MADIA DISCUSSION PAPER 10 8437 ,TMI Smmz mnk (Mumo JUAN GAVIRIA VISHVA BINDLISH UMA LELE MANAGING ______ AGRICULTURAL _____ DEVELOPMENTI_____ IN AF \ FOREWORD The MADIA study and the papers comprising this MADIA Discussion Paper Series are important both for their content and the process of diagnosis and analysis that was used in the conduct of the study. The MADIA research project has been consultative, nonideological, and based on the collection and analysis of a substantial amount of concrete information on specific topics to draw policy lessons; it represents a unique blend of country-oriented analysis with a cross-country perspective. The conclusions of the studies emphasize the fundamental importance of a sound macroeconomic environment for ensuring the broad-based development of agriculture, and at the same time stress the need for achieving several difficult balances: among macroeconomic, sectoral, and location-specific factors that determine the growth of agricultural output; between the development of food and export crops; and between the immediate impact and long-run development of human and institutional capital. The papers also highlight the complementarity of and the need to maintain a balance between the private and public sectors; and further the need to recognize that both price and nonprice incentives are critical to achieving sustainable growth in output. The findings of the MADIA study presented in the papers were discussed at a symposium of senior African and donor policymakers and analysts funded by USAID in June 1989 at Annapolis, Maryland. The participants recommended that donors and African governments should move expeditiously to implement many of the study's valuable lessons. The symposium also concluded that the process used in carrying out the MADIA study must continue if a stronger, more effective consensus among donors and governments is to be achieved on the ways to proceed in resuming broad-based growth in African agriculture. The World Bank is committed to assisting African countries in developing long-term strategies of agricultural development and in translating the MADIA findings into the Bank's operational programs. Stanley Fischer Edward V K. laycox Vice President Development Economics Vice President and Chief Economist Africa Regional Office MADIA DISCUSSION PAPER 10 THE RURAL ROAD QUESTION AND NIGERIA'S AGRICULTURAL DEVELOPMENT JUAN GAVIRIA VISHVA BINDLISH UMA LELE THE WORLD BANK - WASHINGTON, D.C. _U Copyright © 1989 All rights reserved The International Bank for Reconstruction Manufactured in the United States of America and Development/THE WORLD BANK First printing December 1989 1818 H Street. N.W. Washington, D.C. 20433, U.S.A. MADIA Discussion Papers are circulated to encourage discussion and ment. at the address shown in the copyright notice above. The World Bank comment and to communicate the results of the Bank's work quickly to the encourages dissemination of its work and will normally give permission development community, citation and the use of these papers should take promptly and when the reproduction is for noncommercial purposes. with- account of their provisional character. Because of the informality and to out asking a fee. Permission to photocopy portions for classroom use is not present the results of research with the least possible delay, the manuscript required, though notification of such use having been made will be has not been prepared in accordance with the procedures appropriate to appreciated. formal printed texts, and the World Bank accepts no responsibility for The complete backlist of publications from the World Bank is shown in errors. The findings. interpretations, and conclusions expressed in this the annual Index of Publications. which contains an alphabetical title list and paper are entirely those of the authorls) and should not be attributed in any indexes of subjects, authors, and countries and regions. The latest edition is manner to the World Bank, to its affiliated organizations, or to members of available free of charge from the Publications Sales Unit, Department F. The its Board of Executive Directors or the countries they represent. World Bank. 1818 H Street. NWW, Washington, D.C. 20433. U.S.A. or from The material in this publication is copyrighted. Requests for permission Publications. The World Bank, 66. avenue d lena, 751 16 Paris, France. to reproduce portions of it should be sent to Director, Publications Depart- luan Gaviria and Vishva Bindlish are consultants of Agricultural Policy in the Africa Technical Department at the World Bank. Uma Lele is the manager of Agricultural Policy in the Africa Technical Department at the World Bank. Library of Congress Cataloging-in-Publication Data Gaviria, Juan, 1961- The rural road question and Nigeria's agricultural development / Juan Gaviria, Vishva Bindlish, and Uma Lele. (MADIA discussion paper; 10) Includes bibliographical references. 1. Rural roads-Nigeria. 2. Agriculture-Economic aspects- Nigeria. I. Bindlish, Vishva. II. Lele, Uma. III. Title. IV. Series. HE336.R85G38 1989 338.1'09669-dc2O 89-22749 ISBN 0-8213-1326-6 Contents Introduction to the Rural Road Question In Nigeria ..................... 5 Rural Road Infrastructure in a Regional Context ........................ 7 The Role of Rural Transport ........................................ 7 The Economic Context of Rural Transport ............................. 7 The Evaluation of Benefits ......................................... 7 Regional Context ................................................. 9 Assessment of Existing Rural Road Densities .......................... 9 Rural Roads, Population Densities, and Land Availability ............... 11 Rural Infrastructure and Market Integration .......................... 11 Construction and Maintenance of Rural Road Infrastructure In Bank Projects in Nigeria .............................................. 12 Experience with Main Road Network Projects ........................ 12 Institutional Arrangements for Rural Roads ........................... 12 Rural Road Interventions .......................................... 12 Bank's Perception of the Need for Rural Roads ....................... 12 Changing Views at Appraisal on Construction Targets .................. 12 Increased Attention to Maintenance at Appraisal ..................... 14 Construction Achievements are Only Measured in Kilometers ........... 14 Road Specifications ............................................ 14 Targets Against Actual Construction ............................... 14 Traffic Increases ............................................... 16 Indirect Benefits ............................................... 16 Very Poor Maintenance Achievements ............................... 16 On the Issue of Contractors and Force Account ....................... 17 Broad Assessment of Future Needs for Expansion, Rehabilitation, and Maintenance of the Rural Road Network in Nigeria .................. 18 Calculation of Target Road Densities in Nigeria ....................... 18 Food Sector Report ............................................ 18 Bank Appraisal and Completion Reports .......................... 18 Comparison With Existing Densities in Neighboring Cameroon ........ 19 Comparison with India in 1960 ................................... 19 Selection of Rural Road Target Densities for Nigeria ................. 19 Rehabilitation and Maintenance Cost Estimates ...................... 20 Assessment of Investment Needs to Reach Target Densities ..... ....... 21 Comparison of Needs and DFRRI Investments ....................... 22 Institutional Considerations and Sustainability ........................ 23 Notes ...................................................... 24 Bibliography ..................................................... 26 Annex 1. Rural Road Benefits ....................................... 27 Annex 2. Review of the Road Components In ADPs .................... 28 2.1. Review of Views on Construction Targets at Appraisal ..... ......... 28 2.2. Review of Views on Maintenance at Appraisal ..................... 29 2.3. Review of Rural Road Standards ................................ 30 2.4. Traffic Counts ................................................ 31 2.5. Review of Indirect Benefits ..................................... 31 2.6. Review of Maintenance Achievements ........................... 31 Annex 3. Cameroon Rural Road Target Densities ...................... 33 Annex 4. Rural Road Densities In India ............................... 34 Annex Notes ..................................................... 34 3 Abbreviations ADF Agricultural Development Fund ADPs Agricultural Development Projects ADT Average Daily Traffic BSADP Bauchi State Agricultural Development Project DFRRI Directorate of Food, Roads and Rural Infrastructure FACU Federal Agricultural Coordinating Unit FHD Federal Highway Department km2 square kilometers kms. kilometers KNARDA Kano Agricultural Rural Development Authority LGAs Local Government Authorities LGCs Local Government Councils MOW Ministry of Works MSADP Multi-state Agricultural Development Project N Naira NDE National Directorate of Employment N'OOO Thousands of Nairas PCRs Project Completion Reports PPARs Project Performance Audit Reports PPWP Pilot Public Works Programme UNDP/ILO SAR Staff Appraisal Report SARDA Sokoto Agricultural Rural Development Authority SFRP Cameroon Second Feeder Roads Project sq. square vpd vehicles per day This paper has been done as part of a joint effort of the Sub-Saharan Africa Transport Program and MADIA. We would like to acknowledge the detailed comments and assistance during the preparation of a mission in lune 1990 provided by Peter Long. We have also benefited from comments by Asif Faiz, lohn Riverson, A.O. Falusi (Head of FACU), R.D. Balogun (Rural Roads Engineer, FACU), Sudhir Wanmali from IFPRI, and Prof. Dele Olowu (Head of the Public Administration Department, Obafemi Awolowo University, Ife-ife, Nigeria) on an earlier draft. 4 Introduction The Government of Nigeria and the World Bank have long have long been recognized to be of great importance. The noted the poor condition and lack of all-weather roads in recognition of such indirect benefits can be used in future rural areas. Nigeria's Fourth National Plan (198i-85), for rural road planning to address each region's economic example, pointed out that one of the critical problems of potential, and each region's demand for and future growth Nigeria's agricultural sector is the inadequate or nonexist- of agricultural production. ent supporting physical infrastructure such as rural roads, The maintenance of both ADP and non-ADP rural roads is storage and marketing facilities, and water supply. In so poor generally that the problem of inadequate infras- response, the government and the World Bank have tructure remains. Until recently, for example, no ADP had dedicated large portions of rural investments to road succeeded in involving a single Local Government Authority construction and rehabilitation. Between 1976 and 1987, the (LGA) in maintenance of roads. This is due to the continual Agricultural Development Projects (ADPs) constructed or disregard of the need to build effective maintenance rehabilitated 9,300 kilometers of rural roads, at the cost of capacity at the LGA and state level governments. Indeed approximately 300 million Naira.' While most projects have Lele et al. (1989) show that much of the institutional either fulfilled or exceeded the appraisal targets of road capacity built by the Bank has been at the federal level, kilometers to rehabilitate, the maintenance achievements and outside the existing governmental structures. Capacity have fallen short of targets in all projects. Rapid road at the state and local government levels continues to be deterioration has decreased expected project benefits. very weak. Construction and rehabilitation of rural roads is now Since 1986 the federal government has allocated substan- undertaken in three ways: (I) by the Directorate of Food, tial amounts for construction and rehabilitation of rural roads. Roads, and Rural Infrastructure (DFRRI) in the president's However, the very poor achievements of the state directo- office which emphasizes new road construction, (2) by the rates of DFRRI have not matched expectations and have second and third multistate ADPs (legally constituted as prompted major criticisms. This paper concludes by stress- state level entities, financed by state and federal govern- ing the urgent need to search for effective mechanisms ments, and supervised by federal level units),2 which have including reform, technical assistance, and coordination of shifted emphasis from new construction, typical of earlier federal directorates (DFRRI and NDE), state entities (includ- enclave and statewide projects, to rehabilitation and ing the rural road capacity of ADPs), and local government maintenance of existing roads, and (3) a few self-help institutions (ultimately responsible for the administration of projects initiated by local communities.3 The establishment rural roads), to build institutional capacity at all levels but of the DFRRI at such a high level points to the fact that the particularly in LGAs. lack of rural roads is seen as a very important constraint to The paper comprises a discussion of the extent of the agriculture and the response of the government is expected rural roads network and the consequences of its growth on to have a positive impact on rural life. The lack of roads is Nigerian agriculture during the 1970 to 1988 period. Follow- so acute that the road construction output of the ADPs has ing, a review of the experience of construction and mainte- been in many cases the single most appreciated result by nance of rural roads in Bank projects is presented. Further, the affected communities.4 an assessment is made of future needs for rehabilitation This paper shows that insufficient work has been done to and maintenance of the network, with calculations regarding evaluate the indirect benefits of such investments (includ- the resources necessary to develop such a network. Finally, ing but not limited to multiplier effects, increased marketed institutional aspects, sustainability, and reform are surplus, availability of inputs, and market integration) which considered. 5 Rural Road Infrastructure in a Regional Context The Role of Rural 'llansport5 rural areas (Blaikie 1977; Carnemark 1979)." (See Rural Road The great importance attached to feeder roads by the Benefits review-Annex 1.) Government of Nigeria is reflected in the establishment in In this line, ex-post studies in various countries have 1986 of the Directorate of Food, Roads, and Rural Infrastruc- demonstrated the following: ture (DFRRI) in the president's office. The budget alloca- I. Infrastructure affects agricultural production through tions and actual expenditures for the directorate since 1986 prices, diffusion of technology, and the use of inputs have been substantial. As Table I shows, expenditures have (Ahmed and Hossain 1988; Anderson 1982; Binswanger exceeded 10 percent of all capital expenditures net of 1989; Richards 1984). foreign financing.6 As well-intentioned as the program is, 2. The degree of commercialization of agriculture measured the lasting effect of the DFRRI and the ADP approach to as the ratio of marketed agricultural output to total addressing the problem should be examined. agricultural output is inversely related to transport costs to and from regional centers, other things remaining Table 1 constant (Airey 1980). Directorate of Food, Roads, and Rural Infrastructure Budget 3. Agricultural supply of cash crops responds positively to as a Percent of Nigeria's Capital Expenditures Net of a reduction in transportation costs as the net producer Foreign Financing price increases, other things remaining constant (Carne- mark 1984). 1986 1987 1988 4. Agricultural supply increases are larger in areas with a Budget Allocation 10.2 1 8.0 95 timely supply of agricultural inputs and credit available, Budget Outcome 11.2 10.8 NA other things remaining constant. 5. With unequal land distribution, most benefits are Source: EIU 1988; World Bank 1988a. accrued by larger farmers who are able to increase production according to the new modified relative prices of factors of production, which result from changes in The Economic Context of Rural Transport transport costs. Further inequality in asset distribution Basic transport infrastructure and services are needed in will occur but lowest income groups might have greater order to guarantee adequate freight transport of agricultural employment opportunities if production is increased, inputs and outputs and nonagricultural consumption goods. other things remaining constant (Narayana et al. 1988). They are also essential to increase personal mobility of 6. Infrastructural development improves access to institu- rural households (including that of agents who affect the tional credit and contributes positively to shift the information and incentives available to these households) allocation of credit from nonproductive to productive and attract investments to rural areas (for agricultural and activities and increases demand for credit in nonagricul- nonagricultural activities). All increase their demands for tural activities (Ahmed and Hossain 1988; Binswanger transport as economic development takes place. 1989). As in most of Sub-Saharan Africa, Nigeria's public 7. Interregional transport subsidies (e.g., panterritorial investments in rural transport have concentrated on pricing of agricultural products) results in increases in improving road infrastructure, mainly highways and not agricultural supply in distant areas, taxing areas which feeder roads. Few projects have attempted approaches that are closer to the regional markets, and subsidizing focus on increasing the availability of adequate means of distant ones (which might have lower costs than trans- transport or improving the efficiency of institutions.7 At porting food to them)(Ndulu 1979). present, efforts in rural areas focus on the rehabilitation and maintenance of existing rural road networks. These improvements are expected to reduce transport costs The Evaluation of Benefits which will modify the relative returns to labor and land, and How to evaluate the benefits listed above in practice and therefore affect the demand and supply of agricultural how to let these influence the choice of investments has products and consumer goods. been an important problem. For instance, the evaluation of In Nigeria as elsewhere, rural road improvements have rural road benefits has been a major cause of concern and long been associated with improvements in productivity, inquiry in the World Bank. Following the Bank's advice, rural increased communication and administrative control, and roads in many countries have been prescreened using an enhancement of rural life (Ogundana 1973). Numerous social and political factors, with the final selection made on inquiries in other countries have also recurrently docu- the basis of economic criteria (Beenhakker 1983). In this mented, among others, the following effects: increase in line, most ex-ante evaluations of Bank road projects in specialization (Mitchell 1977),8 increase in trading activities recent years have been done at the micro level with an (Moerman 1968),9 promotion of migratory movements (Airey appraisal of consumer or producer surplus, or a combina- 1980; Hegen 1966; Okada 1978),10 improved market opera- tion of both.12 tion (Lele 1968), and promotion of the development of In the first case, producer surplus is estimated as a towns and urban areas maintained with surplus from the function of expected increases in production. This method 7 is used in evaluations of low volume traffic roads (below 100 The ex-ante exercises were supposed to be refined over vehicles per day) when most benefits are expected to be in time. This has not happened in part because the ex-post the form of increased output and increased access to the evaluations have been very deficient. After completion of transportation system. Second, the consumer surplus the ADPs most reports have evaluated program success by method quantifies savings in user costs in the form of the number of kilometers rehabilitated and the average reduced vehicle operating costs, a method used in partic- costs per kilometer, and have not considered indirect ular for roads with traffic volumes over 100 vehicles per day. effects. Regardless of who conducts the exercise, the ex- These two methods, however, do not capture the indirect post evaluation should be complete since it sets a prece- nature of most benefits described in the previous section. dence by which future road projects can improve the This is in part due to the limitations of the methods, and selection of investments. also because "in practice, the data required to carry out the The evaluation of the road component in the ADPs has evaluation would probably require more man-days than been a source of debate between the Bank and the that required to build the roads themselves" (Edmonds Nigerian government.'7 In a 1985 letter, the Head of FACU 1983). addressed the methodology used by the Bank to compute In the early ADPs, no rigorous ex-ante evaluation of roads economic rates of return for the ADPs, noting that ". . .the took place. A recent supervision states, for example, that: infrastructure programme can support itself in terms of . none of the northern ADPs Istatewide ADPsl financial and economic costs and benefits (through pro- employs any system for appraisal of road investment. ducer surplus, road user cost savings and secondary Road programs generally are compiled from agricul- multiplier effects in processing and marketing). The infras- tural and LGA ILocal Government Areal recommenda- tructure programme could therefore be left to stand on its tions with little or no estimation of financial benefit or own and therefore be excluded from the overall cost- ranking.13 benefit analysis." For the recent multistate ADPs, however, a road selection Lele has argued elsewhere that there is a need for exercise required by the