Third Highway Project -Tanzania

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 4031 PROJECT PERFORMANCE AUDIT REPORT TANZANIA THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) June 30, 1982 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contenti may not otherwise be disclosed without World Bank authorization.  FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT TANZANIA THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) TABLE OF CONTENTS Page No. Basic Data Sheet. ......... ............................. iii Highlights........e........ *..e.*...... ................. v PROJECT PERFORMANCE AUDIT MEMORANDUM I. Background.... .......... ** *** ooo .................. 1 II. The Project................................ ............. 2 Identification and Appraisal........................... 2 IDA Project Processing...o............................ 4 Project Implementation (1972-1974)..................... 5 Project Implementation (1974-1978)..................... 8 II . Ea u to . . ..1. . . . . . . . .. . . . . . . ... 10 Annex: African Development Bank Comments........................ 14 PROJECT COMPLETION REPORT Summary and Concltisions...............e................ 16 I. Introduction...... ...................................... 24 II. Project Preparation and Appraisal...................... 26 III. Project Implementation and Costs........................ 30 IV. Institutional Development and Borrower Performance...... 41 V. Economic Reevaluat:ion................................... 44 VI. The Role of the Bank Group...............e . ............ 50 VII. Conclusions............................................. 51 Annex: Sources of Information o... e.e.o..o.o.............. ...... 52 Tables 2.01 Appraisal Design Standards, Mtwara-Masasi Road...o...... 53 3.01 Feeder Road Betterment Program.......................... 54 3.02 Actual and Appraisal Estimates of Project Costs......... 55 3.03 Schedule of Disbursements.............................. 56 3.04 Actual and Expected Project Implementation.-........... 57 5.01 Mtwara-Masasi Road, Construction and Supervision Costs...o..... . .. .... ....... .. .. ......... 58 5.02 Mtwara-Masasi Road, Actual and Forecast Traffic.....o.. 59 5.03 Mtwara-Masasi Road, Detailed 1978/79 Traffic............ 60 5.04 Mtwara-Masasi Road, Vehicle Operating Costs............. 61 5.05 Geita and Mara Feeder Roads, Betterment and Supervision Costs.................................... 62 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  - 1 - PROJECT PERFORMANCE AUDIT REPORT TANZANIA THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) PREFACE The following is a performance audit on the Third Highway Project in Tanzania for which Credit 265-TA (US$6.5 million) was made in August, 1971. The Credit became insufficient to cover the increased costs of the project and was supplemented in February 1975, by the use of a surplus US$1.9 million of Loan 586-TA, originally granted to assist finance the Second Highway Project (PPAR, dated June 1982). The project was expected to be completed in June 1974, and the credit account to be closed in December 1974. However, the closing date was extended long after that date and the Credit was not fully disbursed until February 1978. At that time, all remaining funds were inadvertently applied to accrued expenditures, even though one element of the project remained unfinished. Disbursements of the US$1.9 million from Loan 586-TA continued until September 1979. The main component of the project, paving of the Mtwara-Masasi road (200 km), later reduced to just Mtwara-Mingoyo (82 km), was completed in August 1978, at costs about 215% above appraisal estimates (about 185% if just Mtwara-Mingoyo is considered). One of the feeder road improvement components, #that for the Mara region, was completed in March 1977. Works on the other, in the Geita peninsula, were stopped in December 1978, following the inadvertent full disbursement of the Credit and the consequent lack of funding for them. At the time, only 75% of the Geita roads had been completed. Preinvestment studies included in the project were carried out by consultants. Although the consultants used acceptable procedures, they expressed little confidence in their results and recommendations because the studies were executed with little cooperation from the regional authorities. The training element that had been built into the feeder roads component was unsuccessful, and the continuation of the staffing and training program initiated in the First Project (1964) was not carried out. The attached Project Completion Report (PCR) was prepared by the Eastern Africa Region's Highway Projects Division without the assistance of the Borrower. The PCR reviews in a comprehensive and critical manner implementation of the project. For this reason, the audit has purposely avoided, as much as possible, repeating the content of the PCR. An Operations Evaluation Department (OED) mission visited Tanzania in July 1981, to discuss the project with Government officials. The assis- tance of the Government and the Ministry of Works, in particular, is grate- fully acknowledged.  - ii - OED has reviewed the Appraisal and President's Reports, the Minutes of the Executive *Director's Meeting at which the project was approved, Bank records and files and the PCR. Based on these and the field visit, the audit concludes that the PCR covers most of the issues arising from implementing the project. However, the audit considers that certain salient issues, although touched upon in the PCR, ought to be emphasized and highlighted. This is done in the Audit Memorandum. The draft audit report was sent to the Borrower and the African Development Bank for comments; the African Development Bank indicated it had no comments on the report (see Annex) and no comments were received from the Borrower.  - iii - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET TANZANIA THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) KEY PROJECT DATA Appraisal Actual or Item Estimate Reestimate Total Project Cost (US$ million) 9.50 25.18 Overrun (%) - 165 Credit Amount (US$ million) 6.50 6.501a Disbursed - 6.50 Cancelled - 0 Repaid to May 31, 1981 - 0 Outstanding to May 31, 1981 - 6.5L/b Project Completion Date 06/74 11/78&- Proportion Completed by Expected Date 100 0 Proportion of Time Overrun - 150 Economic Rate of Return (%) 14 <4 /d Cumulative Estimates and Actual Disbursements_e (US$ Million) FY72 FY73 FY74 FY75 FY76 FY77 FY78 FY79 FY80 Estimate 0.4 3.1 6.3 6.5 7.5 8.4 - - - Actual - 0.2 0.2 2.3 3.9 6.3 7.0 7.3 8.4 Actual/ - 6 3 35 52 75 83 87 100 Estimated (%) OTHER PROJECT DATA Item Original Plan Actual First Mention in Files or Timetable 03/20/69 Government's Application 01/03/70 Appraisal 10/69 11/22/70 Negotiations completed 05/70 05/28/71 Board Approval 06/70 07/27/71 Credit Agreement Date 07/70 08/06/71 Effectiveness Date 10/12/71 Closing Date 12/31/74 12/31/79/f Borrower United Republic o Tanzania Executing Agency Ministry of Works% Fiscal Year of Borrower July 1 - June 30 Follow-on Project Name Fourth Highway Project (Maintenance) Credit Number Credit 507-TA Amount (US$ million) 10.2 Credit Agreement DatE0 08/21/74  - iv - MISSION DATA/h Month/ No. of No. of Staff- Date of Mission Year Weeks/- Persons weeks Report Identification 05/69 1.0 1 1.0 06/69 Preparation 06/69 1.0 2 2.0 07/79 Preappraisal 10/69 0.5 1 0.5 12/69 Preappraisal 12/69 0.5 2 1.0 12/69 Preappraisal 06/70 0.5 2 1.0 07/70 Appraisal 11/70 2.7 4 11.0 04/71 Supervision I 08/71 0.3 1 0.3 09/71 Supervision II 03/72 0.7 2 1.4 04/72 Supervision III 06/72 1.0 1 1.0 07/72 Supervision IV 11/72 2.0 1 2.0 12/72 Supervision V 02/73 1.0 1 1.0 03/73 Supervision VI 08/73 0.7 3 2.1 10/73 Supervision VII 05/74 0.8 1 0.8 07/74 Supervision VIII 10/74 0.5 2 1.0 11/74 Supervision IX 04/75 0.5 1 0.5 06/75 Supervision X 02/75 0.4 2 0.8 02/75 Supervision XI 03/75 0.4 2 0.8 04/75 Supervision XII 09/75 0.8 1 0.8 10/75 Supervision XIII 11/75 1.0 1 1.0 01/76 Supervision XIV 03/76 0.3 2 0.6 05/76 Supervision XV 10/76 0.3 2 0.6 11/76 Supervision XVI 03/77 1.0 1 1.0 05/77 Supervision XVII 10/77 0.6 2 1.2 12/77 Supervision XVIII 10/78 0.6 2 1.2 12/78 Supervision XIX 03/79 0.2 2 0.4 04/79 Supervision XX 06/79 0.5 1 0.5 06/79 CURRENCY EXCHANGE RATES Name of Currency (Abbreviation) Tanzania Shilling (TSh.) Appraisal Year Average US$ 1 = TSh 7.14 Intervening Year Average US$ 1 = TSh 7.70 Completion Year Average US$ 1 = TSh 7.44 /a US$1.9 million surplus funds from Loan 586-TA were used with the approval of the Board tq finance cost overruns of this project. /b Includes US$0.01 million exchange adjustments. /c Project was not fully completed. In November 1978 work on the Geita feeder roads component, the remaining unfinished element, was stopped. Disbursements continued, however, for committed expenditures until September 1979. /d The rate of return for the Mtwara-Masasi highway was 4%, and the rate of return for the two feeder road components is estimated to be negative (PPAM, para. 37). /e Includes disbursement of US$1.9 million from Loan 586-TA used to finance cost overruns of this project. /f The Credit was fully disbursed inadvertently in February 1978. Loan 586-TA was fully dli.sbursed in September 1979. 4/ Prior to 1974, Ministry of Communications, Transport and Labor. /h This project had important manpower inputs from the Regional Mission in Eastern Africa (RMEA). This is not included in the mission data. /i Missions often dealt with more than one project, figure shown i. estimate of time si)ent on this Droiect.  - v - PROJECT PERFORMANCE AUDIT REPORT TANZANIA THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) HIGHLIGHTS The project waS not a success, although two of its purposes: paving the road between Mtwara and Masasi and carrying out preinvestment studies, were realized. However, the first was accomplished with significant cost overruns (215%), and the second with significant qualifications (PCR, paras. 3.08 and 3.26). In contrast, the improvement of feeder roads in the Geita peninsula and the Mara region were short lived (PCR, para. 18), training of Tanzanian engineers uet limited success, and the staffing program was not carried out (PCR, para. 13). The goal of stimulating agricultural production (Appraisal Report, para. 5.01) was not reached as cashew nut production in the Mtwara-Masasi area has dropped (despite this and two parallel Bank Group projects designed to more directly assist it), cotton production in the Geita peninsula has also dropped (despite a parallel IDA project to assist increase it), and milk and dairy production in the Mara region has failed to material- ize (despite a parallel Danish Government project aimed at assisting it). Investments in the Mtwara-Masasi road had a very low rate of return (4%) and those in feeder roads had negative rates of return, making the overall rate of return of the project less than 4% (PCR, para. 15 et. seq., PPAM, para. 37). The audit highlights the following points of interest in connection with this project: - the absence of clear objectives led to the implementation of a project misconceived in the early identification stages of the project cycle (PPAM, para. 30); - the absence of an implementation plan for the different components of the project negatively affected its execution and later pre- vented the materialization of a longer term feeder road maintenance impact (PPAM, para. 31); - an overly optimistic assessment of the staffing of the executing agency precluded addressing the problem properly during project preparation and ended up adversely affecting the execution of this as well as subsequent IDA projects (PPAM, paras. 