Malawi - Second and Third Highway Projects

Document of The World Bank FOR OFFICIAL USE ONLY Repod No. 5184 PROJECT PERFORMANCE AUDIT REPORT MAIAWI SECOND AND THIRD HIGHWAY PROJECTS (CREDITS 523-MAI AND 758-MAI) June 29, 1984 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the perfornmance of their offcial duties. Its contents may not otherwise be disclosed without World Bank autiza-ntion. FOR OFFICIAL USE ONLY PROJECT PERFORMANCE ADDIT REPORT MALANI SECOND AND THIRD HIGHWAY PROJECTS (CREDITS 523-NAI AND 758-MAI) TABLE OF CONTENTS Page No. Preface ........................................................... i Basic Data Sheets Credit 523-MAI ............................................... ii Credit 758-MAI ............................................... IV Highlights ........................................................ vi PROJECT PERFORMANCE AUDIT MEMORANDUM I. BACKGROUND AND COMPONENTS ............. 1 II. IMPLEMENTATION .......................................... 2 III. RE-EVALUATION ..............4...........................4 IV. TRANSPORT POLICIES ...................................... 6 PROJECT COMPLETION REPORTS SECOND HIGHWAY PROJECT (CREDIT 523-MAI) I. Introduction ......................................... 11 II. Project Identification, Preparation and Appraisal ....... 31 III. Project Implementation and Cost ............. ...... 13 IV. Institutional Development ......... ...........19 V. Economic Reevaluation .................. ........ 20 VI. Association Performance . .........22---.......22 VII. Conclusions ....................... ............. .. 23 Tables 1-6 24-29 Annex 1 30 Maps - IBRD 15069R IBRD 15070 THIRD HIGHWAY PROJECT (CREDIT 758-MAI) I. Introduction .oo .......... - ...-...... .. 39 II. Project Identification, Preparation and Appraisal ....... 39 III. Project Implementation and Cost .................... 43 IV. Institutional Development .................... .... ..48 V. Economic Reevaluation .................... ........... 48 VI. Association Performance .............................. 50 VII. Conclusions .......................................... 51 Tables 1-6 52-57 Maps - IBRD 15069R IBRD 15070 This document has a restricted distribution and may be used by recipients only in the performance of their official duties Its coatents may not otherwise be disclosed without World Bank authouizato. PROJECT PERFORMANCE AUDIT REPORT MALAWI SECOND AND THIRD HIGHWAY PROJECTS (CREDITS 523-MAI AND 758-MAI) PREFACE This report presents a performance audit of the Malawi Second High- way Project and Third Highway Project for which Credits 523-MAI for US$10 million equivalent and 758-MAI for US$10.50 million equivalent were approved on December 3, 1974 and December 22, 1977, respectively. Both credits were fully disbursed in March 1982. The report consists of a Project Completion Report (PCR) on each of the two projects prepared by the Eastern Africa Regional Office and a Project Performance Audit Memorandum (PPAM) prepared by the Operations Evaluation Department (OED). OED has reviewed the PCRs against the Appraisal and President's Reports, the legal documents and the transcript of the Executive Directors' meeting which considered the Third Highway Project.1/ Project files and documents have also been reviewed and discussions held with Bank operational staff. Further, an OED mission had discussions concerning this project with Malawi authorities in October 1983. The draft Project Performance Audit Report has been sent to the Government for comments. No comments have been received. The audit finds that the PCRs provide a comprehensive, candid and constructive review of the experience and lessons under the projects. The audit notes that the projects, designed to support Malawi's transport policy objectives were, on the whole, successfully implemented. 1/ The Second Highway Project was approved under special procedure. PROJECT PERFORMANCE AUDIT BASIC DATA SHEET MALAWI: SECOND HIGHWAY PROJECT (Credit 523-MAI) KEY PROJECT DATA Appraisal Actual or Item Estimate Current Estimate Total Project Cost (US$ million) 12.50 13.05/a Overrun (%) 4.4 Credit Amount (US$ million) 10.00 10.00 Disbursed 10.00 Cancelled 0.00 Repaid to October 31, 1983 0.00 Outstanding to October 31, 1983 10.00 Date Physical Components Completed 02/77 09/81/b Proportion Completed by Original Completion Date (Z) 100 33 Proportion of Time Overrun (Z) 237 Economic Rate of Return (Z) 14 16 CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ million) FY75 FY76 FY77 FY78 FY79 FY80 FY81 FY82 (1) Appraisal 1.5 5.1 9.6 10.0 10.0 10.0 10.0 10.0 (ii) Actual 1.1 2.0 3.5 6.0 6.7 8.1 9.4 10.0 (ii) as % of (i) 73 39 36 60 67 81 94 100 /a Including project components added during implementation. lb Trunk road constructed under project was opened to traffic in March 1978. Construction was completed in May 1980. September 1981 is completion date for District Road Improvement and Maintenance Program (DRIMP). - iii - OTHER PROJECT DATA Appraisal Item Estimate Actual First Mention in Files 10/69 Government's Application n.a. Negotiations 09/74 Board Approval 12/03/74 Credit Agreement Date 12/19/74 Effectiveness Date 03/21/75 02/21/75 Closing Date 06/30/78 06/30/82 Borrower Republic of Malawi Executing Agency Ministry of Vorks and Supplies Fiscal Year of Borrower April 1 - March 31 Follow-up Project Third Highway Project Credit Number 758-MAI Amount (US$ million) 10.5 Agreement Date 01/18/78 MISSION DATA No. of Date of Item Month/Year Persons Manweeks Report Identification 05170 1 2 06/70 Preparation 07/70 1 1 08/70 Appraisal 03/74 3 7 04/74 Supervision 1 10/75 n.a. n.a. n.a. Supervision 2 01/76 n.a. n.a. n.a. Supervision 3 06/76 2 4 07/76 Supervision 4 11/76 1 1 12/76 Supervision 5 07/77 1 1.5 08/77 Supervision 6 04/78 1 1.5 05/78 Supervision 7 10/78 1 2.5 10/78 Supervision 8 07/79 1 1.5 08/79 Supervision 9 11/79 1 3 12/79 Supervision 10 09/80 1 3 10/80 Supervision 11 06/81 1 1.5 06/81 Completion 02/82 1 L.5 03/82 31.0 COUNTRY EXCHANGE RATE Name of Currency (Abbreviation) Kwacha (Mk) Year: Appraisal Year Average US$1 = Mk 0.83 Intervening Years Average US$1 = 1k 0.86 Completion Year Average US$1 = Mk 0.90 - iv - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET MALAWI: THIRD HIGHWAY PROJECT (Credit 758-MAL) KEY PROJECT DATA Appraisal Actual or Item Estimate Current Estimate Total Project Cost (US$ million) 15.09 14.93/a Underrun (%) - 1 Credit Amount (US$ million) 10.50 10.50 Disbursed 10.50 Cancelled 0 Repaid to October 31, 1983 0 Outstanding to October 31, 1983 10.50 Date Physical Components Completed 04/80 06/81 Proportion Completed by Original Completion Date (Z) 100 80 Proportion of Time Overrun (Z) 0 35 Economic Rate of Return (%) 14 8 CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ million) FY78 FY79 FY80 FY81 FY82 (i) Appraisal 1.8 5.4 10.1 10.5 10.5 (ii) Actual 0 5.2 8.6 10.1 10.5 (ii) as Z of (i) 0 96 85 96 100 /a The Key Project Data Sheet in the PCR shows US$15.08, against $14.93 given in Table 2. -v - OTHER PROJECT DATA Appraisal Item Estimate Actual First Mention in Files 10/75 Government's Application n.a. Negotiations 09/77 Board Approval 12/22/77 Credit Agreement Date 01/18/78 Effectiveness Date 05/18/78 06/29/78 Closing Date 06/30/81 12/31/82 Borrower Republic of Malawi Executing Agency Ministry of Works and Supplies Fiscal Year of Borrower April 1 - March 31 Follow-up Project Fourth Highway Project Credit Number 1099-MAI Amount (US$ million) 33 Agreement Date 02/20/81 MISSION DATA No. of Date of Item Month/Year Persons Manweeks Report Preparation 06/76 1 2 09/76 Preappraisal 11/76 2 4 12/76 Appraisal 02/77 2 5 03/77 Supervision 1 04/78 1 1 05/78 Supervision 2 10/78 1 2 10/78 Supervision 3 07/79 1 1.5 08/79 Supervision 4 11/79 1 1 12/79 Supervision 5 09/80 1 2.5 10/80 Supervision 6 06/81 1 1 06/81 Completion 02/82 1 1.5 03/82 21.5 COUNTRY EXCHANGE RATE Exchange Rate Name of Currency (Abbreviation) Kwacha (Mk) Year: Appraisal Year Average US$1 = Mk 0.90 Intervening Years Average US$1 = Mk 0.83 Completion Year Average US$1 = Mk 0.90 - vi - PROJECT PERFORMANCE AUDIT REPORT MALAWI SECOND AND THIRD HIGHWAY PROJECTS (CREDITS 523-MAI AND 758-MAI) HIGHLIGHTS The principal objectives of Malawi' s transport policies are improvements of access to the sea, north-south integration and upgrading of rural mobility. The two projects were designed to support these objectives and, on the whole, successfully implemented despite delays (PPAM, para. 1). The Second Highway Project suffered a major time overrun. Costs of both projects remained within budgets: civil works under the Second Project were overbudgeted at appraisal, and the scope of one item in the Third Proj- ect was reduced during implementation to absorb higher unit cost (PPAM, paras. 6-8). The Second Highway Project broke new ground by putting a system in place, under which about 5,500 km of district roads are being improved and organizations, methods and procedures are being established to maintain the roads (PPAM, paras. 5, 9-11). The recalculated economic rates of return on construction under the Second Highway Project are slightly above appraisal estimates, and under the Third Project, they are well below. However, in all cases (both SARs and both PCRs), the investment benefits were assessed conservatively, as they exclude benefits from a potential diversion of the import/export trade to the new central-northern Malawi/Tanzanian route once it is complete (PPAM, para. 16). The re-estimates of the economic returns as given in the PCRs are subject to uncertainties with regard to investment benefits due to a weak data base (PPAM, paras. 12-14). Improvements to the quality of transport sector data are being made (PPAM, para. 17-18). Road maintenance budgets became inadequate during the projects' implementation. The matter is being addressed under the Fifth Highway Proj- ect (Loan 2363/Credit 1423/Credit SF6-MAI), approved late 1983 (PPAM, para. 21). -1- PROJECT PERFORMANCE AUDIT MENDRANDUM MALAWI SECOND AND THIRD HIGHWAY PROJECTS (CREDITS 523-MAI AND 758-MAI) I. BACKGROUND AND COMPONENTS 1. Malawi is a landlocked and resource-poor country with regionally imbalanced development and a largely rural population. Transport policies, in this environment, are aimed at the provision of efficient and multiple access to Indian Ocean ports, to safeguard the import and export trade, inte- gration of the less developed Central and Northern Regions with the more developed Southern Region and the supply of the rural population with basic mobility. Both the Second and Third Highway Projects were fully consistent with these policy objectives and were, on the whole, successfully imple- mented. 2. The PCRs written on the two projects provide a comprehensive, can- did and constructive review of the experience and lessons learned. This audit, while confirming the major findings of the PCRs, highlights and elaborates further on some selected topics and puts the two projects into the larger context of ongoing Bank lending for transport. 3. The thrust of the civil works under the two projects was patterned after that under the First Highway Project (Credit 112-MAI), which financed reconstruction/improvement of the trunk road from Zomba (Southern Region) to Lilongwe, the capital (Central Region), a section of the designated 1,100 km long main artery extending north-south over the entire length of the coun- try. The Second Highway Project financed improvement to paved standards of the road from Lilongwe to Kasungu to its north, and the Third Project financed the further northbound extension (improvement) of the road to Jenda, together with strengthening works on the road leading south from Lilongwe to Dedza. The Fourth and Fifth Highway Projects, approved in 1981 and 1983 (Credit 1099-MAI and Loan 2363/Credit 1423/Credit SF6-MAI), continued Bank assistance for new construction/improvement of the north-south axis, which by about 1990 should be fully in place, thus enhancing regional integration and providing another needed outlet to the sea from the northern tip of Malawi to the Tanzanian port of Dar Es Salam. Both routes through Mozambique, pro- viding currently the shortest distances to the sea, are subject to severe constraints. The rail link to the Mozambique port of Beira is virtually closed at present due to insurgency, and the rail link to the other Mozambique port of Nacala suffers from operational bottlenecks, which limit - 2 - capacity to half of Malawi requirements. Outlets from Malawi through the west via Zambia/Tanzania and Zambia/Botswana/South Africa are exceedingly long and costly.1/ 4. At appraisal of the Second Project, there were discussions about whether the project should include the road from Lilongwe to the Zambian border instead of a section of the north-south axis. The issue was settled through the offer and provision of assistance from the United States Agency for International Development (USAID) for the construction of the Lilongwe- Zambian border road.2/ While the Zambian border road was a viable invest- ment, progress on the north-south axis was clearly a high-priority undertaking in the context of Malawi's need for north-south integration and establishing alternative access routes to the sea. 5. The Second Highway Project included a pilot scheme for a District Road Improvement and Maintenance Program (DRIMP), to which, after substantial savings in the construction component were confirmed (PPAM, para. 6), preparation and implementation of DRIMP Phase I and preparation of DRIMP Phase II were added. DRIMP Phase I dealt with the requirements of four dis- tricts, out of a total of 24 districts. DRIMP Phase II, financed under the Fourth Highway Project, covered an additional nine districts, and the balance, eleven districts, are part of DRIMP Phase III financed under the Fifth Highway Project. II. IMPLEMENTATION 6. Implementation of civil works under the Second and Third Highway Projects ran into various problems which are described in the PCRs,3 resulting in a substantial time overrun under the Second and a smaller over- run under the Third Project. Despite these delays, project cost remained within budgets because works under the Second Project were overbudgeted (PPAM, para. 7) and the price of the successful bid was low, and works under the Third Project were reduced (PPAM, para. 8). 7. At appraisal and negotiations of the Second Highway Project, the Bank and the Government did not reach agreement on technical specifications for the (subbase of the) Lilongwe--Kasungu road, which led to the unusual inclusion of a large special contingency item in the cost estimate (US$2.3 1/ A detailed discussion of Malawi's import and export routes is contained in a 1983 consultants report, entitled Transport Sector Assessment, undertaken by Louis Beger International, Inc. on behalf of the Republic of Malawi and the United States Agency for International Development. 2/ PCR, Credit 523-MAI, para. 2.02. 3/ PCR, Credit 523-MAI, paras. 3.01-3.08; and PCR, Credit 758-MAI, paras. 