Bank for each ADP has been greater coordination between the development of agricul- carried out, with assistance from the Federal Agricultural tural potentiai and that of rural road infrastructure. The Coordinating Unit (FACU) and with a combination of both absence of such coordination in many instances results consumer and producer surplus methods. The cost-benefit from a lack of appreciation of the fundamental importance analysis employed by FACU quantifies a stream of costs that of rural roads in the early stages of agricultural develop- includes rehabilitation costs, construction costs, increases ment, including that: in extension service costs,"4 and increases in commercial a many of the benefits of rural roads tend to be indirect service costs.'5 The stream of benefits include vehicle (resulting from multiplier effects) and are difficult to operating cost savings for nonagricultural traffic, benefits to quantify, although there are frequently substantial farmers from producer surplus, and benefits to truckers direct benefits; from reduced transport cost for the incremental crop - there is relatively little recognition of the importance production and fertilizer tonnage. The ranking of roads is of feeder roads for developing markets. There is a strong done by net present value (with 12 percent discount rate) (albeit undocumented) general belief either that and with benefit cost ratios. For the two latest multistate markets in developing countries are already compet- ADPs (MSADPs) the Bank has required that economic rates itive or that if they are not, that infrastructure does not of return be calculated, and that only those roads with an affect mobility, information, or entry of actors into economic rate of return (ERR) above 15 percent be trading, fundamentally; accepted. - given the centralized nature of developing country A simplified and uniform procedure was developed by governments, employment possibilities provided in FACU using a simple spreadsheet. However, its use has only agricultural services by central or state governments slightly improved the actual selection process over previous have greater political payoff, and are therefore more ADPs in the first MSADP A recent evaluation states, for appealing than the maintenance and construction of example, that: roads by local governments (Lele 1987); In Ithe first] MSADP states, FACU has introduced a e the emphasis has been on the provision of roads only formal appraisal system which is ineffective, incorpo- (rather than on provision of means of transport and rating empirical assumptions and insufficient factual institutional development); while road links do rati 16 enhance the accessibility of locations, they do not data. 16 guarantee personal accessibility and therefore fall The ubiquitous use of assumptions in the calculation of short of the full realization of an investment's benefits benefits has made the whole exercise trivial. For example, (Ikporukpo 1987). in absence of adequate traffic counts, the exercise uses a For example, the Project Performance Audit Report fixed traffic generation rate of one vehicle per 750 families (PPAR) for Gusau, Funtua, and Gombe ADPs points to in the area of influence. Furthermore, incremental produc- evidence that "road and dam construction had consider- tion is estimated using fixed crop elasticities equal for all able benefits beyond their contribution to production. regions. No changes in farm gate prices are used in the Many previously-remote villages were made accessible, with evaluations, even though as the Lafia and Ayangba Project the result that consumer prices were reduced and new Completion Report (PCR) suggested, producers received transport businesses sprung up. Increased use of fertilizer higher prices after the roads were improved. Finally, the and other inputs has come about as a result of their benefits to transporters are calculated as the margin increased availability, made possible by the network of obtained over incremental crop output, fixing the margins feeder roads established by the project." The PPAR for at 30 percent for every crop in all regions. Lafia and Ayangba also establishes the importance of 8 indirect benefits of feeder roads, especially in terms of states (see Table 2). Even the densities of rural populations, leading to higher producer prices and incentives. This which we do not have numbers for, but which are affected evidence is in line with that provided for other countries in directly by feeder roads, are likely to be much higher in the the earlier section.]8 South, despite its higher degree of urbanization. The difficulties encountered in the process of quantifying Second, existing data suggests that, except for Borno and such indirect benefits as well as the deficient maintenance Kano, the Northern states have food surpluses, while the of ADP roads after project completion has prevented more Middle Belt and the South, except for Benue, Plateau, and Bank projects from focusing on the development of rural Bendel, seem to incur deficits. Third, the Middle Belt states transport in Nigeria. The Bank's transportation department have the lowest population densities per hectare of arable has not been engaged in any significant feeder road land compared to the rest of the country, and have a large development, and has concentrated only on the main potential for area expansion. But Middle Belt states have roads. Between 1960 and 1988 the Bank committed eight much higher rural road densities (per hectare and per loans amounting to US$470 million for the highway sector. capita) than the North. The lower level of rural road In contrast, rural road development has largely been densities in the North-in terms of kilometers of rural relegated to the Bank's agricultural staff, which have roads per capita and also per square kilometer as Table 3 allocated approximately US$380 million for the rural road shows-has important implications. It is by far the largest components of ADPs in sixteen states.'9 producer of food among the three regions with high rural population densities, second only to those in the Southern Regional Context region. The bulk of rural transport in Nigeria is carried by motorized vehicles on rural roads, with the possible Assessment of Eisting Rural Road Densities exception of the Southern Coast and the Niger Delta where The actual length of all-weather rural roads is difficult to river transport replaces roads. In all instances, when obtain accurately given that rural roads can be washed away products are transported from the farm to the road or during the first rainy season if no proper maintenance is nearest markets, human porterage is common practice on executed. With this in mind, two types of statistics are foot paths and roads. The use of draught animals is only presented: (I) total existing rural roads regardless of significant in some portions of the Middle Belt and condition, and (2) usable or all-weather rural roads. The Northern states. latter is a better indicator of infrastructure available all-year The importance of rural roads in every region is affected round but is rarely available. The importance of all-weather by the region's ability to produce and consume, as well as roads is particularly relevant in the southern rain forest by the size and distribution of its towns and cities under a zone where it rains much of the year. regional competitive environment. This regional context has The current density of rural roads in Nigeria, including been broadly examined in Lele et al. (1989), which points all-weather and others, is almost equal to that of India in out three salient features of agriculture as it relates to 1951, at a time when India had the same population density demand for rural roads. First, the size of the urban sector that Nigeria has at present (see Table 21. The density in and overall population densities in the Southern states are Nigeria of total feeder roads in 1980 amounted to 83 meters much larger than those in the Middle Belt or Northern per square kilometer compared to 80 meters in India in Table 2 Comparative Road Indicators Countries with Population Densities Similar to Those of Nigeria in 1985 Rural Roads Population Total Main and Rural Densities Density Road Secondary Road Length Meters per (persons Network Network Network (meters/ U$ of GNP Developing Country per sq.km.) (km) (km) (km)a sq.km.) (mt./GNP) Low-income Nepal 120 4,700 3,096 1,604 11 0.63 China 110 915,100 254,300 660,800 69 2.09 Pakistan 123 107,673 38,830 68,843 87 1.98 India (1950) 109 399,942 137,108 262,834 80 4.81 Lower-Middle Income Nigeria 106 128,174 53,209 74,965 83 1.13 Thailand 102 150,000 44,534 105,466 205 2.47 Upper-Middle Income Hungary 108 138,185 24,000 114,185 1227 5.33 Portugal 109 52,031 19,031 33,000 358 1.44 Yugoslavia 90 130,000 48,880 81,120 317 1.51 Rumania 96 73,500 14,700 58,800 247 NA Notes: a Includes all rural roads, regardless of state, accounted for in national books. Sources: Derkota 1980; Kingdom of Thailand 1980; Nairn 1981; Government of India 1981; Idachaba 1981; World Bank 1985. 9 1951. Among all countries with similar population densities, in most cases the existing all-weather network was 10 rural road densities in Nigeria contrast favorably only with percent of the Local Government Area (LGA) road network countries that have lower incomes per capita, and with presented in Table 3 (Idachaba 1980).21 Taking into consid- India in 1950 only if total road lengths are considered. The eration improvements made by the ADPs and for the danger with international comparisons is that road lengths purposes of this paper, the existing all-weather network is have to be related to the proportion of total roads calculated as the sum of the kilometers of road rehabili- maintained regularly (all-weather), Nigeria and India (1950), tated and maintained as part of the ADPs until 1985 for for example, had the same road densities, however, the each state (enclave plus statewide ADPs), and one-tenth of proportion of all-weather roads in India (1950) was at least the remaining network. This gives an all-weather rural road 6 times that of Nigeria.20 In any case, even counting all density of 19 meters per square kilometer, which means roads, the rural road density in Nigeria is by far lower than that as much as 77 percent of the roads are difficult to that in Thailand or other upper-middle income countries. If negotiate during the rainy season (see Table 4). As a one compares the meters of rural road per dollar of GNP, proportion of rural roads, the Northern states are found to the figures for Nigeria and Nepal are lowest among these have 50 percent all-weather roads, while in the Southern countries with similar population densities. states only 13 percent of roads are all-weather. Suffice it to The poor condition of the feeder road system in Nigeria say that this is an undesirable situation, especially given was emphasized in the 1980 report of the joint Government that it rains from 8-12 months of the year in the Southern of Nigeria-World Bank food strategy mission. The report region and parts of the Middle Belt. Densities of all- noted that "whilst the country's primary and secondary weather rural roads even in the South, where they are roads are being gradually improved to acceptable stan- highest, remain at very low levels compared, for example, dards, the feeder road system has suffered from years of with the average for India in 1950 of 72 kilometers of road neglect and represents the most serious constraint to per square kilometer (calculated as 90 percent of the agricultural development in Nigeria today" (ldachaba 1980). density reported in Table 2). A survey of roads conducted in several states found that Table 3 Main Road and Feeder Road Densities in Nigeria 1985 Roads Total Estimated Federal Total LGA Road Densities Land Population and State LGA with respect to Length Density Roadsa Area Total Population (sq-km) ('000) (km) (m/km2) (km) (m/km2) (m/person) (1) (2) (3) (4) (5) (6) (7) Northern States Bauchi 65,500 4,176 2,255 34 3,939 60 0.9 Bornob 119,100 5,149 4,049 34 900 8 0.2 Kaduna 69,400 7,039 2,933 42 1,818 26 0.3 Kano 43,700 9,945 2,818 64 3,989 91 0.4 Sokoto 92,000 7,796 3,315 36 3,084 34 0.4 Subtotal 389,700 34,105 15,370 39 13,729 35 0.4 Middle States Benueb 45,500 4,169 2,612 57 3,685 81 0.9 Gongolab 94,500 4,475 3,925 42 3,236 34 0.7 Kwarab 60,100 2,945 3,019 50 2,972 49 1.0 Nigerb 67,300 2,052 2,360 . 35 7,160 106 3.5 Plateau 55,300 3,481 3,991 72 3,497 63 1.0 Subtotal 322,700 17,122 15,907 49 20,550 64 1.2 Southern States Anambra 17,100 6,178 2,287 134 811 47 0.1 Bendel 38,900 4,228 4,812 124 7,079 182 1.7 Cross-River 27,200 5,974 4,128 152 6,504 239 1.1 Imo 11,500 6,309 2,079 181 2,562 223 0.4 Rivers 17,700 2,954 1,058 60 4,000 226 1.4 Lagos 3,510 2,956 630 179 1,723 491 0.6 Ogun 17,200 2,663 1,181 69 6,438 374 2.4 Ondo 20,000 4,689 3,644 182 3,747 187 0.8 Oyo 36,900 8,947 2,113 57 7,821 212 0.9 Subtotal 190,010 44,898 21,932 115 40,685 214 0.9 Total 902,410 96,125 53,209 59 74,965 83 0.8 Notes: aincludes all Local Govemment Area LGA Roads. bin this case LGA road length from 1985 survey for second MSADP. Sources: (1) Rural Infrastructure Project Field Survey; (3),(4), and (5) Idachaba 1980. 10 Table 4 basis of organic and inorganic measures. This is because of All-Weather Rural Roads in Nigeria 1985 (Km) the nature of the environment. High rainfall in the South Ten- ADP Road All- Meters detracts from the use of organic fertilizer. The maintenance percent Construction Weather per of soil fertility requires vegetable cover; therefore the Region of LGA and Rehabilitation Total square km Middle Belt and Northern regions may become the main sources of increased agricultural production in the future, in Northern States 1,400 5,622 7,022 18 order to meet growing demands in the Southern states Middle Belt States 2,050 3,026 5,076 16 which have a large food deficit. Southern States 4,060 817 4,877 26 In order to cope with this growth, the Northern states Total 7,510 9,465 16,975 19 (which have almost the same all-weather road densities but lower total road densities than the Middle Belt) may concentrate on rehabilitation and moderate extensions of the network, to a different extent determined by the Rural Roads, Population Densities, and Land network in each state.23 The Middle Belt, on the other hand, Availability may have substantial increases in its rural road network rural road densities (all-weather and total rural along with continued maintenance of the existing inventory, Htgher rural rad denltles all-wether ad tota rural if rural population densities increase substantially. How- roads) in the Southern region are consistent with the high ever, there may be important constraints for population population densities in this region. As for the Middle Belt mobility in Nigeria including ethnic constraints, about which and the Northern region, their relative ordering in terms of little is known through documents and studies.74 In addi- rural road densities is inconsistent with their population tion, given that both technological changes and area densities with respect to total rural roads, and consistent expansion are more promising in these two regions, future only with all-weather road densities. Whereas the popula- evaluation of roads should emphasize the analysis of tion density in the Northern region is over 50 percent projected conditions of production. higher than in the Middle Belt, the road density is higher in the Middle Belt by roughly the same margin. This further Rural Infrastructure and Market Integration attests to the low level of social and infrastructural devel- In his study of market integration in the Southern state of opment in the Northern region relative to the other two Ogun, Durohaiye (1988) finds that contrary to what is regions, which was noted by Lele et al. (1989). expected, the markets for cowpeas, gari, maize, and rice are Evaluations of rural transport improvements in the future integrated over space (Durojaiye and Aihonso 1988). The could be improved by taking into consideration the factors data used for twelve towns in the state only finds a affecting the prospects for growth in production and nonintegrated market for yams, explaining most of the price consumption.22 In the Southern region, for example, popu- variation between markets by commodity arbitrage (trans- lation densities are already very high, the poorer quality of port and storage). the soils offers fewer technological solutions, and the region However, an analysis using the same methodology based is expected to account for less than 10 percent of the area on data from the Kaduna State ADP in Lele et al. (1989) cultivated. These factors might point to the fact that future shows that in this Northern state, the markets over space rural transport improvements in this region may consist for food crops25 are not integrated, leaving wide margins mainly of rehabilitation and better maintenance practices. explained by factors other than transport and commodity The large size of markets and heavy traffic in the region arbitrage. Another study with data from the first three mean that maintenance operations should be very cost Northern ADPs shows that markets were not integrated efficient in order to support future developments of across space.26 The above points to the fact that food crop agricultural marketing. Evaluations may emphasize the markets in the North are not integrated. However, it cannot effects of better access and information, lower transport be said that the same is true for markets in other regions. costs, and efficient maintenance arrangements in a region Without attempting to generalize, the above leads one to where rural consumption will grow fastest. state that the Northern states in particular lack the As Lele et al. (1989, chapter 3) point out, the scope for minimum infrastructure to improve the flow of price area expansion in Nigeria is largely confined to the Middle information, and the movement of goods and people, which Belt states, and to a lesser extent to the Northern states. ensure market development. Therefore, as stated earlier, Also, there is a general consensus that soil fertility in the there is still a need for network expansion in addition to Middle Belt and Northern states can be maintained on the rehabilitation of the existing roads. 11 Construction and Maintenance of Rural Road Infrastructure in Bank Projects in Nigeria This section looks at how the questions of road construc- Since the main road network handles large volumes of tion and maintenance have been addressed in Bank traffic, the evaluation of benefits is carried out based only projects, focusing primarily on rural road components of on savings in vehicle operating and maintenance costs, ADPs. It argues that while in the early projects the rather than with producer surplus methods which are emphasis was placed on construction, the most recent appropriate for low volume roads. projects are focusing mainly on rehabilitation and mainte- nance because experience has shown that the local author- Institutional Arrangements for Feeder Roads ities lack the capacity to carry out maintenance. Also the In addition to the ADPs financed by the federal and state implementation of road construction and maintenance in governments, and the World Bank, there are three institu- the earlier ADPs by force account units did not strengthen tions involved directly in the provision and maintenance of the development of local implementation capacity, and feeder roads in Nigeria, the LGAs (10 to 25 in each state), hence, the more recent multistate ADPs have required the DFRRI, and the National Directorate for Employment (NDE). use of contractors "in general" for rehabilitation and Both directorates- NDE and DFRRI-were created in 1986 maintenance that requires the use of mechanical equip- as special task forces by the Office of the President, and ment (i.e., rehabilitation, regravelling, and grading have until recently acted on different fronts. The State operations). Ministry of Works and Transport (MOW) also occasionally While this latter emphasis is essential, and therefore leases equipment to the LGAs for emergency repair work. undisputable, it also reflects the lack of emphasis in the The different levels of intervention pose problems in the Bank to date on the broader and deeper administrative and planning process because of difficult coordination, particu- financial issues at the LGA level in particular, which have a larly with the myriad of institutions involved in agricultural fundamental bearing not simply on maintenance, but on development.29 Still it is possible to differentiate between the expansion of the feeder network which is so critical for interventions by the choice of technology. The equipment- expansion of agricultural sector productivity in Nigeria. intensive technology can be administered at the state level The above problem has been accentuated by the divi- (and federal level programs through the states), while the sion of responsibility for rural road projects within the Bank. LGAs can focus on the more rewarding labor-based meth- The transport division has participated in eight highway- ods which are appropriate for routine maintenance of low related projects, concerning federal and state roads, while volume roads. In any case, the coordination between the the agricultural staff have included rural road comiponents various levels, which is weak at present for rural roads, will in all the ADP projects affecting Local Government Council need to be properly developed to ensure the appropriate (LGA) roads. transfer of technology to the LGAs based on standards recognized by the federal and state level authorities. Since ADPs have been established or are being estab- Experience with Main Road Network Projects lished in all states, every state has an ADP rural road office. The main road network has been subject to a substantial The Bank has been studying alternative scenarios for improvement effort in the last 15 years. However, the transfer of responsibilities, in an attempt to establish unexpectedly high increase in traffic of about 30 percent permanent institutional capacity at the state and local per year,2" as well as the high axle loads and light government authority level for planning, implementation, construction,28 have led to rapid deterioration of portions of and maintenance of rural infrastructure. The proposed the main network. The rehabilitation and strengthening of Nigeria Rural Infrastructure Project, for example, includes a the federal network has thus been the focus of the latest proposal by which the state ADPs will retain overall Bank loans to the highway sector. Early road projects had direction and monitoring, while the state MOW will be in important force account components particularly for main- charge of planning and implementation of rehabilitation, tenance. In the most recent projects, construction and mechanical maintenance, and training programs for the maintenance was done by contract. In these instances, LGAs, while the LGAs will continue having the responsibility routine maintenance was not performed well by contractors for routine maintenance. This setup guarantees that the and the present Highway Sector Loan is including a study selection exercise has the participation of the agricultural comparing routine maintenance by contract and force authorities in the ADPs. account. The main problems with this arrangement are the coordination at the federal level, the implementation at the In general, the physical construction components of state and LGA level, and the impact of creating a new projects have been completed satisfactorily, although institution within a weak MOW With respect to the federal often behind schedule. Institution building compo- level coordination, until now the Federal Department of nents have been less successful. The major problem Rural Development has coordinated ADPs, and gives encountered ... was the recruitment of experts to regular advice. Apart from DFRRI and FACU (which have help develop the Federal Highway Department (FHD) coordinated the road selection exercises for the ADPs), ... I which under the proposed Program will be there is no federal agency for rural roads that provides supplied mostly through local consulting firms (World guidelines for planning and implementation of rehabilita- Bank 1988b). tion and maintenance of rural roads. 