Malawi - Highway Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No.946 PROJECT PERFORMANCE AUDIT MEMORANDUM on MALAWI FIRST HIGHWAY PROJECT (CREDIT 112-MAI) December 15, 1975 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. s FOR OFFICIAL USE ONLY TABLE OF CONTENTS Page No. SUJ~MARY .. . . . . . . . . . . . . . . . . . . . . . PROJECT PERFORMANCE AUDIT MEMORANDUM I. The Project .... ............. ......... ......1... II. Implementation and Costs of Road Construction ....... 3 III. Economic Justification .............................. 4 IV. Transport Coordination .............................. 6 V. Institutional Aspects ............................... 7 VI. IDA's Role ......................................... 9 VII. Conclusions ......................................... 9 PROJECT COMPLETION REPORT SUMMARY ............................................. A.1 1. INTRODUCTION ......................... A.2 2. THE PROJECT ...... ........... . A.3 A. General Description ............................ A.3 B. Design Standards ............................... A.4 C. Cost Estimates ................................. A.4 3. PROJECT IMPLEMENTATION AND EXECUTION ................ A.5 A. Construction Works ............................. A.5 - Procurement ................................. A.5 - Contract Modifications ...................... A.6 - Construction Time and Costs ................. A.6 - Performance of Contractors .................. A.7 B. Consultant Services ...........................* A.9 - Supervision of Construction ................ A.9 - Transport Study ............................. A.9 C. Other Covenants of the Credit Agreement ........ A.10 D. Performance of Borrower ........................ A.12 E. Credit Disbursements ........................... A.12 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. TABLE OF CONTENTS (Continued) Page No. 4. REVIEW OF ECONOMIC EVALUATION ..................... A.13 A. Construction Cost Estimates .................. A.14 B. Vehicle Operating Costs ......................... A.14 C. Traffic Forecasts .... ............................ A.15 D. Economic Return .............................. A.16 5. CONCLUSIONS .................................... A.17 TABLES: 1. Road Design Standards 2. Estimated and Actual Project Costs and Implementation Data 3. Zomba-Lilongwe Road Construction Costs 4. Amounts Paid To and Due From Dominion Earthworks, Ltd. 5. Credit Allocation 6. Projected and Actual Average Daily Traffic on Project Road 7. Vehicle Operating Costs, 1967 and 1974 Estimates 8. 1967 Assessments of Vehicle Operating Cost Savings Resulting from Road Improvement 9. Economic Evaluation Data: 1967 and 1974 10. Economic Costs and Benefits CHARTS: 1. Malawi Second Highway Project, Organization of the Ministry of Works and Supplies (WB-8983) 2. Malawi: Organization of the Roads Department: Ministry of Works and Supplies (IBRD-3403) MAP: Malawi Currency Equivalents (Malawi Kwacha) Exchange Rate of Malawi E/Kwacha 1968 Malawi E = US$2.40 1969 " " = US$2.40 1970 " " = US$2.40 1971 MK = US$1.30 1972 MK = US$1.17 1973 MK = US$1.19 Abbreviations Used MWS Ministry of Works and Supplies UDI Unilateral Declaration of Independence Vpd vehicles per day SUMMARY In March 1965, the newly independent Government of Malawi applied for an IDA credit to help finance improvements on two trunk roads in the country. Contacts with the new Government were slow; in the end, IDA granted in September 1966 a US$490,000 Credit (S2-MAI) to help finance detailed studies of the proposed improvements. As a result of the studies made, one of the roads, from Lilongwe to the Zambian border, was deleted from the pro- posal leaving only the 168-mile road from Zomba, then the administrative capital, to Lilongwe, the capital since 1973. This road section is part of the trunk highway linking Blantyre with Lilongwe and then the northern region. As a result of issues raised by IDA during project preparation and appraisal, the credit finally approved included also financing of a compre- hensive study of transport regulation which could serve as a basis for a revamping of the cumbersome licensing system existing at the time. The Government also agreed to the strengthening of the administrative organiza- tion of the Ministry of Works and Supplies (MWS), a training program for the Ministry, and in particular for the Roads Department, the allocation of suffi- cient funds for road maintenance, and improvements in highway planning. The US$13.7 million project finally approved for which the US$11.5 million Credit 112-MAI was finally signed on February 5, 1968, included also the refinancing, following normal Bank practice, of Credit S2-MAI. Actual construction of the Zomba-Lilongwe road commenced in February 1968 and was expected to be completed by September 1970. Construction was completed only in October 1971, but work on some sections was completed closer to schedule. Total actual costs, at US$13.04 million, were prac- tically the same as the appraisal estimate of US$13.11 million, including contingencies. Construction was divided into three contracts. The contractor in charge of the 49-mile section between Dedza and Lilongwe performed very well and work was completed with only minor delays and no overrun. The largest section (90 miles between Livonde and Dedza) was awarded to a consortium consisting of Malawian, French and Mozambique firms. In spite of the con- tractors' internal problems and serious disputes with the consultants, con- struction was completed with only some delay and practically no overrun. The main difficulties and the main causes of delays and overruns occurred in the 29-mile section between Zomba and Liwonde. The contractor's work was unsatisfactory; the company went bankrupt and had to be evicted from the construction site and replaced by another contractor. The new estimate of the economic rate of return for the investments on the Zomba-Lilongwe road is 13% which compares favorably with the 12% estimated at appraisal. Moreover, the new estimate is considered a minimum because of the conservative figures about savings in vehicle operating costs - ii - used in the calculation and the fact that not all of the higher than expected traffic was considered. The road was undoubtedly the first priority in the country at appraisal time and it is now the backbone of the road system. It has helped to decentralize the country by reducing the relative economic weight of the Blantyre region, by helping to increase trade between the central and southern regions and by partially improving transport to the northern region. It has also facilitated the transfer of the national capi- tal to Lilongwe, which is now becoming an important regional center. Finally, the road helped to facilitate the transport of the Zambian exports that use Malawi roads on their way to the Mozambique ports. The study about transport coordination in the country was completed on schedule and resulted in the almost complete deregulation of the road transport industry and an increase in maximum axle-load weights. However, it is not possible to assess the real economic impact of this deregulation because of contradictory evidence which suggests either increasing competi- tion in the road transport industry, or continuation of monopolistic situa- tions existing before. An aspect of the transport sector that remains unclear is the future role of road and rail in the country, particularly the necessary investment in railways. Malawi has made major investments in road and rail facilities and plans a few additional major links, suggesting that some overinvestment might be taking place. The project's institutional objectives have been successfully accomplished. The MWS was reorganized in 1969 in a way that IDA has consid- ered to be sound. The creation of a separate Mechanical Branch in charge of the operation of all equipment for the Ministry has been helpful and has forced the Roads Department to plan its maintenance activities in more detail. The training program has been slow in achieving results and a large propor- tion of the Ministry's professional staff is still expatriate, but it is unlikely that faster progress could have been made in light of Malawi's lack of professionally qualified personnel. Progress in highway planning has been less satisfactory, particularly regarding feeder roads. Practically all the feeder road construction that has taken place has been concentrated on the land development schemes, probably creating an imbalance between the areas under those schemes and the rest of the country, particularly the northern region. In retrospect, IDA seems to have had at appraisal time the right approach towards this project. The road included in the project had the highest priority, IDA's sector view was useful; it emphasized training and managed to keep the size of the project within the capacity of the Roads Department. The exclusion of the Lilongwe-Zambia border road seems in retrospect to have been the right decision. However, IDA's participation seems to have declined in intensity after credit signature. Supervision was not sufficient to help prevent some of the implementation problems and two or three issues at the sector level such as possible imbalances in the feeder road program have not been given sufficient emphasis. PROJECT PERFORMANCE AUDIT MEMORANDUM MALAWI FIRST HIGHWAY PROJECT (CREDIT 112-MAI) This report represents an audit of achievements under the Malawi First Highway Project, for which Credit 112-MAI of February 5, 1968, in the amount of US$11.5 million, was finally disbursed in June 1973. It is based mainly on a review of the attached Project Completion Report (PCR) prepared by the Eastern Africa Regional Office, against other relevant reports and important material from the Bank files, on talks with some of the Bank staff who had been involved in the project, and on the results of a three-day visit to Malawi made to collect additional information and to discuss the evolution and impact of the project with the Malawian authorities. Their excellent help is gratefully acknowledged. I. The Project 1.01 Malawi, formerly Nyasaland and a territory of the Federation of Rhodesia and Nyasaland (1953-63) was not a beneficiary of any of the Bank Group loans to the Federation. After independence, the Government of Malawi applied in March 1965 for a credit from IDA to help finance reconstruction and paving of the trunk road linking Zomba, in the southern region, with Lilongwe, in the central region, the national capital since 1973 (see map). The proposal also included reconstruction of the road from Lilongwe to the Zambian border. Processing of the request proceeded slowly, mainly because of the difficulties faced by the new government administration. In the end IDA granted in September 1966 a US$490,000 Credit (52-MAI) to help finance the detailed engineering of the Zomba-Lilongwe road (180 miles) and study of the design and cost estimates of the Lilongwe-Zambia border road (78 miles). 