3.01-3.08. - 3 - million out of total project cost of US$12.5 million), rather than resched- uling of project processing until the controversy was cleared up. After appropriate engineering tests during project implementation, it was deter- mined that the contingency was not needed, as the Government had always maintained, and the resulting cost savings could be applied to additional project items (including preparation/implementation for DRIMP Phases If II). / 8. The strengthening works under the Third Highway Project were also not fully assessed at appraisal which led to their redesign during implemen- tation coupled with a reduction in road length due to the project's financial ceiling. Bank staff explained that the original design was made about 1977, while the loan became effective in mid-1978, and that further pavement de- terioration had taken place in this period. Instead of 19 k, as envisaged at appraisal, works on only 12 km were performed, with no firm information available on the fate of the remaining 7 km. Despite the exclusion of part of the road link from the project, the entire 19 km have remained open to traffic, though at the cost of heavy maintenance works.5/ 9. Given that DRIMP was an untested undertaking, its implementation under the Second and follow-up Projects, despite obstacles, is widely con- sidered an unqualified success. By late 1981, at the conclusion of DRIMP Phase I, about 1,100 km of district roads were improved, and the entire program (Phases I-III) covering approximately 5,500 km is scheduled to be completed before 1990. Maintenance capacities have been or are being esta- blished concurrently through the creation of maintenance organizations and introduction of appropriate maintenance procedures, which in turn are bene- fiting from the various forms of technical assistance to the Ministry of Works and Supplies (MWS) financed under a succession of well-conceived Bank highway projects for Malawi. The DRIMP improvement works are being carried out with a mix of labor-intensive and intermediate technology with cu:cLently a total of 3,500 laborers and at a cost competitive with alternative tech- niques. Occasional complications arose due to the seasonal unavailability of labor, a major government concern before DRIMP was started, but work sched- ules are now tailored to prevailing conditions in the labor market. 10. Since its inception, DRIMP, with the assistance of a consulting firm, has been organized by MWS, which was put in charge of it because of administrative and engineering capacity and expertise. Upon their improve- ment and some initial testing of maintenance methods and procedures, the dis- trict roads are turned over to the District Councils under the umbrella of the Ministry of Local Governments. The Government and District Councils are fully committed to DRIMP and strongly support the project, as does the rural population, for whom DRIMP is. a demonstrably feasible and keenly welcomed program to better living conditions. 4/ PCR, Credit 523-MAI, paras. 2.04-2.05. 5/ PCR, Credit 758-MAI, paras. 3.05-3.07. - 4- 11. Since a number of years, the Bank, through research, pilot programs and full-scale lending operations, has promoted the use of labor-intensive techniques in implementing civil works in developing countries. DRIMP itself is the beneficiary of the Bank's Research Program and the product of the Bank's promotional efforts, which together with conditions specific to Malawi - very low labor rates, a disciplined work force and strong Government backing, reinforced by consultants specialized in the subject, have made DRIMP an acknowledged success. While the implementation of DRIMP has been documented in the consultants' and Bank's supervision reports, a more in- depth review of DRIMP seems to be desirable to determine whether and under what conditions it could serve as a model for other countries. One phase of DRIMP, the pilot scheme preceding DRIMP Phase I implementation, has been covered in a recent Bank publication.6/ III. RE-EVALUATION 12. The Staff Appraisal Report (SAR) for the Second Highway Project estimated the economic rate of return (ERR) for Lilongwe-Kasungu at 14%, with the ERRs for the three subsections ranging from 10-21%, based on savings in transport cost related to "normal" traffic only. Using the same format of analysis, the PCR re-estimates the overall return at 16% based on returns for the subsections of 11-20%. The slight upward adjustment in overall return is due to the substantial capital cost savings7/ and hi her than anticipated vehicle operating cost (VOC) savings per traffic unit,_/ partly balanced by lower estimates of ("normal") traffic. 13. The SAR's estimates under the Third Highway Project of the economic returne for Kasungu-Jenda and Lilongwe-Dedza were 14% and 13%, respectively. The PCRs' recalculated returns are 8% and 9%, respectively, basically as a result of lower traffic volumes and in the case of Kasungu-Jenda, a decline in the share of heavy vehicles of total traffic. In both the SAR and PCR, the returns were based on cost savings to "normal" traffic only. The actual traffic development is subject to a variety of economic and social factors outside of the control of project management. The experience with the traffic under the two projects illustrates the difficulty of predicting and safeguarding project benefits. 6/ Labor-based Construction Programs. A Practical Guide for Planning and Management, by World Bank Transportation Staff and Consultants, 1983 (Annex D). 7/ The ERR estimate was based on cost including the special contingency for the road subbase referred to above (PPAM, para. 7). 8/ The higher VOC savings were established under a Bank developed VOC model, which was not yet avail8DIe at appraisal. - 5- 14. All recalculated economic returns are subject to considerable n- certainties with regards to past, present and future levels of -normal- traffic and savings in VOC per traffic unit (PPAM, paras. 17-18).1/ The post-evaluation of the Second and Third Highway Projects is a repeat of the post-evaluation of the First Highway Project (1975), which also struggled to establish road traffic estimates and vehicle operating cost.10 15. The PCRs point out that the recalculated economic returns do not include benefits from a potential diversion of the import/export trade to the central/northern Malawi-Tanzania/Dar Es Salam route once it is complete.11/ This is a valid point without which the recalculated ERRs could be mis- leading. Further, in cases where infrastructure provides insurance for sittr- ations of a potential national emergency (closing of alternative import/ export routes), the economic returns based on -normal- or -regular" traffic flows and associated benefits only would also understate the economic bene- fits of the investments. 16. The weak actual traffic development on the project roads, as well as low traffic volumes elsewhere in the country in recent years, brought about by the last recession, have sharpened the perception of Bank staff that the geometric design standards of some Malawi roads may be too high. Savings in capital cost on new trunk roads, in their views, can be realized by lowering these standards without unduly compromising on essential requirements. The audit was given to understand that proper attention will be given in the future to this subject. 17. The recalculation in economic returns reveals deficiencies in road planning and evaluation data. As the PCRs elaborate, there are not only un- certainties about future traffic developments, but also about past and present traffic levels. Traffic counting has been inadequate, even though systematic counts were introduced in 1974, based on a design by the United Kingdom Transport and Road Research Laboratory, and the two projects under audit carried the usual covenants about traffic counting. Malawi transport planners are now reported to be addressing the deficiencies. 18. The proliferation of inconsistent VOC data over the years, also described in the PCRs,12/ is another weakness in road planning and evalua- tion with potentially detrimental effects on transport policies and invest- ment decisions. The Fourth Highway Project provides technical assistance and 9/ PCR, Credit 523-MAI, paras. 5.02-5.03; and PCR, Credit 758-MAI, paras. 5.01-5.05. 10/ Project Performance Audit Memorandum on First Highway Project, Credit No. 112-MAI (SecM75-854). ll/ PCR, Credit 523-MAI, para. 5.03; and PCR, Credit 758-MAI, para. 5.04. 12/ The PCRs make reference to substantial differences in appraisal data from 1974 to 1983 (four projects), which in turn were based on data generated in the country. -6- overseas training for transport planning and coordination, and under the Fifth Highway Project, further technical assistance is being financed. Implementation of these programs is expected to improve the quality of transport planning data. Another contribution toward resolving the issue of inconsistent measures of road user cost will come from the Bank, which by late 1984 intends to publish two Bank Research Reports related to this sub- ject. 19. DRIMP Phase I, financed under the Second Highway Project, did not include a formal economic analysis, presumably because physical planning and execution of works absorbed available resources, though some substitute evaluation procedure was applied. Neither was sufficient information mollected for an economic post-evaluation upon completion of the works. How- ever, under DRIMP Phase II (Fourth Highway Project) and Phase III (Fifth Highway Project), a road evaluation method was instituted, which will permit future er-post analyses. Each individual road included in an ongoing program is required to achieve an economic return of at least 12%, based on VOC savings and induced agricultural production, and to meet defined socio- economic standards. IV. TRANSPORT POLICIES 20. Road maintenance budgets, according to the PCRs, were adequate in 1977-79, but inadequate thereafter, even though they were raised by 10% annually up to 1981.13/ Neither of the two projects committed the Govern- ment to numerically defined funding levels, though they contained a standard clause under which the road network would be "adequately" maintained. The Third Highway Project also required the Government to "exchange views" with the Bank about its program for maintenance of district roads. At the Board meeting considering credit approval for the Third Highway Project, one speaker queried whether the Bank, which was funding only part of Halawi's road infrastructure, should concern itself with maintenance of all roads. The staff replied that some Governments, while receiving high financial cor- tributions for capital projects from foreign aid, may not give sufficient attention to maintenance budgets.14/ 21. Under the Fourth Highway Project, the maintenance problem was subr- sequently studied by consultants, and under the Fifth Project, it was agreed that from 1983184 on budgets for regular maintenance would gradually be raised to reach adequate levels within five years. The maintenance study (Fourth Project) also identified about 660 km of paved roads as candidates for surface treatment and rehabilitation and about 1,730 km of gravel and earth roads for heavy maintenance/improvements, and the Fifth Project provides financing for a large part of these works. 13/ PCR, Credit 523-MAI, paras. 3.21 and 4.02; and PCR, Credit 758-MAI, para. 4.02. 14/ Transcript of Proceedings, December 22, 1977. -7- 22. The PCRs for the two projects under audit point out that one factor in the shortage of maintenance funds was the rapid increase in the length of the improved road network. While tight or inadequate budgets are common place in many countries, the maintenance consequences of road improvements have been rarely analyzed in Bank documents on road maintenance problems in the past, though it is increasingly being done now. The experience of these projects suggests the desirability of the Bank's making routinely an assess- ment of the long-term budgetary conaequences of a country's road development plans and of a Government's ability to fund consequent maintenance budgets, whenever road maintenance is or could become an issue. 23. A measure to safeguard road infrastructure investments, complemen- tary to adequate road maintenance, is the enforcement of axle load regula- tions. There are at present three permanent veightbridges in place in Malawi - two at border points and one in the interior. Until 1976, law enforcement was generally lax,15/ but even at the time of the audit visit to the coun- try, regulations were not vigorously enforced, and the weightbridge in the Interior was closed at nightfall, though the road was open. The Fifth High- way Project, rather than relying on appropriate covenants only as it was done earlier, provides for the positive strengthening of enforcement capabilities, including provision of mobile scales. 24. At the time of appraisal of the Second and Third Highway Projects, the Bank made unsuccessful attempts to reach an understanding with the Government about appropriate railway investments and construction of a new international airport at Lilongwe. The discussion followed the completion of a consultant's study, financed under the First Highway Project, on liberal- izing regulations on road-rail competition and subsequent implementation of several recommendations. The Bank questioned in particular the rationale for major rail network extensions, which had either been completed in recent years or were about to be started, while the Government was determined to go ahead with diversifying its internal and external transport links. The Government's priority was to lessen its dependency on any particular country for foreign trade routes given the political constellation of the area. The feasibilities studies for Salima-Lilongwe and Lilongwe-Zambia rail lines financed by the Canadian International Development Association (CIDA) were not made available to the Bank for review.161 This experience for a period strained the usually cooperative relationship between the Bank and the Government. The investments in question have in the meantime been completed, thus removing a potential obstacle in the Government-Bank dialogue on transport sector issues. Under the on-going Fifth Highway Project, the Government has agreed to a covenant requiring annual joint Government-Bank review of its transport investment programs. 15/ PCR, Credit 523-MAI, para. 3.21; PCR, Credit 758-MAI, para. 3.17; and SAR, Loan 2363/Credit 1423/Credit SF6/NAI, para. 2.09. 16/ PCR, Credit 523-MAI, para. 6.03; and PCR, Credit 758-MAI, para. 6.01. All' -9- MALAWI SECOND HIGHWAY PROJECT - (CREDIT 523-MAI) PROJECT COMPLETION REPORT July 26, 1983 Eastern Africa Projects Transportation II Division ーンり一 ノ諭園がタ介プんK PROJ= COMPLETION REPORT NALWI SECOND HIGHM PROJEICT (CREDIT 523-MI) 1. TVIT DUCTION 1A1 At the identification stage of the Second Highway Project the principal features of Malawi's surface transport system were a main road network which was faiLrly well developed south of Lilongwe, but very undeveloped in the northern half of the country; a largely undeveloped lower class road network; a rail system in the southern half of the country which was primarily used for export-import and transit traffic (from Zambia); and lake transport which mainly served the north. Regional development in the country had been rather unbalanced up to Independence in 1964; most development was concentrated in the southern part of the -country, with a much lower level of economic growth in the central and, in particular, the northern part of the country, resulting in a southward migration of population. Subsequently, during the 1970s one of the prime objectives of the government was to achieve more balanced regional development. As part of this policy the administrative capital was moved some 240 km north from Zomba to Lilongwe, and agricultural projects were undertaken- 'in the northern half of the country. 1.02 To provide better access to the new capital city the Association financed the reconstruction of the road from Zomba to Lilongwe under the First Hiqhway Project. The Second Highway Project included a component to continue the road north from Lilongwe to Kasungu to two lane bitumen standard. Kasungu was part of the main road link Vo Mzuzu, the administrative center of the Northern Region. Sections of the road further north from Kasungu to Mzuzu have been included for upgrading in the Third, Fourth and proposed Fifth Highway Projects. 1.03 The major economic activity in Malawi is agriculture which accounted for 95% of the countrys' exports in 1973. District roads serving agricultural areas were in poor cohdition, mainly because the District Councils responsible for the maintenance of the roads had neither the skills, organization nor the resources to maintain the roads.- Therefore, the Second Highway project included a pilot program in one of the country's 24 districts to determine the most effective way to maintain the roads. 