12 Rural Road Interventions contrast to the Bank reports cited above which had The paper now examines how the questions of rural road criticized the Government of Nigeria for not paying ade- construction and maintenance have been addressed in the quate attention to rural road development, the 1987 road components of ADP projects in Nigeria. The experi- Agricultural Sector Review has criticized the government for ence is reviewed on the basis of project appraisal, comple- the inadequate "institutional arrangements for selecting, tion, mid-term review, and supervision reports. The Bank's designing, building, maintaining and financing these struc- views at appraisal on construction and maintenance are tures" (p.32). According to the review, "the objective now compared with the project implementation experience. In should be to combine maintenance with construction in the the case of the most recent projects, for which there is little new projects, and to establish and/or strengthen mainte- information on the implementation experience, the discus- nance capability to preserve physical assets already built in sion is confined to the appraisal aspect. the completed projects" (p. 24). When the Bank's project experience is considered in its This rhetoric is not new. The lack of financial and entirety, covering all ADPs implemented during the past 13 institutional capacity has long been recognized in the Bank. years, the discussion below shows that the Bank's views on Almost every appraisal report has mentioned it but not the relative emphasis placed on road construction and much changed until the three multistate ADP projects, the maintenance have been reversed. first of which was approved in 1986. The increased focus on rehabilitation and maintenance that has dominated the multistate projects began in 1986. Changing Views at Appraisal on Construction With the establishment at about the same time of the Directorate for Food, Roads, and Rural Infrastructure (DFRRI) in the president's office, with counterpart institu- The World Bank's views on construction targets at appraisal tions in the state governments, the focus of the ADPs on have changed fundamentally in two ways: (I) the emphasis rehabilitation and maintenance was further narrowed. shifted from new road construction in the earlier ADPs to judging from early supervision mission reports for the first rehabilitation and maintenance of existing roads in the multistate ADP, there is a clear division of functions latest statewide ADPs and the three MSADPs, and (2) the between the project and the DFRRI inasmuch as the former implementation was originally done by force account units will undertake only rehabilitation and maintenance work and is now being done "in general" by contractors. (See and the latter only the construction of new roads. Lately, complete review in Annex 2.) however, some ADPs have constructed roads under contract Reflecting these concerns, road construction densities for DFRRI in areas of low priority for agriculture. In some of have also decreased substantially at appraisal from densi- these, the State Directorates have intervened and directed ties of over 100 meters per square kilometer in the earlier that priority should be given to DFRRI programs with the ADPs to densities between 20-30 meters per square consequent delay in the original ADP program. The prob- kilometer (see Table 5). The higher initial densities were lem is exacerbated by serious delays in payments.30 not associated with needs but rather with the number of kilometers that could be built with one road unit or with 40 Bank's Perception of the Need for Rural Roads percent of the budget allocation. The lower densities were Bank reports have noted the lack and poor condition of later adopted to establish rehabilitation (rather than feeder roads in the rural areas of Nigeria. The 1973 Agricultural Sector Survey, for instance, argued that "at Table 5 some point in time, greater attention will have to be paid Actual and Targeted Construction of Rural Roads in to feeder roads" (p.33). It further observed that "where Selected Bank Projects in Nigeria there are feeder roads, their poor condition in many areas Actual restricts the ability of farmers to meet growing demands for Target food and export crops and to reduce the cost of marketing" MetersPKm2 Meters/Km2 (p.33). Thus the survey advised that production programs Kms. ProjectArea Kms. ProjectArea for each crop should include provision for feeder road Early Enclave Projects surveys and construction. Funtua 750 100 521 69 Rural road construction and maintenance has since been Gombe 500 97 706 77 incorporated in all the ADP projects, becoming an impor- Gusau 750 197 750 197 tant complement to the goal of increasing food production. Lafia 600 64 807 86 The Agricultural Sector Review of 1979 stated that "concom- Ayangba 1,300 66 1,667 127 itant to the economic benefit of the ADP feeder road Late Enclave Projects programs there has been a tremendous psychological boost Bida 620 36 429 25 to their farming communities in the ADP areas who begin ilorin 300 25 to feel that farming is receiving the attention due to it" Oyo North 800 65 (p.31). The review noted that in contrast, "development of Ekiti Akoko 500 101 258 50 rural infrastructure has received little attention during the Statewide ADPs current development plan period (1975-1980) in the Niger- Bauchi 1,200 18 ian Government priorities" (p.31). Kano 1,400 33 Nonetheless, because of the inability of local govern- Sokoto 1,700 18 ments to keep up with road maintenance, the Bank is no Kaduna 1,400 21 longer undertaking construction of new feeder roads as Source: Project Appraisal and Project Completion Reports. part of the ADPs, and is instead focusing on the mainte- nance of the existing roads within ADPs. In this vein, by 13 construction) targets in the three MSADPs. Only the second carried out with the lengthman system, where one length- multistate ADP adopted very low densities between 5-9 man is assigned for each three to four kilometers of road, meters per square kilometer, for no apparent reason (see with a headman for every eight lengthmen. The LGCs will Table 6). Moreover, Table 3 suggests that the three states pay the lengthmen and headmen and be reimbursed for included in the second MSADP have different road densi- the work done after supervision from each ADP The case ties; while Niger has a relatively high density of rural roads, for the use of labor-based methods for routine maintenance Gongola and Kwara have low densities. is not made strongly enough. Even though its use in projects appraised by transport divisions in Kenya and Malawi have rendered good results and proven cost Table 6 effective, no rural roads in projects appraised by agricul- Targeted Rehabilitation and Periodic Maintenance of Rural tural divisions have so far employed these methods. Roads in Multi-State ADP Projects in Nigeria Rehabilitation Maintenance Construction Achievements are Only Measured Target Target in Kilometers Meters per Routine as percent Rehabilitation Sq. Km. of Maintenance of all state To analyze achievements in road construction, data are Kms. Project Area Kms. LGA roads needed on lengths constructed, quality of construction, costs incurred for construction and rehabilitation, and First Multistate ADP benefits observed (vehicle operating costs, increases in Anambra 600 35 1,000 100 productivity, and indirect benefits discussed earlier). Avail- Bendel 600 15 1,000 14 able completion and supervision reports do not allow a Benue 600 1 3 1,000 27 complete evaluation of the roads component. Most only Cross River 600 22 1,000 15 conte egalu ation o n the tot. Mber of Imo 600 52 1,000 39 contain aggregate information on the total number of Ogun 600 35 1,000 16 kilometers constructed during the project period, main Plateau 600 11 1,000 29 changes in road specifications compared to appraisal, an Second Multistate ADP average cost per kilometer, and offer vague references to Kwara 550 9 1,760 59 increases in traffic due to the road improvements. Niger 410 6 1,830 25 Based on these limited records, this section presents the Gongola 500 5 950 29 construction achievements of various projects as regards Third Multistate ADP road specifications, reported traffic increases, kilometers Lagos 200 57 540 31 constructed, and indirect benefits. Clearly future supervi- Ondo 750 38 2,400 64 sion reports need to focus on the road component (as the Oyo 880 24 3,050 39 latest thematic supervision of February 1989) emphasizing Rivers 50 3 4,000 100 the quality of road rehabilitated against the SARs design standards, and making a systematic analysis of the traffic and production changes related to the road improvements in order to evaluate them properly. Increased Attention to Maintenance at Appraisal Road specifications The importance attached to maintenance of rural roads has increased over the years as Bank staff realized that any new Divergent standards for acceptance of rural roads con- construction without proper maintenance was not a perma- structed and rehabilitated have always been a matter of nent investment. (See complete review of maintenance at concern in the Bank, however, very little information is appraisal in Annex 2.) available on this subject. Actual road specifications have At appraisal the earlier projects were supposed to varied from standards at appraisal, but generally roads transfer the responsibility for the roads (as soon as they constructed by the ADPs have been regarded as having were finished) to the Local Government Councils (LGCs). high standards. These roads have usually complied with The LGCs were reluctant and unable to perform any and in some cases exceeded the requirements of the effective maintenance given their shortage of financial Ministry of Works. (See the complete review of standards in resources. The later enclave ADPs and statewide projects Annex 2.) began incorporating at appraisal a road maintenance unit that will maintain the roads during the project's life.3 It is Targets against actual construction 6nly the SARs for the multistate ADPs where the need for As shown earlier, the ADPs have substantially increased the more action in this respect is recognized. The Staff network of all- weather rural roads in Nigeria. As Table 7 Appraisal Reports (SARs) for the multistate ADPs ade- shows, the total kilometers of road constructed by all ADPs quately divide the responsibility for maintenance opera- increased until 1986, after which time most of the construc- tions according to type of operation and to financial tion of new roads was replaced by rehabilitation work in the capacity. The SARs assign the responsibility for implemen- multistate ADPs. tation of maintenance with mechanical equipment, and Each project has rendered mixed results. Of the ten partial funding of routine maintenance to the ADPs, while projects that have data on completion or near completion, the LGCs are responsible for the implementation of routine half seem to have been able to construct as many or more maintenance with technical assistance from the ADP The kilometers as the SAR target. However, as discussed in the ADP finances initial training of LGC engineering staff and next section, the maintenance achievements have not provide each LGC with necessary hand tools for routine reached the target in any project, and this has accelerated maintenance. Routine operations might be nianual and road deterioration. 14 Table 7 Annual Rural Road Construction and Rehabilitation Undertaken by ADPs, 1976-871(kms) ADP (Date 76-87 of Approv.) 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987a Total Ayangba (77) 29 570 379 630 59 1,667 Bauchi (81) 0 179 429 177 272 223 1,280 Bida (79) 22 157 91 159 0 429 429 Borno (86) 35 10 0 15 60 120 Ekiti-Akoko (80) 0 62 106 35 54 257 Funtua (74) 24 151 99 139 108 521 Gombe (74) 24 131 199 85 66 406 Gusau (74) 145 259 181 53 112 750 Nlorin (79) 0 0 n.a. 20 52 n.a. 72 Kaduna (84) 118 285 403 Kano (81) 67 232 210 251 180 940 Lafia (77) 201 342 49 71 60 73 796 oyo North (80) 15 81 125 141 362 Sokoto (82) 143 361 427 442 198 931 Total 193 541 508 847 866 994 444 913 1,206 1,194 1,673 723 8,934 Note aTotals for 1987 do not include First MSADP achievements included in Table 9. Source: Project Supervision and Completion Reports (several years). The targets of road construction in the three early the economic returns of irrigation, these fadama roads are enclave ADPs were met in Gusau only. The PPAR suggested likely to have high returns. The latest supervision reports that the better performance in Gusau, compared to Funtua from 1988 indicate that Bauchi completed the target in and Gombe, was due to the employment of an expatriate September 1986, and Sokoto was near completion in mid- full- time road engineer.32 and to the integration of the road 1988. On the other hand, Kano was still behind due to the and dam construction program (which also took place in poor choice of equipment, lack of spare parts, and inade- Gombe) (see Table 5). The road construction program quate selection of spare parts inventory, all of which has turned out to be the most successful component of both reduced equipment availability. Lafia and Ayangba, which were otherwise considered to be Similar problems in implementation of road works are relative failures (with respect to agricultural production reported for the seven states of the first multistate ADP As goals). Table 9 indicates most projects have fallen behind sched- The actual densities of roads achieved in these earlier ule except for Cross River ADP The implementation has projects were the highest achieved in any of the ADPs to Table 8 date. These earlier projects also had a larger share of total Actual and Targeted Expenditures on Rural Roads and Their costs allocated to road investments as a proportion of total Shares in Total ADP Expenditures project costs, as shown in Table 8. These high proportions of funds dedicated to roads appear again only in the Target Actual second and third MSADP project appraisals. One reason Percent of Percent of project authorities may have emphasized road construction N0oo0 Project Costs N'OOO Project Costs in the early enclave projects is because of both the fast disbursements and the urge to start projects at early stages Early Enclave Projects of implementation as stated in the PPAR, and also because Funtua 5,500 17.3 of the perceived need to improve communications infras- Gombe 4,100 17.9 tructure. Given the problems encountered with technolog- Gusau 4,900 19.0 ical packages, the other important component-agricultural Lafia 5,600 12.3 extension-had only a limited impact.33 Ayangba 13,519 24.0 Road construction achievements have varied widely in Late Enclave Projects the subsequent projects for various reasons. The three Bida 4,500 10.8 5,642 10.8 statewide projects, Kano, Bauchi, and Sokoto ADPs, for llorin 3,200 8.7 example, were scheduled to close at the end of 1986, but Oyo North 5,475 13.1 due to slow implementation were extended to 1988. The Ekiti Akoko 2,400 8.4 mid-term review found that road construction-had achieved Statewide ADPs 88 percent of the target in Bauchi, and only 50 percent in Bauchi 24,681 12.8 Kano and Sokoto. The main reasons for the shortfall were Kano 29,567 11.1 cited as the shortage of local funding and the emphasis on Sokoto 25,514 9.3 fadama roads (along the ridges of irrigation canals) in Kano First MSADP 29,100 12.1 and Sokoto which, in the case of Sokoto, proved to be 40 Southern Borno 4,823 12.3 Second MSADP 95,600 18.0 percent more expensive compared to appraisal estimates., Third MSADP 37,250 25.0 Given the importance of irrigation in the Sudan zone, and 15 Table 9 Road Construction and Maintenance 1987-88 Annual Targets and Achievements First MSADP Agricultural Rehabilitation Recurrent Routine Development Target Actual Target Actual Target Actual Project 1987 1988 1987 1988 1987 1988 Benue 150 - - 300 - - 1,200 150 37 Cross River 150 160 NA 400 176 NA 1,000 - NA Imo 179 82 187 450 303 300 1,300 287 230 Ogun no road program Plateau State 350 57 129 840 96 1025 180 - - Anambra 170 64 76 400 - 25 1,000 Bendel 150 32 NA 300 38 NA 1,000 - NA varied from disastrous (in the case of Benue ADP) to were reached or exceeded, while in the rest the achieve- satisfactory in other states. The problems have been ments were somewhere between 50 and 90 percent of the attributed to delays in hiring the roads engineer, and to target. However, in no case to date has the target for lack of equipment and spare parts in the case of force maintenance been achieved, and in most cases mainte- account operations, while procedural complications have nance has been well below the expectations at appraisal. delayed implementation through contractors. Signs of road deterioration have become evident even before the ADPs end. (See the review of maintenance Traffic increases achievements in Annex 2.1 Levels of traffic vary widely among the roads considered for With respect to mechanical maintenance, the targets have rehabilitation in the ADP, with reported counts varying from been too ambitious (see Table 9). In the earlier projects, for 15 to 940 vehicles per day. Volumes critically determine the example, the road construction force account unit usually size of benefits. The wide variation in traffic volumes, concentrated on road construction, devoting very little time therefore, calls for good monitoring of traffic volume to maintenance. When counterpart funds were delayed, the changes before and after project implementation, since maintenance program became a common target for cuts vehicle operating cost savings are substantial on roads with until funds became available. Meanwhile the roads con- average daily traffic (ADT) of over 50 vehicles. Even in the structed by the project have tended to deteriorate. case of roads with low-volume traffic, observed changes in Labor-based methods in general were not considered agricultural traffic are indicators of changes in production attractive during the oil boom in Nigeria. When sufficient and availability of transport services in rural areas. resources were available, equipment- intensive technolo- The PCRs and PPARs, however, suggest that a systematic gies were implemented. Therefore, the advantages of labor- analysis of traffic changes has not been done either before based methods have yet to be tested convincingly. An or after project completion. Traffic counts before the project important development in this field is that the National were performed on some roads in states where the first Directorate of Employment, UNDP, and ILO have recently MSADP is being implemented, due to a special request by started a labor-based demonstration project in Epe (Lagos the Bank, but some counts were made on roads that were State). The engineers from four states (Anambra, Kaduna, later not selected for improvements. (See review of traffic Katsina, and Oyo) are receiving training on the use of labor- counts in Annex 2.) In a few cases the methodoloev used in based methods for road rehabilitation and maintenance. the few traffic counts was inadequate. This is being done in preparation for the Pilot Public Works Programme (PPWP) that is expected to start next year in Indirect benefits selected LGAs in the four states. As stated earlier, the road program component has been A recent visit to the project indicated that the doubts evaluated primarily in terms of the number of kilometers raised by engineers whose previous training favored equip- constructed and maintained, the standards of these roads, ment-intensive applications have been partly dissipated by and in a few cases some observations in traffic change. two factors. First, the training includes a hands-on approach Indirect benefits such as higher producer prices, lower that allows engineers to participate in ongoing labor-based prices of consumer goods, lower operating costs for pas- road construction and maintenance in Epe. Second, the