24, 32, 33; PCR, para. 22); - the combination of executive, training and institution building functions in the same group of- people (consultants) prevented effective execution of these functions on account of incompatibili- ties in the use of time and effort each requires (PCR, para. 3.40, PPAM, para. 33);  - vi - - the non-observance of many IDA guidelines, procedures, principles and policies was deleterious to the efficient design and execution of the project and contributed to the low level of project achieve- ment (PPAM, paras. 34 and 35; PCR, para. 5.15 et. seq. and section VI); - IDA's consistency in yielding to Government on most points of disagreement contributed to the perception that in the case of Tanzania the need to transfer resources had taken precedence over the achievement of reasonable economic returns; thiL adversely affected project control and supervision (PPAM, para. 36). The audit ftlly concurs with the conclusions and lessons drawn by the PCR, and highlights, without repeating them, paragraphs 21 and 22 and 6.01 to 6.04 of the Completion Report.  PROJECT PERFORMANCE AUDIT MEMORANDUM TANZANIA: THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) I. BACKGROUND 1. Bank Group involvement in Tanzanian highways dates to 1964 when the First Highway Project (Credit 48-TA, US$16.9 million) was initiated. That project ran into trouble (PPAR No. 791, June 1975) when in 1965 most senior engineers in the Roads and Aerodromes Division (RAD) of the Ministry of Communications and Works (COMWORKS) left the country. Since few Africans had the education and training necessary to carry out the work, RAD's capacity was seriously impaired. Other problems marred implementation of the project and by 1966 costs were out of control. In 1967, IDA reappraised the project, cut it back significantly and in 1968 provided Tanzania with a supplementary credit (US$3.0 million) to finance the cost overruns. Implementation of the project continued and construction of most of the roads was completed in 1970. However, except for one, all road sections had lower rates of return on completion than expected at appraisal, including construction of the Mingoyo- Masasi Road which was later included for paving in the project which is the subject of this audit. Moreover, three of the six roads had rates of return below 10% (including Mingoyo-Masasi) suggesting that the cutback in 1967 should have been more drastic. Also, the staffing and training program designed to assist RAD overcome its personnel problems had not been imple- mented satisfactorily on completion, and in early 1970, Tanzania was con- sidered in default of specific credit covenants related to it. 2. While the First Project was being carried out with considerable difficulties, the Bank made a second loan to Tanzania in 1969 (US$22.5 mil- lion) to reconstruct parts of the Tan-Zam highway (PPAR dated June 1982). This project was expeditiously concluded by foreign consultants and contrac- tors. The project was suitable for the large scale deployment of foreign personnel which also effectively helped Tanzania overcome its critical short- age of trained staff. The project was completed in 1972 although contractor and consultant's claims were not settled by Tanzania until 1974. 3. Aware of its staffing difficulties, the Government decided to recruit foreign engineers to its staff, especially from India. In 1969, it hired fifteen engineers from that country.. However, in early 1971 almost all had left the Ministry because of lack of advancement opportunities (PPAM 791, June 1975, Tanzania First Highway Project). Furthermore, in August 1971, the Chief Engineer (also of Indian origin) resigned under charges, later dismissed by the courts, that his past decisions had caused the Government financial losses. This brought about the resignation of a number of other senior engineers (most of them foreigners) all of whom had long service in the Government. They were succeeded, in most cases, by junior Tanzanian engineers rapidly promoted to fill the vacant positions. The new appointees were - 2 - inexperienced and were naturally cautious about making decisions, particularly on financial matters. The shortage of skilled personnel in Tanzania in the early 1970's was compounded by the disinclination of foreign contractors and consultants to work in Tanzania, not only because of the Government's position with respect to private contractors and consultants and its slowness in pay- ing, but also because likely professional recruits viewed employment in Tanzania as less attrative than other opportunities. 4. This was the backdrop at the end of the 1960s and early 1970s, against which IDA identified, appraised and approved the Third Highway Pro- ject. IDA was concerned that personnel weaknesses at RAD would affect the implementation of the project, and, at the same time, it was concerned that delays in project preparation were causing the lending program to slip. II. THE PROJECT Identification and Appraisal 5. In early 1969, a mission visited Tanzania, among other things, to identify the possible content of the Third Project. The mission recommended that an important portion of the project be devoted to feeder roads to build on previous efforts under the First Project in the Geita peninsula, and on USAID-financed consultants' work in the Mara and Mbulu regions. The mission also recommended inclusion of a maintenance and betterment program, further feasibility and engineering studies, and continuation of the staffing and training program. However, barely a month after the mission returned to Washington, IDA's Programs Department suggested dropping the feeder roads component in favor of further main road construction (as in the Second Pro- ject), as a means of accelerating appraisal. As a result, another mission went to Tanzania; it recommended that the Geita feeder roads remain in the project, two self-contained minor road construction units be included, the maintenance program be replaced by a self-contained betterment unit for trunk and secondary roads, and studies be included to monitor the economic results of the road improvements, and to review the mechanical and maintenance equip- ment organization. The Government was advised of the necessary steps for an early appraisal in March 1970. 6. In December 1969, however, IDA staff recommended to management that Tanzania be declared in default for failure to execute the staffing and training program in the First Project and that further preparation of the new project be made contingent upon improvements in this area. Another mission was then sent to Tanzania. This mission agreed with the Government that the project would comprise five parts: (a) continuation of the staffing and training program; (b) construction of feeder roads in Geita, Mara and Mbulu; (c) a road betterment program, meaning improvement works on the Masasi- Tunduru-Songea road (450 km); (d) feasibility and engineering studies; and (e) a road maintenance reorganization study. However, the engineering for the - 3 - Geita roads needed reviewing, and consultants had yet to be hired to engineer the Masasi-Songea road, the Mara and Mbulu feeder roads and to assess whether construction by force account or by contract would better suit Tanzanian conditions. Over reservations about RAD's capacity to prepare the project by July 1970, appraisal of the project was rescheduled for that date. Soon thereafter, however, appraisal was postponed to October 1970, to allow the consultants time to review, in particular, the question of how to execute the works. 7. Strong concerns were expressed within IDA over this slippage. In Tanzania, possibly in an effort to accelerate the project, the Government decided that all works should be done by force account against the agreement that the matter would be studied by the consultant. Discussions as to whether the subject should be studied by the consultants, and over the terms of reference to do it, delayed hiring the consultants to June, 1970. 8. In July 1970, another mission went to Tanzania to review progress in project preparation. This mission found that although there were good hopes of finding economically justified works in the Mara and Mbulu areas, the roads in the Geita peninsula would need substantial re-engineering to be economi- cally feasible because of high construction costs. The mission also concluded that serious doubts existed of finding economically justified works on the Masasi-Songea road because of low traffic volumes. Appraisal was then re- scheduled for November 1970. 9. As the appraisal date drew close and it became evident that the studies would not yield meaningful results by that date, and given the dim prospects of an acceptable rate of return for the improvement works, IDA and the Government agreed to replace the improvement works on the Masasi- Songea road by the paving of the Mtwara-Masasi road (about 200 km) which had just been built to gravel standard under the First Project. This possibility had been studied by consultants and found to offer reasonable prospects of an adequate rate of return, a finding that contrasts with the subsequent evalua- tion of the First Project (PPAR 791) where construction of the Mingoyo-Masasi section to gravel standard was shown to have a rate of return of just 6%. But, the border with Mozambique had become strategically important; special opera- tions were being carried out there and IDA itself intended to support a cashew nut production project in the area. The consultant hired to study the Masasi- Songea road was then instructed to concentrate on engineering the paving of the Mtwara-Masasi road including the preparation of soil studies, identifica- tion of sources of'materials, pavement design, preparation of bills of quanti- ties, bidding documents, etc., an enormous task for the time available. The appraisal date was confirmed, however, and the decision on sufficiency of the data was left to the appraisal team. 10. Reviews of the Geita feeder road improvements, meanwhile, determined that increased cotton production in the peninsula, on which the justification of the improvements hinged, would only materialize if, in addition to the roads in the project (about 300 km), another 400 km of minor roads leading to them were also improved. These, however, were to be handled in a cotton production assistance project which was being readied concurrently. In the Mara and Mbulu districts the roads were expected to generate considerable agricultural. development but estimates of returns were not available until the consultant's report was issued. 