1.04 The information for this report was obtained from supervision reports, Association files, consultant progress and final reports on the construction of the Lilongwe-Kasungu road and District Road Improvement and Maintenance Program (DRIMP) comppnents of the project. - 11. PROJECT IDENTIFICATION, PREPARATION AND APPRAISAL. Origin of the Project 2 .01 IDA discussed several alternative roads for possible inclusion in the Second Highway Project during First Highway Project supervision and Second Highway identification missions in 1969. At first, five roads were identified by the Government for possible study, but gradually the number was reduced to four by the Association, given the"low economic priority of one of the roads, and then to two by the Government. The Permanent Mission in Nairobi questioned dropping one of the roads, Chikwawa-.Ngabu-Bangula, since it was regarded as a high priority road in an important agricultural development area. However, the problem was solved when USAID offered to finance the feasibility study for the road. The feasibility study for the other road was financed by the UK. The two roads finally selected for study for possible inclusion in the Second Highway Project were - 12 - Lilongwe-Mzimba, a continuation of the road from Lilongwe to the north, and Lilongwe-Mchinji-Mwami, the main road to the Zambia border. The former road was eventually selected for Association financing (para 2.02). Project Preparation 2.02 At the Association's suggestion, the Government requested UNDP to finance the feasibility study and detailed engineering of the above two roads. The UNDP accepted and the Association agreed to be the executing agency. The feasibility study started in June 1972 and was carried out by Scott, Wilson, Kirkpatrick and Partners (SWKP) in association with the Economist Intelligence Unit. The feasibility studies were completed in May 1973; the results showed that only the Lilongwe-Kasungu section of the Lilongwe-Mzimba road was economically justified. There followed some heated discussions between the Association and the Government about which road to proceed with for the detailed engineering. The Association wanted to go ahead with the Lilongwe-Champira section of the Lilongwe-Mzimba road, but the Government regarded the Lilongwe-Zambia border road as higher priority, even though the feasibility study did not prove the road economically viable. The Government finally agreed to go ahead with detailed engineering for the Lilongwe-Champira road after USAID expressed interest in financing both detailed engineering and construction of the Lilongwe-Zambia border road (the road was eventually financed by USAID and completed by 1979). Traffic on the Lilongwe-Zambia border road subsequently proved to be higher than projected, initially because of Zambian goods being transported through Malawi and later because of Malawian goods having to be transported through Zambia because of problems with transportation through Mozambique (para. 5.01). Appraisal of the Project 2.03 The consultants completed detailed engineering of the Lilongwe-Champira road by early 1974 and the project was appraised in March 1974. Although not in their terms of reference, the appraisal mission suggested that a district road maintenance component be included in the project, after discussions with the Government about the poor condition of the district (or tertiary) roads. The Government agreed to the idea in principle, but wanted first to determine the most appropriate maintenance methods by having a pilot program in one district only, to be later extended to other districts under future projects. Accordingly, US$0.24 million was included in the project to finance consultants to carry out a pilot program. The central district of Kasungu was chosen as the pilot district because of its importance as an agricultural area whose development was being particularly impeded by the poor condition of its feeder road network. 2.04 During the appraisal there was some dispute over the design of the Lilongwe-Kasungu road since the consultants and the Government maintained that available natural gravel was suitable for use'as a base course while the Association maintained that cement-stabilization was necessary. A compromise was reached by not including cement-stabilization in the tender documents but including it in the appraisal cost estimates; also a covenant was included in the credit agreement that the Government would consult with the Association before deciding on the type of base course to be used, following the necessary soil and materials tests to determine the adequacy of available natural gravel. As a result of this provision and other cost increases the economic rate of return for the Lilongwe-Kasungu road fell - 13 - from 21% in the consultant's report to 10% at appraisal (increasing to 14% after the bids were opened because the lowest bid was found to be less than the appraisal estimate). 2.05 The inclusion of the additional US$2.3 million for cement-stabilization and US$0.24 million for the DRIMP pilot program increased the total project costs to US$12.5 million, although the Association had only allocated US$7.0 million for the project. At the Decision Meeting, it was agreed to raise the credit amount to US$8.0 million and by green cover stage the credit amount had been increased to US$10.0 million, or 84% of project costs (net of ta-es). Because no major road work contracts had been awarded in Malawi for some years, estimation of the likely bid prices proved to be very difficult. Since the Government would have had problems financing any cost overruns the Association decided to delay presentation of the project to the Board until after bids had been opened, reviewed by Government and then reviewed by the Association. Credit Negotiations 2.06 Negotiations were held in September 1974 before.bid opening; hence project costs remained uncertain. The Association expressed concern that the enforcement of vehicle dimension and weight regulations was too lax in Malawi and the government agreed to a covenant in the credit agreement to adequately enforce such regulations. The Association also pressed for a proper system of traffic counts given that little was known about traffic trends in the country. The Government therefore agreed to collect the necessary traffic data in order to carry out sound transport planning. Although the road maintenance organisation was quite good in Malawi the Association considered that road maintenance standards were not adequate, mainly because sufficient funds were not provided by the Government. A covenant was therefore included in the credit agreement requiring that sufficient funds be provided for road maintenance to assure proper maintenance of the road network. Other covenants required consultation over the cement-stabilization of the Lilongwe-Kasungu road (para. 2.04), and that after the DRIMP pilot study proposals for reorganizing the development and maintenance of district roads would be discussed with the Association prior to implementation (para. 2.03). Loan Signing and Effectiveness 2.07 Bids for the Lilongwe-Kasungu road were opened in October 1974 and evaluation was completed by November. The credit was presented to the Board on December 3, 1974 and the credit was signed on December 19, 1974. There were no conditions of effectiveness and the credit became effective on February 21, 1975. III. PROJECT IMPLEMENTATION AND COST Physical Implementation 3.01 Lilongwe-Kasunqu Road Construction. Prequalification of contractors for the road construction contract took place during April-July 1974 and six contractors from five countries were prequalified. Bids were opened in October 1974; bids from the six tenders submitted varied from US$7.13 million to US$12.59 million (77% higher). The lowest bid was US$1.5 million less than the next lowest bid (about 20% lower), and was 13% - 14 - below the Association's appraisal estimates. However, the appraisal estimate was not changed before Board presentation, presumably because the Association thought the costs might increase later. The lowest bidder was a Korean company, Kong Yong Enterprises,.which had never worked in Africa before and had limited road-construction experience. Apparently, the company was very keen to diversify and enter the African market. The Association questioned the reliability of the tender, but further investigation by the Roads Department revealed no reason for rejecting the bid. Subsequently, the contract was signed on December 24, 1974 and construction began on April 12, 1975. 3.02 Completion of the main contract was scheduled for April 1977 but because of poor performance by the contractor the complete road was not open to traffic until March 1978, 11 months behind schedule. Only two months of this delay was due to factors beyond the contractor's control, mainly the unavailability of materials and poor weather conditions. The remaining seven months delay was due to financi3l and managerial problems. After completion of the main contract, it still took the contractor an additional 18 months to complete associated works (such as shaping and trimming of side slopes, ditches, roadsides and borrow pits, brick lining of drainage channels, erection of guardrails, fencing, traffic signs, and painting of road markings) and to carry out variation orders (paras. 3.07-08). Road works were finally completed in May 1980. 3.03 The contractor had difficulty in obtainifg explosives for four months during the first half of 1976, delaying production at the quarry and rock excavation on the roadworks. The delivery of plant and spare parts to the site was also delayed and heavier than predicted rainfall impeded constructon activities from December 1975 to April 1976. However, the major causes of slow construction progress were the contractor's cash flow problems, poor availability rates for plant and poor organisation. The company suffered from serious cash flow problems despite assistance from the Malawi Government with short-term advance payments and direct payment to suppliers. According to Kong Yong Enterprises, the company's cash flow problems were compounded by the reduction in May 1976 of the company's line of credit by its Malawi bank plus transfer restrictions imposed by Korea. It would appear, however, that the company was not financially stable. It was eventually taken over in March 1978 by one of the largest holding companies in South Korea, Tae Hwa Construction Co. As a result, financial performance improved and cash flor problems were diminished.- 3.04 The poor level of plant availability was due to the poor condition of equipment which was shipped from Vietnam and Cambodia at the start of the contract and the lack of spare parts. A large number of items were unservicable and many of the replacements wdre purchased second-hand in Malawi and had a poor servicability record. Sixty-seven units of major plant were imported into Malawi at the start of the contract and a further sixty-five units were acquired during the contract. When the road was completed only 22% of the original units and 51% Of the subsequently acquired units were servicable. The Government was able to assist during some of the periods when equipment was short by leasing its own equipment to the contractor. 3.05 Overall, there was inadequate site management and a complete lack of purposeful site organization. The supervising consultants introduced a regular monthly engineers's site visit in June 1976 in order to help improve the contractor's short-term planning. Visits consisted of an - 15 - inspection of works, a review of progress in the previous month with an in-depth investigation of the reasons for shortfalls between projected and actual progress, and a review of the contractor's forecast of progress for the forthcoming month. The contractor improved his capacity for organising his available resources, although continued to fall behind revised construction programs since there were still cash flow problems. 3.06 The Roads Department became extremely concerned about the delay in construction, especially since similiar problems had been encountered with the previous road construction project financed under the First Highway Project. The PPAR1/ for this project considered the Government's decision to evict the contractor and take over his assets as too slow, resulting in considerable delays and cost overruns. Perhaps because of this experience the Government acted quickly, for by March 1976, when the contractor was four months behind schedule, it demanded that Kong Yong Enterprises increase its monthly output to the-level where the increase would represent a work progress of about 5% of the contract amount per month. This was followed by a letter from the Government in April 1976 starting the process of declaring the contract forfeited under the conditions of contract. The Association was rather surprised by this action since the monthly progress reports from the supervising consultants did not discuss the possibility of terminating the contract. However, after discussions with the contractor and the introduction of the supervising consultant's monthly site visits, the Government withdrew the termination order in May 1976. This action was appropriate since the cost of having the construction completed by another contractor would probably have been more expensive than retaining Kong Yong. 3.07 Quality of road construction has been satisfactory. Only four short sections, totalling 1000 feet, showed pavement failure, but these were corrected by the contractor during the maintenance period. Two of the sections were corrected at the expense of the contractor because unsuitable base material had been used, while the other two were financed under the project because the failures had been caused by rising groundwater level, due to recent ngricultural activities fairly close to the road. In one area repairs were not carried out satisfactorily, but were finally corrected in late 1981. During construction the.consultants and Government's claim that no cement-stabilization would be necessary was found to be correct (para. 2.04), since suitable materials were found for the base course, resulting in substantial cost savings (para. 3.16). The contractor carried out work on three variation orders on the contract from September 1978 to November 1979 (during and after the maintenance period). These involved the provision of a roundabout north of the Lingadzi bridge, reconstruction of the road through Mponela, and additional drainage structures. 3.08 Five variation orders were undertaken by other contractors, since Kong Yong could not cope with the additional work load. All of the contractors acted as sub-contractors to Kong Yong to avoid the complications of separate contracts. Local bidding was used since all the sub-contracts were less than US$0.2 million. W & C French (4alawi) Ltd., a UK firm, carried out three of the variation orders involving a 10 cm asphalt concrete overlay for the Lingadzi Jridge and approaches, 1/ Malawi. First Highway Project (Credit 112-MAI). Project Performance Audit Memorandum; December 15, 1975. - 16 - additional drainage works, and a bridge to cross the new Salima-Lilongwe railway. At the start of the road contract the precise location of the railway crossing had not been determined and it was not until mid-1977 that the bridge location and design were finalised. Nello L Teer Co., a USA firm, carried out two variation orders, involving the reconstruction of a 4.8 km section of the Lilongwe-Salima road linking up with the project road and an asphaltic concrete overlay for the remaining 13 km section of road leading to the newly constructed airport north of Lilongwe. Consulting Services for Supervision of Road Construction. 