senger transport, and stimulus to trade of all kinds have trainees are receiving information on how and when it is been mentioned but not evaluated systematically, although appropriate to combine labor and equipment- intensive some PPARs (e.g., Lafia and Ayangba), conclude that methods. The applicability of the labor-based schemes in uncounted road benefits heve been large enough to offset other states is now being examined by the federal govern- the projects low rate of return. (See the review of indirect ment in an effort to increase job creation in rural areas. This benefits in Annex 2.) situation is in contrast with the experience of ADPs. No ADP has succeeded in transferring the roads to the Very Poor Maintenance Achievements LGA for routine maintenance.34 In the past, LGA representation Construction rather than maintenance has had much greater in the executive committees of projects was very limited, political appeal. Also in terms of project achievements, despite the emphasis in the project design at appraisal to construction has greater appeal. In at least half of the ADPs, involve local bodies. The recent appointment of Local the targets for road construction and rehabilitation units Government Councils (LGCs) and an expected increase in 16 resources, however, seems to open a window of opportu- costs for rehabilitation using contractors than by force nity. In an effort to involve LGCs more in the maintenance account. activities, at the beginning of 1989 firm arrangements were The use of contractors is not that widespread for mainte- made between the project managers of Anambra and nance operations. Periodic and recurrent maintenance Plateau and a few selected LGCs (ten in Plateau and two in operations, which in most cases require mechanical equip- Anambra) to undertake training and practice maintenance ment, have been assigned to force account units within the with labor-based methods, and to conduct routine mainte- first MSADP, and depending on contractor availability "they nance. The agreements cover financing for six persons in will be executed by contract," in the second and third each LGA, and the provision of the necessary hand tools for MSADPs.36 Manual routine maintenance is assigned to the the maintenance of 10 to 15 kilometers in each LGA. LGAs, with training and funding assistance from the ADPs. On the Issue of Contractors and Force Account Only the last MSADP describes a routine maintenance program with the lengthman system that has proven cost Despite much recent rhetoric about increased execution effective in Malawi and Kenya (see the section on mainte- through contractors, at present it appears that local contrac- nance at appraisal). Nevertheless, the success of such a tors are available for construction, rehabilitation, and system in other countries has been attributed partly to the recurrent maintenance which require mechanical equipment, implementation of pilot demonstration projects, adequate but very little is known about the amount of contractor technical assistance for the implementation of labor-based capacity available for routine manual maintenance of rural methods, long-term commitments from the government roads. The results of both the pilot agreements between with respect to human and financial resources, and ade- the first MSADP and a few selected LGAs to conduct quate task reporting and monitoring systems. None of routine maintenance, and the ongoing labor-intensive these have been specified in the MSADP appraisal. UNDP/ILO project in selected states will improve knowi- Drawing a comparison between the rural road sector and edge on this matter. the main road sector, it can be seen that the development The development of local contractor capacity for road of local contractors for routine maintenance of rural roads is construction, rehabilitation, and recurrent maintenance was not an easy subject. This often includes two reasons: (I) the not tackled explicitly until the MSADP projects were amount of contracts per kilometer is very low compared to appraised. Even these latter have not dealt with the critical other activities such as mechanical maintenance or rehabil- issue of developing supervision1 capacity. in the early enclave itation (and therefore attracts different types of contractors), projects construction work was done by units within the and (2) the costs of supervision are very high. project. In the three MSADP projects, however, the road It is possible to carry out routine and recurrent rehabilitation component has been designed specifically to maintenance by contract, but the work can be more be performed by local contractors according to the current difficult to specify and monitor. For such operations, it capacity. The SAR for the first MSADP (including Middle may be necessary to specify total amounts of work per Belt and Northern states) reports that 'a study Iwas) kilometer of road for each maintenance activity, with completed in Anambra State to establish the availability of kmeterco roadrs beac maintenance actviy with contractors ifor construction, rehabilitation, and recurrent emergency repa1rs beig paid for at daywork rates maintenancel... and similar studies have been launched in (Robnson 1988). other states" (p.21). Other Middle Belt and Southern states The lengthman system has been proposed in the third also appear to have sufficient contractor capacity, since the The with tm ante hat beentproposexists thin SARs fo the r t s d ad i MSADPs indicate that accountability exists within SARs for the recent second and tinrd MSADPs cntrctors, the community and not only for technical reasons. However, there are an adequate number o inert contrtors enough technical assistance has to be provided to ensure some wth international afiliates, to provide a competive the necessary institutional setups and also in order for environment." standards to be met. in order to encourage contractors to The contracting experience Ifor contruction, rehabilita- use labor-based methods at least in the maintenance, many tion, and recurrent maintenancel of the MSADP I changes have to be introduced beginning with the bidding states is too recent and varied to draw firm conclu- process. The emphasis on equipment makes the qualifica- sions on performance so far. Perhaps the most serious tion process biased toward those contractors with the most problem is the lack of experience in managing equipment. For example, most ADP invitations for prequal- contract work of this nature... Local contractors also ification indicate that rural roads contractors "must have find difficulty in satisfying the terms and qualifications experience in projects involving the construction of feeder for bidding. Many follow old habits Ifrom previous roads using heavy machinery."37 Rather, contractors could practicel. They believe they can avoid agreed bidding be encouraged to use labor-based methods (solely or in procedures,... including unreasonably low pricing combination with equipment-intensive methods) by modi- with the expectation to vary the price during construc- fying the prequalification process. For example, provisions tion. Projects will have to develop the capability to can be included to evaluate the capacity to mobilize local derive the full benefits from competitive bidding.35 labor, previous or proposed training of foremans and gang Reportedly, these practices and the lack of supervision leaders, as well as previous road contracts using labor- capacity in some MSADP I states, have led to higher unit based methods. 17 Broad Assessment of Future Needs for Expansion, Rehabilitation, and Maintenance of the Rural Road Netwoirk in Nigeria In line with the preceding arguments, this section calculates Food Sector Report future needs in terms of expansion, rehabilitation, and In the report of the Food Strategy Mission, Idachaba (1980) maintenance of rural roads, within a regional context argues that an ideal minimum density of roads for Nigeria accounting for each region's economic potential. First, target would be 110 meters per square kilometer of area, however, road densities are obtained from a review of past experi- large investments are needed to achieve such a target. The ence, current project targets, and targets in other countries. present report argues that investments should be concen- Furthermore, needs in terms of kilometers to be con- trated in areas of high agricultural potential and high structed, rehabilitated, and maintained in the future are population densities, so that the desired feeder road presented. Second, cost estimates from project experience densities would be an additional 40 to 80 meters per are presented for construction, rehabilitation, and mainte- square kiiometer, and hence the additional overall road nance activities. Third, an assessment is made of the target density would be 60 meters per square kilometer. resources needed to develop the network according to the various methods presented to calculate target densities, Bank Appraisal and Completion Reports and also priorities according to each region's potential. Target densities estimated at appraisal and achieved at Finally, it is shown that needs can be fulfilled to a great compiletion reflect the capacity of one road construction extent with a more efficient use of present resources unit, but are also measures for desired feeder road available for investment. densities in Bank-related projects. Given the significance of The results point to two major issues. If future investment the rural road component of the ADPs in Nigeria, these is directed to improve a portion of the existing network (i.e., targets are a good procy for possible levels of intensity of ADP targets), most of the rehabilitation work will be done rural road infrastructure. in the Southern and Middle Belt states. However, if the These densities were calculated earlier and are pre- upper bound target densities are used, and therefore sented in Tables 5 and 6. From these tables, the state and population densities are the driving factor, a large amount regional target densities are derived, highlighting important of work and resources are required to develop the network differences as shown in Table 10. As Table 10 shows, the in the Northern region. targets of the early enclaves were high. The targets for other In any case, the needed resourc-es lie somewhere projects are representative of targets used in the latest between the requirements for fulfilling the two targets. The projects. Thus a pattern emerges at the regional level as total requirements estimated are as follows: (1) tor new follows: 1i) 20-30 meters per square kilometer for the construction and rehabilitation, US$105 million for the Northern region, (2) 40-50 meters per square kilometer for Northern states, US$90 million for the Middle Belt states, the Middle Belt region, and (3) 80-100 meters per square and US$195 million for the Southern states; and (2) for kilometer for the Southern region. The above compares annual maintenance an average of USS21 rnillion in the favorably with a national target density of 60 meters per Northern states, US$23 million in the Middle Belt states, square kilometer estimated by ]dachaba.39 and US$37 million in the Southern states. This amounts to a total of US$390 million for construction and rehabilitation Table 10 and US$81 million every year for maintenance. Regional Average Target Densities Meters per square-kilometer Calculation of Target Road Densities in Nigeria Project Method Target Target densities are just one indicator of the development (t) T2) of the rurai road network, which is a good proxy for the Nh - - future investment and recurrent expenditure needs of rural Northern States (Guinea Sarannal) roads. In other countries, e.g., India, target densities of rural Funtua, Gombe, Gusau (Early) 169 169 roads are calculated as a function of the region's area, and Bauchi, Kano, Sokoto, Kaduna ADPs 21 21 the number of cities, towns, and villages of different sizes Middle Belt States (Sudan Savanna) in that area. In India there are high correlations between Lafa, Ayangba (Early Enclave) 84 84 the regon's area and the length of main roads, and Benue, Plateau (Flrst MSADP) 20 40 40 between hi e number of towns in the area by size and the Niger,Kwara,Gongola(SecondMSAOPI 22 15 22 length of rural roads. In Nigeria as well, the correlation Southern States (Tropical Rainforest) between area and main road length is very high as Cross River, Imo, Ogun, Anambra, Bendel (First expected.3" However, data on the number of villages and MSADP) 47 46 47 towns by nurnber of inhabitants is not available. Therefore, Lagos, Ondo, Oyo, Rivers (Third MSADP( 128 56 128 this section simply presents a methodology using compar- isons with ne;khboring Cameroon and India. Source: Staff Appraisal Reports. 18 Comparison With Existing Densities in rable 12 Neighboring Cameroon Population and Road Densities india-1961 The rural road network in neighboring Cameroon is still States Grouped by Population Densities expanding. Nevertheless, an examination of the target Persons Meters densities used in projects in Cameroon (see Annex 3) per square per square provides a good reference, since both countries have State kilometer kilometer regions with comparable ecological characteristics. As seen Group 1 in Table II, the comparison reveals that projects developed Jammu & Kashmir 16 23 in Cameroon in regions with comparable resource endow- Rajasthan 59 58 ments have similar target road densities, even though they Madhya Pradesh 73 33 have significantly lower population densities. Therefore, for Group 2 the analysis proposed, the targets calculated in the pre- Assam 97 136 vious section can be regarded as appropriate lower bound Gujarat 110 51 targets. Onssa 113 88 Mysore 123 139 Maharashtra 128 58 Andra Pradesh 131 72 Table 1 1 Road Target Densities from Bank Projects in Nigeria and Group 3 Cameroon Punjab 166 121 Uttar Pradesh 251 174 Range of Densities Madras 259 141 Population Target Road Bihar 267 242 Region (p/km2) (m/km2) West Bengal 398 368 Kerala 434 189 Nigeria North 43 - 227 20 - 30 Source: see Annex 4. Middle Belt 30 - 92 40 - 50 South 109 - 842 80 -100 Cameroon Selection of Rural Road Target Densities for Nigeria North 9- 50 26- 35 Central 7 8 - 20 The set of lower bound targets is obtained from the targets W. Highlands 70- 95 65 -122 set for the ADPs, while the upper bound targets are set S.Raintorest 4 - 83 3 - 40 from a comparison with India in 1960. Recent estimates prepared for the future Agricultural Development Fund Project (ADF) of kilometers to be rehabilitated corroborate the validity of these ranges with data from several states (Kaduna, Bauchi, Anambra, and Plateau states). Further, the Comparison with India in 1960 target densities used by DFRRI for the first phase of rural India in 1960 had approximately 135 persons per square road construction are included in the ranges specified.' kilometer, which is the population density expected in The lower and upper bounds for targets, calculated with Nigeria in the year 2000. Considering the high densities of respect to past project densities and India's densities, maintained rural roads in India compared to Nigeria (6 coincide with calculating the higher road target densities in times more in the earlier comparison), an upper bound for the Southern states. The same coincidence does not occur target densities can be calculated by looking at the with the calculations for the other two regions. In the densities of rural roads in India in 1960.40 Northern states, the road densities targeted in the ADPs Table 12 includes the densities for most states in India in (lower bound) are less than those in the Middle Belt. 1961, ordered by population densities. Population densities However, the upper bound (from the comparison with of group I correspond to the estimated densities of Middle India), sets higher rural road target densities in the Belt states in the year 2000, group 2 to the Northern states, Northern states than those for the Middle Belt states. The and group 3 to the Southern states. Other similarities in lack of correspondence is explained by the greater impor- addition to population densities occur in the case of group tance given to higher population densities in the calculation 2 states, where extensive regions are dedicated to the of target densities for India. This is likely given that the production of crops similar to those in the Northern states Northern states have a larger number of towns and villages in Nigeria. Therefore, as an upper boundary, the target compared to the Middle Belt. Also the roads would be densities from the comparison with India in 1961 are as used much more in the North because of higher production follows: (I) 50-70 meters per square kilometer for the and greater population, which would imply higher mainte- Northern states, (2) 40-60 for the Middle Belt, and (3) above nance requirements. 120 for the Southern states. Finally, the economic potential of each region is consid- ered here. The regional context described earlier called for extensions and rehabilitation of the rural road network in the North and Middle Belt to allow for increases in areas under cultivation, and at the same time an emphasis on rehabilitation in the Southern states to accommodate large increases particularly of consumption that will require 19 reduced costs of intermediation. It is assumed that the With varying standards and terrain conditions, the cost of ranges calculated for road densities in each region will constructing and maintaining roads varies from state to allow the realization of the potentiai in each region, since state. Table 14 contains averages obtained from appraisal, the same factors included in the estimation of each region's completion reports, and project files for the various poential were mIcluded in the estimation of the target standards. If rehabilitation is considered to cost approxi- densities for the ADPs. in addition, increased population mately half of what a new road costs, the construction cost densities and village densities have been considered in the calculated with ADP figures is US$20,940 per kilometer in case of india. Therefore, these two sets of targets are 1987, which is the same as the figure obtained for the PPWP considered appropriate for the exercise. in the same year.43 However, the difference between these two and the DFRRI is quite high, which calls attention to the fact that with lower standards (construction rather than Rehabilitation and Maintenance Cost Estimates design in this case), the DFRRI roads need much more Road construction and maintenance costs vary according to maintenance. The work standards for the DFRRI roads have specifications, institutional arrangements, technology (labor- been low, and many of the roads constructed are unmain- vs. equipment-intensive), and region (i.e., soil characteristics tainable just after completion.44 While it is possible to be and topography), and of course to variations in exchange precise about road construction costs for Bank projects, it rates, and real costs of labor. This makes the comparison of is not possible in the case of the Federal Directorate of these costs complex.42 In order to calculate approximate Food, Roads, and Rural Infrastructure (DFRRI).45 costs for rehabilitation and maintenance operations, several The trade-off between construction and maintenance is sources were consulted, including DFRRI, UNDP/ILO Pilot important. If average daily traffic is less than 50 vehicles per Public 'Works Programme (PPWP), and Bank project files for day, with low construction standards and little maintenance, Nigeria and neighboring Cameroon. usually the road will have to be rehabilitated every two A survey of rural road specifications used in Nigeria years. However, if the traffic is higher than 50 vehicles per indicates that three main standards prevail, including those day, and no maintenance is performed, an annual recon- used by the Bank ADPs 1which conform to the Ministry of struction has to be scheduled (which is actually happening). Works (MOW) standards), the DFRRI standards which in On the other hand, the ADP roads call for a 7-8 year interval theory vary slightly from the MOW standards, and those of between periodic maintenance, annuai or biannual recur- the Pilot Public WVorks Programme. Table 13 contains the rent maintenance after the rains, and day-to-day routine main characteristics of the road section for design under maintenance. As an illustration, considering a hypothetical the three sets of standards. kilometer of road in the North, and using the ADP costs, the annual costs to have a road operational will be twice as Table 13 high for DFRRI )with low construction standards and little Road Construction Specifications maintenance) versus the current ADP (assuming adequate standards, appropriate maintenance during project life, and Bank Projects little maintenance thereafter). A word of caution about the Feeder Farm usefulness of this comparison, however, is that after a few DFRRI4 Road Road PPWP years of inadequate maintenance, rural roads are easily verageDailyTraffo up to 100 up to 50 washed away, regardless of initial standards (as has been Average Daily reported with ADP roads from the earlier enclave ADPs). Rght o' Way rm.) 10 13 10 13 The costs in Table 14 are national averages. As pointed Min. lormation width in.) 9.5 8 1 0 out before, however, regional variations are important. In (,arrageway width {mr.' 6 6 6 6.5 Base thickness (mm .) 1 50 1 6o 150 max. 300 order to obtain costs for the different regions in Cameroon, Camber (hiks 4 4 the costs used for appraisal of the Second Feeder Roads Drainage adeauate 1.2 m. below grade level Project in Cameroon were examined (see Table 15). These Min. culvert length (m.l 6 & 8 regional data allow the calculation of variations in rehabil- itation and maintenance costs with respect to similar 'Not adhered to because of low allocations per kilometer. regions in Nigeria. Table 14 Average Costs per Kilometer in Nigeria Foreign institution Year Activity Cost per Kilometera Costs % DFRRI 1987 Construction N 9-10,000 US$ 2,330 Second MSADP 1987 Rehabilitation N 45,000 US$ 10,470 80% Periodic Maintenance N 20,000 US$ 4,650 Recurrent Maintenance N 2,600 US$ 610 Routine Maintenance N 1,500 US$ 350 PPWP 1987 Construction Labor-Intensive N 27,000 US$ 6,350 20% Equipment-Intensive N 87,200 US$ 20,510 73% Note: aMarket exchange rates for 1987 from IMF (1987). Costs for DFRRI are not for standard shown in Table 14. Costs for second MSADP and PPWP are from appraisal documents. Source: World Bank 1988c: UNDP/ILO 1988. 