1.02 These studies were completed in July 1967. The main issue derived from them concerned the Lilongwe-Zambia border road. The studies revealed inadequacies in design and soil investigations, increasing original cost esti- mates by about 300%. In addition, traffic counts in 1966/67 gave an average daily traffic over the entire length of 115 vehicles, but only about 20-30 vehicles near the Zambia border. Even assuming a growth in Zambian transit traffic up to 100 vpd over the lifetime of the road, in addition to normal traffic, the estimated rate of return was about 8-9%. At the time, the situa- tion arising out of the Rhodesian UDI and the Zambian decision not to use Rhodesian Railways was discussed in detail; but there was no basis on which to make a reasonable projection of Zambian traffic, or to guarantee a given amount in future years. As a consequence, IDA considered the road uneconomic and suggested that it be dropped from the proposal. The Government accepted the suggestion. In 1973, IDA's position was vindicated by a consultant study which found most sections of the road uneconomic even at that later date. The USAID is now helping to finance a modified version of this project, on the basis of their own independent assessment. - 2 - 1.03 During appraisal and negotiations, IDA was also concerned with several other issues, mostly related to institutional aspects. One of them was transport regulation. Malawi had a complex road licensing system designed to protect the railways. The impression at the time was that the existing regulation, which restricted truck operations considerably, resulted in higher transport costs and low quality of service. In addition, while vehicle dimension and weight regulations were generally adequate, the maxi- mum axle load of seven tons was considered too low and unduly restrictive of the development of road transport. At IDA's suggestion, the Government agreed to include in the proposed project a study of transport regulation in the country which could serve as a basis for a revamping of the transport policy. 1.04 IDA was also concerned about the institutional improvement of the Ministry of Works and Supplies (MWS), and in particular of the Roads Depart- ment. This concern had several facets. First, IDA recognized, probably more strongly than the Malawian Government at the time, that for the sustained development of the road system and the Department, it was necessary to bring an increasing number of qualified Malawians into professional and technical positions. At IDA's suggestion, the Government agreed to launch a program for the education and training of Malawi nationals for professional positions, and to give high priority to the Department's training efforts. In this respect, an additional reason for IDA's suggestion to delete the Lilongwe- Zambia border road from the project was that the size of the project would put too much strain on the capacity of the Roads Department. Second, IDA considered that it was necessary to improve the administrative organization of the MWS in order to adapt it to the higher level of operations envisaged at the time. The Government concurred with this judgment and proposed to act on the subject. Third, IDA was concerned about the increase in mainte- nance requirements after project completion. The Department was at the time responsible for all the primary, secondary and some 400 miles of tertiary roads, but operations were restricted to day-to-day maintenance, mainly because of shortage of competent staff. The Government was aware that main- tenance had to be strengthened and agreed to review maintenance methods in detail. Finally, IDA was interested in making sure that highway planning would take place on a continuing basis, and asked the Government to prepare a road investment program, including feeder roads. 1.05 These issues were solved satisfactorily and the proposed credit was ready to be presented to the Board by August 1967, but delays in obtain- ing the necessary legislation from the Malawi Parliament authorizing the Government to borrow, and the clarification of the consequences for the financing of the project of the devaluation of the Malawi Pound in November 1967, delayed credit signature until February 1968. 1.06 The project finally approved consisted of (for details see PCR, paras. 2.01 to 2.04) (a) the reconstruction and bituminous paving of the highway between Zomba and Lilongwe (174 miles, according to the proposed new alignment), including engineering supervision; (b) a study of transport coordination; and (c) the refunding (following normal Bank practice) of - 3 - Credit S2-MAI for technical studies. Main covenants referred to the need to strengthen the MWS, especially the Roads Department, particularly through the training program; to allocate sufficient funds for road maintenance; and to prepare a plan for road development. The Government also committed itself to a revision of vehicle weight regulations and to "the promotion of more effective competition within the trucking industry, and proper overall transport coordination." II. Implementation and Costs of Road Construction 2.01 Actual construction of the Zomba-Lilongwe road commenced in February 1968 and was expected to be completed by September 1970, with some sections in February and March of that year. Construction was completed only in October 1971, but work on some sections was com- pleted closer to schedule (PCR, Table 2). Total actual costs, at US$13.04 million were practically the same as the appraisal estimate of US$13.11 million (including contingencies). The 10% quantity and 7.5% price contingencies were adequate to cover the changes experienced due to unsatis- factory work by contractors, quantity increases, and the devaluation of the Malawi Pound. Due to the construction delays, consultant services for super- vision had to be extended by 16 months and total consultant costs, at US$0.84 million, were 59% higher than the US$0.53 million originally budgeted (PCR, para. 3.17). 2.02 The main event during construction was the unsatisfactory performance and subsequent bankruptcy of one of the contractors (Dominion Earthworks, of South Africa), which was the main cause for the project's delays and overruns. Construction of the Zomba-Lilongwe road was divided into three sections for the purpose of awarding contracts: 1/ Zomba-Liwonde (29 miles), Liwonde-Dedza (90 miles) and Dedza-Lilongwe (49 miles). Construction of the latter section was awarded to a British firm (14. and C. French) which performed very well and completed the work with only minor delays and no overrun. The largest section, between Liwonde and Dedza, was awarded to a consortium consisting of Malawian, French and Mozambique firms. In spite of rather serious internal problems of organization, difficult relations with the supervision %onsultants and at times inadequate work, construction was completed with only some delay and an overrun of 12% over the original estimate (excluding contingencies). The quality of the work was considered good. The situation in the Zomba- Liwonde section was more complex. The process until the Government decided to evict Dominion Earthworks from the construction site and take over its assets was slow and resulted in considerable delays and cost overruns. In the end, W. and C. French, the original second lowest bidder, took over the works and completed construction satisfactorily. However, the change in contractor resulted in the longest delay and in a cost overrun of 30% over the original estimate, including contingencies. I/ The actual length of the road was 168 miles. 2.03 Due to the problems described above, disbursements from the Credit lagged about one year behind schedule. The closing date had to be extended twice, until December 31, 1973, but disbursements were finally completed in June 1973. Still pending in the project are a US$570,000 suit filed by the Government against Dominion Earthworks and a dispute between the Government and an insurance company about a US$168,000 security bond related to this contract. 2.04 The Roads Department handled the problems that arose during construction largely by itself. It tried to help the contractors and con- sultants, attempting to solve the disputes that arose between them, and did not ask for damages when extending the contracts for several months. The Department followed the right course of action when it evicted Dominion Earthworks from the construction site although, in retrospect, it appears that the measure could have been taken earlier. IDA, however, had a minor role in the solution of these problems. Notwithstanding the quarterly progress reports sent by the consultants, in addition to letters from the Secretary of the Ministry of Works and Supplies (which, for example, noted that Dominion Earthworks was already 21 weeks late in April 1969), the first supervisory mission visited the country only in October 1969, over a year and a half since the project works actually started. Had IDA sent a super- vision mission earlier, it possibly could have promoted a speedier solution to these problems. III. Economic Justification 3.01 At appraisal, it was conservatively estimated that the investment on the Zomba-Lilongwe Road would have a return of 12% on the basis of savings in vehicle operating costs and in road maintenance costs, which were roughly estimated as a proportion of total quantified benefits. However, the Appraisal Report does not contain the detailed information on which this estimate is based nor the assumptions used in the calculation. The new estimate of the rate of return included in the PCR is 13%, which is also a conservative estimate (PCR, Chapter IV), because it does not fully consider the fact that actual traffic has been about 50% higher than the appraisal forecast (with the exception of the section between Dedza and Lilongwe, where actual traffic was only 9% higher), 1/ and because it is based on a new calcu- lation of savings in vehicle operating costs, especially prepared for the 1/ The PCR estimate of the rate of return has been based mainly on appraisal figures. These higher traffic levels have not been fully taken into account. A difficulty is that 1974 traffic figures are not directly comparable with 1968 or 1970 figures because of a change in the system of traffic counts. The 1974 data are based on a special study done by the Road Research Laboratory of the UK, which, for the level of traffic on this road, results in a margin of error of about 20%. The figures available for 1968-70 were not based on a scientific sampling method and as such it is not possible to know the margin of error in them. - 5 - Second Road Project in Malawi (Credit 523-MAI), which indicated savings smaller than those used at appraisal, but which even later estimates suggest to have been too cautious (PCR, para. 4.05). 