11. Appraisal took place as scheduled in November 1970 and although traffic counts and expected traffic growth rates, as well as test results of soils and materials, and pavement design, in short the final engineering for the Mtwara-Masasi road, were not available to the mission, it concluded that the information was sufficient to complete the appraisal (PCR, para. 2.04). The mission also concluded that the Geita and Mara roads would yield accept- able rates of return if just minor construction, betterments and maintenance were done, and that the Mbulu roads would better be handled outside the project. The project as appraised included: (a) paving the Mtwara-Masasi (200 km) road; (b) improvement of feeder roads in the Geita (310 km) and Mara (165 km) districts, to be done by betterment and maintenance units (BMUs) established and equipped under the project, operated by departmental forces but coordinated and supervised by consultants hired under the project; (c) preinvestment studies; and (d) continuation of the staffing and training program (PCR, para. 1.05). Consultants would provide three teams of experts, one for the overall coordination of the feeder roads program and two for the supervision of the BMUs. The Government, for its part, would appoint suitably qualified engineers to act as counterparts to the teams who would be trained in the execution of the different activities. IDA Project Processing 12. The missing information on the Mtwara-Masasi road, including cost estimates, was sent to Washington later. As was subsequently learned though, the information was based on preliminary results only and not on final engi- neering (PCR, para. 3.03). When this information was processed, it revealed that paving the end section of the road, Nanganga to Masasi (60 km of the 200 km), was not economically justified as it yielded only an 8% rate of return (PCR, para. 2.04). The low returns were caused by low traffic volumes. The section was deleted from the project and IDA's Loan Committee approved a project proposal that only included paving between Mtwara and Nanganga (140 km). Negotiations for the credit were held in May 1971, at which time the Tanzanian delegation pressed strongly for the reinclusion of the Nanganga- Masasi section despite its doubtful economic justification. The Transporta- tion Projects Department was consequently asked to reconsider the economic merits of paving the section. The Department reviewed the case and concluded that the rate of return would increase to 9% if: (a) only direct construction costs were counted, i.e., if preliminary, mobilization, withdrawal and over- head costs were excluded; (b) certain unquantified benefits were generated by the project; (c) traffic was assumed to grow at more than 5%; and (d) unquan- tified benefits were to accrue on account of an accelerated economic and social development of the area (PCR, para. 2.05). IDA's Programs Department was satisfied that on these grounds the road section could be reinstated and the Chairman of the Loan Committee concurred. - 5 - 13. Shortly before presenting the project for Board approval IDA's Transportation Projects Department was informed by the Regional Mission in Nairobi of the difficulties the Cotton Project in Geita peninsula was running into, its likely delays and poor prospects of a significant impact. However, the Department considered that, despite its earlier finding (para. 10), the highway program did not depend on the cotton project, and could, therefore, continue to be processed independently. [The Cotton Production Project did not materialize until 197L. (Credit 454-TA, US$17.5 million) and cotton produc- tion has since decreased rather than increased.] 14. The Appraisal Report did not discuss these risks; neither did it mention the performance of the Government under the staffing and training program of the First Project; nor did it allude to a rather significant difference of opinion between IDA and the Government on hiring consultants for the coordination and supervision of the feeder roads component. The Credit was approved in July 1971, with general praise for the project. However, when one Executive Director expressed concern about the staffing situation at the Ministry, he was informed that all previous projects had contributed significantly to training local staff: plant operators, mechanics and the likes. Project Implementation (1972-1974) 15. The Credit Agreement was signed in August 1971, and the Project became effective in October 1971. However, it took three years to commence the works. When the credit was signed the consultant was still working on,the engineering studies for the Mtwara-Masasi road and only preliminary results were available (PCR, para. 3.03). For this reason, although drafting of the tender documents was started in August 1971, they were not completed until early 1973. There were major delays because the initial draft of the tender documents did not comply with IDA Guidelines and because the engineering had to be completed. However, despite the delay, the bill of quantities used at tender was still not supported by complete final engineering, but by just nominal estimates of the amount of earthworks (PCR, para. 3.03 and 3.29). Tender documents were approved by the Bank in January 1973 without final engineering, and were then delayed until August 1973, because the Government had not finalized the contract with the consultant for the supervision of the works (PCR, para. 3.02). The contract for the supervision was not signed until November 1973, after protracted negotiations over fees and conditions, and in the face of some Government officials' reluctance to employ consultants despite credit covenants to this effect. 16. Hiring the contractor took a similarly extended period of time. Applications for prequalification were called in early 1972; in October 1972, the Government submitted a first list of six prequalified contractors, of which only four were acceptable to IDA. However, the mentioned delays in preparing the tender documents prevented calling for tenders. When the tender documents were agreed upon a year later, in August 1973, the Government submitted a new list of prequalified contractors. This list included the - 6 - four on the first list in addition to three others who had been prequalified even though they had applied after the closing date for applications (PCR, para. 3.03). Of the seven prequalified contractors, four made site visits and only two submitted bids in November 1973. Bids were about 100% above apprai- sal cost estimates (PCR, para. 3.03). The large difference in costs was attributed to the lack of firm cost data at appraisal and the absence of final engineering at the time. The high bid prices were linked to the continued lack of complete detailed engineering (PCR, paras. 3.03 and 3.29), the impact of inflationary pressures which occurred in Tanzania between the estimated bidding date (early 1972) and the actual bid date (late 1973), and the remote- ness of the area and its closeness to the on-going civil war in Mozambique. Detailed engineering still remained to be completed three years after apprai- sal and the Government and IDA agreed to have it done during construction, just ahead of the contractors work front. 17. The high prices led to a reassessment by IDA of the economic merits of the project (PCR, para 3.04). The reappraisal showed (PCR, para. 5.03) that two of the four sections into which the road had been divided had rates of return of 10% (Mtwara-Mingoyo and Mingoyo-Mtama), one had an 8% rate of return (Mtama-Nanganga), and the other a 2% rate of return (Nanganga-Masasi, the same section excluded at appraisal). Yet, as the PCR demonstrates (paras. 5.09 et. seq.), the figures used in the reappraisal contained a number of mis- takes which went undetected because of extremely lax review procedures.1/ The PCR shows that, had these mistakes been identified and corrected, the reappraisal would have shown the first three sections of the road with rates of return well under 10% (Mtawa-Mingoyo-Mtama-Nanganga, 6%, 7%, and 4% respec- tively), and the last section with a negative rate of return (Nanganga- Masasi, -2%). 18. Although the rates of return even under the mistaken basis of the reappraisal were marginal, IDA management decided to agree to awarding the contract to the lowest bidder and continue with the project, albeit with a reduced IDA participation (on account of the higher costs) and a recommenda- tion to the Government to endeavor to reduce costs by reducing the design standard of the Nanganga-Masasi section (PCR, para. 3.04). In reply, the Government advised IDA in February 1974, that it would not reduce the standard for the last section and that it would secure additional financing to cover the cost overruns. 19. The Government then set out to negotiate a US$4.8 million loan from the African Development Bank (AfDB). When the AfDB loan was granted, in October 1974, IDA's participation in the project was cut back to assisting only the Mtwara-Mingoyo section of the road (about 80 km), while the remaining 1/ The Eastern Africa Projects Department Highways Division agrees that the review procedures for the reappraisal were extremely lax. It has indi- cated that a recurrence of the mistakes is highly unlikely as review procedures have been tightened considerably. - 7 - three sections .(Mingoyo-Mtama-Nanganga-Masasi) became the AfDB's project. In January 1975, the legal description of the Project was changed to reflect the reduced scope and, in addition, the Loan Agreement for the Second Highway Project (Loan 586-TA) was modified to use a US$1.9 million surplus in it to assist paying for the cost overruns. IDA disbursements were then correspond- ingly increased from the 70% of costs agreed at appraisal for the full road to 85% of construction costs for the reduced project (PCR, para. 3.28). 