased agricultural production in these areas as early as possible, it was decided to provide interim low standard roads by carrying out culverting and spot improvement works at minimum cost for roads which could not be covered under the 7 years program. The feeder roads for culverting and spot improvement will be selected according to criteria to be agreed by DFR, USAID and IDA. The types of work to be done will be limited to providing minimum access so that agricultural produce can be transported to markets at reasonable cost. Up to 5,000 culverts with spot improvements on some 720 km will be financed by USAID. 4.08 Road Maintenance and Workshop Equipment. Tools and Spare Parts. DFR will continue to carry out by force account (MOREMAT units) most of the pothole filiing, spot patching and recurrent blading operations. DFR will soon have enough motor graders and operators to carry out recurrent maintenance works on a larger scale. The MOREMAT units are making valuable contributions to feeder road improvements and are considered the most efficient way to execute recurrent maintenance. DFR needs limited additional equipment such as tippers for long distance haulage and farm tractor-trailer units for short distance haulage of gravel; 4-wheel drive vehicles and motor cycles for field supervisory staff; workshop and training equipment; road maintenance and workshop tools; spare parts for both existing and new equipment, and mobile radios and office equipment. These items will be financed by IDA. The list of equipment and vehicles to be financed under the project has been developed by DFR and agreed by IDA. Annex 4-1 contains a description of DFR's equipment status.  - 19 - 4.09 CQnS e s The project will provide funds through IDA and a Japanese grant for engineering design (200 m/m) and supervision of contracts (230 m/m) for all rehabilitation and maintenance works. For design work, international firms have been encouraged to form joint ventures with local and foreign firms. A contract management specialist (56 m/m) will support DFR in project procurement and implementation through IDA financing. During negotiations, it was agrcA that appointment of the contract management specialist will be a condition of credit effectiveness. DFR's institutional capacity will also be strengthened through (i) execution of an organization and management study under Japanese grant funding (18 m/m) (AnnexA