1/ 3.02 The appraisal also contained rate of return estimates for several road sections which varied from 11 to 13%. The PCR estimate indicates that the return on the Zomba-Liwonde section fell below 11% due to higher construc- tion costs and the lower level of unit savings in vehicle operating costs. However, it is doubtful that given the main purpose of the road, which is to provide a better connection between the southern and the central regions of the country, and the difficulties of stage construction for a relatively small project such as this one, it is really meaningful to have separate rates of return for different road sections. In retrospect, it does not seem practical to have postponed the improvement of the Zomba-Liwonde section. 3.03 The Zomba-Lilongwe road (part of the trunk highway joining Blantyre and Lilongwe) is undoubtedly the most important highway in Malawi. At apprais- al, IDA was aware of this importance and expected that the highway would help in the development of the Lilongwe area, and even of the northern region. Several partial indications suggest that the highway may have helped in decentralizing the country by reducing the relative economic weight of the Blantyre region, by helping to increase trade between the central and the southern region, and by partially improving transport to the northern region, which is the least developed area in the country. Another new development which, although known at appraisal, was not given sufficient weight, was the transfer of the country's national capital to Lilongwe in 1973. It is clear that this transfer would have been more difficult and costly without the road. Lilongwe itself is becoming an important regional center, as indicated by the number of new industrial establishments and by the fact that a number of trading companies previously located only in Blantyre have opened branches in Lilongwe to serve the central and northern portions of the country. The new road has also helped other IDA-assisted projects, such as the Lilongwe Land Development Scheme, which has become a major supplier of maize to the southern region. Other land development schemes are planned in the area of influence of this road. 3.04 Finally, the road has also facilitated the transport of Zambian export goods. Although this trade has fluctuated drastically from year to year (no figures are available about the total number of Zambian vehicles using this road since its opening; a survey made in May 1974 registered 71 commercial vehicles per day, about 30% of all commercial vehicles using the road at the time), it is clear that at times a large number of Zambian vehicles have used the road on their way to the ports in Mozambique. 1/ A closer analysis of the existing figures on vehicle operating costs in Malawi together with some new estimates made this year by the Roads Department suggest that the savings used in the appraisal of Credit 523-MAI might be too conservative. Thus, at least from this angle, the PCR estimate is perhaps too low. - 6 - 3.05 Two other factors have influenced the impact of this road. One is the deregulation of the road transport industry, and the other is the effect that the improvement and construction of railway lines now underway will have on traffic on this road. These two points are discussed in the following section. IV. Transport Coordination 4.01 At appraisal, IDA was concerned that transport coordination in Malawi might lead to a misallocation of resources. The road transport indus- try was heavily regulated through a licensing system covering the number of vehicles in operation and the routes on which they could operate. In 1967-68, the Government had even increased the degree of control by fixing very low maximum axle loads, which resulted in the use of small and probably uneconomic trucks. The main purpose of this regulation was to protect the railways (PCR 3.18). Also, Malawi was very concerned about its lack of infrastructure and had rather ambitious plans for major investments in road and rail transport, which apparently had been planned without an adequate analysis of the expected role of each mode in the future evolution of the Malawian economy. 4.02 To help solve these problems, IDA suggested, and the Government accepted, the inclusion in the project of a thorough study of transport regulation and its consequences. The study, which was carried out by the Economist Intelligence Unit (EIU), concluded that Malawi should deregulate its road transport industry as much as possible, including raising axle load limits; that railway pricing policies should be more closely related to economic factors; and that the statistical services for transport operations should be improved. In retrospect, the EIU study appears to have been of good quality and its recommendations useful. 4.03 The Government has implemented practically all of the study's recommendations, with the exception of the point related to railway pricing which has not experienced major changes, in part reflecting the fact that most of the railway lines are still not in direct competition with areas where road transport has improved substantially, and the fact that railways deal mainly with foreign trade. 4.04 However, there is not sufficient evidence to assess the economic impact of the deregulation measures. To start with, the EIU study concluded in 1969 that the distortions derived from the old regulatory system were less than what was thought at appraisal time, partly because the economic realities had already forced a more flexible transport system. Second, the structure of the road transport industry in Malawi is unusual in the sense that two large trucking companies control an important proportion of the market, particularly on the main routes. Third, the demand structure for trucking services is also unusual because the country's agricultural market- ing board (ADMARC) controls an important proportion of what is a small and thin market (practically all the transport of agriculture products) and is able to negotiate standard rates with the big trucking companies or with - 7 - large groups of truckers. These characteristics of the market have resulted in rates per ton-km that are similar in many parts of the country, and as such do not reflect costs. In addition, the few scattered pieces of infor- mation available about the evolution of trucking rates (basically from the large truckers) in the last four years on the Zomba-Lilongwe road suggest that the rates per ton-km in real terms have not declined much, although the size of vehicles, at least of those owned by the large trucking concerns, has increased considerably. This kind of information could be used to support either the position that the market was already very competitive or that the big trucking companies have kept most of the benefits of deregulation and of the improved road. Thus, the evidence is sufficiently contradictory to make it impossible in this audit to assess the impact of deregulation or the distribution of benefits of the investment. IDA should suggest to the Malawi Government a more thorough analysis of these issues. 4.05 An aspect of the transport sector that remains unclear is the future role of road and rail in the country, particularly the necessary investments in railways. The EIU study concluded that the rehabilitation project of the rail line from Balaka to Salima proposed by the Government was, in principle, not justified and required further study. In addition, there were doubts at the time that a new railway line linking the main line eastwards through Liwonde to the Mozambique system, to connect Malawi with the port of Nacala, had high priority. Since then, investments in railways have proceeded at a fast pace. The section between Balaka and Salima is being rehabilitated (in retrospect, this investment will probably be justified mainly because of two large industrial developments in the northern part of the country, sugar and pulp and paper, that will use lake transport until they reach the railhead near Salima). The new line towards Nacala has been built and in addition the country has under construction a new railway line between Salima and Lilongwe, and a project for construction of another line between Lilongwe and the Zambian border is at an advanced state of prepara- tion. Thus, Malawi will have a first class road and rail link between Blantyre, Lilongwe and the Zambian border. Given the size of the Malawi economy and the type of goods transported, there are grounds to believe that the country will have an oversupply of transport services in that corridor. IDA has been concerned about this apparent overinvestment but has not been able to convince the Government to study the issue in more detail. V. Institutional Aspects 5.01 IDA was interested in the capabilities and level of efficiency of the MWS and of the Roads Department in two ways: (1) it was concerned that the size of the project would not strain the Roads Department's capabilities, and (2) it recognized the need to improve their operations in several ways. The program of institutional improvement comprised four areas: (a) strength- ening of the administrative organization of the Ministry of Works and Supplies; (b) a training program for the Ministry, and in particular for the Roads Department; (c) need to allocate sufficient funds for road mainte- nance; and (d) improvements in highway planning (PCR, para. 3.22). - 8 - 5.02 The Ministry of Works and Supplies was reorganized in 1969 in a way that IDA has considered to be sound (PCR, para. 3.23). The most impor- tant change, from the point of view of road development, is the creation of a separate Mechanical Branch which is in charge of the operation of all equipment for the Ministry. This arrangement has proved to be satisfactory to the Roads Department, with the exception that the Branch does not buy specialized equipment--an operation that has still to be carried out by the Roads Department, and which is relatively time consuming since it is the only one of this type left in the Department. 