20. Parallel to this a similar chain of events was unfolding on the feeder road components. At appraisal it had been agreed, and later confirmed during negotiations, that the feeder roads component of the project would be done by force account but that the works would be coordinated by consultants and that each of the implementing units (BMUs) would also be supervised by the consultant. The Credit Agreement contained specific covenants to this effect. Yet, soon after the project became effective in October 1971, the Government questioned the need for the consultants. The Government wished to proceed without their assistance arguing that its in-house capacity was adequate, that by taking full responsibility it could train its own people better, and that the money saved could be used on additional work (PCR, para. 3.12). By late 1971, however, most senior RAD engineers (including the Chief Engineer) had left the Ministry and the staff had been significantly weakened. Protracted discussions took place with IDA over whether or not to hire consultants, and with the consultant over fees and conditions. At the time, the consultant hired to assist with the preinvestment studies was still on site, and the Government had IDA's agreement to have him coordinate the BMUs and supervise the execution of the works. However, contract negotiations came to a stand- still. IDA pointed out: repeatedly that the reason for using a consultant was to supplement COMWORKS capacity and avoid having to draw on its scarce trained staff. Yet, it was not until late 1973 that a contract was signed with the consultant. By then, however, recruiting difficulties prevented the consul- tant from forming a new team. Manning a full team was only possible in December of 1974 (PCR, para. 3.13). 21. In early 1972, however, shortly after credit effectiveness, the Government had, without consulting IDA, called for bids to supply the equip- ment for the BMUs. When concurrence with the contracts was sought from IDA and the documents were reviewed, they were found to be inconsistent with IDA Guidelines as well as with procedures agreed with the Government. Yet, because of the imminent, expiration of the bids IDA agreed to proceed with the award (PCR, para. 3.14). As a result, the equipment arrived long before the BMUs were formed and ready to work, and had to be held at the port of Dar Es Salaam where it remained exposed and subject to considerable pilferage. Most of it had to be rehabilitated before it could be used by the BMUs when they began working in late 1974 (PCR, para. 3.16). 22. To sum up, in late 1974, three years after credit effectiveness, during which time..a number of IDA policies, guidelines and procedures were compromised, after eight supervision missions, and despite heavy involvement by the Resident Representative in Dar Es Salaam and the Regional Mission to Eastern Africa, the project had made little progress beyond the selection of - 8 - consultants and the signing of a construction contract. The project's main component had been reduced from the original 200 km to 80 km, but IDA's participation in it had increased from 70% to 85% of costs. Another Bank Group project had been appraised, the Fourth Highway Project, and in August 1974 another credit for US$10.2 million had been approved. Project Implementation (1974-1978) 23. The construction contract for Mtwara-Masasi specified completion in 24 months at a total cost of TSh 82.4 million. This was subsequently extended to 37 months due to significant increases in work, later to 41 months because of unusually heavy rains, and eventually the contractor required 12 more months to complete the work. The road was finished in August 1978, at a total cost of TSh 136.3 million (215% above appraisal estimates and 65% above contract costs). An important reason for the increase in costs was the increase in the geometric standards of the road (PCR, para. 3.08). Because the works were contracted without final engineering, the consultant was to complete it as work progressed; instead, the consultant redesigned the road to significantly higher standards (PCR, para. 3.31 et. seq.). The width of the road was enlarged and the vertical alignment significantly improved causing the earthworks to increase much beyond what was necessary to bring the road to the standard agreed at appraisal. In turn, the contractor found himself ill-prepared to cope with the higher volumes of work and fell behind schedule. Finally, progress was delayed because the consultant's resident engineer lacked authority, closure of the Kenya-Tanzania border made procuring supplies slow, and the imposition of tight security as a result of the civil war in Mozambique made access to the area difficult. Partly for this last reason the expanded scope of work went undetected by IDA until January 1976, when most of the earthworks had been completed (PCR, para. 3.33). When IDA brought it to the attention of the Government it blamed the consultant. The Consultant's contract, in turn, expired in February 1976, and despite the Government's claim that he was obligated to continue until completion of the works, the consultant withdrew his services in April 1976, ostensibly because no agree- ment could be reached on increased fees (PCR, para. 3.10). Despite the doubts raised by the Government on the consultant's performance and despite the firm's withdrawal against the Government's desire, the Government awarded it another major consulting contract in the power sector shortly thereafter (PCR, para. 3.37). 24. The consultant's team was replaced, with IDA's agreement, by a Ministry team (PCR, para. 3.11). However-, to staff the team engineers had to be taken from the small numbers of experienced people who were working on other tasks, in particular on implementing the Fourth Highway Project. The team performed well and the construction was completed in 1978, but the maintenance works of the Fourth Project were badly affected by the severe lack of local staff. 25. It had taken from 1972 to 1974 to set up the arrangements to initi- ate the feeder roads component of the project. Implementation after 1974 also proceeded with many difficulties and delays. The schedule for completing the - 9 - 165 km of feeder roads in Mara was 24 months, and for the 310 km in Geita 30 months. However, about a year and a half after commencement of the works only 20% had been completed (PCR, para. 3.18). This slow progress was attri- buted to the consultant's management, the weak and insufficiently qualified local counterpart staff, and the team's inability to deal with the local branches of the industrial organization and the ruling national party, both of which influenced the hiring, termination and other management decisions of the BMUs. Also, the team could not cope with the powerful regional author- ities who diverted the equipment to other uses (PCR, para. 3.17). Further, the lack of spare parts and materials caused by slow procurement procedures, and the slow decision making at the Ministry of Works, together with long supply lines and lack of transport, were also partly to blame for the slow progress. The performance of the consultants was increasingly judged unsatis- factory, especially that of the Geita team, which had more difficulties with the local political and industrial organizations. In November 1975, the resident engineer of the Geita BMU was withdrawn and the engineer in charge of Mara was made responsible for both Geita and Mara. The roads in Mara were completed in March 1977, about 12 months after the expected schedule at commencement of works, and those in Geita were stopped in December 1978, with only 75% of the length completed (PCR, para. 3.19), after the inadvertent, full disbursement of the credit funds.i When no more credit funds became available the Govenment ceased to provide budgetary allocations for the works. No decisive action was taken by IDA, no commitment was obtained from the Government to complete the works, and, contrary to agreements to the effect that the equipment would be dedicated to the maintenance of the roads improved, the equipment was dispersed to other works (PCR, para. 3.20). 26. Maintenance of the roads in Mara and Geita has since reverted to the local authorities and, as was witnessed by an OED mission, are now almost impassable even in the dry season, likely no better a condition than before the project (PCR, para. 3.22). 27. The project also included feasibility and engineering studies for 250 km of high priority roads and continuation of the staffing and training program initiated in the First Highway Project. After a preliminary assess- ment of priorities, the Govenment requested to use the study funds to appraise feeder roads in the Kilombero and the Kilimanjaro areas. Both studies were 1/ Funds in a Credit are disbursed on the basis of an agreed percentage of costs, estimated at appraisal to ensure continued funding until comple- tion of the works. When works are delayed and costs increase, it is normal procedure to reduce this percentage to maintain a continued IDA participation until the end of contruction. In this case, however, IDA intended to reduce the disbursement percentage, but because of Government delays in processing the agreement, funds in the Credit were disbursed at the old rate, thus prematurely exhausting credit funds. The audit found no explanation why the Government could not adjust its financial flows to ensure completion of the works when, in fact, it had received, for all practical effects, advance reimbursement, as if works had been completed. - 10 - completed by consultants, but both consultants expressed lack of confidence in their results and recommendations because the regional authorities, who are responsible for feeder roads, were not consulted in the decision to study the roads and consequently refused to cooperate with the consultants (PCR, para. 3.24 et. seq.). The Kilombero study has since been used by the EEC and Swiss aid programs, and the Kilimanjaro study for a USAID-assisted program. Con- tinua tion of the staffing and training program was not carried out and the funds were eventually used against other project items (PCR, para. 4.01 et. seq.). 28. In January 1979, IDA approved another credit to Tanzania (876-TA, US$20.5 million) to assist implement the Fifth Highway Project. III. EVALUATION 29. The project set out to (a) pave a main road (Mtwara-Masasi); this was accomplished albeit with significant cost overruns; (b) recuperate feeder roads in two agricultural areas (Geita and Mara); short-lived improvements were obtained; (c) carry out preinvestment studies, these were completed and later used in bilateral aid programs (EEC and Swiss Government and USAID); (d) improve RAD's manpower situation; this was not implemented; and (e) train Tanzanian engineers; on this, very limited success can be claimed as few people were trained while the BMUs were active. The project's objective of stimulating agricultural development by paving the Mtwara-Masasi road and improving the feeder roads (Appraisal Report, para. 5.01), was clearly not achieved: the production of cashewnuts grown in the area of influence of the Mtwara-Masasi road has decreased (despite this project and two parallel Bank Group agricultural projects designed to directly assist production: Loan-1014, 1974 and Credit-801, 1978); the production of cotton in the Geita peninsula has decreased (despite this project and another IDA agricultural project designed to help increase it: Credit-454, 1974); and although there is only fragmentary information on the milk and dairy production in the Mara region (which was assisted by a Danish Government project), indications are that increased production also failed to materialize. 