20 Table 15 rehabilitation and maintenance costs. The estimated con- Cameroon Second Feeder Roads Project (1988) struction costs per kilometer for the three regions in Construction and Maintenance Cost Nigeria are shown in Table 16. The activities included in Region Activity Cost per Km Table 16 have the periodicity indicated in Table 17, which varies by region (because of soil quality, rainfall, and Northern Rehabilitation US$ 17,000 average daily traffic). Plains Maintenance US$ 410 Central Rehabilitation US$ 18,000 Assessment of Investment Needs to Reach Savanna Maintenance US$ 410 larget Densities Western Rehabilitation US$ 28,000 To calculate the investment needed to reach the desired Highlands Maintenance US$ 620 road densities, the difference between the existing all- Tropical Rehabilitation US$ 26,000 weather road density and the target densities is calculated Rainforest Maintenance US$ 620 for both the targets set for ADPs and the upper bound Total Rehabilitation US$ 24,000 targets calculated earlier from comparison with other Maintenance US$ 500 countries (see Table 18). The differences calculated in Table Costs exclude design and contingencies. 18 represent additional kilometers of all-weather roads Market exchange rate for 1988 from IMF(l 988) needed per unit area of each region specified. The proposed extensions to the network may be in the form of rehabilitation and new construction. Each is calculated Rather than comparing the costs directly, it is interesting using the information on existing all-weather roads and to extract the costs of rehabilitation, which are 60 percent total roads presented earlier in Tables 3 and 4. All roads are higher in the tropical rainforest area compared to the assumed to be regularly maintained. The total length of the Northern Plains. Also the maintenance costs are higher by desired all-weather network is presented in Table 19. It about 50 percent in the South. The regions in Cameroon should be noted that the length expected under the higher with terrain similar to that in the Middle Belt and Northern targets set is almost the length of the total existing network, states in Nigeria have lower costs, given the lower rainfall, which as stated before has only a small percentage of all- and better soils for roads. Mountainous terrain like that in weather roads. The cost to rehabilitate the needed addi- the Western Highlands in Cameroon result in higher tional roads to meet the targets are included in Table 20. Table 16 Estimated Average Rehabilitation and Construction Costs per Region Cost per Kilometer Maintenance Region Construction Rehabilitation Periodic Recurrent Routine Northern US$15,000 US$ 7,500 US$ 3,300 US$ 550 US$ 320 Middle Belt US$17,000 US$ 8,500 US$ 3,800 US$ 580 US$ 350 Southern US$25,000 US$12,800 US$ 5,000 US$ 630 US$ 380 Table 17 Table 19 Periodicity of Maintenance Operations Length of Target All-Weather Rural Road Network Region Periodic Recurrent Routine Region Northem 7-8 years annual day-to-day Northern Middle Belt Southern Total Middle Belt 7 years annual day-to-day Under ADPs Targets: Southern 6 years annual day-to-day Existing All-Weather 7,000 5,100 4,900 17,000 or biannual New Construction - - - - Rehabilitation 2,700 9,700 12,500 24,900 Maintenance 9,700 14,800 17,400 41,900 Table 18 Under upper Bound Targets: Existing All-Weather 7,000 5,100 4,900 17,000 Difference Between Target and Existing Road Densities New Construction 9,700 - - 9,700 Road Densities (m/km2) Rehabilitation 6,700 11,300 18,200 36,200 Maintenance 23,400 16,400 23,100 62,900 Upper Differences ADPs Bound Upper Region Existing Targets Targets ADPs Bound Northern States 18 25 60 7 42 Middle Belt States 15 45 50 30 35 Southern States 24 90 120 66 96 21 Table 20 roads, which is in line with the present apportionment Costs to Upgrade and Maintain Network by Region among rural infrastructure components, the total funds (US$'000) available for roads in the next five years will be US$425 Region million. This falls between the lower and upper bound needs calculated in the previous section only for construc- Northern Middle Belt Southern Total tion and rehabilitation. In addition, the only funds for rural Under ADPs Targets: roads are the ADP road investments (which amount to New Construction - - - - approximately US$35 million for the next 4 years), and the Rehabilitation 20,250 82,450 160,000 262,700 LGA appropriations. Maintenance (per year) 12,440 21,800 32,070 66,310 The colossal effort needed to build maintenance capacity Under Upper Bound Targets: at the LGA level will require more resources than those New Construction 145,500 - - 145,500 available from DFRRI, ADPs, and present appropriations to Rehabilitation 50,250 96,050 232,960 379,260 LGAs, if technical assistance and effective institution Maintenance (per year) 30,010 24,160 42,570 96,740 building is to materialize. Therefore, if the government's commitment to large construction of rural infrastructure is unaltered, and supplementary funds are made available to build local maintenance capacity, then investment levels would not be far from the ones estimated here for future Comparison of Needs and DFRRI Investnents needs. The Federal Directorate of Food, Roads, and Rural Infras- Other points should be taken into account. The construc- tructure (DFRRI) received N 500 million in each of the last tion output and quality of the DFRRI roads has not been two years (1987-881 for rural infrastructure programs- the best, and therefore the density of all-weather rural approximately US$120 million in 1987 and US$85 million in roads might not improve at all under the present arrange- 1988-a large portion of which was dedicated to road ments. The low standards reported for the roads con- construction. While the funds available approximate road structed have also adversely affected the longevity of road construction and rehabilitation needs, the lack of funds for improvements. The state directorates of DFRRI are represen- maintenance is still to be resolved. Thus, even if abundant tatives at the level of the federal government. Once the resources are dedicated to extend and rehabilitate the roads are finished, as is the case of ADP roads, inadequate rural road network, there would still be a compelling need provisions are made for their upgrading and maintenance. to improve the maintenance capacity at the local level to LGAs with scarce resources are usually not able to take over ensure that such investments have a long-lasting effect. and improve the roads to a "motorable" condition, partic- DFRRI's target is to construct 60,000 kilometers of rural ularly since this will absorb all local resources. Also, the roads in six years. If recent funding levels are maintained, wide scope of the DFRRI might imply that it has large a total of approximately US$708 million will be allocated to overheads above that calculated for state-level projects the directorate in the next six years.46 Assuming that about (from the ADP data), which will largely increase the financial 60 percent of these funds are allocated to construction of requirements to improve rural road availability. 22 Institutdonal Considerations and Sustainability Since DFRRI is already a major federal level entity, the rural ADPs giving advice to the LGAs in the selection and road planning capacity of the directorate might be bol- evaluation process. Every two years each LGA will select stered through technical assistance. Some of FACU's expe- and evaluate those roads that need rehabilitation and rience with road selection exercises in the ADPs could be periodic maintenance providing this information for transferred as part of the technical assistance. With approval to a new rural roads division in the state MOW. improved planning capacity at the federal level, the Upon approval, the LGAs will proceed with the rehabilita- directorate could address some of the interregional tion and mechanical maintenance program by contract and resource allocation issues related to regional resource the routine maintenance by the lengthman system. The endowments and the potential sources of demand and main issue here will be the development of capacity at the production. Micro-level road selection exercises might LGA level in order to expect this to happen. Contractors will furnish better results if performed by local level govern- be selected from a pool of contractors prequalified by the ments with assistance from FACU. Further, the regional rural roads division of the state MOW for this purpose. The character of rural road programs should be coordinated contractor selection process will be subject to approval by with the interregional agricultural strategies, which might the state MOW Contractor supervision has to emphasized. only happen loosely under present conditions. LGAs and the state MOW will definitely need technical Increased coordination within the federal government between NDE assistance in the various aspects regarding contractor and DFRRI appears to have many potential benefits. Both supervision including costing, evaluation, and quantification directorates are presently involved, to a different extent, in of road works, qualification of contractors, bidding practices, construction and rehabilitation of rural roads. Since NDE's and equipment comparison and evaluation. participation in rural roads is just starting, the establish- Within each LGA, simplified road selection exercises may ment of an interdirectorate committee could coordinate the be conducted using systematic vehicle counts and agricul- road selection exercises, share some of DFRRI's previous tural data from the state ADPs. The evaluation exercise experience, disseminate NDE's experience with the imple- could also involve at least initially the local community, mentation of labor-based methods, and also organize joint which to a certain extent can give an indication of the need training of DFRRI and ADP engineers. for roads and the availability of local contractors and labor The states have little experience with rural road infras- force for labor-intensive maintenance. tructure, and have plenty to worry about with state roads. Finally, given that at present LGAs do not have either the Until recently, rural road infrastructure has been the technical or financial resources to do this, the implementa- concern of the ADPs and the LGAs, and meanwhile funds tion could be done progressively as stated in the Proposed for rural roads have been channeled mostly through the Rural Infrastructure Project. A technical rural roads division ADPs. In a way these funds have bypassed the traditional could be created in each state MOW primarily in charge of state institutional structure. Since 1986 the funds available technical supervision, technical advice to LGAs, and con- to the state directorates of DFRRI from the federal govern- tractor prequalification. This division could be formed with ment, and the increased funds to LGCs have increased some of the capacity of the state directorates of DFRRI and substantially the total funds available for investment in rural road divisions of ADPs. After this, two to three LGAs roads (the resources from ADPs were allocated for some could be selected in each state to introduce this imple- states every year). Rural road investments have been more mentation scheme. attractive to states when the funds were to be administered Unless some steps are taken soon with very explicit by them. capacity building objectives, the problems of rural road A setup proposed here might be one in which each state maintenance and expansion might continue. is involved in planning rural road activities through the 23 Notes 1. This includes the achievements and disbursements for the first 7. The reorganization of Colombian rural roads administration multistate ADP as detailed in Tables 7-9. in 1972 has often been cited as a case of successful implementa- 2. The ADPs have had three major stages. Initially there were tion of organizational changes. A national fund for rural roads nine enclave projects which covered areas smaller in extension (FNCV-Fondo Nacional de Caminos Vecinafes) has existed since 1960, than a state. The early enclave projects (started between 1974-77) but until the early 1970s it functioned as a secondary entity in the included Funtua, Gombe. Gusau, Lafia, and Ayangba ADP and the Ministry of Public Works (MOPT) structure. In 1972 FNCV was late enclave projects (started between 1979-80) included Bida, reorganized as an autonomous organization under the MOPT and Ilorin, Oyo North, and Ekiti Akoko ADP After 1981 statewide charged with planning, construction, and rehabilitation of rural projects followed in the states of Bauchi, Sokoto, Kano, Kaduna, roads. The institution enjoys a reputation for efficiency and sound and Borno. Finally two multistate projects (MSADPs) have been management, which is confirmed by the good performance of the implemented (in Southern and Middle Belt states), which along Bank's ongoing Rural Roads Project. Over 60 percent of rural roads with the approved third MSADP (in Lagos, Ondo, Oyo, and Rivers in Colombia have been constructed either totally or partially by states) will mean that each state has a statewide ADR An FNCV. The network of rural roads has increased notably at an Agricultural Development Fund project is being prepared to annual rate of 10 percent (World Bank 1986; SAR Colombia, Rural support the ADPs in the future. Sector Transport Project, Latin America Office). 3. For example the redesigned Borno State ADP has allocated FNCV's activities are funded through a combination of transfers resources to provide cost-free materials, loan tools, and give from the national budget on account of earmarked taxes on oil advice and supervision to villages for self-help installation of products (53.1 percent in 1985), special budgetary appropriations drifts and improvement of tracks. Communities have shown (14.8 percent), own resources from contracts to build roads on considerable enthusiasm to improve their access by providing behalf of other institutions (2.7 percent), and multilateral and free labor and local material. bilateral loans (28.2 percent). 4. "The road programme was the flagship of the project... It is Another example is the proposed organization and manage- always the first thing people mention when asked about the ment reorganization of the rural road sector in Tanzania under project" (Ayangba ADP PCR); "The agricultural services did not present study by the Government of Tanzania and the Bank. It perform as expected. . . On the infrastructural side the project was mtends to improve the maintenance of the very deteriorated more successful . . ."(Bida ADP PCR). "Project extension activities district road network in Tanzania. were the weakest element in the ADP strategy of roads, input 8. For example increased specialization in rice production is distribution, and extension... Increased use of fertilizer and other reported by Mitchell (1977) as a result of a new road in an area inputsproject" (Gusau ADP PCR). with no other roads. 5. Types of Rural Transport 9. Better transport increased the use of market places in Rural transport will be divided in two broad categories: primary Thailand when an earth road from Chiang Rai to Chiang Kham and secondary rural transport. Primary rural transport includes local reduced travel time from three days to three hours, and altered level transport from the farm gate to or from primary markets, the regional rice trade. See Moerman 1968. small villages, and small towns. Vehicles and pedestrians use in 10. Airey (1980) documents that after improved school roads this case rural local roads and rural collector roads with average increased education opportunities, out migration increased traffics of up to 50 vehicles per day. (The definitions of rural local indirectly. Areas have received migrants, as documented by roads, rural collector roads, and rural arterial roads are the result Hegen (1966). Also intraregional movements in search of work or of the "Rural Roads Evaluation Conference," which convened in better economic opportunities are documented in Indonesia by Harpers Ferry, West Virginia, in November 1980. In Nigeria they Okada (1978). correspond to farm access roads, feeder roads, and Local 11. Carnemark (1979) reports that the construction of a feeder Government Authority roads according to the Ministry of Works road linking a village in Yucatan, Mexico with the main highway categories.) Rural local roads provide direct access from farms to tended to strengthen the village as the service center of the primary markets and small villages, while rural collector roads surrounding villages. Blaikie (1977) reports the same for villages in connect villages to small towns with local markets and basic West Central Nepal. services. Local government and institutions are the ones usually 12. For a complete review of procedures see Carnemark 1984; concerned with the development and maintenance of this Bovill 1978; and Hine 1982. infrastructure. 13. February 1989 Supervision Report, p. 2. Memo from L. Secondary rural transport connects small towns to the main Campbell to A. 19 Supervision -Feeder roads road network (primary and secondary roads) and regional and to A. Seth on "Thematic Supervision-Feeder Roads markets. Higher traffic volumes are common but not a rule, and and Water Supply Components of ADPs." in most cases transport is performed with the use of motorized 14. Pro-rata of total extension costs by proportion of extension vehicles on rural arterial roads. While primary transport costs are agents in zone of influence to total number of agents. assumed directly by the producer, the secondary transport costs 15. Pro-rata of total commercial service cost by proportion of are usually larger and incurred by the marketing agent and incremental fertilizer tonnage into zone of influence as a propor- therefore indirectly by the producer. tion of the total incremental fertilizer tonnage. 6. Approximately 19,000 kilometers of road were constructed 16. Page 2, February 1989 Supervision Report. Memo from L. between 1986 and 1987, which represents an estimated invest- Campbell to A. Seth on "Thematic Supervision-Feeder Roads ment of N 190 million for the first two-year phase, or N 95 million and Water Supply Components of ADPs." per year. In 1986 the ADPs constructed approximately 1,602 17. Letter from F.S. Idachaba, Head of Federal Agricultural kilometers, which at N 10,000 per kilometer (average cost from Coordinating Unit, to Alan Denness, lune 24, 1985. Bida ADP Project Completion Report) give a total of N 16 million 18. See Devres 1980; Airey 1984; and Howe 1984. in construction and rehabilitation only. Therefore, the investment in roads has augmented substantially with the creation of DFRRI. 19. Including the first and second MSADP approved quantities as detailed in Tables 7-9. 24 20. Statistics from tne Public Works Department and local 40. This is considering the length of rural roads as the sum of bodies show that at least 90 percent of all rural roads were other district roads and village roads, both of which serve the included for maintenance in 1961 (see Central Road Research function of rural roads as defined earlier in Nigeria. Institute, History of Road Development in India, New Delhi, 19631. 41. DFRRI's targets for the first of three phases was obtained Estimates for Nigeria presented later indicate that no more than from DFRRI's office in Lagos. The targets for the three phases 15 percent of the LGA network is all-weather. combined were calculated assuming the same regional targets for 21. This unbalances the comparison with India in 1951, given each phase. Therefore, the targets obtained for each region were that the length reported in Table 2 for that country includes only as follows: Northern states 45 meters per square kilometer, those roads on which maintenance operations were reported in Middle Belt states 54 meters per square kilometer, and Southern 1951, and therefore the expected proportion of all-weather roads states 130 meters per kilometer. is probably far above the 10 percent reported for Nigeria. 42. For example the comparison of construction unit costs from 22. Chapters 3 and 5 in Lele et al. 11989) examine interactions appraisal and completion reports have the following problems. at the regional level between human and livestock populations, Construction costs reported at appraisal refer to proposed ecology, land availability, agricultural potential, ethnicity, social standards, while the costs reported at completion usually refer to and economic organization, government policies and investments, an average of standards achieved in a vast area (a state in Nigeria education and health care supply, and income growth. is quite large), are usually not reported, and are different from the 23. Chapter 5 in Lele et al. (1989) points out how the need for standards set at appraisal. land expansion in the North in the next years will require larger Similar problems appear in the comparison of road construc- amounts of land than that which is available. tion and maintenance unit costs between different organizations 24. For further discussion see chapter 5 of Lele et al. 1989. and between countries. Under different institutional arrange- ments, different costs are reported including different proportions 25. Including millet, sorghum, cassava, rice, and cowpeas. of total overhead. Donor projects usually report the total cost with 26. Delgado as quoted in Lele et al. 1989. an overhead representative only of projects which are added to 27. The fast rate of growth of traffic on main roads during the an existing institutional framework. On the other hand, govern- 1970s slowed down during the early 1980s, and finally declined ment figures include different levels of total overhead, usually during the late 1980s. higher than that of donor projects. 