5.03 The training program, which aimed at a fast localization of the Department, has proceeded slowly but is unlikely that it could have moved faster in light of the lack of national expertise. In 1974, two engineering positions in the Roads Department were filled by nationals. The Ministry has eight trainees abroad financed by different agencies and they are expected to take over additional positions now held by expatriates. However, senior positions will inevitably continue to be held by expatriates for some time. Training of technical personnel and skilled workers is satisfactory and of good quality. This training takes place at the Ministry training schools in Zomba, where 146 persons graduated in 1973 (PCR, para. 3.25). At present, all of the technical staff are Malawians. 5.04 Maintenance standards are satisfactory, although there is still a lack of funds to do adequate regravelling or resealing. The few failures, such as the one that occurred on the road section between Lilongwe and Dedza, have been taken care of adequately. Overloading does not seem to have been a problem in Malawi, although some does occur with the Zambian trucks, but no analysis has been made of its effects. The maintenance programs were improved considerably in the last years. In a way, the creation of the separate Mechanical Branch forced the Roads Department to prepare detailed maintenance programs as a basis for their equipment requirements--a positive result of the institutional change. Finally, construction and maintenance are separate activities and no mix-ups take place. 5.05 Progress in highway and transport planning has been less satis- factory, in spite of efforts made at improving the quality of development plans, particularly in highways (PCR, paras. 3.27 to 3.29). We have already mentioned the possible overinvestment in transport facilities; to that we should add that the planning of feeder roads seems to be inadequate. Until 1970 there was a feeder road plan but since then practically all the feeder roads that have been built or improved in the country have been directly related to specific land development schemes, such as the Bank-assisted Lilongwe and Shire programs. It is obvious that these schemes should have priority, but the excessive concentration on them might have resulted in a lack of necessary investment in other parts of the country, particularly the northern region. There is at present a rolling three-year highway prog- ram, but only part of the investments included in the plan have been subject to a satisfactory economic appraisal. -9- VI. IDA's Role 6.01 In retrospect, IDA seems to have had at appraisal time the right approach regarding this project. The road selected for inclusion in the project had the highest priority at the time. IDA did have a sector view which resulted in the inclusion of the EIU study in the project and the sub- sequent modifications in the regulation of the road transport industry. It emphasized training and managed to keep the size of the project within bound- aries that were reasonable for the capacity of the Roads Department. The exclusion of the Lilongwe-Zambia border road, for example, seems in retro- spect to have been the right decision. 6.02 The main criticism of IDA's role comes at implementation time. IDA did not supervise the project adequately. The first mission visited the country too late to prevent the problems raised by the poor performance of two of the contractors. Similarly, IDA could have helped more actively the Malawians to solve the problems raised by the bankruptcy of one of the contractors. The issue of probable overinvestments in transport is more complex. IDA did strongly raise the subject, but without much success. In retrospect, IDA should have suggested around 1971-72 the preparation of a transport investment plan in which necessary investments would have been clearly determined. VII. Conclusions 7.01 All of the project objectives were achieved in spite of difficulties mainly due to the poor performance of contractors, and the resultant delays and overruns. The road from Zomba to Lilongwe was successfully completed and is now the backbone of the country's transport network. The rate of return of the investment is at least 13%. The transport coordination study financed under the project resulted in an important change in Malawian policy dealing with the road transport industry. The full impact of the change in policy is still unclear but deregulation will probably have a positive effect on the country's allocation of resources. The institutional efforts, al- though modest, were well-conceived and seemed to have produced good results. 7.02 IDA made a positive contribution in the design of the project and in bringing a sectoral view that has been extremely useful. It also high- lighted the need for institutional improvements, particularly training and localization. IDA's participation seems to have declined in intensity after credit signature. Supervision was not sufficient and two or three issues at the sector level, such as possible regional imbalance in feeder road investments, have not been given sufficient emphasis.  A. 1 PROJECT COMPLETION REPORT MALAWI FIRST HIGHWAY PROJECT (Cr. 113-MAI) SUMMARY On February 5, 1968 a Credit Agreement was signed to help finance the reconstruction and supervision of construction of a 168-mile main road to two-lane bituminous standards; consulting services for a study of road transport licensing regulations and road-rail coordination; and the refunding of the Highway Engineering Project under Credit S2-MAI. Construction works were completed in November 1971 and the last disbursement from the Credit account was made in June 1973, two years later than originally scheduled. In spite of substantial delays in construction works, due to the rather weak performance of one contractor and the bankruptcy of another, there were no significant cost overruns. No changes were made in the list of goods. Performance of the Borrower was satisfactory. Notwithstanding the less satisfactory performance of the contractors and consulting engineers, the quality of the construction works is of high standard, and the road con- structed under this project is being well maintained. The objectives of the project have been achieved. Re-evaluation of the project road, including recalculation of the economic return based on available data and certain estimates, confirms the justification for, and economic viability of its construction. A. 2 PROJECT COMPLETION REPORT MALAWI FIRST HIGHWAY PROJECT (Cr. 113-MAI) 1. INTRODUCTION 1.01 The Government of Malawi applied in early 1965 for a credit from the International Development Association (IDA) to finance a project the country had initiated and prepared. It comprised the reconstruction and bituminous paving of the country's principal north-south road which links Zomba, the former national capital in the Southern Region, with Lilongwe, the national capital since 1973 in the Central Region, and the reconstruction of the primary road west from Lilongwe to the border with Zambia. The pro- posed project was based on a 1964 United States Agency for International Development-finance transport survey and Malawi Ministry of Works and Supplies (MWS) feasibility studies. 1.02 In response to this application, in September 1966 IDA provided a US$490,000 equivalent credit (S2-MAI) which included financing for a detailed engineering of the Zomba-Lilongwe road (168 miles) and a review of the design and cost estimates of the Lilongwe-Zambia border road (78 miles) which had been prepared by the Roads Department of MWS. This was the Bank Group's first lending operation in the transportation sector in Malawi. 1.03 In July 1967, the consultants, Brian Colquhoun and Partners (UK), completed the detailed engineering of the Zomba-Lilongwe road. Their review of the Lilongwe-Zambia road design revealed inadequacies in design and soil investigation requiring additional pre-construction engineering work. Further, review of the preliminary cost estimates for reconstructing the road confirmed IDA's doubts about the economic justification of improving the road at that time (doubts which were reconfirmed in a United Nation Development Programme- financed feasibility study by Scott Wilson Kirkpatrick and Partners in 1973). Therefore, during appraisal in April 1967 it was decided to drop the Lilongwe- Zambia border road from the proposed project and to include construction only of the Zomba-Lilongwe road. It was also decided to include in the project consulting services for a transport study and the refunding of Credit S2-MAI. The project was presented to the Board in January 1968 and the Credit Agree- ment (Credit 112-MAI) signed on February 5, 1968. 1.04 The present report is based on: (i) quarterly progress reports Nos. 4-18 prepared by the supervising consultants (Brian Colquhoun and Partners); A. 3 (ii) findings of IDA supervision missions undertaken during and after project execution; (iii) highway division files; and (iv) traffic counts carried out by the Malawian National Statistical Office in 1974. 