30. The background and history of this project show that many factors interacted to produce this result. The following are among the most evident, purposely listed in.no hierarchical order. Firstly, the project was miscon- ceived in the early identification and preparation stages. The fact that none of the identification and preparation missions could agree on the content of the project, either with the Government or among themselves, and that missions to find economically justified works had to be organized, is indicative of an absence of clear purpose. The most important component of the project, the paving of the Mtwara-Masasi road, was a last minute compromise reached despite the fact that it was clear that the road did not need paving either from a vehicle operating cost savings or an engineering point of view. Similarly, the lack of integration between the feeder road components and agricultural development shows little concern for the objectives of the project. The lack - 11 - of clearly identified objectives and targets is further evidenced in the purported relationship between transport improvements and community and social development used at appraisal to justify the inclusion of the Nanganga-Masasi section (para. 12) and cited during implementation to justify certain actions such as awarding a construction contract for clearly uneconomic works (para. 18). 31. Secondly, the absence of an agreed implementation plan must, in retrospect, be viewed as a serious shortcoming. For the main road, final engineering was not ready at appraisal nor for a long time afterwards, and without it, it is impossible to plan, much less execute the work. For the feeder road components, there was no plan on how to organize the BMUs, on their composition and operation, and there was no clear agreement on their hierarchical or structural relationship, particularly vis-a-vis the local authorities. Feeder and minor roads were the jurisdiction of local author- ities but they were not consulted nor did they participate in the structuring of the program. For these reasons, organizing the BMUs was slow, even after the consultants were mobilized; local authorities interfered with the manning and operating of the units; and the permanency of the units was questioned and they were finally disbanded, thereby precluding their continued use for future maintenance purposes. In these circumstances neither the use of foreign consultants and contractors on the Mtwara-Masasi road, nor the use of consul- tants to coordinate and manage the feeder roads program, nor the concentration of IDA staff on the project could have been effective in carrying out the project expeditiously. 32. Thirdly, there is little doubt that the staffing weaknesses in COMWORKS, and more specifically the RAD, were an important contributory factor to the poor results. This weakness was well known to IDA at the time the credit was made since the Government's poor performance under the staffing and training program in the First Project was well documented in various reports. Yet, manpower problems were not adeptly addressed during project preparation. IDA apparently assumed that the project could be executed satisfactorily inspite of this weakness, an overly optimistic assessment. 33. Further, it does not appear that IDA contributed much to strengthen- ing RAD during project implementation. The only active training element in the project was that associated with the BMUs. However, an effective training through the BMUs did not occur because of (a) the unreceptive attitude, lack of interest and low.caliber of the counterparts, and (b) the conflict between the need to devote time to keep the work moving and the time needed to train the local staff (PCR, para. 3.40). Ten years later, the same weakness per- sists and problems of understaffing and inadequate staffing are cited as common to all Ministry of Works Divisions (Appraisal Report, Fifth Highway Project, December 1978). 34. Fourthly, the non-observance of many IDA guidelines, principles, policies and procedures was deleterious to the efficient design and execution of the project and interfered with the achievement of any implied or otherwise - 12 - perceived objectives (PCR, section VI). Normally observed procedures and policies which were not applied in preparing and implementing the project include: IDA's agreement to appraise a project, the principal component of which was not ready for appraisal; IDA's decision to follow through anyway after having appraised the project and determined that the main component was obviously just marginally justified; IDA's agreement to reinclude in the project a road section which had previously been deleted because it was uneconomical on the basis of a one percentage point higher rate of return and some tenuously defined unquantified benefits; IDA's agreement to tender the civil works on the Mtwara-Masasi road without final engineering and therefore without designs for contractors to make bids on, or adequate cost information on which to evaluate the bids; IDA's agreement to award the contract for the Mtwara-Masasi road and proceed with the project (using unspent balances from another loan to help meet part of the increased costs) instead of revising it and drastically reducing construction standards after bids for the execution of the works showed that costs were more than double the appraisal estimate, and that the already marginal rate of return of the project had dropped by almost 50%, (including one road section with just a 2% rate of return). Not only did these actions violate principles and guidelines but they signaled the Government the wrong message, encouraging it to proceed with a project known to be uneconomical. 35. Similarly, IDA's agreement to award the contract for the supply of the equipment for the BMUs knowing that procurement procedures had not been followed and that tender documents violated IDA guidelines, and evidently out of phase with the establishment of the BMUs, set in motion prematurely the execution of this component and allowed delivery of the equipment too early, its consequent damage and the need to rehabilitate it before it could be used. The inadvertent over-disbursement of credit funds in the later stages of the project also had serious consequences in shutting down the works before they were finished. 36. The question is posed how it came about that so many IDA policies and procedures were disregarded. The defects in the appraisal process were perhaps related to a decision to complete the credit quickly. The rather relaxed standards applied in the administration of this project appear to reflect a consistent pattern from identification onward which created the perception of great leniency on the part of IDA in its dealings with the Tanzanian Government (PCR, para. 6.03). As a result, controls were relaxed and standards of analysis were not kept, as exemplified by the major mistakes in the 1974 reappraisal which led to wrong recommendations and actions. Similarly, the standards of supervision were insufficient. This resulted in IDA's failure to become aware, until after most earthworks on the Mtwara- Masasi road were completed, that they had been done to considerably higher standards than those agreed. IDA's continued assistance for highways in Tanzania despite the serious difficulties experienced by this and other on-going projects, did not help to dispel this perception of leniency. - 13 - 37. The audit concurs with the economic evaluation of the project presented in the PCR (section V), which concludes that the Mtwara-Masasi road as a whole, as well as the Mtwara-Mingoyo section specifically (the only section IDA ultimately assisted in financing), were indeed poor investments with just a 4% rate of return as opposed to 14% and 16%, respectively, at appraisal. The two feeder road components had such short-lived impacts before the roads reverted to an impassable condition, that the rate of return on these investments is judged to be negative. The overall rate of return of the project is therefore less than 4%. 38. The audit concludes that this project's experience highlights the difficulties that can be created by pressures arising from forcefully pressed lending targets and undue laxity in the application of normal requirements.i/ These difficulties led in this case to an ineffectual use of scarce resources and to a project with a very low, possibly negative, rate of return. 1/ The Eastern Africa Regional Office is of the opinion that when it comes to assessing facts, pressure should not be a factor.limiting the quality and depth of the analysis, and that staff should be able to isolate the technical process from the pressures. The audit fully shares this position. - 14 - WORLDBNK641450 Annex 3312 AFDEV CI WORLD BANK WASHINGTON DC EE8761 - FOR : MR. S.S. KAPUR - DIRECTOR, OPERATIONS EVALUATION DEPARTMENT - SUBJECT : TANZANIA : PERFORMANCE AUDIT REPORT O' THIRD HIGHWAY PROJECT - (A) THANK YOU FOR THE DRAFT COPY OF THE ABOVE STOP U!LFDRTUNATELY THERE HAS BEEN DELAY IN RECEIPT OF THE REPORT STOP (B) WE HAVE READ THE REPORT UITH INTEREST AND NOTED THE DETA,ILS OF THE DIFFICULTIES ENCOUNTERED ON THE PROJECT STOP WE HAVE HO COi1ENTS TO ADD TO THE REPORT STOP REGARDS G. DOSSOU A F D EV COL. EE8761 - 15/6/82 AT 14.06 3312 AFDEV CIl WORLDBNKo4145r- 3612 AFDEV CI - 15 - TANZANIA PROJECT COMPLETION REPORT CREDIT 265-TA/LOAN 586-TA - THIRD HIGHWAY PROJECT June 30, 1980 Highways Projects Division Eastern Africa Regional Office  - 16 - TANZANIA THIRD HIGHWAY PROJECT - CREDIT 265-TA/LOAN 586-TA PROJECT COMPLETION REPORT SUMMARY AND CONCLUSIONS Background 1. On July 27, 1971, the Board approved a Third Highway Project for Tanzania, authorizing IDA Credit 265-TA for US$6.5 million, which was later supplemented by a transfer of a US$1.9 million surplus from Loan 586-TA originally approved for the Second Highway Project. These funds were to help finance: (a) the construction of 200 km of the Mtwara-Mingoyo-Masasi road in southern TanzaniLa; (b) the improvement and maintenance of 475 km of feeder roads in Geita district (310 km) and Mara region (165 km); and (c) preinvestment studies, consultant costs for supervision, technical assistance and training. 2. Total project cost was estimated at appraisal at US$9.5 million of which IDA was to finance the foreign exchange cost of US$6.5 million and the Government would finance the US$3.0 million equivalent. 