28. These were aggravated by a combination of thin pavements, 43. The per kilometer costs of road construction with equip- construction practices which did not meet specifications, and ment-intensive and labor-based methods were reduced after delayed strengthening. devaluation as the following table indicates: 29. There are three levels of intervention: federal, state, and local government. New construction is carried out by the federal Construction Costs per Kilometer government through DFRRI, rehabilitation and mechanical main- 1983 1987 tenance mainly at the state level by the ADPs, and routine (U$1 = N 067) (U$19 N 4.25) maintenance primarily by the LGAs with assistance from the ADPs and NDE Ifederal level) in a pilot program. Equipment-intensive Kaduna ADP 30. February 1989, Supervision Report. Memo from L. Campbell Local Costs Us 24,160 = N 16,000 Us 3.760 = N 16,000 to A. Seth on Thematic Supervision-Feeder Roads and Water Foreign Costs U$ 12,440 = N 8,240 U$12,440 = N 52,900 Supply Components of ADPs.' Total Us 36,600 = N 24,240 US 16,200 = N 68,900 31. Even though appraisal mentioned the road maintenance Labor-Intensive PPWP unit, it was not implemented in most statewide ADPs. Local Costs US 32.220 = N 21,600 Us 5.080 = N 21,600 32. "Suitable roads engineers were never found for Funtua and Foreign Costs Us 1,270 = N 850 Us 1,270 = N 5,400 Gombe," PPAR, p. 58. Total US 33,490 = N 22,450 US 6,350 = N 27,000 33. See Lele 1989. 34. February 1989, Supervision Report. Memo from L. Campbell to A. Seth on "Thematic Supervision-Feeder Roads and Water 44. World Bank 1988c; and "DFRRI: Facts and Fiction," in News- Supply Components of ADPs," p. 2. watch, March 28, 1988. 35. February 1989 Supervision Report. Memo from L. Campbell 45. The costs per kilometer for DFRRI in Table 14 are set by the to A. Seth on "Thematic Supervision-Feeder Roads and Water Directorate to allocate funds to each state, and include only "out Supply Components of ADPs." of pocket" costs excluding equipment depreciation. However, the 36. Periodic maintenance (approximately every 5-7 years) actual cost of road construction is not available, particularly since includes among others the following activities: regravelling, there have been some problems with the reports on the compacting, grading, reshaping of the road's camber, and ditch construction achievements in each state. A recent article inter- repair. Recurrent maintenance (once or twice a year depending on viewed the chairman of an LGA in Bendel state saying that only traffic, soil, and rainfall patterns) includes mechanical grading. traces are found of the roads claimed to have been constructed 37. Daily Times, May 30, 1989. in his LGA (Newswatch, March 28, 1988). Certain areas have been 38. A, simple regression model was estimated for length of bul4dozed and others graded butthe roadsare not usable. federal roads in each region as a function of area in each region. 46. Above upper bound targets but almost identical to the The results give good fit (R2 = 0.89) and highly significant target road length to be maintained regularly. coefficients. 39. These targets are intended to be lower bound densities. The fact that targets should be higher for the North than for the Middle Belt, given its higher population densities and much greater agricultural production and rural population, will be taken into consideration in the calculation of upper bound densities in the following sections. 25 Bibliography Ahmed, R. and M. Hossain. 1988. Infrastructure and Development of the Idachaba, ES. 1980. The Green Revolution: A Food Production Plan. Rural Economy of Bangladesh. International Food Policy Lagos, Nigeria: Federal Ministry of Agriculture. Research Institute for USAID, Dhaka. Idachaba, ES. 1981. Rural Infrastructures in Nigeria. Federal Depart- Airey, A. 1980. The Role of Feeder Roads in Promoting Rural Change in ment of Rural Development, Ibadan University Press. Eastern Sierra Leone. London: Ministry of Overseas Develop- Ikporukpo, C.O. 1987. "The Management of the Rural Transporta- ment (ODA). tion Problem." Quarterly Journal of Administration (January). Ife- Airey, T. 1984. Feeder R"2ds and Rural Development in Africa: The Case Ife: University of Ife. of Sierra Leone. Centre for Development Studies, Univ. of International Monetary Fund, 1987. Government Financial Statistics Swansea. Yearbook. Volume XI, Washington. Anderson, G.W. and C.G. Vandevoort. 1982. Rural Roads Evaluation Kingdom of Thailand. 1980. Studies of National and Provincial Road Summary Report. A.l.D. Program Evaluation Report No. 5. Network in Thailand. Louis Berger Int. Washington, D.C.: USAID. Lele, U 1987. "Agriculture and Infrastructure." Paper presented at Beenhakker, H. 1983. Economic Appraisal of Rural Roads, Simplified the Symposium on Transportation and Structural Adjust- Operational Procedures for Screening and Appraisal. World Bank ment organized by the World Bank Transportation Depart- Staff Working Paper, 610. Washington, D.C.: World Bank. ment, May 6-8, 1987. Binswanger, H. et al. 1989. The Impact of Infrastructure and Financial 1. 968. "The Traders of Sholapur" in J. Mellor et al., Developing Institutions on Agricultural Output and Investment in India. The Rural India. Ithaca: Cornell University Press. World Bank, WPS 163. et al. 1989. Nigeria's Economic Development, Agriculture's Role, and Blaikie, P. et al. 1977. The Effects of Roads in West Central NepaL Part World Bank Assistance, 1961-88: Lessons for the Future. MADIA 1, Summary. Ministry of Overseas Development. Paper. Washington, D.C.: World Bank. Bovill, D.l.N. 1978. Rural Road Appraisal Methods for Developing Mitchell, B. and X Rakotonirina. 1977. The Impact of the Andapa- Countries. TRRL Supplementary Report 395, Crowthorne, UK. Sambava Road: a Socio-Economic Study of the Andapa Basin, Carnemark, C. 1984. The Economic Analysis of Rural Road Projects. Madagascar. Transportation Department. World Bank Staff Working Papers, No. 241. Washington, D.C.: Moerman M. 1968. Agricultural Change and Peasant Choice in a Thai World Bank. Village. Berkeley: University of California Press. _ 1979. Mexico: Case Study of a Rural Roads Program and its Nairn, B. 1981. Financing Roads and Bridges. Government of Indone- Interface with the Local Communities. Washington, D.C.: World sia, Sectoral Study No. 4. Bank. Narayana, N.S. et al. 1988. "Rural Works Programs in India: Costs Derkota, B. 1980. Government Expenditure on Road Transport in Nepal. and Benefits." Journal of Development Economics (September). Centre for Economic Development, Katmandu. Ndulu, B. 1979. The Role of Transport in Agricultural Production. Devres Inc. 1980. Socio Economic and Environmental Impacts of Low Unpublished dissertation, Northwestern University. Volume Rural Roads, A Review of the Literature. A.I.D. Program Ogundana, B. 1973. "The Transport Constraint on Rural Develop- Evaluation Discussion Paper No. 7. ment in Nigeria" in Rural Development in Nigeria. Proceedings Duroiaiye, B.O. and Aihonsu, 1. 1988. "Market Integration and of the 1972 Annual Conference of the Nigerian Economic Seasonal Prices of Foodstuffs: A Case of Ogun State, Society, University of Ibadan. Nigeria." Food Policy 13, No. 4 (November). Okada, F: 1978. Rural Works Project Padat Karya Gaya Baru: Socio- Economist Intelligence Unit. 1988. Nigeria Country Profile. United Economic Assessment. USAID Kingdom. Richards, P. 1984. "The Economic Context of Rural Roads" in P. Edmonds, G.A. 1983. "Rural Transport Policy in Developing Richards, ed., Rural Roads and Poverty Alleviaton. Intermediate Countries" in Highway Investment in Developing Countries. Technology Publications. Proceedings of a Conference of the Commission of the Robinson, R. 1988. A View of Road Maintenance Economics, Policy and European Communities. London: Thomas Telford Ltd. Management in Developing Countries. TRRL Research Report Federal Republic of Nigeria. 1981. Fourth National Development Plan 145, Crowthorne, United Kingdom. 1981-1985. Lagos: Federal Ministry of National Planning. The World Bank. SARs, PCRs, PPARs, various transport and Government of India. 1981. Bask Road Statistics of India. New Delhi: agricultural projects, 1970-1989. Transport Research Division. The World Bank. 1985. Road Deterioration in Developing Countries. A Hine, I.L. 1982. Road Planning for Rural Development in Developing World Bank Policy Study. Washington, D.C.: World Bank. Countries: A Review of Country Experience. TRRL Laboratory The World Bank. 1988a. Country Economic Memorandum. Africa Report 1046. Crowthorne, United Kingdom. Region. Hegen, E. 1966. Highways into the Upper Amazon Basin, Pioneer Lands The World Bank. 1988b. Nigeria Highway Sector Loan. SAR, Western in Southern Colombia, Ecuador and Northern Peru. Gainesville: Africa Department. University of Florida Press. The World Bank. 1988c. Nigeria: Proposed Rural Infrastructure Project Howe, 1. 1984. "The Impact of Rural Roads on Poverty Alleviation: ProTect Brief. A Review of the Literature" in P Richards, ed., Rural Roads PrO Brie. and Poverty Alleviation. London: Intermediate Technology UNDP/ILO. 1988. Pilot Public Works Programme. Volume 1, Main Publications. Preparation Report National Directorate of Emptoyment. 26 Annex 1. Rural Road Benefits Bingswanger et al. (1989), for example, presented a compre- to very strong competition, and therefore has the potential hensive model in which the impact of roads and other rural to increase the strength of rural-urban linkages. Transporta- infrastructure on agricultural productivity is studied. This tion improvements can thus bring on a move away from impact is examined as it affects production, both in village level subsistence toward greater specialization of conjunction with other factors as well as separately. The urban and rural areas. Also within rural areas so far as framework presented in Figure I explores also the impact different crops are concerned, a process occurs which can of rural infrastructure on private investments and on be seen as increasing modernization of agriculture (Rich- increased access to credit.' An extensive analysis was ards 1984). conducted in 85 districts in India with data for 1961 to 1981. When improvements in road transport (e.g., a road Improved road investment was found to increase agricul- rehabilitation) change the relative returns from labor and tural output (with roads contributing directly for seven land, the potential for increases in production and con- percent each to the growth of agricultural output and sumption are greater if both land and labor availability are fertilizer demand), with roads also contributing significantly distributed equitably, and if institutions (e.g., markets) allow to bank expansion. The conslusions suggest that the major for such changes. Larger farmers are more capable of effects are not via an impact on private investment but mobilizing factors of production in response to changes, rather through improved marketing opportunities and and are therefore more able to appropriate a greater reduced transaction costs of all sorts. portion of the benefits. Where levels of poverty and All these changes, however, vary according to local available resources are such that effective demand for conditions. Exchange of agricultural products increase at means of transport is low and few intermediate means of existing levels of technology or at improved levels when not transport are available, the best road will have little more only land but also complementary resources are available merit than a footpath (Richards 1984). through an improved quality of agricultural research (e.g., Finally, local conditions (e.g., labor market) influence the adaptative trials), increased access to extension, and choice of road construction and maintenance technology increased supply of inputs. At low levels of technology which in turn affects the extent of impacts of road improve- price-based changes are likely to do less for increases in ment projects. There is a whole range of technologies exchange than changes in technology. Indeed evidence varying from equipment-intensive to labor-intensive meth- suggests that in resource poor areas where scope for ods, all of which have different effects depending on their technical change is limited the majority of increase in use of resources. The experience with labor intensive demand for transport is for personal travel (Hine 1982). construction methods has, for example, demonstrated that One of the main effects of improved transport is its role low volume roads can be constructed and maintained as an accessory to agricultural modernization programs that wherever labor is abundant and available, even seasonally, encourage the introduction of mechanization, the use of increasing rural incomes even with the existence of institu- chemical fertilizers, and high yield varieties. Adequate tional inefficiencies. Construction technology and the levels freight transport is necessary to ensure timely arrival of of agricultural production affect construction standards. inputs at low transport costs. However, not only agricultural With lower standards, for example, more regular mainte- inputs find their way in. Rural areas are also opened to an nance is usually required, and better local institutional inflow of consumer goods proportional to the purchasing capacity is therefore desirable to maintain the road. Hence, power, increasing the demand for imported and urban the standards used are important determinants of the produced goods, and expanding the market for national extent to which the rural population can actually use the manufactures. This usually makes local artisans vulnerable related infrastructure. Figure 1 Agricultural Opportunities Private _ -Agrocnimatic eeInvestmentsst Endowments (2) , and Input Use -New Technology (4 |_i i(5 Financa (7 Pr and | Policies Rural (b) (111 FInfrastructure _ (1) 2 27 Annex 2. Review of the Road Components in ADPs 2.1. Review of Views on Construction Targets at alignments which would be needed for the fadama roads, Appraisal secondary roads would have to be built to provide access In the early enclave projects, rural road construction to villages cut off from the main roads as a result of these received a strong emphasis along with input distribution alignments.' Thus, it was proposed that "the project would and extension in as much as it was seen to facilitate these assist LGCs in undertaking these road networks with their activities and provide marketing opportunities for the own resources. In this way, LGCs would benefit from project expected increases in agricultural production. As the PCR expertise and gain valuable experience and a sense of for Gusau, Funtua, and Gombe noted, there were very few involvement in carrying out their own road construction rural roads in the areas covered by these projects at the program." Finally, the appraisal reports for all three time of their initiation. Moreover, the roads that did exist statewide ADPs indicated that the annual rural road were not only in poor condition but they could be used program "including justification, length and location" would only during the dry season. Thus, a declared goal of these be prepared by each project in consultation with the state projects was ". . . to bring most rural farmers within two ministries and LGAs,"and would be subject to an annual kilometers of a road" (PPAR for Gusau, Funtua, and Gombe, action plan to be agreed with the Bank prior to the p. 28). As a result, the highest targets of rural road densities beginning of each calendar year." of all the ADP projects were set, as seen in Table 5. These In the SAR for the Southern Borno ADP, the Bank has targets were subsequently replaced by less ambitious been forthright about the undesirability of undertaking new targets in the larger statewide projects. rural road construction in LGA jurisdictions because of the Since the conception of these ADPs in the early 1970s, inability of the LGAs to maintain them. Both SARs for the the existing local capacity for rural road building was Southern Borno ADP (June 23, 1986) and the first MSADP identified as a persistent constraint. However, as Asif Faiz (june 3, 1986) were released at about the same time, and noted on an earlier draft "in these three projects there yet in the case of the MSADP the Bank has seen a role for never was a serious attempt to mobilize small contractors the LGAs as discussed below. In fact, in the SAR for the and local residents as was done in the Rural Roads Access Southern Borno ADP, the Bank alluded to the futility of Programme in Kenya." The construction task was entrusted constructing new roads in LGA jurisdictions. As that SAR (p. to field construction units managed and operated by the 9) observed, projects themselves. Experience with existing ADPs has shown that local Lafia and Ayangba ADPs set lower construction targets, government agencies assigned the responsibility for and with the same reasoning of shortage of local building maintaining new roads constructed under the projects capacity, the actual construction was performed by project have neither the financial nor physical resources to construction units. In this context the PPAR for Lafia and fulfill this function. A similar situation exists in Borno Ayangba (p.20) cites a revealing criticism, namely that State; the existing Local Government Authority (LGA) "unless the road components were used to build up local road network is not being maintained and there road expertise through local contracting, these would hardly appears no likelihood that this situation will improve. leave road building institutional development any heritage Consequently, further additions to this system are not at all." justified. In the first statewide ADPs (Bauchi, Kano, and Sokoto), Thus, the road program envisaged by the Bank under the the proposed road-building intensity was much less than Southern Borno project focuses on tracks ". . . which are that actually achieved in the earlier enclave projects. For maintained by local communities and has not assumed that instance, the density of new rural roads targeted for the LGAs would fulfill this role" (SAR for Southern Borno ADP, Bauchi ADP (BSADP) was 18 meters per sqare kilometer of p. 5). In brief, the main objectives of this program are to the project area compared to 77 meters per square undertake improvements on about 1,000 kilometers of kilometer actually achieved at Gombe, the enclave project tracks built and maintained by local communities so as to that preceded Bauchi (see Table 5). in part this is because make them more accessible during the wet season. In this, only surfaced, all-weather roads which could withstand wear the project would cooperate with the local communities to and tear better were to be built in all three statewide ADPs. identify the required improvements as well as provide As in the case of the enclave projects, road construction materials and assistance for the actual work. The project was to be performed by the projects themselves. Given the would also perform spot improvements on about 150 large size of the projects that was incumbent upon their kilometers of LGA feeder roads, and construct 135 kilome- statewide nature, each project was divided into zones. Each ters of new access tracks to the fadamas. These new tracks of these zones was to have its own road construction unit would be maintained by the project during the project supervised by a zonal roads engineer and a mechanical implementation period and for four years after the comple- engineer. The recruitment of engineers, as well as the tion of the project; thereafter, the responsibility for their purchase of road construction and maintenance equipment maintenance would be transferred to local communities. (through ICB), was to be done internationally. In line with the shift of emphasis in the statewide ADPs Although it is not clear from a reading of the appraisal from road construction to road rehabilitation and mainte- report for Kano whether this applies to it also, at least in nance, the first Multistate ADP (covering 7 states) focused the case of Bauchi and Sokoto, the road programs were to on rehabilitation and maintenance from the outset. How- be phased so as to allow for preconstruction training given ever, at about the same time that the MSADP commenced, the shortage of experienced, local road construction the Directorate of Food, Roads, and Rural infrastructure operators and mechanics. In addition, the appraisal report (DFRRI) was also established in the president's office with for Sokoto (p. 38) indicated that because of the watershed counterpart institutions in the state governments. judging 28 from early supervision mission reports for the MSADP, there clearly limits the amount of work done under force account is a clear division of functions between the project and the to 20 percent, as a means to ensure maximum efforts to DFRRI inasmuch as the former will undertake only rehabil- develop the work through contractors. In addition to the itation and maintenance work and the latter only the cost benefit analysis estimating as benefits the road user construction of new roads. For instance, a supervision savings and increases in production, the SAR states that the mission report for the Anambra state component of the roads to be rehabilitated should have traffic of more than MSADP (February 18, 1987) noted that "it was agreed that 25 vehicles per day. there is no conflict between the road program envisaged by Most of the above discussion has been about densities the ADP and the DFRRI. The DFRRI would concentrate on at the state level. Much less is known about the distribution construction of new roads and the ADP would focus on of target densities within subareas of each state, at the LGA rehabilitation and maintenance of existing roads." level. Unfortunately, with available data this exercise is not In much the same way as for the earlier ADPs, an possible. As an example, an examination of the plans for objective of the first MSADP road program, as stated in the the first two years of the Benue ADP (part of the First