2. THE PROJECT A. General Description 2.01 The project consisted of: (a) consultant-supervised reconstruction to bitumen standards of the Zomba-Lilongwe road (168 miles); (b) a study of road transport licensing regulations and road-rail coordination by consultants; and (c) the refunding of the Highway Engineering Project under Credit S2-MAI. (a) Reconstruction of the Zomba-Lilongwe Road and Supervision of Construction 2.02 The Zomba-Lilongwe road is part of the north-south artery which connects most principal Malawian centers from the Mozambique border in the south to the Zambia border in the north, including Blantyre, Zomba, Lilongwe, Mzimba, Rumphi, and Chitipa. The road section from Zomba south to Blantyre, the southern region's commercial center, had already been improved and the pavement widened from a single lane to two in 1960. From Zomba north to Liwonde (28.6 miles), the existing road was single lane bituminous-paved. On the Liwonde to Lilongwe section (139.4 miles) vertical and horizontal alignment was generally poor, as was the gravel or earth surface. Through the rugged escarpment between Ncheu and Dedza, the road was winding, badly aligned, and difficult to maintain. Under the project, the Zomba-Lilongwe road has been reconstructed to improved standards including two-lane bituminous paving. A weighbridge was installed near Balaka. The engineering consultants, Brian Colquhoun and Partners, who had prepared the detailed engineering supervised the construction. (b) Consulting Services for Transport Study 2.03 Under the project the study of road transport licensing regulations and road-rail coordination were undertaken by consultants. Due to the shortage of qualified personnel, Government was unable to carry out these studies without the use of consultant services. A.4 (c) Refunding of the Highway Engineering Project (Credit S2-MAI) 2.04 Credit S2-MAI had included a detailed engineering of the Zomba- Lilongwe road section, a review of the Lilongwe-Zambia detailed engineering and the design and cost estimates on which the project under consideration was based. In accordance with the Association's operational policies regard- ing the financing of detailed engineering, refunding of the Credit was included in the new project. B. Design Standards 2.05 The detailed design standards adopted at the time of appraisal and to which the Zomba-Lilongwe road was built are shown in Table 1. They included, for the first time in this country and apparently with good results, an un- usually thin layer (3/4") of bituminous concrete surfacing. The geometric standards are judged adequate to accommodate forecasted traffic volumes over the service life of the road (20 years). C. Cost Estimates 2.06 At the time of appraisal the total project costs were estimated as follows: US$ '000 (i) Construction 10,596 (ii) Supervision of construction 528 (iii) Studies 96 (iv) Contingencies (a) Quantity variations 10% on item (i) 1,060 (b) Price variations 7.5% on items (i) - (iv)(a) 922 Subtotal 13,202 (v) Refunding of Credit S2-MAI 490 Total 13,692 2.07 The construction cost estimates were based on bid prices received for all works tendered prior to Board presentation of the project. Estimates for construction supervision were based on the consultant's price proposal. Estimates for a study of transport regulation and coordination were based on the cost of about 30 man-months. The contingency allowances of 10% for quantity variations and 7.5% for price escalation were considered adequate in view of, respectively, the quality of the detailed engineering and expected price increases in Malawi for such works. A.5 2.08 The final costs of the project were basically in line with appraisal estimates. Estimates and actual costs are compared in Table 2 and discussed in paras. 3.08 and 3.09. 3. PROJECT IMPLEMENTATION AND EXECUTION A. Construction Works Procurement 3.01 There were 3 developments affecting procurement: (a) award of contracts; (b) devaluation of the Malawi pound; and (c) bankruptcy of Dominion Earthworks, Ltd. (a) Award of Contracts 3.02 The construction works for the project road were divided into three lots, and by the time the Credit Agreement was signed bids had been received for all three after international competitive bidding in accordance with Bank Group standard procedures. The lowest bidders were: - Zomba-Liwonde (28.6 miles) - Dominion Earthworks, Ltd. - Liwonde-Dedza (90 miles) - General Construction, Dragages, Ermoque Consortium (GECDEC); and - Dedza-Lilongwe (49.4 miles) - W. & C. French (Overseas), Ltd. 3.03 In an attempt to secure contracts for the Zomba-Liwonde and Dedza- Lilongwe sections, GECDEC offered "a further considerable reduction" in their already-submitted bid for Liwonde-Dedza road, for which a contract had not yet been awarded. While having previously proposed to award the contracts to the lowest bidders, Government cabled the Association for approval to award the total construction works to GECDEC, although the post bidding re- bate offered by this firm was marginal (US$30,000 on a US$4.5 million contract) and only slightly below the offers made by W. & C. French and Dominion Earthworks. The Association considered GECDEC's offer of price reduction unethical and prejudicial to international competitive bidding and rejected Government's request. (b) Devaluation of the Malawi Pound 3.04 Another procurement problem arose as a result of the devaluation of the Malawi Pound on November 20, 1967, that is, between the time of bid opening and award of contracts. Government suggested evaluating the bids on the basis of the original exchange rate and including a clause in con- tract documents providing that additional costs above the tendered amount would be payable only if they could be proven to be due to the pound de- valuation; the Association agreed to this procedure. Later, during project A.6 execution, contractors were compensated for the pound devaluation by a lump sum (Table 3) to which the Association had agreed. (c) Bankruptcy of Dominion Earthworks, Ltd. 3.05 After two years of construction on the Zomba-Liwonde section, during which time only 40% of the contract work was completed, Dominion Earthworks went bankrupt (para. 3.10). The Association agreed for Government to negotiate a contract with the bidder who had originally submitted the second lowest bid for the road, W. & C. French. The company was reluctant to take over work started by another contractor, so Government accommodated it by allowing use of the equipment left by Dominion Earthworks, without charge. Contract Modifications 3.06 No major variation orders were issued during construction and the road has been built to the original specifications and standards. Construction Time and Costs 3.07 During appraisal it was estimated that all construction works would be completed by the end of 1970. As Table 2 demonstrates, all contracts had to be extended substantially; the last contract was completed in November 1971. Although part of this delay is attributable to unusually heavy rains in the construction area, the major reasons are the contractors' unsatisfactory performance (para. 3.10-3.13) and the lack of coordination between consultants and contractors. 3.08 In spite of the time overruns and the fact that one contractor went bankrupt and had to be replaced by another, total project costs remained within initial estimates (Table 2). The 10% contingency for quantity variation and the 7.5% contingency for price inflation covered all cost increases, including those caused by currency devaluation and the additional costs of re-contracting the Zomba-Liwonde Section. 3.09 Within the individual items, the following cost increases over appraisal estimates may be noted: - Zomba-Liwonde section from US$1.75 million to US$2.68 million, or 53%, mainly due to the necessary change of contractor; - Liwonde-Dedza section from US$5.60 million to US$6.26 million, or 12%, due to currency devaluation and quantity increases; - Dedza-Lilongwe section from US$3.26 million to US$3.29 million, or 1%; and A.7 - Construction supervision from US$528,000 to US$868,300, or 65%, mainly due to the increase in construction time necessitating 46 months of supervision services instead of the 32 initially scheduled. Performance of Contractors Dominion Earthworks, Ltd. (Zomba-Liwonde) 3.10 By the time 90% of the original contract period was over, this South African contractor had only performed 40% of the construction tendered. Performance was consistently slow. Progress was hampered by numerous site staff changes and worn-out equipment, which also detrimentally affected construction quality. Frequent meetings between consultant administrators and the contractor did not remedy the situation. Toward the end of 1969, 20 months after contract award, Dominion Earthworks ran into financial difficulties and clearly could not carry out his contractual obligations. Government then cancelled the contract on January 2, 1970, expelled the contractor from the site, and took over his assets. The works were later completed by W. & C. French Ltd. (paras. 3.05 and 3.15). General Construction, Dragages, Ermoque Consortium (Liwonde-Dedza) 3.11 This consortium, consisting of a Malawian contractor (General Construction), a French contractor (Societe Francaise de Dragages et de Travaux Publics), and a Mozambique firm (Ermoque Empreiteiros), seemed to have the ideal prerequisites for successfully completing the awarded con- struction works. General Construction could provide the local knowledge, Ermoque the engineering knowledge and experience, and Dragages the financial backing, accounting services, and site administration. However, the Con- sortium ran into difficulties from the beginning. Each partner relied on the others in dealing with problems that arose. The Project Manager was unable to knit together into a team the collection of people of varying nationalities and backgrounds. Also, the consortium headquarters was slow to recognize field staff problems and slow to find solutions. 3.12 Fourteen months following commencement date, the value of work completed was only 15.5% of the contract amount although about half of the total 27-month contractual construction time had elapsed. Reports prepared by the supervising consultants give some impression of the contractor's performance, for example: "GECDEC have made a very poor showing to date. This has been due to poor direction, poor supervision and an inability to recognize and rectify problems as they arise." "The Contractor has neglected the planning and programming of the job, and this, if not rectified virtually immediately, A.8 will have serious repercussions. Already he is finding himself boxed-in due to lack of detailed forward planning, and if this continues, will affect production seriously." "Production of acceptable compacted layers is woefully slow. Repeated recompaction is still the rule rather than the exception." Other shortcomings were identified: - site staff's lack of highway construction experience; - inability to construct drainage works in correct sequence; - disinclination to discuss problems frankly with supervising consultants and to take their advice, resulting in disharmony between Resident Engineer and contractor's staff; - lack of reaction to Resident Engineer's instructions; - construction of sub-base over rejected subgrade; - use of unapproved gravel as base material; and - excessive equipment down-time. 3.13 By the end of the contract period as stipulated in the tender doc- uments, only 70% of the works had been completed. Government agreed to contract extensions amounting to 14 months. For several reasons, only US$5,520 in damages, instead of the full liquidated damages stipulated in the contract, were imposed. The reduction was based, on the one hand, on consideration of those parts of the works that consultants certified as complete or opened to traffic, and, on the other hand, probably also on the fact that supervising consultants were partly responsible for the time extension because they were unable to establish an adequate working relationship with the contractor (para. 3.16). W. & C. French (Overseas), Ltd. (Dedza-Lilongwe and Zomba-Liwonde) 3.14 After a slow start on the Dedza-Lilongwe section due to heavy rains and unexpectedly slow delivery of new plant, "the contractor soon produced the work, both in quality and quantity, which he had envisaged and undertaken in his tender ... The quality of the work throughout was satisfactory and relationships between the Resident Engineering staff and the contractor harmonious". This evaluation, in the supervising consultants' confidential report on this local contractor's performance, was endorsed both by MWS and in Bank supervision reports. Nevertheless, the contract A. 9 period was extended for seven months, probably not because of a real need to do so but rather to put W. & C. French on an equal basis with the other contractors. 