3. The original Credit Agreement was signed on August 6, 1971 with a Closing Date of December 31, 1974. The Credit became effective on October 12, 1971. Mtwara-Mingoyo-Masasi Road 4. During appraisal, the lowest trafficked 58 km section of this road was found to be uneconomic and was deleted from the project before presentation to the Loan Committee. However, during Credit negotiations, the Government's negotiating team pressed strongly for its re-inclusion. The Bank Group, after reviewing the economics of this section and as a result of adjusting the economic rate of return upwards by 1% and also referring to unquantifiable additional benefits, agreed to its re-inclusion. When bids for this component came in at more than twice the appraisal estimate, the Association reappraised the project and concluded that the standards on this same 58 km section should be reduced and that the disbursement percentages in this project element should also be reduced. Once again, the Government pressed strongly for the retention of the original design standards and again the Association agreed, but to reflect the increased construction cost, reduced disbursement percen- tages from 70% to 32%. Subsequently, the African Development Bank (AfDB) allocated the 'equivalent: of US$4.8 million to help finance the Mingoyo-Masasi (118 km) section. The Bank Group amended its project to include a reduced civil works component of only 82 km between Mtwara and Mingoyo for which - 17 - the estimated total cost at the time was US$6.5 million equivalent based on the lowest bid. Moreover, in January 1975, the Bank agreed to transfer a US$1.9 million undisbursed balance from the Second Highway Project to the Third Highway Project to help meet the cost overrun, and the disbursement percentages for this project component was increased from 32% to 85%. 5. Prequalification and international competitive bidding (ICB) procedures for this road were in accordance with Bank Group Guidelines. Bids for the construction contract were called, and the contract was awarded as a single contract in March 1974. Construction supervision was carried out by the consultants who had previously carried out detailed engineering. The supervision contract was not explicit on provisions and conditions for extending the contract time in case of time overruns for the construction contract. At the expiration of the original supervision contract, the consultants withdrew in April 1976, following a dispute with the Government over a pay rate increase. The Government then deployed some of its own scarce experienced staff to supervise the completion of the contract. 6. Work quality for the completed road was good. However, the construction work which, at the time of appraisal was expected to take 21 months and be complete by December 1973, took 53 months and was not completed until August 1978, four and a half years behind appraisal schedule. This was also 29 months longer than the original contract completion time of 24 months. The delays in substantial completion were mainly due to a start-up period nearly four times longer and a construction period more than two and a half times longer than expected. The longer start-up was caused by Government's slowness in negotiating consultants' fees and conditions which delayed finalization of the consultant contract. The longer-than-expected construction period resulted from: (a) additional works required to correct significant road deterioration during the long start-up period; (b) higher road design standards incorporated in engineering carried out during construction; (c) consultants' slowness in reaching technical decisions; (d) contractor's lack of preparedness and financial difficulties; (e) unusually heavy rains during construction; and (f) the hindrance caused to material supplies by the closure of the Kenyan-Tanzanian border In addition, some of the delay was caused by the fact that engineering designs and surveys were not fully detailed and had to be completed by the consultants supervising the contract. Geita and Mara Feeder Roads 7. The Association's First Highway Project included feasibility and detailed engineering studies for the feeder roads in the Geita district and a separate USAID project financed a study of feeder roads in six rural areas, including the Mara region. During project preparation for the Third Highway Project, Bank Group staff rightly concluded that the design standards were inappropriately high in view of the projected traffic. After reviewing the designs, agreement was reached with Goverment that unengineered gravel standard roads would be adequate and that construction would be by force account. - 18 - 8. Another issue during negotiations and start-up concerned the management and supervisLon of the units that would be set up to carry out the program. The Government wanted to use its own staff whereas Bank Group staff favored the use of consultants to supervise the units and train local staff. The Bank Group's views prevailed and were agreed to at negotiations. In view of the Governmenl'- reluctance to use r.onsultants in earlier projects and its apparent lack of cortmitment to the Bank Group's position in this project, it should not have been surprising that despite its agreement to use consultants, the Government was slow to finalize consultant contracts and reverted to its earlier position, that it undertake the work without consultants. After protracted discussions with both the Bank Group and the selected consultants, the contract was signed in late 1973 for a project start by mid-1974. However, because of recruiting difficulties, the consultants' team was not assembled until December 1974. 9. Despite the fact that equipment for the Betterment and Maintenance Units (BMUs) carrying out this work was to be procured under ICB procedures with which Government had had previous experience under earlier projects, the Government prepared bidding documents and called international bids without reference to the Bank Group. The resultant bid documents and tenders received, did not comply completely with Bank Group procurement Guidelines, but after some deliberation, the Bank Group accepted the bidding process as satisfying the spirit if not the letter of the Guidelines and permitted equipment pro- curement to proceed. After the equipment arrived it was held at the port of Dar es Salaam until the BMUs were ready to start work. As mentioned below, there were delays in mobilizing the units, which caused some deterioration in the idle, weather-exposed equipment and this, together with pilferage of parts and tyres which occurred, meant that some of the equipment had to be rehabilitated before it could start work. 10. The BMU work started late and progressed slowly because of: (a) delay in employing consultants; (b) delay in posting local staff counterparts; (c) lack of and slow supply of spare parts and materials; (d) interference by, and lack of cooperation from, the politically influential regional authorities in operational as will as disciplinary matters; (e) lack of local funds towards the end of the project; and (f) diversion of men and equipment away from the project roads. In addition, a significant factor in the slow progress of the work at Geita was the unsatisfactory performance of the consultant's resident engineer. This arose despite the consulting firm's good reputation and satisfactory past performance and also in spite of the fact that the resident engineer had appropriate professional qualifications and experience. The problem was caused by the engineer's uncooperative attitude, lack of motivation and incompatibility with local people. 11. At date of writing about 70 km of the Geita roads remain to be completed and work has been at a standstill since November 1978. The com- pleted roads have been handed back for maintenance to the regions which lack financial, equipment and manpower resources, and accordingly the roads have deteriorated rapidly, in some cases to the poor condition they were in prior to improvement. - 19 - Preinvestment Studies 12. The feasibility studies of the 250 km of feeder roads were satisfactorily carried out by consultants selected in accordance with Credit requirements. Although the draft final reports were received on time, due to delays in receiving Government comments, the final reports were only completed in December 1976 for Kilimanjaro and in September 1977 for Kilombero. The Kilombero study was taken into account by the EEC and the Swiss Government aid agency in preparing development aid projects for the Morogoro Region. USAID has used the Kilimanjaro report as an input for preparing an aid project for that region. Technical Assistance and Training 13. Technical experts to assist the executing Ministry had been recruited and financed under the First Highway Project and were to be conti- nued for a further two years under the Third Highway Project. However, the two experts concerned left the Ministry before the project started. The Ministry did not replace them nor did it use the funds for alternative technical assistance and training purposes despite the Bank's urging. The Credit funds for this component were eventually disbursed against expenditures for other project elements when the Credit was inadvertently fully disbursed (see para. 14). Project Costs and Disbursements 14. The total project cost was US$25.2 million compared to the appraisal estimate of US$9.5 million. The rate of disbursementwas slow in line with the slow rate of progress of the project. Only US$1.7 million or 26% of the Credit had been disbursed by the original Closing Date. Disbursements from the Credit for the civil works component were limited to US$3.6 million which was fully disbursed by December 1976; at that time disbursements continued for this component from the US$1.9 million allocated from Loan 586-TA and this was fully disbursed in September 1979. Disbursements for all other elements of the project were made from the Credit and these continued until February 1978 when the Credit was inadvertently fully disbursed before the feeder roads elements were completed, following a Bank Group mistake of overlooking a December 1977 decision to reduce the disbursement rate of this component. The Closing Dates of both the Credit and the Loan were extended several times and finally occurred on December 31, 1978. Economic Re-evaluation 15. The economic re-evaluation of the project considers the Mtwara- Mingoyo-Masasi road and the feeder roads components in Geita and Mara. 16. Mtwara-Mingoyo-Masasi Road: The following table summarizes the internal rates of return (IRR) of the four sections making up the Mtwara- Mingoyo-Masasi road and reflects the appraisal estimates, the reappraisal estimates, the "corrected" reappraisal estimates, i.e., the