3.09 The same consultants, SWKP, that had carried out the feasibility study and detailed engineering for the road were appointed as supervising consultants in December 1974. The performance of the consultants was good, particularly given the difficulties encountered with the contractor (pares 3.03-3.05). Every effort was made to assist the contractor overcome his problems: the regular site meetings set up by the consultants (para. 3.05) probably helped to prevent even further delays in construction of the road. The consultants' contract was extended because of the protracted construction period. District Road Improvement and Maintenance Program. 3.10 The district road network totalled some 5000 km in the 24 districts of Malawi. The objective of the pilot program in Kasungu district was to determine how to build up maintenance capacity at the district level and to improve roads that had deteriorated. Labor-intensive methods were to be tried given the very low cost of la?jr in the rural areas (MkO.3 per day in 1978). Actual maintenance and improvement operations began in Kasungu district in September 1976. About 96 km of roads were maintained during the first two years of the program, and some improvements were made, particularly for drainage and bridge works. Activities were sometimes impeded by the unavailability'of labor in the districts, particularly during the tobacco harvest. 3.11 Two major problems emerged from experience on the pilot program. The first was the lack of an adequate organisation and qualified personnel at the district level to take over and continue operations. The second was the shortage of funds at the district level for improvement and maintenance operations. These funds had to be supplemented by budget allocations from central Government. MWS's training center at Zomba was able to meet some of the training requirements for road foremen, mechanics and operators. In addition, accounting personnel were able to be trained in other training facilities. Organisational improvements were to be recommended by SWP as part of their study of the extension of the pilot program to other districts. The study was financed with cost savings from the construction component of the project (para. 3.16). Their final report was prepared by December 197B and outlined a road improvement and iaintenance program for the first phase extension of DRIMP to six additional units. IDA agreed that the bulk of the cost savings from the project be used to finance Phase I of DRIMP which would include the six additional units and continuation of the Kasungu unit. The Credit Agreement was amended accordingly on June 9, 1978 (Annex 1). 3.12 The consultant's recommendation was for labor-intensive methods for all maintenance (one laborer was to be allocated an average of 3.5 km of road to maintain) and most improvement activities. Some criticism was - 17 - expressed by certain staff in the Bank over the consultants recommendations to use grader-based methods for road formation work, particularly since experience in other parts of Africa had proved labor-based methods as effective for this type of activity. IDA proposed that, as part of Phase I of the program, a demonstration project be set up in order to examine comparative costs and outputs of labor-based and grader-based methods of road improvement. 3.13 The demonstration project was set up in July 1979 an the 45 km Thuchila- Phalombe road in Mulanje district and lasted one year. The results showed that costs of the two methods were similiar after taking into account supervisory and amortization costs. However, it also proved that labor-intensive methods could be used quite effectively for formation works as long as there was good organisation and supervision as well as a good task work system. No difficulties were found in recruiting labor. 3.14 Establishment of three improvement units in Mzimba, Salima, and Mulanje took longer than expected and the units were not fully equipped until December 1979. Fuel shortages in late 1979 (para. 5.01) and the onset of the rainy season delayed actual start-up of the improvement units until April 1980. Only four units, including the Kasungu unit, were eventually included in Phase 1, because of an increase in DRIMP and other project component costs (para. 3.18). The remaining three units were established in the second phase under the Fourth Highway Project. Improvements of 1100 km of district roads were completed by September 1981. Each improvement site was supervised by a foreman with a casual labor force of up to 120. One of the biggest problems was the shortage of foremen, so that provision had to be made for training of foremen for Phase II of the program. Also, additional expatriate engineers had to be hired because of the shortage of local engineers. Improvement costs for 1980 were about US$3,160/km and maintenance costs US$250/km. 3.15 Maintenance units were established and maintenance of the roads was good, but there was still a continuing problem of funding, even with central Government allocations. Therefore, the Fourth Highway Project financed incremental maintenance costs on a declining basis over the first three years of regular maintenance after handover of the roads to the district councils. Only minor changes were recommended for Phase II by the consultants compared to Phase I. Grader-based methods for formation work was still recommended in most circumstances, partly because labor availability was still a problem at certain times of the year. However, one unit in Mulanje showed that formation works could be carried out by labor-intensive methods at a lower cost than*equipment-intensive methods. Therefore, labor-intensive methods were increasingly used under Phase II of DRIMP. Project Costs 3.16 A comparison of actual and estimated project costs is made in Table 2. Lilongwe-Kasungu road construction costs (including contingencies but excluding variation orders) were 60% lower than estimated in the appraisal report, mainly because (a) cement-stabilization amounting to US$2.3 million proved to not be necessary (para. 3.07) and (b) the bid by Kong Yong was US$2.4 million lower than the appraisal estimate even without cement stabilization. Additional expenditures on variation orders amounted to US$1.5 million. - 18 - 3.17 In 1978 cost savings were used to finance part of the following expenditures: (i) US$0.25 million for the supervision of construction, representing a 60 increase over appraisal estimates;.(ii) US$1.5 million for variation orders to the Lilongwe-Kasungu road construction contract (parea 3.07-08); (iii) US$0.14 million for a consultant study to prepare a first phase extension of DRIMP (para. 3.11); (v) US$3.55 million for implementation of a first phase extension of the DRIMP pilot program to a further three districts (pars. 3.11); (v) US$0.02 million for a socio-economic pilot study of Mangochi district in preparation for DRIMP's second phase (pars. 5.05). 3.18 Schedules I and II of the Credit Agreement were amended in June 1978 to reflect most of the above additions (Annex 1), except that the original intention in 1978 was to extend DRIMP to six additional districts. However, Phase 1 of DRIMP had to be cut back to three additional districts (para. 3.14) because of unforeseen variation orders during the maintenance period, cost overruns on the extension of DRIMP, and 1978 and 1979. A breakdown of the increase in costs from 1978-1982 can be seen below: Increase in Project Costs from 1978-1982 UJS$ 000 (W) Lilongwe-Kasungu road construction: overlay to airport access road 145.75 drainage repairs 147.00 (ii) DRIMP: technical assistance for improvement units 137.50 technical assistance for labor-intensive project 126.35 improvement and maintenance cost increases 400.00 (iii) loss on exchange rate 437.50 TOTAL US$1,394.10 Source: Bank supervision reports Disbursements 3.19 The disbursement schedule for the project (Table 3) shows that actual disbursements continually lagged behind appraisal estimates. Only 46% of the credit proceeds were disbursed by September 1977, the date by which the loan was to have been completely disbursed as estimated at appraisal. Credit proceeds were not completely disbursed until March 1982. Disbursements were much slower than estimated because of slow and inexpensive construction of the Lilongwe-Kasungu road (pars. 3.02) and the inclusion of an additional project component, Phase I of DRIMP, financed with project cost savings (pars. 3.11). Phase I was not able to proceed until the pilot program had been substantially completed. Closing Date 3.20 The closing date for the credit was extended once, from 3une 30, 1978 to June 30, 1982 to allow completion of the Lilongwe-Kasungu road but mostly for the implementation of Phase I of DRIMP. - 19 - Credit Covenants 3.21 Most of the covenants of the credit agreement (para. 2.06) were adequately complied with. Governsent consulted with the Association on whether to use a cement-stabilized base for the Lilongwe-Kasungu road and on the extension of DRIMP to other districts. Axle load regulations were enforced with three weighbridges built in strategic locations. For some time these weighbridges were not operated but after some pressure from the Association they were-eventually opened and existing legislation-was enforced. The TRRL designed a traffic census system for Malawi in 19742/ and assisted with the training of Transport Statistics staff. A prograw of periodic counts covering all the country and continuous counts in selected locations was started in 1974. Automatic counters were also purchased and put into use. However, there have been problems with the reliability of the counts, since the traffic census reports show considerable fluctuations in traffic volumes. Lastly, with respect to the maintenance covenant, the Government increased road maintenance budget allocations by 10% annually up to 1981 (grants to district councils for district road maintenance increased by 50% 1978/1979), while allocations for public works were either stagnant or decreased. Maintenance of the main and aecondary road network improved and was considered satisfactory from 1977-1979. Thereafter, budgetary increases could not keep up with high inflation and the rapid increase in the length of the improved network and. by 1980 funds were barely sufficient for routine maintenance. Subsequently, the maintenance problem was studied under the Fourth Highway Project (para. 4.02). IV. INSTITUTIONAL DEVELOPMENT 4.01 The Roads Department of the Ministry of Works and Supplies was assisted by consultants in preparing and executing the project, given the limited resources available within the department. At the beginning of the project there were only seven engineers in the Roads Department, all of whom were expatriate, but by 1978 there were eleven civil engineers and one transport economist, of whom only four were expatriate. By 1982 this number had increased to 15 staff of whom only 7 were expatriate. Nevertheless, dependence on consultants for preparation and implementation of projects will continue for some time until more staff and experience are available. 4.02 Maintenance of main and secondary roads was considered adequate until 1980 when inadequate funding and declining performance of the Maintenance Department began to give cause for concern. Therefore, the Fourth Highway Prrject included a maintenanoe study to determine maintenance needs and the technical, financial and manpower resources required to meet those needs. The istudy was carried out by consultants, Renardet S.A., Geneva, and completed in 1983. The Fifth Highway Project includes implementation of a four year maintenance program for the main and secondary road network. 4.03 From the identification stage of the Second Highway Project the Association had been concerned about transport planning in Malawi, both in 2/ Transport and Road Research Laboratory. G.D Jacobs and D.H. Blackmore: Traffic Census Design in Malawi. 1974. - 20 - terms of the lack of manpower resources to carry out planning activities and the actual transport projects undertaken by the Government, particularly railway line extensions (para. 6.03). Since the first problem had not been resolved by 1978 a technical assistance component was included in the Fourth Highway Project to strengthen the transport planning unit of the Economic Planning Division (in the Office of the President) through provision of a transport economist and overseas training of local staff. 4.04 The district councils had a core of road employees, but their training as well as administration of road improvement and maintenance were weak. DRIMP was effective in building up the capacity of the district councils included in Phase 1, although problems were encountered with availability of trained supervisors (para. 3.14) and financing of maintenance activities (para. 3.14). It is unlikely that these problems will be entirely resolved for a few years since it takes time to build up a core of trained staff and local sources of funds for maintenance have still to be found. V. ECONOMIC REEVALUATION Lilongwe-Kasungu Road 5.01 The project road is a section of the Ml road which runs the length of the country and is of national importance insofar as it serves to unify the entire country - it will eventually form part of an international route to Dar-es-Salaam via Karonga. The Association has subsequently financed the upgrading and improvement of 254 km of the Ml road under the Third and Fourth Highway Projects and is proposing to upgrade a further 53 km under the proposed Fifth Highway Project. Little is known about the impact of the project on the population of 600,000 residing in the area served by the road. Most are small farmers growing maize, groundnuts, tobacco and other crops. No doubt some of the marketable surplus has been shipped to Lilongwe, given the increase in the city's population. However, Malawi has suffered quite badly economically during the last few years, not only because of the world-wide recession but also due to internal transportation problems in Mozambique which have disrupted import-export trade. Fuel and other vital supplies have been frequently cut off and other, more expensive, transport routes have been used, such as through South Africa, Zimbabwe and Zambia. The economic impact of this has obviously been felt in all sectors, including the agricultural sector. 5.02 Traffic forecasts are extremely difficult to make considering the unreliable nature of the traffic count figures and fluctuations in traffic levels arising from fuel shortages. 1982-83 traffic figures show a substantial decline in traffic (Table 4), with a 500 reduction on two sections of the project road. On the other hand, 1978 figures show a substantial increase in traffic over appraisal projections, and 1980 traffic counts for two of the three sections show an even larger increase (only partial data is available for the third section). An estimate of 1984 traffic levels has, therefore, been made on the basis of the mid-point of traffic between the high of 1980 and the low of 1982-83. A growth rate of 5% p.a. is projected from 1984 onwards compared with 9-8% at appraisal. For the third section, Lilongwe to Dowa intersection, 1978 traffic figures were used as the base for 1984, given the lack of data for the intervening -21- years. The 5s growth rate is based on consultant3/ and Association4/ estimates because main road traffic in Malawi, for lower levels ofceconomic growth are now projected for the future than during the 1970s. Traffic on the road varies from one section to another, as indicated in Table 4, with the section nearest Lilongwe having the highest volume. 