3.15 Under the contract for the Zomba-Liwonde section taken over by W. & C. French from Dominion Earthworks, performance was likewise satis- factory. B. Consultant Services Supervision of Construction 3.16 Brian Colquhoun and Partners, who had prepared the detailed engineer- ing of the Zomba-Lilongwe road, were employed to supervise its construction on terms approved by the Association. The few changes in contract specifi- cations and the moderate quantity variations are evidence of accurate detailed engineering. However, Government faulted the consultants for "observed in- adequacy in on-site administration and decisive supervisory control" which, together with "the hitherto inability of the contractors to adhere to their obligations," delayed construction progress. 3.17 The cost estimate for construction supervision services had been US$528,000 equivalent, while the actual cost was US$868,300. This 65% increase resulted mostly from the 16 additional months of supervision needed to complete the works. On the basis of available information, it is not possible to determine the degree of consultant versus contractor blame for the delays. Transport Study 3.18 In 1968, Malawi's road transport industry was considerably restricted by licensing regulations. These regulations had unfavorable economic conse- quences since general restrictions on for-hire trucking operations meant that most road haulage was conducted by private account trucking at costs higher than commercial trucking because of lower load factors and lower vehicle utilization. Further, the policy of limiting trucking operations to particular areas held back competition and led to overcapacity in some places, inasmuch as temporary surplus capacity could not seek employment elsewhere. In addition, the regulations were designed to favor the railway over road trans- port. Without preferential treatment, railways would have been vulnerable to road competition, especially between Balaka and Salima (about 150 km), because of the high costs of rail operations. Aggravating the situation even further, the railway rate structure did not reflect operational costs but was based roughly on the value of goods carried. 3.19 At negotiations it was agreed to carry out a thorough study of the situation. Two consulting firms were invited to submit proposals for a road transport regulations study based on terms of reference prepared by the A. 10 Association and reviewed by Government. The Economic Intelligence Unit (EIU) was selected by Government and approved by the Association. EIU carried out the study in early 1969 and submitted a draft report in July 1969, which Government and the Association found to be generally satisfactory. The principal recommendations of the study were: (a) the vehicle axle load limit should be raised from 7 to 8 tons; (b) all restrictions and controls on entering the road transport industry and geographically or commercially expanding existing operations should be abolished; (c) the railway rates should be reformed; and (d) a system of statistical data collection should be instituted to provide pertinent road transport information. 3.20 The Government has implemented the recommendations (a) and (b) and has recently started to implement (d) by setting up a statistical unit for regular and systematic traffic counting. The railway rate structure, however, is basically unchanged. 3.21 The lifting of trucking operations restrictions has favorably affected the industry. The goods vehicle fleet increased by 30% between 1969 and 1973 (the latest date for which figures are available). Extensive competition has developed among truckers, and equipment is probably better utilized. In general, the trucking industry is in reasonably good financial condition. Despite some loss of railroad traffic to the road transport industry, rail traffic has continued to grow, showing an 8.6% ton-km increase between 1970 and 1971 and a 6.6% gain in the following year over 1971. C. Other Covenants of the Credit Agreement 3.22 During Credit negotiations, understandings were reached between Government and the Association, and confirmed by a sideletter to the Credit Agreement, that Government would: (a) strengthen the administrative organization of MWS; (b) intensify the professional training program; (c) allocate sufficient funds for the maintenance of the road system; and (d) work out a general program for road development. A. 11 (a) Organization of MWS 3.23 Government reorganized the Ministry of Works and Supplies in 1969 (see Charts). During appraisal of the Second Highway Project in March 1974, MWS' functional organization was found to be sound. (b) Professional Training 3.24 In one of the first instances of directly promoting localization, IDA made a review and expansion of the Malawian education and training pro- gram for professional road administration positions a covenant of the Credit Agreement in a sideletter. In 1969, the number of engineering positions increased from 34 to 45 and the number of filled posts rose from 20 to 33. Since then, the number of engineering positions has further increased to 55, of which 37 are filled. In 1968, all professional posts were held by ex- patriates, but by 1974 six were held by nationals. The ratio of Malawian engineers to expatriates in MWS, while low, should improve considerably in coming years. Government has an active program for selecting, educating, and training nationals for senior technical positions. Further, Malawi Polytechnic is developing its capability in the area, so the need to send engineers abroad for training will gradually diminish. 3.25 The training of lower echelon personnel, such as road supervisors, foremen, equipment operators, and mechanics, is continuing with satisfactory results at MWS's training school in Zomba. Whereas 79 persons were trained in some aspect of road work in 1966, 146 were trained in 1973. 1/ (c) Funds for Road Maintenance 3.26 Although the general level of RD road maintenance is satisfactory, there are notable exceptions where regravelling or resealing is overdue. It is true that the budget allocations for road maintenance have been progressively increased during the last few years, but they do not appear to have kept pace with the growth of traffic and inflationary trends. During credit negotiations for the Second Highway Project this point was again stressed by the Association and agreement was reached that sufficient maintenance funds would be allocated to assure proper maintenance of the classified highway network. (d) Road Development Program 3.27 In February 1968, Government submitted a program for road development which was approved by the Association in April 1968. In recent years, it has been improved and adapted to Government's present transport policy on a sector- wide basis. Its aims are to: 1/ MWS, Training Branch, 1974. A. 12 (i) improve the administrative, social and economic integration of the country by linking Malawi's three regions with reliable all-weather connections; (ii) encourage agricultural development by improving access to rural areas; and (iii) to assure efficient and reliable access to the sea for import/export transport. 3.28 Toward these ends Government is undertaking an extensive transport development program, encompassing planned investments in rail, lake and air transport as well as roads. Presently underway, with Bank Group financing assistance (Cr. 523-MAI), is construction of the Lilongwe-Kasungu road (73 miles), which will improve the link between the southern and central regions. Under the same credit, a District Council road maintenance pilot program is being carried out. Under construction, with USAID-financing, are the Chikwawa-Bangula road (51 miles), which improves access to the southern region, and the Lilongwe-Zambian border road (67 miles). Govern- ment is seeking financing from UNDP with the Bank as executing agency to study a road link between Karonga and the border with Tanzania. 3.29 Coordination of transport planning is being improved by employing an experienced transport economist. A one-year transport planning study carried out in 1974 by a UK-financed consultant was aimed at assessing the development priorities in the sector for the period 1975-1985. D. Performance of Borrower 3.30 The reconstruction of the Zomba-Lilongwe road was the most important project in the highway sector ever to have been undertaken in Malawi. Govern- ment has overseen its execution diligently and efficiently with a straight- forward and even-handed approach to contractors and consultants. With the exception of a short interlude over the awarding of contracts for construction of the Zomba-Liwonde and Dedza-Lilongwe road sections, relations between Government and the Association were decidedly satisfactory. Government has met substantially all obligations assumed under the Credit Agreement, although a greater effort to train Malawian professionals would have been desirable and Government should have provided adequate budget allocations for road maintenance. E. Disbursements 3.31 Due to construction delays, the time expended to settle disputes about contractors' claims, and the forfeiture of Dominion Earthworks, dis- bursements from the Credit account lagged on the average about 12 months behind projections, and the closing date had to be extended twice, first from June 30, 1971 to December 31, 1972, and then to December 31, 1973. A. 13 US$ million Calendar year 68 69 70 71 72 73 Estimated Cumulative Disbursements 5.0 9.5 11.5 11.5 11.5 11.5 Actual Cumulative Disbursements 2.45 6.04 8.91 9.98 10.7 11.5 3.32 Disbursement was made at 84% of total construction costs and 85% of total costs of consulting services in the form of reimbursements to Government against a certificate of payment to contractor or consultant. This means that the Association disbursed against all payments made to the contractors. 