reappraisal estimates corrected for a total project cost error of 40% made during the reappraisal, and finally the re-evaluated "actual" estimates. The IRRs on the cost side take into account construction and supervision costs and on the benefit side total vehicle operating cost (VOC) savings resulting from the road improvement. Total VOC savings in turn are based on unit VOC savings, i.e., savings per vehicle, and the volume of traffic using the road. - 20 - Estimated Internal Rates of Return (%) 1.971 1974 "Corrected" 1974 1980 Re-evaluation Road Section Appr. Reappr. Reappraisal "Actual" I Mtwara-Mingoyo 16 10 6 4 II Mingoyo-Mtama 15 10 7 7 III Mtama-Nanganga 13 8 4 4 IV Nanganga-Masasi 9 2 -2 1 All sections 14 8 4 4 17. The economics of the project road are poor, with an overall IRR of 4% and IRRs for individual sections ranging between 7% and 1%. These "actual" IRRs are about the same as those of the "corrected" 1974 reappraisal which reflect a more accurate and comprehensive estimate of the ex-ante economic cost (construction and supervision) of the project road. Though the IRRs are about the same in the two calculations, the underlying data are somewhat different. In the re-evaluation, the impact of higher construction and supervision costs is offset by higher unit VOC savings while the traffic volumes in both calculations are about the same. In hindsight it would seem that if the Bank in 1974 had undertaken a "proper" reappraisal it would probably have concluded (a) not to finance, at that time, construction of the road at the specified design standards, and (b) to look into the economics of the road based on lower design standards. 18. Geita and Mara Feeder Roads: The feeder roads improved under the project in some cases have already deteriorated and in others are dete- riorating rapidly to the poor condition they were in before (see para. 11). In economic terms this means that the investment made in this improvement is largely wasted and produces a negative rate of return. Project Objectives 19. The project's objectives were: (a) to construct an economically viable section of trunk road to serve the southern regions of the country; (b) to improve and maintain feeder roads in two economically important areas in the north; (c) to develop institutional strength and capacity for future betterment and maintenance; and (d) to provide a basis for future road investments by carrying out feasibility and engineering studies. 20. The construction of the trunk road in the south constitutes a technical and physical tchievement of the first objective. However, as described in para. 17 above, the economic objective was not achieved. Regarding the objective for feeder roads, although most of the roads were improved, due to the lack of subsequent maintenance, much of the value of the work has been lost and hence the economic objectives of this component were not achieved either. Further, the institutional - 21 - development objective of the feeder roads component was not achieved because the training was unsuccessful, and the betterment and maintenance units were discontinued. Feasibility studies were carried out successfully and the resulting reports have since been used in making investment decisions by other aid agencies. Lessons 21. Lessons arising from the experiences of this project reconfirm lessons already learned from other Bank Group projects in a number of countries. They are: (a) to ensure accurate specification of the amount and type of work to be done and reliable cost estimates, detailed engineering of civil works should be complete before appraisal; this has been standard practice for Bank Group projects, and this project has reconfirmed the soundness of this practice, especially in countries like Tanzania with weak institutions and histories of difficult working conditions for consultants and contractors; and (b) irrespective of a consultant firm's good reputation and past performance, the crucial criterion for its future performance in a new project is the experience, ability and calibre of the specific staff assigned to that project. 22. In addition, a number of other lessons can be learned from this project: (a) occasionally the Bank Group will be subjected to pressures from Borrowers to retreat from its usual rigorous policies, guidelines and procedures and accordingly there is a need for an explicit Bank Group strategy to deal with those situations. From this project, two options for such a strategy are suggested: (i) the Bank Group should stand firm and hot give in to the pressures to reduce its standards; or (ii) where it yields to pressures, the Bank Group should expect increased risks and reduced objectives and be quite explicit in its analysis and presentation of the new results expected; (b) the reappraisal of a project, undertaken because of changed circumstances since the time of appraisal, deserves as much attention and scrutiny as the original appraisal; - 22 - (c) the recommendai:ions and conclusions of such a reappraisal deserve the same respect as those of the original appraisal even if they imply a drastic revamping or abandoning of the original project concept; (d) consultant contracts for supervision of civil works contracts should be specific about obligations and conditions for continuing beyond the original contract termination date if the civil works contract is not complete by that time; (e) in Tanzania, start-up periods have been long and continue to be protracted for more recent projects; due allowance should be made in the design, analysis and scheduling of future projects in that country; (f) technical assistance in executive line positions can be extremely effective for carrying out a given work program provided the individual expert is experienced, competent, motivated and compatible with the local people. Irrespective of experience and competence, without motivation and compatibility an expatriate expert in an executive role may be less effective than a relatively inexperienced but motivated local; (g) where projects include the dual objectives of carrying out a program of work as well as institution building and training, separate technical assistance is needed for helping with each objective. This project has shown that for the carrying out of a work program, the direct executive role is effective. However, for the institution building and training objective, the only conclusive lesson arising from this project is that the people who are responsible for carrying out a work program cannot effectively carry out a training and institution building function as well; (h) where institution building is an objective of a project, it is crucial that the institution to be developed.is one which will retain continuing responsibility and involvement in the functions for which it is trained; that is, if a feeder road BMU is formed, it should be part of an organization with continuing responsibilities and programs in feeder roads; (i) Government's institutional weaknesses and deficiencies contribute to ineffectiveness of expatriate experts. Provision of expert technical assistance at project level does not guarantee effective and efficient - 23 - implementation if institutional weaknesses mean that necessary back-up and support services away from the project site are poor or do not exist. Institutional development is required throughout the entire organization not just at project level. The fact that this project suffered because of readily identifiable weaknesses away from the project site indicates that inadequate attention was given to this issue in the design of the project. The design of future projects should, therefore, include some feature such as a project implementation unit or project expeditor to identify, follow up and resolve bureaucratic and operational bottlenecks; and (j) upgrading of feeder roads can be undertaken by BMUs but unless regional authorities' road maintenance ability and capacity are improved, the upgraded roads will deteriorate rapidly; the Fifth Highway Project has a pilot program for improving regional road maintenance and several bilateral aid agencies also have projects to assist regional institution building in this area. - 24 - I. INTRODUCTION First and Second Highway Projects 1.01 For many yeara, Tanzania's transport policy was directed almost exclusively towards the requirements of external trade, and emphasis was therefore on railway development with roads seen primarily as a complement and feeder to the rail system. While traffic associated with imports and exports continues to feature high in Tanzania's investment priorities, transport policy since the late 1960s has included increasing emphasis on regional development. This led to an ambitious program to improve primary roads and more recently to improve roads serving rural areas. In addition,Tanzania has assumed a major role in helping its land-locked neighbors to improve their access to the sea. 1.02 The Bank'Group's First and Second Highway Projects helped with this new phase of road development. Primary roads, both national and inter- national,were progressively upgraded and the problems and needs of rural roads were identified and studied. Under Bank influence, the Government began introducing sound engineering data and economic criteria in preparing its development plans; much needed attention was given to immediate and long-term staffing requirements; and emphasis was directed to previously neglected maintenance needs. 1.03 Under the First Highway Project in 1964, (Credit 48-TA for US$16.9 million and Supplementary Credit 115-TA for US$3 million), the detailed engineering and construction of eight main road sections were undertaken, as well.as detailed engineering of feeder roads on Geita Peninsula and a five-year program of staffing and training for the Roads and Aerodromes Division (RAD) of the then Ministry of Communications, Transport and Labor.l/ 1.04 The Second Highway Project in 1969 (Credit 142-TA of US$8 million and Loan 586-TA of US$7 million) continued the program of constructing trunk roads. Under that project, Tanzania constructed to engineered standards about 500 km of the Tan-Zam Highway, the main road connection between Zambia I/ In 1975 the Ministry of Communications, Transport and Labor was divided into two separate Ministries, the Ministry of Communications and Transport and the Ministry of Works (MOW). This project was assigned to MOW for execution and this report will refer to that Ministry rather than RAD. - 25 - and the Tanzanian port of Dar es Salaam. When this work was completed, the Bank Group agreed to use a remaining US$1.9 million Loan balance to supplement the Third Highway Project Credit to help meet cost overruns for the bitumi- nizing of a main road in the southern part of the country. Third Highway Project 1.05 The Third Highway Project (Credit 265-TA for US$6.5 million and carry over of Loan 586-TA of US$1.9 million), the subject of this Project Completion Report, originally included 200 km of main road in the southern part of the country, but this was reduced to 82 km during project implemen- tation. The project also included improvement of about 475 km of feeder roads in the Geita district (310 km) and Mara region (165 km) as well as prein- vestment studies, consultant costs for supervising road construction and improvement and technical assistance for a staffing and training program. 