project SAR, is to provide farmers with all-weather access to MSADP) indicates that there is a more or less balanced service centers for the purchase of inputs and the sale of distribution of roads among the 23 LGAs in the state.3 Each output. But in addition, the SAR (p. 21) assigns to this year's program includes approximately 10 of the 23 LGAs in program two other objectives which are of an institutional the state. The only concentration of roads to be rehabili- nature, i.e., to ". . . begin to build a permanent capacity in tated (as a proportion of total kilometers) is observed in the LGCs to maintain their roads, and . . . strengthen the those LGAs which are in a radius of 65 kilometers from Nigerian institutional capability to prepare, implement, and Makurdi, the state's capital. monitor execution of a feeder road program." The aim of the rehabilitation and spot improvement 2.2. Review of Views on Maintenance at program is to improve road success by undertaking work Appraisal especially along those stretches of roads most prone to The maintenance of the rural roads constructed under the disrepair (e.g., water crossings and areas with poor drain- early enclave projects, as the SARs indicated, was to be the age). But unlike the earlier ADPs which have preferred to responsibility of the local government authorities (LGAs). In perform road construction and maintenance tasks them- this context, the SAR for Funtua (Annex 4, p. 41 observed selves through project units, "the preferred method of that while "some maintenance would be undertaken by the execution is by contract." The SAR for the first MSADP project during the development period, the allocation of noted that studies were underway in the various states construction plant for maintenance would not be allowed to included in the project to assess the availability of local interfere with output of new road; thus once a new road has contractors, and to consider the relative efficiency of using been built maintenance responsibility would be handed them as opposed to the project undertaking the work itself. over to the Katsina Local Authority." Funds were to be It was suggested that only in cases where local contractors allocated under the projects for the purchase and operation were not available would the project carry out the work of maintenance equipment. Also, in the case of the first itself. three projects, it was proposed that staff of the local The project SAR states that ". . . the participating states, government authorities would be seconded to the projects with the exception of Plateau and part of Benue, have for "training and operational control." adequate ... LGA roads constructed to varying standards." Once again, in the Lafia and Ayangba ADPs also, road In spite of this, the road rehabilitation targets were set at maintenance was to be entrusted to the LGAs. However, 600 kilometers spread over four years in each of the seven unlike in Gusau, Funtua, and Gombe where the intent at MSADP states, and no further provisions were made to appraisal was to hand over the roads to the LGAs almost as improve, if needed, the densities in Plateau and Benue. As soon as they were built-i.e., even while the projects were Table 6 shows, the resulting densities of road rehabilitation still functioning-this time it was to be done on completion vary widely from state to state but are in line with densities of the projects. In relation to this, the SARs for Lafia and for road construction in previous ADPs (see Table 5). The Ayangba (pp. 11-12 and p. 11, respectively) noted in an road selection task was assigned to the Infrastructural identical fashion that "the project would work closely with Services Division to be created in each state ADP, with LGC to develop an adequate road maintenance capacity assistance from FACU. One of these programs prepared for that could be operated by the LGC at the end of the Benue state criticizes the road rehabilitation target, set project period." Indeed, during the period of their exist- equal for all states, because in order to achieve the annual ence, these two projects were expected to not only rehabilitation target included in the SAR, some of the maintain the roads constructed by them but also the selected investments will have economic rates of return nonproject roads in their jurisdictions, and for the first time below the 15 percent required.2 the SAR for Ayangba contained targets for routine and The second MSADP includes the states of Niger, Kwara, periodic maintenance. and Gongola which have some of the lowest densities of The staff appraisal reports for Bauchi, Kano, and Sokoto rural roads in the country. The SAR sets the lowest targets were completed in early 1981. On the other hand, the of roads to rehabilitate per square kilometer of project area Ayangba and Lafia projects did not close until 1984. Thus (see Table 6). Road rehabilitation as in the previous MSADP few lessons pertaining to road maintenance would have was to be carried out "in general" by contractors. While no been available, and available lessons would have been apparent reason for these lower densities is included in the limited largely to those from the completed trio of initial SAR, the rehabilitation targets of the third MSADP return to ADPs (Gusau, Funtua, and Gombe). All the same, it is the level of the first MSADP (as shown in Table 6). The views interesting to note that some of the maintenance arrange- on rehabilitation arrangements at appraisal are the same as ments proposed by the SARs for Bauchi, Kano, and Sokoto in the second MSADP with one minor exception. The SAR conformed to the recommendations contained in the 29 project completion reports for Lafia and Ayangba. indicated that (p. 22) "the project is seen as only the first For instance, according to the SARs for Kano and Sokoto, phase of a longer term program and additional support a road maintenance unit, quite distinct from the road could be provided under a second phase." In this context, construction unit in terms of being assigned its own funding needs and the progress in the development of a equipment and staff, was to be set up in each project zone locdl road maintenance capacity were to be reviewed every (this was recommended by the project completion report year on a joint basis by the state governments concerned, for Lafia also). Although distinct road maintenance units the LGCs, the project management, and FACU engineers. were proposed in the case of Kano and Sokoto, this was not The SAR for the second MSADP (May 1988), as in the case so in the case of Bauchi.4 of the first MSADP, also divides the responsibility for The SAR for Bauchi indicated that zonal construction maintenance operations. In this case the project funds units would maintain the new roads for a period of one year recurrent maintenance and portions of routine mainte- after construction before relinquishing them to the LGCs for nance. Recurrent maintenance will be performed by a force future maintenance. Nevertheless, the SAR also made it account mechanical maintenance unit in each state, which clear that if the need arose, the Bauchi road construction will be for rental to the LGCs, but will not be subject to units would be prepared to maintain these roads for the contract. The costs of these units inclusive of overhead entire period of the project. Because of their ultimate costs will be monitored by FACU and compared to private responsibility for road maintenance, it was agreed upon contractor costs. The LGCs and ADP would jointly agree on between the Bauchi state government and the Bank during an annual programme, defining the extent of force account negotiations that the LGCs would provide funds and staff and contract works. The project will finance initial training for all road maintenance work carried out by the project (no of LGC engineering staff and provide each LGC with such financial arrangement was mentioned by the SARs for necessary hand tools for routine maintenance. Routine Kano and Sokoto). As opposed to Bauchi, in Sokoto's case, operations would be manual and carried out with the all roads built under the project were expected from the lengthman system. One lengthman would be assigned for outset to be maintained by the proposed zonal mainte- each three to four kilometers of road, and a headman for nance units through the entire period of the project. On the every eight lengthmen. The LGCs would pay the lengthmen other hand, in the case of Kano, it was indicated that traffic and headmen and be reimbursed for the work done after monitoring studies were to be performed on the project a supervision from each ADP The maintenance targets in roads, and the responsibility for the maintenance of those the second MSADP cover a larger proportion of existing roads used by more than 100 vehicles a day was to be roads (see Table 6). transferred to the State Ministry of Works and Highways; Finally, the SAR for the third MSADP (draft October 1988) although it is not clear from the SAR, it seems that those assigns, as in previous projects, recurrent maintenance to project roads carrying less than 100 vehicles a day were to be done by contractors on an annual basis awarded under be maintained by the project maintenance units for the competitive bidding, and routine maintenance under the entire project period. LGCs. The maintenance targets for this project are higher in All three statewide ADPs were to provide training to LGC absolute numbers per state, than all targets in preceding staff in order to prepare them for their future responsibil- projects (see Table 6). As a proportion of all rural roads in ities with respect to road maintenance. In this context, the the state, the targets are similar to that of the previous SAR for Sokoto specifically referred to the inclusion of a MSADP The SAR has no specific mention of the actions road maintenance training engineer in that project to recommended from each LGC with reference to the rest of identify the training needs of the LGC staff. Finally, in the the network. case of all three projects, the SARs suggested that on 2.3. Review of Rural Road Standards completion of the projects the zonal road maintenance units (the road construction unit in Bauchi's case) could The early enclave projects intended to build roads accord- serve as the basis for the creation of joint zonal LGC boards ing to one uniform standard of minimum specifications. for road maintenance and construction. The view was that These roads were supposed to be unsurfaced, with a right this would not only be more efficient than each LGC having of way of 12 meters, and a carriage-width of 5 meters. Soon a separate maintenance unit but it would also vitiate the after initiation, the PPAR states, the three projects decided need to divide among LGC's the zonal equipment accumu- to build only surfaced roads since "the roads will generate lated under the project (again, this was recommended by much traffic and the vehicles using them were heavier than the PCR for Lafia also). expected."' Other than this change, the specifications Unlike the previous projects, the SAR for the first MSADP followed closely the SAR. divides responsibility for maintenance operations. Recur- The Lafia ADP changed deliberately the standards after rent maintenance,5 was assigned to the project, and initiation, increasing the carriage-width to 6 meters, and periodic maintenance6 to the LGCs. As the SAR states, "in constructing two types of roads to meet different levels of each state one mobile unit is established to carry out traffic: (I) a type I road with laterite surface and culverts for biannual recurrent maintenance on 400 kilometers of roads. roads with high traffic, and (2) a type 11 road with no The LGCs would be closely involved in their planning and surfacing or culverts. In this project as in Ayangba no control implementation. There Lil also be close liaison with LGCs of surfacing thickness or quality of surfacing was conducted. to establish routine maintenance capability on about 1,000 Specifications in these enclave projects complied with km of roads in each state." The training of LGC personnel the project design, but the completion report states that as well as supervision and equipment was to be procured there was a wide range for their acceptance. For example, by the projectd the carriage-width increased one meter in the five early Some uncertainty was expressed by the SAR about the enclave projects, while it was reduced in the Ekiti-Akoko ability of the LGCs to maintain roads after the project ADP The same is true of the camber, i.e., the slope of the period given their financial problems. It was, however, section between the road center line and the edge, which varied between 2 and 5 percent in these ADPs. 30 Finally, the multistate ADP projects have adopted stan- MSADP in which "site inquiries by the appraisal mission in dards that are equal to those of the Ministry of Works, and agricultural areas generally indicated average daily traffics that adapt to the variability of traffic among different rural (ADT) of 30 to 35 vehicles on market days and normal of 15 roads. These standards have three versions, according to to 20." The differences in traffic volumes are encountered the level of traffic, including: (1) farm access roads (up to 50 simply because often roads that carry larger volumes of vehicles per day), (2) feeder roads (up to 100), and (3) LGA traffic and should be classified as secondary roads for roads. There is no specific mention as to how much these administrative reasons end up being classified as a rural standards have been achieved, and how quality control is road to be included in the rehabilitation program. This is supervised in the more recent projects. common given that the definition of rural road is too broad, The actual standards achieved in the first MSADP vary particularly if no traffic data is collected. unnecessarily between projects and exceed the standards set.8 As an illustration, the actual pavement width of some 2.5. Review of Indirct Benefits of the roads constructed varied between 7 and 10 meters The ex-post evaluation in the PPAR for Gusau, Funtua, and against a suggested width of 4.5 meters, a difference that Gombe considered the road construction program to have leads to a substantial increase in rehabilitation and been successful as it both ". . . provided each project with maintenance costs. In the future the Bank expects to correct the infrastructure required to implement project activities this, reimbursing against standards previously set and and ... very important transportation access to the village therefore ignoring excesses. farmers" (p.31). Indeed, in the context of its calculations of the ex-post economic rates of return the PCR (p.48) 2.4. lTafflc Counts observed that the computed rates could be significantly First, road impact evaluation should vary given the wide raised by the inclusion of indirect benefits in addition to variation of traffic levels discussed below (see section 2.3.). the direct ones associated with production increases."" Second, a review of the completion and appraisal docu- The most notable among these benefits were felt to be ments shows that traffic counts in the earlier projects were "the benefits derived from road construction which are only done sporadically, and never in a way that allowed a manifested in the extended life of motor vehicles, lower systematic evaluation of changes in road use. The PPAR for prices of consumer goods, lower operating costs for Ipas- Gusau, Funtua, and Gombe (p.32) mentions only one road sengerl transport, landl value of time saved Ito count where 400 movements were observed in a road in passengersl." one week. In the same way, the PPAR for Lafia and Ayangba The PPAR for Lafia and Ayangba reflected also on the mentions that "all the formal economic analyses . . . deficiency of the economic analyses performed for these overlooked the principal contribution of their feeder roads: projects to take account of the benefits of the constructed namely road user cost savings." The PCR for Ekiti-Akoko rural roads. In addition to overlooking road user cost states that "benefits from roads were not recognized" in the benefits, as mentioned in the previous section, "all the evaluation. Post project traffic was collected for the Bida formal economic analyses . .. overlooked the stimulus to ADP roads. Since these roads are now part of the Niger production and trade of all kinds" (p.8). The PPAR conse- state ADP project, these traffic levels will be used for the quently decided to retain the reestimated project ERRs of future project evaluation. The Bank criticizes the methodol- 6 percent for both Lafia and Ayangba because ". . . ogy of the counts stating that the "data is suspect."9 uncounted road benefits could well offset the discount in For the more recent ADPs, FACU has collected traffic data crop production benefits introduced by the Bank in the on some roads before and after improvements. However, PCR overview." If the ex-post rate of return is 6 percent and comparisons for the same road have not been possible in the technology and extension components are considered most cases, because the surveys are conducted on roads to have yielded no return, then the share of roads in the which are later not included in the road programs, or have estimated 6 percent rate of return is obviously much higher not been included in the road programs to date. For than 6 percent. example, the road selection for Benue State (First MSADP) Finally, a recent supervision mission reports that as a states that "traffic levels on roads rehabilitated are not result of road rehabilitation in the Kaduna ADP a survey known for most roads, despite the fact that a traffic count "showed lal reduction in the cost actually charged for study was commissioned by FACU. However, the roads transportation of farm produce by pickup or minibus 133 covered in the survey do not coincide with the areas of and 48 percent, respectivelyl . . . with corresponding influence of the candidate roads in the first year plan in reduction in passenger fares."'2 these states."l0 Probably in view of the difficulties encoun- tered in conducting road counts, the Bank hired a consult- 2.6. Review of Maintenance Achievements ant to conduct appropriate counts on the roads to be The view at the time of the appraisal of the early enclave included in the second MSADP in Kwara and Niger state. projects, as noted above, was that all types of maintenance Available traffic counts indicate that traffic levels vary were the responsibility of the LGAs. However, in terms of widely. For the BIDA ADP roads included in the Niger State the actual experience of the projects in relation to this ADP (first MSADP), observed traffic varied from 15 to 124 question, the PPAR (p. 40) noted that the LGAs ". . . have vehicles per day, which clearly denotes the need to include been deficient in funding and staff resources, and have had road user savings in the busier roads, priorities which do not include maintenance of project-built From the consultants counts for the second MSADP, the facilities" (PPAR, p. 40). More specifically (p. 40), results yield even larger variations of traffic in the sup- At Gusau, LGCs have begun contributing to the costs posedly rural roads. In Niger state, five out of ten roads of the maintenance being carried out by the project surveyed had daily traffic ranging from 174 to 658 vehicles unit, while at Gombe the LGCs have been doing some per day, while in Kwara state two roads had traffic of 350 maintenance work themselves, though this has not and 970 vehicles per day. These large volumes of traffic are been adequate. LGCs in the Funtua area have not yet quite distinct from those expected in the PPAR for the first accepted responsibility for maintenance. 