5.03 The basis for the economic analysis is an assessment of the benefits for expected traffic and costs with and without the project over a 20 year period. The capital cost includes both construction and supervision costs, net of taxes. The principal quantifiable benefits are savings in vehicle operating costs (voc) and road maintenance costs. Difficulties were encountered in estimating voc since substantial differences were found between estimates of voc used in the 1974, 1977, 1980, and 1983 highway appraisals. The 1980 and 1983 appraisal estimates of voc were based on the Highway Design and Maintenance Model eatimates, but even these two sets of voc differed substantially, even after inflation was taken into account. It was decided to use the 1983 Fifth Highway appraisal estimates of voc, deflated to 1977 prices (Table 5). Total economic costs and benefits of the road gave an overall rate of return for the Lilongwe-Kasungu road of 16% compared to 14% estimated at appraisal. Economic returns on the three individual sections of the road indicated in Table 6 are as follows: Lilongwe to S32 intersection, 20%; S32 intersection to Mponela 12% and Mponela to Kasungu, 11%, compared to appraisal estimates of 21% for Lilongwe-S32 intersection and 10% for the other two sections. Despite the lower traffic levels, the rate of return is higher than estimated at appraisal because of lower than estimated construction costs (60% of appraisal estimate) and higher estimates of voc savings, as a result of using the Highway Design and Maintenance Model. The rate of return based on reevaluation does not take into account potential benefits attributable to the diversion of export/import traffic from relatively costly routes through Zambia, Zimbabwe and the Republic of South Africa to a more economical route via the main road north from Lilongwe a link to Tanzania and the port of Dar es Salaam. DRIMP 5.04 The appraisal report made no attempt to do an economic analysis of DRIMP since the original project only included a pilot project. Obviously, little thought had been given to measuring the economic benefits of the program, since the terms of reference contained no economic analysis requirements. For Phase 1, roads were selected by a fairly simple procedure. Access to social services and volume of agricultural surplus were assessed along with the strategic importance of the road within the highway network. From these ratings an overall ranking was made and compared to the subjective rankings of various concerned local and central government officials to arrive at a final ranking. No subsequent information is available on the roads finally selected. 5.05 The Association subsequently became more concerned with the economics of feeder road improvement and maintenance programs, so that the 3/ Scott, Wilson, Kirkpatrick and Partners. Luwawa-Mowe Feasibility Study 1983. 4/ IBRD. Appraisal of the Fifth Highway Project (Yellow cover). 1983. -22 - SWKP were requested to do a preliminary trial study using cost-benefit analysis of roads in Mangochi district, one of the districts to be included in Phase II of DRIMP. The study was financed out of the cost savings of the project. As a result of the study, the selection of roads was based on cost-benefit analysis using as benefits voc savings for nonagricultural traffic and increases in agricultural production made possible by easier access to farms and markets. VI. ASSOCIATION PERFORMANCE 6.01 Cost savings of US$4.7 million were the result of the overestimating of the cost of road construction. A large part of the overestimate (US$2.3 million) occurred because the Association insisted on making provision for cement-stabilization of the base, even though Government engineers and the consultants insisted that locally available materials were adequate. The Association's main concern was to make sure that the Government would not have to bear any cost overruns in the event that cement-stabilization was necessary,. given the very limited resources of the country. Given the circumstances, the Association was probably right to act prudently and include provision for the additional work. 6.02 The Project was adequately supervised, with-an average of two supervision missions a year, not to mention two appraisals during project implementation. With respect to road construction delays there is probably little the Association could have done to speed up the contractor'a progress. Adequate support was given for the development of DRIMP, and the use of the cost savings to finance Phase I of the program proved justified, especially when other public works projects had to stop during periods of fuel shortages while DRIMP continued to function. 6.03 At appraisal the Association expressed concern over the number of railway projects which had either been constructed in recent years (the Nkaya-Nayuci rail line linking up to the port of Nacala in Mozambique), or which were at the planning stage (the Salima-Lilongwe rail line and the Lilongwe-Zambia border rail line). It is difficult to judge whether the Association could have put more pressure on the Government to get a reconsideration of certain transport projects whose economic justification seemed doubtful. The situation was a delicate one for the Association given the Government's determination to go ahead with the railway projects and the involvement of bilateral donors in financing the projects. CIDA financed the Salima-Lilongwe and Lilongwe-Zambia border rail lines and claimed that they were economically viable projects: since the Association did not have access to the feasibility studies it could hardly categorically state that the lines were not feasible, but could only express "reservations" about them. The Governmentas priority was to lessen their dependency on any one particular country for foreign trade routes, especially given the political context of the area. To a certain extent the Malawians fear of having the Beira railway route cut off has been vindicated (para. 5.01), but the Nacala route has also proved to be equally unreliable because of poor condition of the track within Mozambique, thus defeating the main purpose of having an alternative route to the sea. The Lilongwe-Zambia border rail link has not carried much traffic since most of the road traffic going through Zambia continues by road in Malawi instead of transshipping to rail. -23- VII. CONCLUSIONS 7.01 The project was satisfactorily completed, albeit with some delays. Fortunately, the long delay in completing the Lilongwe-Kasungu road did not result in substantial cost overruns as is usually the case: in fact, the contract was completed well below original estimates. Much of the credit for this can probably go to the supervising consultants who were able to improve the management of the contract through regular and intensive site meetings. The Bank had expressed reservations about the competence of the contractor at the time of bid evaluation (para. 3.01), but it was not in a position to reject the bid, given that the contractor passed the formal evaluation procedures. Perhaps this is an example of it being very difficult to put the Bank's experience into a formal evaluation system. 7.02 The project road proved economically justified on the basis of voc and maintenance savings alone. If transportation problems continue in Mozambique (para. 5.01), the road may well become even more important if it is used as a major northern transport route for Malawi's import and export trade. With the cost savings from construction of the Lilongwe-Kasungu road it was possible to establish DRIMP, which has proved effective at maintaining roads in the rural areas. When many other operations were brought to a standstill, the labor-intensive improvement works on the district roads were able to continue because of their very low fuel requirements. MALAWI SECOND HIGHWAY PROJECT (CREDIT 523 - MAI) PROJECT COMPLETION REPORT ACTUAL AND ECPECTED PROJECT IMPLEMENTATION PERCENT OF WORK COMPLETED CONTRACTOR/ DATES OF BY EXPECTED CONSULTANT BID RECEIPT CONTRACT AWARD BEGINNING OF WORK COMPLETION OF WORK COMPLETION PROJECT COMPONENT NATIONALITY ACTUAL EXPECTED ACTUAL EXPECTD ACIJAL EXPECTED ACTUAL EXPE DATE A. Construction Lilongwe-Kasungu Road (114 km) Korean 10/74 10/74 1/75 12/74 4/75 3/75 9/78 2/77 . 33 B. Variation Ordejg Korean 9/78 10/79 To Contract A - UK 4/77 5/0 USA 12/78 5/60 C. Supervision of A and B UK 12/74 11/74 4/75 3/75 6/80 2/77 33 D. DRIMP Pilot Progra 5/75 3/75 Procurement 7/75 7/75 6/76 Improvement Works UK 7/76 * 9/78 Late 76 20 Z. Preparation . DRIMP Phase I UK 5/77 5/77 3/78 12/776/ F. 'DRIMP Phase I Execution 7/78 Procurement 7/78 3/79 Labor Intensive Works 79 8/80 7/o Improvement UK 4/80 8/79k 9/81 9/8D-' 0. Preparation DRUIP Phase II UK 1/80 2/80 1/ Eight variation orders undertaken at various times between stipulated dates. b/ Expected at time of project revision. lourcei Supervision reports, appraisal report and consultant progress and final reports. June 1983 HIALAWI SECONO HIGHWAY PnojEcT (CREOIT 523-MWl PROJRCT COMPLETION REPORT ACTUALOAND APPRAISAL ESTIMATES OF PROJECT COSTS (UhS AND US) Equivalent - Cillionn) ACTUAL COST AS PROPORTION OF APPRAISAL ESTIMATE OF COSTO/ ESTIMATE oF CONTRACT LOCAL FOREIGN TOTAL COST (%) WITH CONTRACT PROJECT COMPONENT ACTUAL COST AMOUNT WTqWfTRT WRTWITRUT WITH WITHOUT CONTINGENCIES AMOUNT () --------------- WIN ENC --------------- Original ProJect Components A. (1) Construction b Lilongwe-Kasungu 7.08 7.12 4.83 3.77 6.95 5.42 11.77 9.19 60 99 (11) Additions to Contract 1.28E/ B. Supervision of A 0.67h/ 0.33/ 0.06 0.05 0.36 0.28 0.42 0.33 160 203 C. DRIMP Pilot Program 0231h/ 0.10 0.08 0.20 0.16 0.31 0.24 100 - Sub-Total 9.34 4.99 3.90 7.51 5.86 12.50 9.76 Components Added To Project, 1978 Costs Estimated For Credit Amendmentb/ D. Preparation DRIMP Phase I 0.14-1 .02 .11 .13 E. Implementation Phase I 3.55c/ 1.10 2.50 3.6De/ F. Preparation DRIMP Phase 1I 0.02-" 0.02 0.02 Sub-Total 3.71 1.12 2.63 3.75 TOTAL 13.05 5.02 8.49 13.51 E/ At the exchange rate of US$1 - MkO.83 b/ At the exchange rate of US$1 = MkO.90 of At the exchange rate of US$1 - MkO.84 df At the exchange rate of US$1 - MkO.88 I/ Estimate for six additional DRIMP units but only three were financed under the project. 1/ At the exchange rate of US$1 a KkO.81 June 1983 - 26 - Table 3 MALAW! SECOND HIGHAY PROJECT (CREDIT 523-MAI) PROJECT COMPLETION REPORT Disbursement Schedule Actual Disbursement As a Percentage of Cumulative Disbursements Appraisal Estimate Up to Latest Qtr. or New Estimate or Disbursements as a Percentage of Appraisal Estimate IBRD/IDA for Future Quarters Fiscal Year Actual Total Appraisal (1:2) x 100 or k marter Disbursement Estimate (4:2) x 100 1974/75 June 30 1,100 1,500 73 1975/76 Sept. 30 1,330 2,100 63 Dec. 31 1,530 3,500 43 March 31 1,810 4,500 40 June 30 2,030 5,100 40 1976/77 Sept. 30 2,480 6,300 39 Dec. 31 3,000 7,800 38 March 31 3,200 7,800 36 June 30 3,510 9,600 37 1977/78 Sept. 30 4,550 10,000 46 Dec. 31 5,200 52 March 31 5,800 58 June 30 6,000 60 1978/79 Sept. 30 6,240 62 Dec. 31 6,330 63 March 31 6,650 67 June 30 6,730 67 1979/80 Sept. 30 6,940 69 Dec. 31 7,340 73 March 31 7,791 78 June 30 8,075 81 1980/81 Sept. 30 8,392 84 Dec. 31 8,703 87 March 31 9,126 91 June 30 9,440 94 1981/82 Sept. 30 9,709 97 Dec. 31 9,821 98 March 30 9,975 1/ 100 Closing Date 6/30/78 100 On March 5, 1982 MALAWI SECOND HIGHWAY PROJECT (CREDI 523-MAI) PROJECT COMPLETION REPORT ACTUAL AND FORECAST TRAFFIC (VPD) LILONGWE-532 INTERSECTION (29 KM) S32 INTERSECTION - MOPONELA (27 KM) MPONEKA - KASUNGU (70 KM) 1978 1982 1978 1982 1978 1982 VEHICLE TYPE IEfTMgT ACTL rORECAST ATA1MO T ACTUAL FOREECAST ACTUAL FORECAST ACTUAL FORECAST Care and Vans 233 - 114 - 122 - 78 - 110 - 55 - Land Rovers 71 25 - 37 - 16 - 40 - 13 - Buses 17 - 5 - 9 - 4 - 9 - 4 - Trucks: 7 Ton 113 - 53 - 60 - .42 - 88 - 21 - 14 Ton 16 - 5 - 6 - 2 - 7 - 2 - 27.5 Ton 9 - 11 - 7 - 8 - 6 - 6 - TOTAL 459 327 213 462 241 199 150 281 260 195 101 275 Appraisal Forecast PCR Forecast Growth Ratel 1978 - 1986 9% 1986 - 1996 8% 1984 - 1996 52 Sources Road Department traffic counts and IDA estimates June 1983 - 28 - Table 5 MALAWI SECOND HIGHWAY PROJECT (CREDIT 523-MAI) PROJECT COMPLETION REPORT Comparison of Vehicle Operating Costs (Tambala per km) Vehicle type Poor Gravel Bitumen 1977 data 1974 datal/ 1977 data 1974 datal! Car 16 8 9 6 Van 10 7 Landrover 26 12 14 9 Bus 55 35 35 26 Trucks: 7 ton 26 28 19 19 14 ton 41 - 28 - 27 ton 96 64 62 44 1/ Appraisal estimate Source: Draft appraisal report for Fifth Highway Project. April 1983. Costs deflated to 1977 prices. - 29 - Table 6 MLAMI SECM0 HIGHMAY PR=CT (EDIT 523-W) PRCT CMETION PRT Reevctiin ad appraisal Estimates of Economic Rates of Retur Major Cam ts Proportion of Rate of return () Total InvesbEnt Reevaluation estimate Apprai~ Estimate Lilxge4asugu: 70 16 14 Lilonge-S2 20 21 S52-bnela 12 10 mponela-Kasgu 11 10 Sour: IDA estimates - - ••ø 744; 31 aga i of 2 -Me World Gank 1818 H SceI NW. Waswgfa D.C. 30433 U.S.A. Tcphonh: 002) 3936X* CsWes: dITUAM0AD The ronoreble D. T. Matenje Unster of Firce r. 0. Box 30049 Llon:se 3 Re: Credit ro. 523 1I (Sc-ò~d ILIC;r.ay rroject) nenudt of :evelo-:nt Credic Aprec.ent' Uar Igr. M-inister: With respect to the request of the Govermsent of Xalavi that savinns under the above-n-ed Credi be utilized for pu=porcs of the Second Et;hray rrojecz, the A4socd~-tion proposes that the Developmen: rcdit Ågcemcnt for the rroject as amen.,d by a letter dated January 16, 1976 betwccn the Reublic of Malawl ard tha Acsociation le further asended as set out bclov: Ci) Schedulc . The folloming new -tablo -hould be substituted for the. tblac Sa paragraph 1: Anount of tc Credt allocatcd -Eprscd in % of Expe-nditure= Cater.ooy olL-ar £ruiva1l.) to be financed C.) Civil vorks t=der Part A of the Project 6,325,000 80 (2) Consultants' services: (a) • Part A of the roject 555,U00 ) 100% of foreiga (b) Part 3 of the expenditures .Project 110,000 ) (3) Development andI man- tenance c! dljtrICC ronds: (a) Part Dl of tha Project 200,900 ) 70% ) () Part 12 of the ) Project . 2.500.000 ) The Monorable -. T. Hatenje - 32 - June 9.. 1978 Annex 1 Page 2 of 2 Amount of the Credit allocated (Exprecssed in Z o( rxpenditures. Category Dollar Ecuivalent) to be financed (continued) (4) Unallocated 310,000 Total 10,000,000" (2) Schedule 2 The existing )?art 3 of the Projeet should be ren=.bered as Tart 3 (1) and the following new cri 3 (2) should be added: "(2) Extension of the pilot progra to be established under Part 3 (1) of the Project both uithin Lasungu District and to such other dstrict.s in which such extension is - justified by the study to oe undertaken under the said Part of the Project." Closin: Date 2hq Association also proposes the establishment of June 30, 1982 as the new Closin= Date for the Project. ledse confi= 7ZZr a5reemCnt tZ the hZ-e p=eoalz:' countersig=fng the enclosed copy of this letter and returnin; the same to us. Ver truly yours, By Michael H. -vehen Director, Country Programs Department Eastern Africa Regional Office Republic of I!alavi Byr 'C Cr e ril Y n TFOURTH HIGH-WAY.PROJEC -TRANSPORTATION NETWQRK ti pa %Songws ...q.......Pro|ect road,constçuctioin. .... Project.-rådd, Imýýövimeos Kardga Roade conotructc underCre is 112tM44 -O• qAMR A 523-MAI and 7581-MAII. KILOMETERS 0 20 40 6 Pavd road unde consruction MIES 0 20 40 60 -----Paved roads, proposed lumba - Principal gravel and earth roads Roade being studied %,,.A~~~ -s---RollWays Livingsoniao -- RollqW'undier construction PNYII<A EhiUe la e """"Steamer routes PLATEAU 4 Airports with scheduled se rvices NR TJPH Riveri R U M P Hl'~Rvr R National capital LY u5rwe ® District administrative centers iG O y - trict boundar. 