3.33 In the case of Dominion Earthworks, the equipment he forfeited was sold by Government (US$91,000), the money received, including the Association quota, was kept by Government and probably not used for project purposes. 3.34 Government is still disputing the security bond (US$168,000) with the insurance company and has filed suit against Dominion Earthworks for the extra cost of work due to his forfeiture (US$570,000). Appropriate measures should be taken for the re-use in other Bank projects of the money that may eventually be recovered (Table 4). 4. REVIEW OF ECONOMIC EVALUATION 4.01 The principal quantified benefit from improving the Zomba-Lilongwe road is a vehicle operating cost saving. In addition, a road maintenance cost saving, representing about 12% of total quantified benefits, is attributable to the improvement. At appraisal, a 12% economic rate of return was calculated based on these projected savings and estimated con- struction costs. The basic data used for this calculation are given in Tables 8 and 9. 4.02 The project has been re-evaluated using updated figures although lack of specific data in the Appraisal Report and division files has made it necessary to use certain assumptions. In particular, the original traffic forecasts on which calculations were made were not given, nor was the basis for these forecasts provided in sufficient detail to reconstruct them. Also, while economic returns were calculated for each of four road sections, neither the original construction cost estimates nor the actual costs are available for each of the sections. 4.03 There have not been any major changes in economic trends in the area served by the road since appraisal that would lead to substantially different conclusions about the area's economic potential. Certain data, A.14 however, which became available after the road was improved do not fully substantiate the original quantification of benefits. The resulting margins of error occur primarily in the direction of overestimation. There does not, however, appear to be sufficient cause to doubt the original conclusion that the road improvement was economically justified. The basic data are reviewed below in the light of experience. A. Construction Cost Estimates 4.04 The construction cost estimates were prepared after bids had been received by Government. Actual cost figures supplied by the Malawi Government are presented in terms of the original contracts. As indicated in Table 2, actual project costs were, in total, only 7% higher than the estimates but the relationships differed greatly for the three road segments: Actual Actual as % of Original Estimate (US$'000) Cost Estimate Quantity Construction Supervision Variation Total Total Zomba-Liwonde 1,747.2 87.1 175.1 2,009.4 2,874.4 143 Liwonde-Dedza 5,594.4 287.8 560.2 6,433.4 6,701.2 104 Dedza-Lilongwe 3,254.4 162.1 325.7 3,742.2 3,519.4 94 Overall 10,596.0 528.0 1,061.0 12,185.0 13,095.0 107 In the case of the Zomba-Liwonde road, the actual cost was 43% higher than the estimate and this large difference has an adverse effect on the economic return for this section, as described in para. 4.09. B. Vehicle Operating Costs 4.05 The vehicle operating cost estimates in the original analysis appear high in comparison with estimates, based on more detailed data, adopted for evaluating the second road improvement project in Malawi (Credit 523-MAI): A.15 Estimated Savings in Vehicle Operating Costs (US cents per vehicle mile) Credit 112-MAI Credit 523-MAI Improvement to two- Light Heavy Light Heavy lane bitumen standards Vehicles Vehicles Vehicles Vehicles from: (cars/vans) (trucks/buses) (cars/vans) (trucks/buses) Earth or poor gravel (rolling terrain) 7.7 14.6 2.2 10.6 Earth or poor gravel (mountainous) 9.8 22.6 - - Improved gravel - - 1.0 3.4 Single lane bitumen 3.4 8.9 - - 523-MAI estimates were made in 1974 (Table 7) and have been deflated to allow for inflation during the period 1967-1974. These data, which are in line with the de Weille unit costs of 1966, 1/ indicate that the 112-MAI estimated savings for light vehicles were three times the comparable deflated savings estimated in the 523-MAI appraisal. This difference is particularly signi- ficant because light vehicles constituted approximately two-thirds of the traffic forecast. For heavy vehicles, the 112-MAI estimated savings are also considerably higher than the corresponding deflated savings estimated in the later project. There is, of course, the difficulty of equating the respective classifications of highway conditions for the two projects, but it is believed that the above comparisons are sufficiently valid to justify the conclusion that the earlier appraisal overestimated the potential unit operating cost benefits. The effect of these overestimations of unit costs are offset, to some extent, by the effect of traffic underestimation, as discussed below. C. Traffic Forecasts 4.06 As stated above, neither the original traffic forecasts nor the data on which they were developed are available (Table 6). In the original presentation, it was forecast that normal truck traffic would increase 6-7% annually and that automobile traffic would increase at "a somewhat faster rate" which was not specified. Generated traffic was estimated on the basis that total traffic would increase about 30% during the first four years following the opening of the road. Diverted traffic amounting to 8-10 trucks per day was estimated. If one assumes that, in the original 1/ Quantification of Road User Savings, Jan de Weille (Johns Hopkins Press, 1966), P.31. A. 16 forecast, all normal and diverted traffic were forecast to grow at 6.5% p.a., the 1974 levels of total traffic would be as shown in the first column below: Assumed % Increase of Forecast Actual Actual over ADT ADT Forecast ADT Zomba-Liwonde 533 820 54 Liwonde-Ncheu 362 560 55 Ncheu-Dedza 350 530 51 Dedza-Lilongwe 501 520 9 Actual traffic, shown in the second column, was more than 50% higher than the assumed forecast on three of the road sections. It appears that the original forecasts represent a considerable underestimation of the first few years' traffic after road improvement. 1/ It is, of course, too early to draw any firm conclusions as to the validity of the forecasts over the 20-year economic life of the road. D. Economic Return 4.07 The economic returns of the project in both the original evaluation and this re-evaluation are nearly the same, in spite of using in the present evaluation substantially lower unit savings in vehicle operating costs than in the former. Unfortunately, the basis for the original calculation is not shown in the appraisal report and, therefore, was not available for review. In the earlier evaluation, an economic return of 12% was calculated for the overall road improvement, and a re-evaluation of the project in- dicates an economic return of 13% (Table 10). The latter percentage was based on the lower unit saving in vehicle operating costs shown in the table in para. 4.05. The traffic forecasts underlying this economic return are based on traffic data in the original presentation and an assumed 6.5% p.a. growth rate for normal traffic. Since the original presentation gives no indication of the tax element in the construction cost, the 5.6% tax element found in the Second Highway Project was assumed in deriving the economic cost of construction of the Zomba-Lilongwe road. 4.08 The traffic forecast used as a basis for calculating the 13% return did not reflect the influence of recent actual traffic growth. If account had been taken of the relatively high actual traffic growth in the first few years after improvement (para. 4.06), the economic return would have been higher than 13%. The high growth rate in these initial years will probably 1/ For additional details on projected and actual ADT, see Table 6. A. 17 not be sustained, however, over the long term. Nevertheless, the long term growth rate apparently used in the original forecast may well be too low. In the Second Highway Project the basic growth rate employed in forecasting traffic was 8-9%. 4.09 For individual sections of the road, the economic return calculated in the original evaluation ranged from 11% on the Zomba-Liwonde section to 13% on the Ncheu-Dedza section. In the re-evaluation, the return fell below the previously calculated return only for the Zomba-Liwonde section. It fell below 11% largely as a result of two factors. First, the construction cost was 43% higher than the original estimate (para. 4.04), and, secondly, a lower level of unit savings in vehicle operating costs was substituted in the calculation (para. 4.05). It is, of course, difficult now to relate the actual condition of this road section without improvement to particular levels of vehicle operating costs. With regard to the high construction cost, there were presumably no indications during the original evaluation that contract- ing problems would develop and cause costs to rise sharply above estimates. In general, there is insufficient evidence to dispute the original conclusion that the improvement of this section of the road was economically justified. 5. CONCLUSIONS 5.01 The economic data in the appraisal report was not presented in sufficient detail to permit more than a general review of the economic rates of return. Nevertheless, a re-evaluation of the economic returns was undertaken using the available data and introducing certain estimates and assumptions. Recalculating the economic returns on this basis confirms the original judgement that improving the Zomba-Lilongwe road was economically justified. The re-evaluation of this project was handicapped by the lack of adequate economic data in the original presentation. In future, every effort should be made to include in appraisal reports the necessary elements for checking and re-evaluating the original economic justification of a project. 