31 In the case of Gusau, the intention seems to have been that devote to maintenance. the LGAs would contribute financially to the project for the Thus, the PPAR for Lafia and Ayangba (p. 3) concluded maintenance work until such a time that a zonal mainte- that the limited efforts of these projects with respect to nance unit could be created to serve the needs of a maintenance, "together with the absence of specific post- number of LGAs. What in fact happened was that the project arrangements for maintenance and the poor project road construction unit continued to exist and resources of the Local Government Authorities (LGAs), have perform the maintenance functions even after the project left roads in a deteriorating condition soon after project itself closed, i.e., until the start of the Sokoto statewide ADP completion." In relation to this, the PCR for Lafia (p. 71) which then took over the responsibility for maintenance of provided the following advice for future projects: the roads constructed at Gusau. The feeder road maintenance problem is completely The road component was considered to constitute the predictable and should be taken into account in only successful aspect of Ayangba and Lafia and substantial designing future projects. First, feeder roads should benefits were associated with it during the project period. be built to higher standards initially to minimize The PPAR (p. 8) for these projects expressed uncertainty subsequent maintenance requirements. . . Second, about the future flow of such benefits "because of the projects should set up separate maintenance sections unresolved problems of maintenance." As noted above, it within the roads unit which could continue with state had been proposed at appraisal that the projects would or local government funding after the initial project collaborate with the LGAs on maintenance in order to investment period; keeping a functioning unit prepare them to assume full responsibility for it on together makes more sense than dividing equipment completion of the projects. But as the PCR for Lafia (p. 71) among a number of LGC's. observed, ". . . joint maintenance program with LCG's have The road maintenance arrangements that had been not proved possible. The LCG's ... simply lack the staff and envisaged in the SARs for the statewide projects were resources to participate effectively." The PCR for Ayangba proving to be unrealistic in all cases as indicated by the (p. 143) further noted that "The SAR proposal that the roads midterm reviews. For instance, in the case of Bauchi, the should be handed over to the LGCs is clearly impractical decision at appraisal to hand over roads to the LGCs, after and the only sensible solution would seem to be the they had been maintained by the project itself for an initial project unit remaining operational and funded for mainte- period of one year, did not materialize. The project found nance." This indeed happened at Gusau as mentioned that the LGCs did not have the funds to do this and earlier. consequently initiated arrangements whereby it acted as a Concern about post-project maintenance of roads came "maintenance contractor" to LGCs. to the fore early in the case of Lafia and Ayangba because In the same way, the intention at appraisal in the case of signs of deterioration emerged almost as soon as the Kano to entrust the maintenance responsibility for all roads projects were completed. Certain features of the road carr- ing over 100 vehicles a day to the State Ministry of construction programs pursued by these projects contrib- Works and Highways was also proving difficult to imple- uted to this early deterioration. The features in question ment. As the midterm review for Kano (p. 19) noted "The can be summarized on the basis of the PPAR for these transferring of responsibility for maintenance of roads is projects: usually a slow and often difficult task particularly when the First, the projects were deficient in meeting their road agencies accepting the responsibility do not have adequate maintenance targets although they exceeded their targets equipment and funding." Thus, as was already being for road construction. For instance, at Lafia, little mainte- initiated at Bauchi, the midterm review for Kano recom- nance was attempted in the first three years, while in the mended that the project enter into contractual agreements last two years, the financial difficulties faced by the project for performing the required maintenance work itself and be restricted the amount of maintenance that could be paid for it. Lastly, against the appraisal view that the roads undertaken. constructed at Sokoto would be relinquished to the LGAs Second, it appears that in exceeding their road construc- for maintenance on completion of the project, the midterm tion targets, the projects might have sacrificed quality. As review for Sokoto (p. 15) noted that ". . . project officials are the PCR for Ayangba (p. 142) noted ". . . the rapid progress" aware that LGAs will not be able to take over the mainte- made by the road construction program "was not without a nance because of lack of resources." It was indicated that cost. Technical standards were changed and the resulting the project was making arrangements to do that mainte- quality is not as good as should have been attained." On nance work itself. the other hand, in the case of Lafia, as mentioned in the The reports of recent Bank supervision missions provide last section, a significant portion of the total road length some idea of the experience of these statewide ADPs with constructed consisted of unsurfaced roads. Interestingly respect to road maintenance since the aforementioned enough, the PCR for Lafia (p. 71) pointed out that given the midterm reviews were completed (i.e., May/Decem- excess capacity of the project road-building plant, all roads ber1985). Following the general conclusion of the midterm could have been surfaced at little extra cost. reviews that the LGCs did not have the capability to carry Finally, despite the intention at appraisal to establish a out the required maintenance, these supervision reports separate road maintenance unit in each project, as distinct underscore a growing concern within the Bank about the from a road construction unit, this was not done in practice. future maintenance of the project-constructed roads. In this The road construction unit was expected to meet the vein, a supervision mission report for Bauchi (December 3, maintenance requirements also. This might explain why, for 1986, p. 4), while acknowledging the project management's instance, at Lafia, very little maintenance work was under- view that the project could exceed its road construction taken in the first three years of the project; it was during target by a considerable margin (i.e., by even more than these years that the majority of the roads were built and what was anticipated at the time of its midterm review), the road construction unit probably had little time to reiterated the Bank's concern that the capital asset 32 developed under Gombe ADP and BSADP should not be sacrificedl to an over-ambitious new construction Annex 3. Cameroon Rural Road programme." Supervision mission reports for Kano and Sokoto suggest Target Densities that in these projects, which appeared unlikely at the time of the midterm reviews to achieve their road construction targets before closure, the emphasis had shifted from For comparison purposes the regions in Cameroon are construction to maintenance. To quote a supervision broadly as follows: the Northern Plains (North and Far mission report for Kano (December 3, 1986, p. 6), "It was North provinces) have predominantly Sudan Savanna vege- agreed by the State Government and KNARDA that the tation, rainfall between 400-1,000 millimeters, and popula- authority would in future concentrate its efforts on the tion densities of 25 persons per square kilometer. In the rehabilitation and maintenance of roads and not on new Central Savanna (Adamaoua province), where Sudano- construction." Similarly, in a supervision mission report for Sahelian Savanna predominates, rainfall averages 1,400- Sokoto (December 3, 1986, p. 9), it was observed that ". . . 1,700 millimeters per year, and the population densities are the mission fully supported management's emphasis on low (about 7 persons per square kilometer). The Western road maintenance as the main focus of the project's road Highlands, with Sudano-Guinean vegetation, fertile soils, program." This same supervision mission report for Sokoto heavy rainfall between 1,700-3,000 millimeters, and popula- however, also suggests that the intention at appraisal (1981i tion densities of 82 persons per square kilometer. Finally, to establish distinct zonal units for construction and the Tropical Rainforest (i.e., East, Center, South, Littoral, and maintenance had not been implemented as of December Southwest provinces) has a dense vegetation cover, which 1986. For the report stated that "SARDA is proposing to gives way to sparser Guinea forest and wooded savanna as form four road maintenance units from the existing road one moves away from the coast. This region has heavy construction units, and four new units from equipment to rainfall of over 3,000 millimeters a year in the Coastal be procured . . ." (p. 9). region, and population densities that vary between 80 Given that in addition to a lack of expertise, funding persons per square kilometer in Littoral, to 20 for the region constraints were also considered to constitute an important as a whole. reason for the inability of local authorities to maintain Several projects were reviewed, covering all the provinces roads, the aforementioned supervision mission reports for in the country. The target densities were calculated as in Kano and Sokoto as well as Bauchi indicated a willingness the previous section (see Table below). According to these on the Bank's part to finance some of the recurrent costs calculations, the targets in both countries fall in the same associated with the maintenance work undertaken by the range, between 30 and 120 meters per square kilometer. As projects. Indeed, in Sokoto's case, the supervision mission in the previous analysis of road target densities in Nigeria (p. 10) affirmed that "recently Bank's management has one finds a close correspondence with population densi- agreed that the Bank would be willing to finance on a ties, and accordingly one can see that the highest targets declining basis part of the non-salary recurrent costs of are the ones for the Western Highlands, followed by the SARDAs road maintenance and rehabilitation program." Northern Plains and the Tropical Rainforest. The lowest Nevertheless, all of these supervision mission reports targets are assigned in the sparsely populated Central stressed that Savanna. ...in general the Bank expected the State or Federal Government to cover the recurrent costs of operations Cameroon Regional Average Target Densities like road rehabilitation and maintenance. This was to Meters per square-kilometer ensure that when the Bank loan funds were exhausted the local authorities had already been providing for Project Area Road Target recurrent expenditures and thus the chance of disrup- Kilometers Densities tion in maintenance and rehabilitation was minimized. Northern Plains The achievements of the first MSADP are included in Projet Centre-Nord SODECOTON 24,200 645 26 Table 9. This project was the first to divide maintenance North and Far North Provinces operations, and even though the actual kilometers main- SFRPa 102,600 3,590 35 tained are well below targets in all states, the number of Central Savanna kilometers with recurrent maintenance in one year are Livestock Project 8,000 150 20 higher compared to that of most previous ADPs. Adamaoua Province SFRP 61,990 480 8 "Routine maintenance [in 19871 is practically non-existent Western Highlands in most states at this stage of the project." The ADPs and Western Highlands Project 13,890 900 65 LGAs could not agree on the percentage of expenses to be West and Northwest Provinces reimbursed by the ADP to the LGA and local communities. SFRP 31,190 3,820 122 Meanwhile, the LGAs did not carry out maintenance Southern Tropical Rainforest activities citing lack of funding availability for salaries. This ZAPI East Rural Dev. Project 13,500 540 40 points out the recurrent lack of financial ability of the LGAs Cocoa Project 25,000 950 40 which is still an unresolved critical point under the present Center, South, Littoral, Southwest proposed arrangements. Provinces SFRP 161,260 6,400 40 East Province SFRP 108,900 310 3 33 Annex 4. Rural Road Densitdes in India Roads in India in 1961 Total Roads Area Popu- Total Meters Meters Other Village Rural Meters Meters lation Roads per square per District Roads Roads per square per State sq.km. km kilometer person Roads (6)+(7) kilometer person (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Andra Pradesh 274,674 35,977,999 53,818 196 1.5 2,862 16,985 19,847 72 0.6 Assam 121,984 11,860,059 29,139 239 2.5 529 16,078 16,607 136 1.4 Bihar 174,083 46,457,042 80,555 463 1.7 13,777 28,308 42,085 242 0.9 Gujarat 186,879 20,621,283 24,344 130 1.2 5,281 4,276 9,557 51 0.5 Maharashtra 307,909 39,504,294 50,760 165 1.3 4,310 13,671 17,981 58 0.5 Jammu & Kashmir 222,802 3,583,585 10,461 47 2.9 633 4,561 5,194 23 1.4 Kerala 38,858 16,875,199 19,422 500 1.2 533 6,828 7,361 189 0.4 Madhya Pradesh 443,434 32,394,375 47,229 107 1.5 2,572 11,988 14,560 33 0.4 Madras 129,842 33,650,917 46,448 358 1.4 4,889 13,471 18,360 141 0.5 Mysore 191,976 23,547,081 62,275 324 2.6 3,697 22,912 26,609 139 1.1 Orissa 155,819 17,565,645 31,114 200 1.8 3,038 10,617 13,655 88 0.8 Punjab 121,947 20,298,151 30,190 248 1.5 4,923 9,846 14,769 121 0.7 Rajasthan 342,268 20,146,173 40,982 120 2.0 7,875 11,823 19,698 58 1.0 Uttar Pradesh 293,846 73,752,914 98,304 335 1.3 7,995 43,233 51,228 174 0.7 West Bengal 87,873 34,967,634 64,491 734 1.8 6,200 26,165 32,365 368 0.9 Delhi 1,484 2,644,058 Himachal Pradesh 28,177 1,348,982 Manipur 355 177 532 Tripura 10,453 1,141,492 675 675 Andaman & Nicobar Islands 8,327 63,348 160 19 2.5 100 100 All India 2,919,821 436,424,429 705,002 241 1.6 69,569 241,614 311,183 107 0.7 Note: Total roads in column (3) include national highways, state highways, major district roads, other district roads, and village roads. District roads serve areas of production and markets in the district connnecting them with one another or with main highways. Village roads connect villages and groups of villages with one another and with the nearest district road or highway. Source: (1) and (2) India, Central Statistical Organization 1961. Statistical Abstract. (3) and (6) Central Road Research Institute 1963. "History of Road Development in India," New Delhi. Annex Notes Annex I and building fadama access roads stressing the importance of roads in facilitating irrigation. I. "Better agroclimatic opportunities such as better rainfall, a higher moisture holding capacity of the soil and better irrigation 2. FACU, Project Year Two, Feeder Road Improvement Programme (Benue potential directly affect agricultural output (relation 1). Better ADP), Ibadan, Nigeria, 1988. opportunities also increase the economic return to private farm 3. FACU, Project Year One. Feeder Road Improvement Programme, investments such as tractors, draft animals or pump sets (relation Federal Department of Rural Development, Federal Agricultural 2). The greater private profitability of agriculture in well endowed Coordinating Unit, Ibadan, 1982; and FACU, Project Year Two. Feeder regions induces farmers to press governments for increased Road Improvement Programme, Federal Department of Rural Devel- investment in the supportive infrastructure (relation 3). Financial opment, Federal Agricultural Coordinating Unit, Ibadan, 1988. institutions find it more profitable to locate in environments 4. It is not clear why arrangements differed among the three where a good agroclimate and rapid technical change lead to a projects especially given that they were all appraised at about the substantial demand for agricultural investment and working same time. capital and a high repayment capacity (relation 4) and where good 5. Mechanical grading performed once or twice per year infrastructure reduces their cost of intermediation (relation 5). depending on traffic levels. Private agricultural investment and input use is more profitable the better the agricultural opportunities (relation 2), the better 6. Includes manual pothole repair, drain and ditch clearing. the government infrastructure (relation 61, the cheaper the cost of 7. Surfaced roads refer to a road with 10 to 15 millimeters of financial services (relation 7), and the more favorable price and laterite surface which ensures an all-weather condition. interest policies are which are pursued by the government 8. February 1989 Supervision Report. Memo from L. Campbell (relation 8). Exactly the same factors affect the output supply to A. Seth on "Thematic Supervision-Feeder Roads and Water (relations 9,10,11). This means that agricultural opportunities must Supply Components of ADPs." be translated into public and private investment efforts to affect 9. PPAR Second MSADP agricultural output. The traditional production function approach 10. FACU, Project Year Two, Feeder Road Improvement Programme has attempted to estimate the direct impacts of capital stocks (Benue ADP) c Ibadan r Nigeria, 1988. (investment) and input use on output (relation 12), ignoring much of the factors discussed here and all the simultaneity problems" 11. Ex-post calculated rates of return were 16, 18, and 24 (Binswanger 1989). percent for Gusau, Funtua, and Gombe, respectively, including all project components. Annex 2 12. World Bank, Kaduna State AD., Aide-Memoire, December 1988. 1. Given the emphasis placed on small-scale irrigation in the Annex 3 statewide ADPs, an objective assigned to the road construction 1. Proposed Second Feeder Roads Project. The road kilometers programs at Kano (KNARDAI and Sokoto c /;isted of improving in this case are the total kilometers to be maintained every year. 34 List of MADIA Discussion Paper Series Lele, Uma. "Agricultural Growth, Domestic Policies, the External Environment, and Assistance to Africa: Lessons of a Quarter Century." MADIA Discussion Paper No. 1. Washington, D.C.: World Bank, 1989. Lele, Uma. "Managing Agricultural Development in Africa: Three Articles on Lessons from Experience." MADIA Discussion Paper No. 2. Washington, D.C.: World Bank, 1989. Lele, Uma and L. Richard Meyers. "Growth and Structural Change in East Africa: Domestic Policies, Agricultural Performance, and World Bank Assistance, 1963-86." MADIA Discussion Paper No. 3. Washington, D.C.: World Bank, 1989. Lele, Uma and Steven Stone. "Population Pressure, the Environment, and Agricultural Intensification: Variations on the Boserup Hypothesis." MADIA Discussion Paper No. 4. Washington, D.C.: World Bank, 1989. Lele, Uma, Robert Christiansen, and Kundhavi Kadiresan. "Fertilizer Policy in Africa: Lessons from Development Programs and Adjustment Lending, 1970-87." MADIA Discussion Paper No. 5. Washington, D.C.: World Bank, 1989. Lele, Uma and Manmohan Agarwal. "Smallholder and Large-Scale Agriculture: Are There Trade-offs in Growth and Equity?" MADIA Discussion Paper No. 6. Washington, D.C.: World Bank, 1989. Lele, Uma, Nicolas van de Walle, and Mathurin Gbetibouo. "Cotton in Africa: An Analysis of Differences in Performance." MADIA Discussion Paper No. 7. Washington, D.C.: World Bank, 1989. Idachaba, F. S. "State-Federal Relations in Nigerian Agriculture." MADIA Discussion Paper No. 8. Washington, D.C.: World Bank, 1989. Lele, Uma. "Structural Adjustment, Agricultural Development, and the Poor: Lessons from the Malawian Experience." MADIA Discussion Paper No. 9. Washington, D.C.: World Bank, 1989. Gaviria, Juan, Vishva Bindlish, and Uma Lele. "The Rural Road Question and Nigeria's Agricultural Development." MADIA Discussion Paper No. 10. Washington, D.C.: World Bank, 1989. Lele, Uma and Robert Christiansen. "Markets, Marketing Boards, and Cooperatives in Africa: Issues in Adjustment Policy." MADIA Discussion Paper No. 11. Washington, D.C.: World Bank, 1989. Jammeh, Sidi, Mathurin Gbetibouo, and Uma Lele. "Building Agricultural Research Capacity in Senegal." MADIA Discussion Paper No. 12. Washington, D.C.: World Bank, 1989. Lele, Uma, Mathurin Gbetibouo, and Paul Fishstein. "Planning for Food Security in Africa: Lessons and Policy Implications, 1960-88." MADIA Discussion Paper No. 13. Washington, D.C.: World Bank, 1989. Lele, Uma, Bill Kinsey, and Antonia Obeya. "Building Agricultural Research Capacity in Africa: Policy Lessons from the MADIA Countries." MADIA Discussion Paper No. 14. Washington, D.C.: World Bank, 1989. 35 THE MADIA STUDY Although many generalizations have been made about the agricultural crisis in Africa, relatively few detailed country and cross-country studies of African agriculture based on systematic data analysis have been conducted. Similarly, although foreign aid has constituted a large part of total government expenditures in Africa for close to fifteen years, there has been little analysis of the role of external assistance in African countries that goes beyond political criticism of official assistance or the alleged self- serving objectives of donors. The impetus for the study "Managing Agricultural Development in Africa" IMADIAI was to begin the process of filling this gap and to explain the nature and sources of the agricultural crisis, particularly the extent to which it originated in resource endow- ments, historical and contemporary events, external and internal policies, and the economic and political environment. The MADIA study involved detailed analysis of six African countries- Kenya, Malawi, Tanzania, Cameroon, Nigeria, and Senegal. In addition to the World Bank, seven donors, USAID, UKODA, DANIDA, SIDA, the French and German governments, and the EEC participated in the study. The analysis of country policies and performance during the last 20-25 years was carried out with the benefit of substantial input from the governments and nationals of each of the countries represented. The study had three main areas of focus: (1) the relationship between domestic macroeconomic and agricultural policy and agricultural performance, (2) donors' role in the development of agriculture, and 131 the politics of agricultural policy. The MADIA study was the result of encouragement and support from many people. Anne Krueger, former Vice President for Economic Research Staff in the World Bank, encouraged the establishment of these studies on aid and development in 1984. Gregory Ingram, former Director of the Development Research Department, provided unstinting support for the study. During the reorganization of the World Bank in 1986, the strong support from Benjamin King, then acting Vice President for Economic Research Staff, proved invaluable. Barber Conable, President of the World Bank, and Mr. Edward V K. Jaycox, Vice President for the Africa Region, have played a key role by ensuring support for the study's completion, as did Stanley Fischer, the Vice President for Development Economics. Yves Rovani, Director General of the Operations Evaluation Department, was particularly helpful as the MADIA study drew heavily on the works of OED. A special debt of gratitude is owed to the World Bank's Research Committee, which provided the initial funding for the study, and to the MADIA Steering Committee. In particular the strong support of the chair of the Steering Committee, Stephen O'Brien, has been of critical importance. Finally, without the active and continued encouragement of many African policymakers and donor officials, including numerous colleagues in the World Bank, this study would not have provided new perspectives. This support has taken the form of numerous reactions to written and oral presentations, and refinement of the analysis to identify the areas of consensus and continuing controversy. The World Bank Headquarters 1818 H Street, N.W. W Washington, D.C. 20433, U.S.A. 3 Telephone: (202) 477-1234 Facsimile: (202) 477-6391 Telex: WUI 64145 WORLDBANK RCA 248423 WORLDBK Cable Address: INTBAFRAD WASHINGTONDC rD CD European Office 66, avenue d'lena 75116 Paris, France D Telephone: 11) 40.69.30.00 Facsimile: (1) 47.20.19.66 Telex: 842-620628 0 Tokyo Office O 0- Kokusai Building 0 1-1, Marunouchi 3-chome C Chiyoda-ku, Tokyo 100, Japan Telephone: (3) 214-5001 0 Facsimile: (3) 214-3657 Telex: 78 1-26838 CD n 0-82 1 3-1326-6