1- - Regional boundaries Ekwenål ni International boundaries Mba Bay - - ---* f T1 %V ý N K Hý OUNTA NS A Y 1 z A M B I A Mzimb M 2CHI u DISTRICT ROAD IMPROVEMENT AND MAINTENANC1 PROGRAM (DRIMP) .enda Firt phase areas /(. Y WY/(CR.523-MAI) Socond phase *roas Cr758MA Dwangwa KHOTAKOTA ~ Ihookoa K A U N G Uhoko MOZAMBIQUE C E A . aCabral ... . ......Ntchisi' Bonga NTCHISI h oN3-MAI - Chipata (,,%MCH NJI ö" D 0 LIUMA iingal 合 IJ, ....... ... . - - - - - - - - - - fil til kJj 0 IBRD 15070 3sSEPTEMR 1980 ci 1a Ekwend 1INYANGWA ROREST MZ RESERVE- - Tirnbiri Mbowe - --KALWE F.Rlj NHT A -4- TNIKHW I FR H W R C irjc rodpHntrdo FORKEST- RESERVE CHI SIRA FR. -a RUVUO F.R. { MALAW i FOURTH HIGHWAY PROJECT "° PROJECT ROADS Project rod, construction ••••••••Project road. improvement K RUW.I. Road constructed under Cr758 MAi - Paved road ----Poved roads under construction Earth and gravel roads: - -Main Secondary ------ District and others . Regional heodquarters § District headquarters 0 Towns, villages Airports Rivers Areas subject to flooding /marsIes and swamps Forest reserves ------international boundaries 34-w ni - 37 - MALAWI THIRD HIGHWAY PROJECT - (CREDIT 75B-MAI) PROJECT COMPLETION REPORT JULY 26, 1983 Eastern Africa Projects Transportation II Division w I - 39 - PROJECT COMPLETION REPORT MALAWI THIRD BIGEWAY PROJECT (CREDIT 758-MAX) I. INTRODUCTION 1.01 The M1 road in Malawi is the country's south-north spine road which runs through most of the principal centers, from the Mozambique border in the south through Blantyre, Zomba,.Lilongwe, Mzimba, Rumphi and Karonga, to Chitipa on the Zambian border in the north. Up to the 1970s the section from Zomba to the north was earth/gravel and impassable for part of the wet season. The First Highway Project financed the reconstruction of the section from Zomba to Lilongwe, while the Second Highway Project financed the upgrading of the section from Lilongwe to Kasungu. At the time of appraisal of the Third Highway Project, financing had been obtained to upgrade the Rumphi-Chiweta- Karonga section (177 km) with KfW and EDF funding, and the Ekwendeni-Rumphi section (31 km) was being considered for ADB funding (the section was constructed in 1980-1981). The only remaining section from Zomba to Karonga to be improved was from Kasungu to Ekwendeni. Subsequently, the Third Highway Project included the upgrading of the section from Kasungu to Jenda and a feasibility study of the Jenda-Mzuzu section, part of which is being upgraded to paved standard under the Fourth Highway Project and part of which is being considered for upgrading under the Fifth Highway Project. 1.02 The upgrading of the northern section of the Ml road was part of the Government's policy to improve regional balance, since most development had been concentrated in the southern half of the country. To this end, the administrative capital had been moved to Lilongwe in the Central Region and several rural development projects had been undertaken around Mzimba, about 60 km north of Jenda. Since political disturbances in countries surrounding Malawi have affected the transport of import and export traffic the northern half of the Ml road has taken on new significance, since it can provide an alternative route through Tanzania once the northernmost section from Karonga to the Tanzanian border is completed. 1.03 The information for this report was obtained from supervision reports, Association files, consultant progress reports for the road construction components and the Jenda-Mzuzu feasibility study. II. PROJECT IDENTIFICATION, PREPARATION AND APPRAISAL Origin of the Project 2.01 Once the Second Highway Project had got under way the Association and the Government set about identifying components to be included in the Third Highway Project. The Karonga-Tanzania border road was considered as one alternative, but UNDP could not finance the feasibility study as requested by Government because of lack of funds. The other alternative identified in early 1976 was the upgrading of the Kasungu-Champira road, a continuation of the section of the Ml (Lilongwe-Kasungu) financed under the Second Highway Project. The section had been included in the original feasibility study and detailed engineering for the Second Highway Project, but had eventually been rejected because of its low economic rate of return and limited financing. The feasibility study and detailed engineering had been carried out by Scott Wilson Kirkpatrick and Partners (SWKP) from 1972-1974 and had been financed by the UNDP with the Association as executing agency. The economic rate of return had last been estimated in 1973, therefore, the Association agreed to Government's request to review - 40 - the economic viability of the road. If proved viable the road was to be included in the project. If not, a secondary road construction and improvement program was to be considered (this was eventually financed by CIDA). 2.02 The Association and the Government had also discussed the extension of the District Roads Improvement and Maintenance Pilot Program (DRIMP), financed under the Second Highway Project, to other districts. The pilot program was to determine the best method of improving and maintaining the feeder roads (district roads), which had been neglected and were generally in poor condition. The district councils were responsible for maintaining the district roads but lacked the skills, organization and resources to adequately carry out the work. The pilot program was being carried out by consultants, SWKP, in one district, Kasungu: one of the original intentions was to extend the program under the Third Highway Project. Preparation of the Project 2.03 By the time of preappraisal in November 1976 the economic viability of the Kasungu-Champira road had improved. This was because there were currently planned agricultural development projects in the Mzimba area, not scheduled for implementation at the time of the 1973 feasibility study. Also, traffic data showed growth in traffic more rapid than projected in the feasibility study, as well as there being higher vehicle operating costs (voc) as a result of the petroleum Price increases of 1974. Nevertheless, during preappraisal feasibility of upgrading the road to paved standard was still marginal. At IDA's request, new traffic counts were carried out in the Kasungu-Mzimba area, and estimates of traffic generated by new agricultural schemes were reassessed. 2.04 The only section of the Ml north-south road from Karonga southward to Mzuzu that did not have financing for detailed engineering was the section from Mzuzu to Champira. Therefore, a feasibility study and detailed engineering (if justified) was proposed for inclusion in the project. The Government also indicated their interest in obtaining IDA financing for a feasibility study of the Karonga-Chitipa-Tunduma road in the extreme north of the country. Overseas training of transport economists was also proposed for inclusion in the project, given the shortage of economists in the Economic Planning Division of the Office of the President, which was responsible for transport planning and coordination. 2.05 DRIMP maintcnance and improvement operations started in Kasungu district in July 1976. By March 1977 the Ministry of Works and Supplies (MWS) had submitted to the Association a detailed report on operations and results of the program to the end of 1976. About 95 km of mainly drainage improvements had been carried out on different sections of district roads and sufficient experience had been gained to establish the feasibility of extending the program to other districts. However, due to the considerable amount of trained manpower required both from the district authorities to implement the program and from the MWS to direct and supervise the program, it was clear that it would not be possible to extend the program to all of Malawi's 24 districts as originally envisaged. It was, therefore, decided to include the extension of the program to only six to eight districts under the project. - 41 - Appraisal of the Project 2.06 The project was appraised from February 15 to March 4, 1977. After reviewing the traffic figures the consultants had reestimated the rate of return of the Kasungu-Jenda (or Champira) at 15%. Therefore, the road was included in the project. The Karonga-Chitipa feasibility study was dropped from the project because CIDA was considering financing the study: even if CIDA financing had fallen through there would have been sufficient cost savings under the Second Highway Project to finance the study. Later in 1977 the UNDP agreed to fund the feasibility study but eventually had to drop the study because of its financial crisis. The overseas training of economists was also dropped from the project since the Government maintained that invitations from various sources for overseas training was well in excess of the number of qualified candidates. However, for some reason these requests never materialised, so that overseas training of transport economists was eventually included in the Fourth Highway Project, together with technical assistance for the Economic Planning Division. The feasibility study for.the Jenda-Mzuzu road remained in the project. 2.07 During appraisal it was decided to extend the DRIMP to eight other districts. The appraisal mission estimated that there would be sufficient cost savings for the financing of five district units under the Second Highway Project, leaving three to be financed under the Third Highway Project. During appraisal the Government informed the Association that two sections of the Lilongwe-Dedza road (reconstructed under the First Highway Project), were showing signs of serious structural deterioration and required strengthening. Traffic had increased 50% more than anticipated at appraisal, making the design thickness inadequate for the traffic load and resulting in subgrade failure in a black cotton soil area. Strengthening of the road sections was therefore proposed for inclusion in-the project, and the design was to be carried out by MWS's Design Department. 2.08 Total costs of the project were estimated at US$13.75 million. The Government was most concerned not to finance more than 20% of total costs, especially since it was required to meet the recurrent costs of DRIMP, amounting to about US$65,000 per unit year. These costs had not been originally included in the project cost estimates, since they were not considered capital cost items. The possibility of funding from the OPEC Special Fund was discussed, for OPEC was willing to make available up to US$1.8 million to Malawi. Parallel financing was preferable to all concerned, and the Association suggested that ptrengthening of sections of the Lilongwe-Dedza road and the Jenda-Mzuzu feasibility study were suitable components for parallel financing arrangements. This was discussed with representatives of OPEC, and agreement in principle was obtained. No final decision was to be taken until after tenders had been evaluated in order to have a more exact estimate of OPEC financing requirements. The Association had decided to postpone Board presentation of the project until after bids were received because recent civil works bid prices in Malawi had been unexpectedly high and the Government was concerned a out having to finance possible cost overruns. Besides, it was important to proceed with invitations to tender for the construction components of the project before Board presentation because a delay until afterwards would have meant that commencement of construction work would have coincided with the start of the rainy season, making for logistical problems in construction. - 42 - 2.09 By Green Cover stage of the appraisal DRIMP operating costs had been included in project costs, increasing total costs to US$13.6 million. This was done on the advice of CPS who maintained that CPM 2.5 provided for the classification of operating costs during an initial development period as a capital cost where IDA may be involved in financing of these costs. CPS considered that exclusion of operating costs because they were entirely funded by the Government would be inconsistent with the intent of CPM 2.5. 2.10 The only other major issue raised within the Association was the high share (46%) of planned Government capital expenditure during FY1978-1980. This was attributable to major investments in rail extensions and a new international airport for Lilongwe. The appraisal noted the Association's reservations about ongoing and planned rail extensions, but it was felt that there was little more the Association could do and that the project should not be delayed given its high priority and the fact that the project components did not duplicate other transportation investments. The Association had already discussed with Government the one remaining rail project to be implemented (Lilongwe-Zambia border), but the Government was determined to proceed with it and had already obtained financing from CIDA on very favorable terms. The matter was raised once again at the Bank's Annual Meetings in October with the Minister of Finance, but the only concession granted by the Government was an agreement in principle to exchange information with the Bank with respect to tkansportation investments. However, the Bank was never given the opportunity to review the feasibility studies for the Lilongwe-Zambia border rail line extension or the new international airport. Credit Negotiations 2.11 Negotiations were held from September 12 to 16, 1977. However, in late July there was an Association mission to Malawi to supervise the Second Highway Project and to discuss the Third Highway Project credit documents, so that many of the issues were discussed in the field. It was found that the cost savings for the Second Highway Project would be greater than originally estimated and it was therefore decided to finance the first phase extension of DRIMP with these savings and to delete the DRIMP component from the Third Highway Project completely. Subsequently, the Second Highway Credit Agreement was amended in July 1978 to include the financing of six DRIMP units with the available cost savings. However, costs turned out to be higher than projected, so that in the end only three units were established: a further nine units were established under the Fourth Highway Project. The elimination of the DRIMP component from the project was partially offset by an 80% increase in cost estimates for the strengthening of sections of the Lilongwe-Dedza road. Total Project costs were now estimated at US$13.2 million. The IDA contribution would be US$10 million, leaving the Government US$3.2 million to finance, including taxes of US$0.7 million. During negotiations it was agreed that OPEC funding would be requested for the two sections of strengthening plus a third six km section of the same road. An OPEC contribution of US$1.8 million would reduce the Government contribution to US$1.9 million, since it did not want to contribute more than US$2.0 million, excluding taxes, to the project. 2.12 Further discussions on transport investments (para. 2.10) both before and during negotiations resulted in both sides maintaining their - 43 - original position. The Government objected to part of a clause in the draft credit agreement requiring it to discuss budget allocations for road maintenance annually with the Association. The Government claimed that since maintenance had improved over the past few years and was assessed by IDA missions as satisfactory such a clause was unnecessary. Finally, a compromise was reached and the clause was changed to a requirement that the Government exchange views with the Association about its program for maintenance of district roads, including the financing of such a program. The rest of the clause was a standard maintenance clause requiring the Government to adequately maintain the classified road network. 2.13 The Government agreed to a covenant in the Credit Agreement requiring it to consult and exchange views with the Association on the results of the feasibility study for the Jenda-Mzuzu road prior to proceeding with detailed engineering for the road. Other covenants included in the credit agreement related to the enforcement of axle-load regulations and maintenance of an appropiate traffic counting system. Board Presentation, Loan Signing and Effectiveness 2.14 Between negotiations and Board presentation tenders for the construction of the Kasungu-Jenda road were evaluated. The lowest tender price was US$0.56 million higher than estimated, indluding contingencies, resulting in a total project cost of US$14.2 million (excluding taxes). Since the maximum amount of OPEC funding (US$1.8 million) was already accounted for, the Association agreed to provide a further US$0.5 million for a total credit contribution of US$10.5 million. 