5.02 Reconstruction of the Zomba-Lilongwe road was the most important project in the highway sector ever to have been undertaken in Malawi, one of the first instances in which professional training of nationals was stipulated in an IDA Credit Agreement, and the first highway project fi- anced by the Bank Group there. On the other hand, there was no Bank supervision mission until a year and a half after the signing of the Credit Agreement; by then, serious difficulties with the contractors had developed. An earlier supervision mission might have detected the problem and instigated corrective action which could have reduced the lengthy construction time. East Africa Regional Office June 1975  TABLE 1 MALAWI Credit 112-MAI First Highway Project Project Completion Report Design Standards ---------Terrain---------- Flat Rolling Mountain Width of right of way (feet) 200 200 200 Design speed (mph) 6o So 40 Stopping sight distance (feet) 475 350 275 Minimum radius, horizontal (feet) 1,150 800 500 Maximum grades (percent) 5 6 8 Bituminous surfaced width (feet) 22 22 22 Roadway width, general (feet) 32 32 32 Roadway width on embankments higher than 5 feet (feet) 36 36 36 Pavement type Stabilized sub-base and base with double bituminous surface treatment or bituminous concrete Pavement design wheel load 9,000 lbs equivalent Structures: Width, carriageway (feet) Span less than 30 feet Equal to roadway width Span more than 30 feet 24 24 24 Width, footpath (feet) 2 x 2 Loading AASHO 20 -s16 Source: Credit 112-MAI Appraisal, 1968. June 1975 ,,rcta 1l?-rRI First h1.stway Croiect Froiect Comulution Report fltimaed and ActmAl Poject Ipats ai Imlementation Data IDA Details Road Consultants Contractor Contract Costs Date Completion Date Remarks Category Length No. Appraisal Actual Commenced Scheduled Actual Estimates Costs miles US$ 000 I Reconstruction and bituminous paving of M between: Zomba-Liwonds 31 Brian Colquhoun Dominion Earth- 3/67 ) 1,139.9 April 25, 1968 March 19, 1970 Contractor expelled from site on Jan. and Partners works ) 2, 1970 1,747.2 W & C French 3A/67) 1,543.9 Aug. 13, 1970 Oct. 13, 1971 Oct, 13, 1971 Livondo-Dadma 94 GECDEC 1/67 5,594.4 6,256.8 Feb. 1, 1968 Sept. 1, 1970 Oct. 30, 1971 Dadas-Lilongwa $5 " W & C French 2/67 3,254.4 3,286.0 April 25, 1968 Feb. 17, 1970 Sept. 21, 1970 11 Supervision of con- 180 Brian Colquhoun 528 868.3 February 1968 September 1970 November 1971 struction works and Partners III Transport Coor- Economic Intel- dination study ligence Unit 96 78.7 Janua 1969 February 1970 Februao 1570 IV Refinancing of 490 404.4 Credit 82-MAI V Contingencias for: a Quantity variation 1,061 10% on I b Price escalation 922 7.3% on I, II, III, Va TOTAL: i2,&92 13,378.3 JUNE 1975 MALAWI Credit 112-MAI First Highway Project Project Completion Report Zomba-Lilongwe Road Construction Costs uS$ UUU Road Section Zonba-Liwonde Liwonde-Dedza Dedza-Lilongwe Contract Nr 3/67 1/ 3A/67 1/67 2/67 Contractor Dominion Earthw. W.&C. French GECDEC W.$C. French Contract Amount 1,681.1 2/ 1,301.6 5,511.6 3,119.9 Quantity Variation - 541.2 J + 70.3 + 278.1 - 94.7 Price Inflation + 76.1 + 41.3 + 34.0 Currency Devaluation + 95.9 + 431.3 + 226.8 Liquidated Damages - 5.5 Final Amount 1,139.9 h/ 1,543.9 6,256.8 3,286.0 Total Construction Costs .12,195.2 1 See Table 4 Initial amount of contract 3/ Does not represent quantity variation, but difference between initial contract amount & total cost of contract ITI Represents amount paid to contractor minus money received for sale of equipment plus direct expenses JUNE 1975 TABLE 4 MALAWI Credit 112-MAI First Highway Project Project Completion Report Amounts Paid To And Due From Dominion Earthworks According to a memorandum from the Secretary of the MWS to the Secretary of the Treasury dated April 24, 1971, the amount paid to Dominion Earth- works can be calculated as follows: Revised Estimate Contract 3/67 MK 805,783 $ 966,940 Add indemnity bond MK 140,042 $ 168,110 Add sale of equipment MK 80,000 $ 96,000 Deduct direct expenses mK 43,332 $ 52,000 Amount paid to Dominion Earthworks: $1,179,050 Deduct actual sale of equipment (letter treasury dated $ 91,150 October 3, 1975) MK 75,959 Add direct expenses $ 52,000 Total cost of contract 3/67 $1,139,900 Amounts still disputed (letter MWS dated Dec. 9, 1974) Security bond mK 140,092 $ 168,110 Extra cost contract 3A/67 mK 474,766 $ 569,719 Total amount outstanding $ 737,829 JUNE 1975 TABLE 5 MAIAWI Credit 112-MAI First Highway Project Project Completion Report Credit Allocation (US$) IDA Detail Original Final Disburse- Category Allocation Allocation ment % I Construction 9,637,000 lo,314,723.84 84 IT Supervision 676,000 715,015.38 85 ITT Studies 80,000 65,84$.04 85 TV Refunding 490,000 404,415.74 100 V Contingencies 617,000 11,500,o00 11,500,000.00 JUNE 1975 MALAWI Credit 112-MAI First Highway Project Project Completion Report Projected and Actual Average Daily Traffic on Project Road Road Section Average Daily Traffic Annual Traffic Total Light Vehicles Heavy Vehicles Growth Over 7 Years Appraisal Projection Actual Appraisal Projection Actual Appraisal Actual Appraisal Actual 1966-Z67 at Appraisal 1973 7 1966-67 at Appraisal 197374 Estimate to 1973/74 2/ to 1973/74 2/ 177 1 Zomba-Liwonde 198 363 600 88 170 220 9.3 16 533 820 Liwonde-Ncheu 133 244 350 61 118 210 9.3 16 362 560 Ncheu-Dedza 128 233 320 60 117 210 9.3 16 350 530 Dedza-Lilongwe 175 335 290 83 166 230 10.0 11 501 520 1/ Normal traffic growth: 6.5% Generated traffic: 30% within first four years Diverted traffic: 10 heavy vehicles per day 2] Projection made at appraisal with traffic growth indicated under 1/ Source: Appraisal Report Malawi Highways I and Traffic Observations 1974 Malawi National Statistical Office JUNE 1975 MALAWI Credit 112-MAI First Highway Project Project Completion Report Vehicle Operating Costs, 1967 and 1974 Estimates Road Surface Vehicle Operating Costs Light vehicles = Passenger cars Heavy vehicles 7 ton truck Appr. HWI Appr. 1 II Ratio Appr. HWI Appr. HWII Ratio 1967 1974 Appr. HWI Appr. HWII 1967 1974 Appr. HWI Appr. HWII US$ US$ % % UT UT % % Bitumen 0.071 0.114 100 100 0.120 0.373 100 100 Gravel 0.085 0.120 120 105 0.157 o.428 130 115 Earth 0.121 0.150 170 131 0.204 o.543 170 145 Source: Appraisal Reports Malawi Highways I and Highways II JUNE 1975 TABLE 8 MALAWI CREDIT 112-MAI - FIRST HIGHWAY PROJECT PROJECT COMPLETION REPORT 1966 Assessments of Vehicle Operating Cost Savings After Road Improvement 12 ton truck & 12 ton 7 ton 5 ton Passenger Passenger 10 ton trailer truck truck truck Bus Car (Pence per Vehicle Mile) Zomba-Liwonde 35.6 26.0 15.7 13.5 16.4 8.5 Liwonde-Ncheu 43.5 30.L 18.1 15.3 18.8 10.3 Ncheu-Dedza 52.4 36.0 22.4 19.0 23.2 12.1 Dedza-Lilongwe 43.5 30.4 18.1 15.3 18.8 10.3 In each case it is assumed that operating costs would be reduced to: 27.5 19.1 12.0 10.2 12.7 7.1 except Ncheu-Dedza, where because of steep climbs, the operating costs would not likely fall below: 30.0 21.0 13.0 11.0 13.5 8.0 Source: 1966 Appraisal mission estimate on basis of interviews with transporters JUNE 1975 MALAWI CREDIT 112-MAI - FIRST HIGHWAY PROJECT PROJECT COMPLETION REPORT Economic Evaluation Data Construction Costs ADT Annual Traffic Growth % Internal Rate Road Section Length US$ 000 Appraisal Since Completion of Return Miles Appraisal Actual 1967 1974 Estimate of Road Appraisal % Zomba-Liwonde 31 2,009.4 2,874.4 286 820 16 11 Liwonde-Nchen 47 6,433-L1/ 61701.21/ 194 560 16 11-13 Nchen-Dedza 47 188 530 16 13 Dedza-Lilongwe 55 3,742.2 3,519.4 258 520 11 11-13 Total 180 12,185.0 13,095.0 12 1/ Combined cost for Liworde-Ncheu and Ncheu-Dedza road sections. 2/ Traffic growth for trucks was estimated at 6-7$ annually and for automobiles was indicated to be "a somewhat faster rate". (Automobiles represented about two-thirds of total traffic on all road sections). It was also estimated that there would be an additional 30% of induced traffic during the first four years. Moreover, it was forecast that the equivalent of 8-10 trucks per day would be diverted from the railway, JUNE 1Q75 TABLE 10 MALAWI CREDIT 112-MAI - FIRST HIGHWAY PROJECT PROJECT COMPLETION REPORT Re-calculation of hconomic Costs and Benefits (US$ 000) Year Capital Savings in Vehicle Savng in Total net Costs 1/ Operating Costs nMaintenance Costs Benefits 1968 3,115 - 3,115 1969 3,870 - 3,870 1970 2,84L - 2,8LL 1971 2,533 356 49 - 2,128 1972 1,199 163 1,362 1973 1,299 177 1,476 1974 1,416 193 1,609 1975 1,556 212 1,768 1976 1,646 225 1,871 1977 1,753 239 1,992 1978 1,853 253 2,106 1979 1,967 268 2,235 1980 2,080 284 2,364 1981 2,207 301 2,508 1982 2,33 320 2,663 1983 2,482 338 2,820 1984 2,629 359 2,988 1985 2,785 380 3,165 1986 2,958 Lo3 3,361 1987 3,145 429 3,574 1988 3,337 455 3,792 1989 3,547 L86 L,031 1990 3,770 514 4,284 1991 2,620 357 2,977 Economic Return 13% Economic costs derived by (a) combining actual financial costs of improvement and supervision and (b) reducing this sum by estimated 5.6% for taxes and duties. 2/ Calculated on the basis of (a) normal, diverted and generated ADT; (b) vehicle operating costs per mile and (c) length of road. 3/ Calculated on basis of original estimate that the savings were equivalent to 12% of total benefits. j/ Total net benefits reflecting savings in vehicle operating costs, savings in maintenance costs and capital costs. JUNE 1975 MALAWI SECOND HIGHWAY PROJECT ORGANIZATION OF THE MINISTRY OF WORKS AND SUPPLIES MINISTER PERMANENT SECRETARY DEPUrTYSECRETARY EGNE ADMINISTRATION RAEGN RNETORECTOR OINECTOR OFFCERSEC TRY F BRDIGSOF ROAOS OF DESIGN ACCOU TS A MINIS RATI N T ST RESIG SU VYANA DS A I GP A IGDR C SW T R P L EO D DEVELEHIME O H IRE MDINGNANONLANNCTECUE REGIONAL ADMINISTRATION ---NOTHRNEGONCENTRA.LREGION SOUTHERN REGIONI EGINEEL ADMINISTRATION ADMINISTRATION ADMINISTRATION (MZUZUI (LILONGWE) (BLANTRE) .REGIONAL REGIONAL 1EGIONAL WOAKSHOP AND WORKSROP AND WORKSHOP AND RNRIECENTER HIRECENTER HIRE CENTER RALDNG ROADS WTEM SUPPLIES CONSTRACEiON MAINTENANCE Wad Bon, - 153W  MALAWI: ORGANIZATION OF THE ROADS DEPARTMENT MINISTRY OF WORKS AND SUPPLIES COMMISSIONER FOR WORKS [ANDSUPPLI ES ROADS DEPARTME NT OTHER DEPARTMENTS DIRECTOR OF ROADS] ENGINEERING TECHNICAL ASSISTANCE CONSULTANTS PROGRAMS PLANNING BRANCH DESIGN BRANCH CONSTRUCTION AND MAINTENANCE BRANCH MATERIALS DESIGN SOUTHERN CENTRAL NORTHERN LABORATORY OFFICE REGION REGION REGION World Bank - 15299  MALAWI HIGHWAY SYSTEM S-' To Maimba Kora Kota Kosungu -0 20 40 60 \ LMILES Dowa rMary chtnmeSelr Senga 00 >0 Lslongwe Chkao7 Monkey <9 Dedza Golompti Mne - Fort Kosinje -' Johnston Primary bituminized roads Lo*e\- - MI ombe' Other primary roads - Secondary roads - -AFRCA RONcheuCAPrCO- --------- Other roads+ - 0 Project road L0od Major airports N- Rai lways --ciw ++Proposed railwayZOB --- Marshes MwanzaF/s Lr wo Escarpment -ie 0 Chilk ..C iradzulu Doressooo oBLAN YRE| ,. olornbe uc za O Chikwawa- (. Cholo LubumbaiShi,f L M /aw Area covered \ by map- ra Ncicala 7 P, .FremChiromo LuskaBlantyr Zo nbo Salisbury R H OD SIA g- 31 .Nsonje BulawayoO . . \ 0 /50OUTH \oBer L----r----AFRICAý -------.-- - "E"_T er AUGUST 1967 IBRD-2092R