2.15 The project was presented to the Board on December 22, 1977. Through an oversight, IDA had failed to ask the Government if they wanted the additional US$0.5 million, but the Government was quite happy to accept the additional funding when informed that US$10.5 million had been approved by the Board. OPEC approved parallel financing of US$1.8 million on December 15, 1977. The credit documents were signed on January 18, 1978 and the credit became effective on June 29, 1978, six weeks behind schedule, because of delays in Government submitting required documents to the Association and OPEC. III. PROJECT IMPLEMENTATION AND COST Physical Implementation 3.01 Kasungu-Jenda Road Construction. Eight contractors from six countries were prequalified for the road constructi6n contract in July/October, 1977. Bids were opened in November 1977; tenders from the six contractors varied from US$9.07 million to US$11.8 million (31% higher). The lowest bid was US$0.38 million less than the next highest bid (about 4% lower). The appraisal estimates were changed to reflect the increased costs before Board presentation (para. 2.08). The lowest bidder was from a joint venture of two South African companies, Reef Lefebvre and Basil Read, Pty, registered in Malawi as BRL Construction Company. The contract was signed on January 24, 1978 and the contractor was on site by April 1978. 3.02 Completion of the main contract was scheduled for December 31, 1979, but construction was not completed until six months later, in June 1980: including variation orders construction was not completed until December 1980. Very little of the delay was caused by the contractor; overall his performance was very good. Much of the delay was due to fuel shortages which were sporadic from late 1978 to June 1979, followed by serious shortages until November 1979. The problem was mainly the result of transport difficulties in Mozambique; fuel supplies were unpredictable, but for what reason was not clear - it could have been insurgent activity or just poor organisation on the part of the railway company. Other reasons for fuel shortages were Malawi Railway's shortage of rolling stock and limited facilities for stockpiling in Beira. Extremely heavy rains delayed the start of construction works for one month and an unusually heavy rainy season during 1978-1979 also helped to delay progress. The contractor pressed claims for idle time during fuel shortages of equipment and expatriate and skilled staff for an amount of US$1.5 million, but after negotiations the amount finally awarded was US$0.875 million. 3.03 In October 1980 a herd of buffalo which normally roamed in Kasungu National Park were cut off from the park because of'construction of the Dwangwa Bridge. Since they had been the cause of three human casualties MWS wanted to tranquilise and move them back to the park at a cost of US$6,250, which was requested from the credit proceeds. This evidently caused some confusion in the Association, because it was unclear whether the Credit Agreement allowed the financing of such activities. At first, the Association replied that there would have to be an amendment to the Credit Agreement (without mentioning the word "buffalo"!), but later the Association lawyers pronounced that the existing wording of the agreement was such that the relocation of buffaloes was a matter of "interpretation". Finally, the buffaloes were relocated with proceeds from the credit. 3.04 Three variation orders were undertaken by the contractor: these -were a spur to Kasungu airport, improvements to the main street of Kasungu, and the reconstruction of Dwangwa Biidge. These additions to the contract were financed out of cost savings on the main road contract (para. 3.14). 3.05 Strengthening Works. MWS employed consultants, SWKP, to review the Ministry's design of strengthening works for sections of the Lilongwe-Dedza road and to supervise construction. During their review the consultants detected cracks and cracking patterns in parts of one of the sections to be overlaid at the end of the rainy season. Further investigations were carried out involving excavating pits through the pavement layers, sampling and testing the materials found, and checking the water table levels in the vicinity of the pits. The results showed that deterioation had accelerated to the point that 5.25 km required reconstruction (rather than overlays, as originally envisaged) and 17.25 km required strengthening, making a total of 22.5 km, 3.5 km longer than estimated at appraisal. These changes required a totally new design which was not completed until late 1978, a delay of about one year. -45 - 3.06 Since financing was limited the consultants selected priority sections of road to be included in tender documents. Only 1.25 km of reconstruction and 14.3 km of strengthening were included. Priority was given to strengthening those sections of the road where deterioation was apparent, but had not reached an advanced stage. Heavy maintenance was recommended for those sections of road where reconstruction was necessary. 3.07 Once the design was complete the consultants began the process of prequalifying contractors. However, there was another delay of three months because of confusion about Bank Group rules for prequalification of contractors in cases when the Bank administers the funds of a cofinancing agency, such as the OPEC Special Fund. MWS proposed disqualifying a contractor from prequalifying because of his nationality, in accordance with OPEC lending conditions. However, since the Association was administering the OPEC fund, prequalification of contractors had to follow Bank procurement procedures. Seven firms were prequalified, but only two submitted bids, given the small size of the contract. Bids were opened in August 1979; tenders from the two firms were US$2.14 million and US$3.36 million. The cost of the strengthening works had increased considerably over appraisal estimates because of the additional work required, higher than expected inflation and a 10% devaluation of the US dollar against the Kwacha. The Association suggested that the Government approach the OPEC Special Fund for additional financing. The Government indicated that it would, but afterwards changed its mind and never did request more funds - for what reason is not apparent. Meanwhile, MWS negotiated with the lowest bidder, W.C. French (Malawi) Ltd., a UK firm, for reduced works to match available funds; a contract was signed in February 1980 for US$1.47 with an option of reinstating full scope of works should more funds be available at a later stage. Reconstruction of all sections was dropped and strengthening was reduced to just over nine km from the 14.3 km originally tendered. 3.08 There was a further delay because both of the firms submitting bids had stipulated starting strengthening works in March 1980, at the beginning of the dry season, instead of September/October 1979, near the beginning of the wet season in December. The official commencement date was March 12, 1980 with the completion date set for the end of September. Works as stipulated in the contract were completed by the end of October, one month behind schedule because the contractor experienced problems in designing a suitable mix for the asphaltic concrete overlays and because his asphalt plants required ey*ensive overhauls before work could begin. By completion of the contract it became apparent that there would be some cost savings from construction of the Kasungu-Jenda road, therefore, the contractor was retained to strengthen a further three km of road. The extra work was delayed until the end of the rainy season, only beginning in February 1981 and ending on schedule by June of the same year. 3.09 Quality of construction has been good. There were some failures of the binder course on some downhill sections of the road caused by very heavy vehicles breaking up newly-laid asphalt. Additional factors may have been dust from traffic prior to laying and diesel spillage from broken-down vehicles, both of which could have led to poor adhesion between the binder course asphalt and the underlying surface. All affected areas were cut out and replaced and have given no further trouble. One fatal accident -46- occurred in August 1980 when a heavy cement truck and trailer ignored the traffic control measures and lost control while attempting to negotiate, at high speed, the one-way single-lane route past the construction area. Three men lost their lives when the truck jack-knifed into the paving operation. A condition of the contract was that the contractor maintain traffic on the existing carriageway and not use diversions. The supervising consultants concluded that, in hindsight, it would have been preferable to have used diversions, given the low level of traffic control observance in Malawi. Diversions may also have proved to have been a less expensive method of traffic control. 3.10 Supervision of Construction. SWP designed and supervised both road construction components. The performance of the consultants was excellent, particularly with respect to the many changes that took place with the road strengthening works. The consultants contract was extended for both contracts because of protracted construction works. 3.11 Feasibility Study and Detailed Engineering of the Jenda-Mzuzu Road. In October 1977, four firms from two countries were invited to submit proposals for the feasibility study of the Jenda-Mzuzu road. The evaluation was completed by March 197B, and SWKP was awarded the contract. Work started on July 10, 1978 and an Inception Report was submitted on September 8. The draft final report was meant to be completed by December, but there was a delay of just over two months with the report submitted in early March 1979. The delay was due to difficulties the consultants encountered with running the Bank's Highway Design and Maintenance Model, which they were using to analyse different maintenance and construction standard options. 3.12 The study identified two possible alternative routes: an eastern route passing over the Viphya plateau and a western route passing to the west of Viphya and with a more dense settlement pattern and ongoing and planned agricultural development projects. The feasibility study recommended the eastern route because of the planned Viphya Pulp Mill Project, which would generate considerable traffic. Since this particular project had been subject to many delays in the past the Association requested that the consultants carry out sensitivity analysis to determine which route would be selected if t1e Viphya Pulp Mill project was delayed by five to ten years. The consultants found that the eastern route would still be preferable if the Viphya project was delayed for five years and only marginally so if delayed for ten years. 3.13 Detailed engineering for the eastern route was given the go-ahead in September 1979: only 54 km was carried out since 89 km had already been designed under the Viphya Pulpwood Primary Road Network Project. Quantities were prepared in a form to facilitate the preparation of any of the possible sections of the route into individual contracts. Detailed engineering and tender documents were completed by October 1980; subsequently, as part of the Fourth Highway Project some of the sections were only upgraded to a low gravel standard because it was not economically viable to pave them. Some of these sections are now proposed for upgrading to paved standard under the proposed Fifth Highway Project. - 47 - Project Costs 3.14 A comparison of actual and estim.uted costs is made in 'able 2. Kasungu-Jenda road construction costs were only 2% lower than estimated at appraisal (including contingencies), despite claims for idle time during fuel shortages and additional works (paras 3.02-3.04). This was possible because quantity (physical) increases amounted to only 2% vis-a-vis 10% estimated at appraisal and price increases were also less than estimated. The savings of US$340,000 were used to finance (1) upgrading the road through Kasungu; (2) building a spur to Kasungu airport; (3) Dwangwa bridge; and (4) cost overruns of about U525,000 for supervision of construction of the road and US$320,000 for strengthening works on the Lilongwe-Dedza road (Para. 3.08). Cost overruns for the latter component were 3% higher than estimated at appraisal but the actual length of road strengthened was 37% less than estimated (paras 3.05-3.08). Costs for the feasibility study and detailed engineering of the Jenda-Mzuzu road were 6% lower than estimated at appraisal. The Credit Agreement was amended in June 1981 to allow the Association to disburse against road strengthening works (given that this component was originally to have been totally financed by parallel OPEC funding) and to reallocate Credit proceeds accordingly. Disbursements 3.15 The disbursement schedule for the project (Table 3) shows that actual disbursements were ahead of appraisal estimates for the first year, falling to 86% of appraisal estimates by the beginning of the third year and then increasing again. All of the OPEC Special Fund proceeds and 96% of the credit were disbursed by March 1981, the date by which all the proceeds were to have been completely disbursed according to the appraisal estimate. Credit proceeds were totally disbursed in March 1982. Disbursements were slower than projected because of delays in project implementation. At the beginning of the project the contractor for the Kasungu-3enda road became very concerned about being paid, since there was a three month delay disbursing against his mobilization advance. This occurred partly because the date of credit effectiveness was delayed (para. 2.15), there were computer problems in the Bank's Controllers Department, and because Controllers did not have the necessary authorized signatures. However, disbursement was made to the contractor in August 1978. Closing Date 3.16 The closing date for the credit was extended once, from June 30, 1981 to December 31, 1982 to allow completion Pf the strengthening works. Credit Covenants 3.17 Most of the covenants of the Credit Agreement (paras 2.12-2.13) were adequately complied with. Adequate consultations were carried out on the results of the Jenda-Mzuzu feasibility study before detailed engineering began. Axle load regulations were enforced with three weighbridges built in strategic locations. For some time these weighbridges were not operated, but after some pressure from IDA they were eventually opened and legislation enforced. The TRRL designed a traffic -48- census system for Malawi in 1974 and assisted with the training of Transport Statistics Unit staff. A program of periodic counts covering all the country and continuous counts in selected locations was started in 1974. Automatic counters were also purchased and put into use. However, there have been problems with the reliability of the counts, since the traffic census reports show considerable fluctuations in traffic volumes. 3.18 Lastly, with respect to the maintenance covenant the Government increased road maintenance budget allocations by 10% annually up to 1981 (grants to District Councils for district road maintenance increased by 50% 1978/1979), while allocations for public works were either stagnant or decreased. Maintenance of the main and secondary road network improved and was considered satisfactory from 1977-1979. Thereafter, budgetary increases could not keep up with high inflation and the rapid increase in the length of the improved network; by 1980 funds were barely sufficient for routine maintenance. Subsequently, the maintenance problem was studied under the Fourth Highway Project (para. 4.02). Adequate exchange of views took place with respect to the program and budget for maintenance of the district roads. IV. INSTITUTIONAL DEVELOPMENT 4.01 The Roads Department of MWS was assisted by consultants in preparing and executing the project, because of MWS's limited resources. Nevertheless, the Roads Department has been able to increase its staff since the beginning of the project and to gradually replace expatriate personnel with local engineers. At appraisal there were ten engineers in the Roads Department, five of whom were expatriate. By 1982 this number had increased to 15 engineers and two economists, of whom seven were expatriate. Nevertheless, dependence on consultants for preparation and implementation of projects will continue for some time until more staff with adequate experience are available. 4.02 Maintenance of main and secondary roads was considered adequate until late 1979, when inadequate funding and declining performance of the Maintenance Department began to give cause for concern. Therefore, the Fourth Highway Project included a maintenance study to determine --- maintenance needs and the technical,'financial and manpower resources required to meet those needs. The study was carried out by consultants, Renardet S.A., Geneva, and completed in 1983. The Fifth Highway Project includes implementation of a four year maintenance program for the main and secondary road network. V. ECONOMIC REEVALUATION Kasungu-Jenda Road 5.01 The Kasungu-3enda road is a section of the Ml road which runs the length of the country and is of national importance insofar as it serves to unify the entire country; it will eventually form part of an international route to Dar es Salaam via Karonga. The Association has subsequently financed the upgrading and improvement of 169 km of the Ml road under the Fourth Highway Project and is proposing to upgrade a further 53 km under the proposed Fifth Highway Project. Malawi has suffered quite badly - 49 - economically during the last few years, not only because of the world-wide recession but also due to transportation problems within Mozambique disrupting import-export trade. Fuel and other vital supplies have been frequently cut-and other, more expensive, transport routes have been used, such as through South Africa, Zimbabwe and Zambia. The economic impact of this has obviously been felt in all sectors, including the agricultural sector. 5.02 The economic downturn is reflected in the traffic figures for the Kasungu-3enda road as compared to appraisal estimates, which were based on rates of traffic growth in the early and mid 1970s. As shown in Table 4, traffic levels for 1980 are higher than projected at appraisal, 150 vpd compared to the appraisal estimate of 122 vpd. However, the fuel shortages arising from transportation problems in Mozambique (para. 3.02) and economic recession resulted in a decline in traffic throughout the country from 1980-1983. Figures were only available for the northern-most section of the road for 1982: this section had a 33% decline in traffic from 1980. Therefore, traffic for 1982 for the whole road was estimated at 66% of the 1980 level. Traffic forecasts are extremely difficult to make considering the fluctuations in traffic levels over the past five years. However, to project traffic from 1984 onwards traffic levels between the high of 1980 and the low of 1982-1983 have been taken and used as the base for a future growth of 5% per annum. This projection is based on consultantl/ and Association2/ estimates for main road traffic in Malawi, given the projection 'f lower levels of economic growth than during the 1970s. These projections compare with original appraisal estimates of 9'. traffic growth until 1989 and 8% thereafter. 5.03 The basis for the economic analysis is an assessment of the benefits for expected traffic and costs with and without the project over a 20 year period. The capital cost includes both construction and supervision costs, net of taxes. The principal quantifiable benefits are savings in voc and road maintenance costs. Diff2culties were encountered in estimating voc since substantial differences were found between estimates of voc used in the 1974, 1977, 1980, and 1983 highway appraisals. The 1980 and 1983 appraisal estimates were based on the Highway Design and Maintenance Model estimates, but even these two sets of voc differed substantially, even after inflation was taken into account. It was decided to use the 1983 Fifth Highway appraisal estimates of voc, deflated to 1977 prices (Table 5).- Total econmomic costs and benefits of the road gave a rate of return of 8Z compared to 14% at appraisal. 5.04 The road compcnent shows a low rate of return mainly because of lower than projected traffic levels, and a different composition of the traffic, Justification of the road at appraisal depended very much on the large number of heavy trucks projected, comprising 44% of total traffic. However, the limited traffic counts available show heavy trucks represented only 70 of total traffic (Table 4). The rate of return based on reevaluation does not take into account potential benefits attributable to the diversion of export/import traffic from relatively costly routes through Zambia, Zimbabwe and the Republic of South Africa to a more economical route via the main road north from Lilongwe, a link to Tanzania and the port of Dar es Salaam. 1 SWP. Luwawa-Hbowe Feasibility Study. 1983 2/ IBRD. Appraisal of Fifth Highway Project (Yellow cover). 1983. - 50 - Strengthening Works 5.05 At appraisal it was estimated that unless strengthening was carried out on 19 km of the Lilongwe-Dedza road, complete reconstruction would be necessary. As it turned out strengthening was only carried out on 12 km of the road because of the increase in costs (para. 3.07). Information is not available on what has happened to those sections not strengthened. Traffic figures for the Lilongwe-Dedza road show that 1981 levels were below those projected at appraisal - 444 vpd compared to 509 vpd projected. No traffic counts were available after 1981. Therefore, it has been assumed that traffic declined by 40% from 1981 to 1982/83, given the economic decline throughout the country. The 1984 level of traffic was projected taking into account the actual traffic in 1981 and the subsequent general traffic decline in the country; thereafter, the same 5% growth rate was applied as in the case of the Kasungu-Jenda road. 5.06 The appraisal considered that reconstruction would be necessary in 1984 if strengthening was not carried out. However, this was using a far higher growth of traffic than has materialised or is projected for the future. Therefore, reconstruction is now considered not necessary until 1992. This would give a rate of return of 9% as compared to 13% estimated at appraisal. An alternative method is to assume that no reconstruction would take place if strengthening was not carried oUt, but that voc would gradually increase as the road deteriorates. Assuming voc savings would increase by 10% per annum the rate of return would still be 9%. VI. ASSOCIATION PERFORMANCE 6.01. At appraisal the Association expressed concern over the number of railway projects which had either been constructed in recent years (the Nkaya-Nayuci rail line linking up to the port of Nacala in Mozambique), or which were at the planning stage (the Salima-Lilongwe rail line and the Lilongwe-Zambia border rail line). It is difficult to judge whether the Association could have put more pressure on the Government to get a reconsideration of certain transport projects whose economic justification seemed doubtful. The situation was a delicate one for the Association, given the Government's determination to go ahead with the-railway projects and the involvement of bilateral donors in financing the projects. CIDA financed the Salima-Lilongwe and Lilongwe-Zambia border rail lines and claimed that they were economically viable projects: since the Association did not have access to the feasibility studies it could hardly categorically state that the lines were not f6asible, but could only express "reservations" about them. The Government's priority was to lessen their dependency on any one particular country for foreign trade routes, especially given the political context of the area. To a certain extent the Malawians fear of having the Beira railway route cut off has been vindicated (para. 5.01), but the alternative routes have not proved to be very viable. Only small amounts of traffic have been moved over the Nacala route because of the poor condition of the railway line in Mozambique: also, the Nacala port facilities are somewhat limited. The Lilongwe-Zambia - 51 - border rail link also has not carried much traffic, since most of the road traffic going through Zambia continues by road in Malawi instead of transshipping to rail. 6.02 The project was adequately supervised, with an average of two supervision missions a year, as well as one appraisal during project implementation. VII. CONCLUSIONS 7.01 Physical components of the project were satisfactorily carried out. The economic justification of the Kasungu-Jenda road was marginal, even at pre-appraisal (para. 2.03). The Association's request for the consultants to review the traffic figures, particularly those generated by agricultural developments in the area, may well have led to over-optimistic projections of heavy traffic. However, the benefits of the project only took into account voc and maintenance cost savings. There are undoubtedly unquantifiable benefits of linking up the central and northern parts of the country to the more prosperous south and now the road may become of great importance if transport problems continue in Mozambique. If the main justification of the road is one of regional integration and access to a northern international transport route, then the only question remaining is whether Malawi needed to have two paved roads up to Mzuzu in the north, one running along the west side of Lake Malawi (not all paved) and the road of which the project road forms a part, the Ml. 7.02 The extent to which the strengthening of sections of the Lilongwe- Dedza road is justified is difficult to assess without more data on the condition of the road. In the past "guestimates" of when reconstruction would be necessary without strengthening have not been that reliable, erring on both the optimistic and pessimistic side. In the future a better estimate of when strengthening or reconstruction will be necessary can be made using the Highway Design and Maintenance Model. MALAWI THIRD HIGHWAY PROJECT (CREDIT 758 - HAI) PRn3ECT COMPLETION REPORT ACTUAL AND EXPECTED PROJECT IMPLEMENTATION PERCENT OF WORK COMPLETED CONTRACtnR/ DATES OF BY EXPECTED CONSULTANT BID RECEIPT CONTRACT AWARD BEGINNING OF WORK COMPLETION-OF WORK COMPLETION PRUIECT COMPONENT NATIONALITY ACTUAl EC ACTUAL EXPECTf ACTUAL EXPErED ACTUAL EXPECTED DATE A. Construction Kasungu-Jenda Road (08 km) S. African 12/77 12/77 1/78 2/78 4/78 4/78 6/80 12/79 80 B. Reconstruction Dwanga Bridge (Addition to Contract A) S. African 4/80 12/80 C. Strengthening and Reconstruction Sections Lilonge-Dedza Road (12 km) UK 8/79 3/78 2/80 4/78 3/80 5/78 6/81 12/78 0 1 Ln D. Supervision of A and 8 UK 2/77 2/77 12/77 11/77 12/77 12/77 1/81 4/80 65 1 S. Design and Supervision of C UK 6/77 6/77 5/77 5/77 6/81 12/78 5 F. Jenda-Hauzu Road: Feasibility Study UK 4/78 1/78 6/78 3/78 7/78 5/78 8/79 3/79 90 Detailed Engineering/ 11/79 4/79 9/79 4/79 10/80 3/80 60 Tender Documents Sources Supervision reports# consultant progress reports and appraisal report June 1983 MALAWI THIRD HIGHWAY PROJECT (CREDIT 758-HAI) PROJECT COMPLETION REPORT ACTUAL AND APPRAISAL ESTIMATES OF PROJECT COSTS (US$ AND US$ Equivalent - Millions) ACTUAL COST AS PROPORTION OF APPRAISAL APPRAISAL ESTIMATE OF COSTS/ ESTIMATE OF CONTRACT LOCAL FOREIGN TOTAL COST () WITH CONTRACT PROJECT COMPONENT ACTUAL COST AMOUNT WITH WITHOUf WITH WITHOUT WITH WITHOUT CONTINGENCIES AMOUNT () ---------------CONTINGENCIES--------------- A. Construction Kaoungu-Jenda (88ka) 11.5b/ 9.05a/ 4.94 3.80 6.78 5.25 11.72 9.05 98 127 B. Strengthening and Reconstruction of Sections Lilongwe-Dedua (12 km) 1.94/ d/ 1.474/ 0.54 0.48 1.34 1.12 1.88 1.60 103 132 C. Supervision of A 0.61h/ 0.431/ 0.14 0.09 0.47 0.36 0.61 0.45 100 142 1 D. Design and Supervision of B 0.15d/ 0.13d 0.01 0.01 0.09 0.07 0.10 0.08 150 115 w E. Feasibility Study and 0.734/ 0.311/ 0.18 0.11 0.60 0.45 0.78 0.56 94 101 Detailed Engineering for 0.4ik/ Jenda-zuzu TOTAL 14.93 11.80 5.81 4.49 9.28 7.25 15.09 11.74 af At the exchange rate of US$1 - 0.89 k/ At the exchange rate of US$1 * 0.82 cf Reduced to 12Km, compared to l9Km estimated at appraisal. d/ At the exchange rate of US$1 - 0.85 e/ At the exchange rate of US$1 - 0.81 June 198a - 54 - Table 3 THD HIQIMY PROET (CÆIT 5234 I) PAO:EW CORETD FEPORT Diebureent dch1e Actual Didbur As a Percentate of CTuulative DiEburseents Pgrapiml Estimae up to Latest Qtr. or New Estimate or Didursmemnts as a Percetage of pprai Estimte IBFD/IDA for Future Qu=rters Fical Year ctual Total Apprai~1 (1:2)xlO or Quater Disa nt Estimate (4:2)x100 1978 Sept. 30 Dec. 31 Ma~c 31 Jo 30 1,80 0 1979 Sept. 30 2,600 2,200 118 Dec. 31 3,430 2,80 123 March 31 4,680 4,00 117 Jure 30 5,190 5400 % 19E0 Sept. 30 6,100 6,8 90 Dec. 31 7,283 8,200 89 Mard~ 31 8,081 9,2m 88 1"e 30 8,652 10,100 86 1981 Sept. 30 9,613 10,100 95 Dec. 31 9,735 10,100 % Mach 31 10,052 10,50 % Jre 30 10,134 10,500 97 Sept. 30 10,259 98 Dec. 31 10,451 99 March 30 10~ 1/ 100 Clasing Date 6/3V81 Juv 1983 HAI.AWI THIRD HIGHWAY PROJECT (CREIT 758 - MAI) PROJECT COMPLETION REPORT ACTUAL AND FORECAST TRAFFIC KASUNGU-3ENDA (88 KM) LILONGWE- - DEDZA (12 KM) 1980 1982 1981 1984 VEHICLE TYPE ACTUAL FORECAST ACTUAL FORECAST ACTUAL FORECAST ACTUAL FORECAST Cara and Vans 46 19 30 23 190 - 154 - Land Rovere 25 26 17 31 58 - 47 - Buses 5 6 3 7 22 - 18 - Truckes 7 Ton 63 23 42 27 ( - ( - 14 Ton 9 (59 5 (70 (174 - (141 - 27.5 Ton 3 L 2 L L TOTAL 150 133 99 158 444 509 360 659 Appraisal Forecast PCR Forecast Growth Rate: 1980 - 1989 9z 1990 - 1999 8 1984 - 2000 352 Source: Road Department traffic counts and IDA estisates June 1983 HALAWI IHIRD HIGHWAY PRO3ECT (CREDIT 758-MAI) PROJECT COMPLETION REPORT COMPARISONS OF VEHICLE OPERATING COSTS (TAMBALA - PER KH)*. Vehicle Type EARTH/GRAVEL POOR BITUMEN BITUMEN and Cost Iteits New Inro. - 1979 Appraisal - 1977 New Info. - 1980 Appraisal - 191 New Info. 1979 Appraisal - 1977 Ca. ( 8 ( 6 ( 5 24 13 12 Val 11 ( 9 ( 8 Land Rover 42 14 21 11 18 10 Bull 85 30 53 21 46 19 Tricke 7 Ton 41 36 ( 24 25 20 14 Ton 50 (8 (40 (54 36 (46 27.5 Ton 148 ( 541 ( 0% Sotree Draft appraisal report Fifth Highway Project 1983. Coste deflated to 1979 or 1980 prices. *100 Taubala M Nkl.0 June 1983 '- Ln ；江！〕，,-:&; -! ro ro dhnc MALAWI FOURTH HIGHWAY PROJECT' GONChitipö TRANSPORTATION NETWORK lp Sonow Project road.comtructic>n .......... Proicct road, improvemerta Karån ag Ronds constructod under Crodita 112-MAI, A N A 523-MAI and 758-MAI KILOMETERS 0 20 40 60 Povadroads ýO 190 Paved road& under construction MILES 0 io 4,0 60 Poved road%, proposed bilumbo Principal grovel and earth road* .......... Rocidå being studied Rallways eý Livingsiontao Raiiway under construction NlilkA Chiweti ............. Sreaffler routes 10ý W6WO Airportå with scheduled services /V R U M P, Rivere N Rumph National copital Iv iuorwe 0 District adrninistrative centers 1, .: ..- - - District boundarics RE l usisya I ) ... ~ ~ .. Regional boundories Ekwandeni % international boundaries M z Aibowe .»hato Bol~ ~ N K H A T A ýÅOLINT N5 8 A y Z A MY B I A ýMzimbä' hikangawa -12@ CHýsultfy"lb, rv; %-. u'. DISTRICT ROAD IMPROVEMENT AND MAINTENANCE PROGRAM (DRIMPI Jend Firar phase arem (C R. 523 -MAI) Second phose areas hwangwa At '04,SKHOTAKOT )Nkhotalkom K A 5K N G U Kofungu MOZAMBIQUE VM Cabral Nfchýsl' Renga . t N r C HI51 I pata MCH N D 0 S LIMA 1 Chi Dowa f il .,en20: A 森 ---------- uj LL: CO óg :2.7 W4 .... .... .. lý4 -4 C.,V rw 2ZI! IBRD 15070 kn -.•4 SEPTEMBER 1980 LUNYANGWA F.R. Timbiri - bowe - KALWE F.R NHATA BAY sri F.ROJECT ROADS s. ~----raect road, construction ----?roject road, improvement KUWIL RRoad constructed under Cr758 MAJ ----Paved roadstunder constructian Earth and gravel ods: MazoMnia ----District and athers ®ä Regional headquorters § District headquarters o Tons, viRlages Areas subect to flooding/marshes and swamps eFrestrsv --International boundaries ScOooo31