1.06 The works included in the project, though varying considerably in their character, were all geared to the general objective of stimulating agricultural development. The road to be constructed in the south forms part of the regional highway linking the port of Mtwara and the entire southern fringe of the country to the Tan-Zam Highway. However, most of its traffic is local and concerned with the movement of the cashew nut crop grown in the area assisted under a Bank Group project. The feeder roads in the Geita district provide an outlet for the important cotton crop grown in that area, while the Mara region roads are essential to the general agricultural section as well as an important milk-producing scheme established with Danish aid. 1.07 The Government made a formal request for a Third Highway Project on January 3, 1970. The project was appraised in November 1970, approved by the Board on July 27, 1971 and signed on August 6, 1971. The project became effective on October 12, 1971. Loan 586-TA (Second Highway Project) was-formally amended on January 22, 1975 to reallocate the US$1.9 million surplus to the main road construction component. 1.08 The sources of information on which this PCR is based are listed in Annex A. - 26 - II. PROJECT PREPARATION AND APPRAISAL Mtwara-Masasi Road 2.01 The Third Highway Project evolved from a number of pre-investment studies carried out in ranzania between 1968 and 1970. In 1968 USAID-financed consultants carried out a feasibility study and recommended paving of the Mtwara-Masasi road (200 km) in the southern regions of the country. The first 82 km of this road between Mtwara and Mingoyo had been constructed to engi- neered gravel standard between 1959 and 1962 and the first 16 km from Mtwara to Mikindani were given a bituminous surface dressing. The balance of the road from Mingoyo to Masasi (118 km) had been constructed to engineered gravel standard between 1964 and 1968 under the Bank Group's First Highway Project (Credits 42-TA and 115-TA). 2.02 The original concept of the Third Highway Project had included a component to improve to all weather standard the 451.km section of road between Masasi and Songea if found justified by a consultant study which was retroactively financed under the project. By August 1970 agreement had been reached between the Government and IDA on the study's terms of reference and the consultants to be used. However, in August 1970 the Government and IDA agreed that, as the appraisal was scheduled for November 1970, the study would not be sufficiently advanced for the work to be appraised. The Government then extended the consultant contract to include and give first priority to final engineering for the Mtwara-Mingoyo-Masasi road with the objective of having enough data and information available by November to enable the appraisal mission to review the cost and justification for paving this road. 2.03 IDA staff had misgivings and felt that there was insufficient time to update economic data and to substantially complete detailed engi- neering, pavement design, a detailed soils and materials survey and cost estimates based on a bill of quantities. IDA agreed to include paving of the Mtwara-Masasi road in the appraisal, but the decision on the adequacy of the data available was left to the appraisal team. 2.04 The appraisal mission concluded that the study was sufficiently advanced to carry out an appraisal and to include this road in the project. Some necessary data such as updated traffic counts, forecast traffic growth rates and laboratory test results of soils and materials were not available during the appraisal mission,but these were completed and sent to Washington for use in preparing detailed project costs and the economic evaluation. The appraisal mission concluded that paving of the road between Nanganga and Masasi was not justified since the internal rate of return was only 8%, and this section of road was deleted from the project as appraised. 2.05 The project as presented to the Loan Committee excluded this road, but during negotiations the Tanzanian delegation pressed strongly for its re-inclusion and caused IDA to reconsider the road's economic merits. The road was finally re-included in the project on the grounds that - 27 - (a) the rate of return increased from 8% to 9% if its costs were viewed as marginal J/ to the rest of the project; (b) although unquantified, the project would generate benefits other than the vehicle operating cost savings included in the evaluation; (c) traffic growth rates would probably exceed the 5% assumed; (d) the road would speed up the economic and social development of the relatively undeveloped southern areas; and (e) Masasi, a market and administrative centerwas a logical terminal point. Geita and Mara Feeder Roads 2.06 The Bank's First Highway Project included detailed engineering studies for the construction of secondary roads in the cotton-growing areas of the Geita district in the Mwanza region. Estimated costs of this work exceeded the estimates from earlier feasibility studies which had shown construction to be justified. In 1970 the responsible Government ministry together with a research unit from the University of Dar es Salaam reviewed the economic justification of these roads and found that construction to high engineering standards was not justified. However, the review established that for the low traffic volumes involved, the roads could be upgraded to acceptable levels of service through minor construction and betterment works and improved maintenance, and this work was included for appraisal under the project. 2.07 In 1969, USAID financed a consultant study of the feasibility for the betterment of feeder roads in six agricultural areas. In 1970 the Government,in consultation with the Association, arranged for a con- sultant to carry out preliminary engineering for the betterment of about 300 km of feeder roads in two of these regions, the Mara and Mbulu regions. The Mara roads were found to be suitable for a pilot betterment program which was included for appraisal under this project. Issue of Implementation 2.08 The appraisal team found that in the Ministry of Communications, Transport and Labor, which was to be the executing agency, there was a lack of trained, experienced staff who could carry out the supervision and training contemplated. This was a continuing problem from the time of the First Highway Project when provision was made for the recruitment of experts to strengthen the Ministry. However, the recruitment program under that project was considerably behind schedule and, at the time of appraisal of the Third Highway Project, had not achieved its objectives due to the reluctance of the Government to recruit expatriate experts (see Project Performance Audit Report No. 791 for the First Highway Project). In view of the lack of staff, 1/ Disregarding all preliminary, establishment, overhead and withdrawal costs of the contract which would be borne by the other road sections. - 28 - this project was appraised on the basis that consultants would be used for the supervision of the construction of the Mtwara-Mingoyo-Masasi road. 2.09 During the appraisal, Ministry staff diifered with Bank Group staff on the operation of the feeder road betterment and maintenance units. The Ministry wanted to use its own staff for implementation with consultants only providing the preinvestment studies and designs. In view of the Ministry's relatively weak staff, the appraisal team insisted on using consultants for supervision and training of local staff. Accordinglyprovision was made in the project for consultants to help plan operations,lay out the works, supervise execution of construction and monitor and analyze the results of experience gained under the project. The consultant teams were to train Tanzanians in the various operational and management functions involved so that they could continue these functions independently after the consultants left. Project Description 2.10 The project consisted of: (a) Road Construction (i) the paving of 200 km of the Southern trunk road between Mtwara, the port serving the southern regions, and Masasi, an inland agricultural and market center. (During implementation this was reduced to 82 km from Mtwara to Mingoyo); (ii) preparation of final design and bidding documents and construc- tion supervision for the Mtwara-Masasi road; (b) Betterment and Maintenance of Feeder Roads (i) the betterment and maintenance of about 475 km of agricultural feeder roads and maintenance in the Geita district of the Mwanza region and in the Mara region in northern Tanzania; and (ii) supervision and management of the betterment and maintenance units and training of local staff. (c) Technical Assistance - continuation of the staffing and training program initiated under the First Highway Project; and (d) Preinvestment Studies - engineering and other studies for the preparation of this project and a further 250 km of roads which could be included in future highway projects. (i) Paving of the Mtwara-Mingoyo-Masasi Road 2.11 The project road forms part of the regional highway linking the port of Mtwara and the entire southern fringe of the country inland to the Tan-Zam Highway near Mbeya. Most of the traffic is agricultural - 29 - and is principally concerned with the substantial cashew nut crop grown in the area. The first 16 km to Mikindani had a poorly maintained bitumen pavement and narrow shoulders. The project provided funds for the repair of this road, reconstruction and realignment of a short section of bridge approach road across a swampy section of terrain, widening and strengthening of shoulders and improvements to the road's drainage. The balance of the road was to be rehabilitated to its previously constructed engineered gravel standard, the sub-base strengthened where necessary and a new cement stabi- lized base course and double bituminous surface treatment added. The design standards of the road are shown in Table 2.01. (ii) Detailed Engineering and Construction Supervision 2.12 Funds were included to meet the costs of survey,design and bidding documents financed retroactively and for supervision of road construction by consulting engineers. (b) Feeder Roads Work (i) Betterment and Maintenance 2.13 Funds were provided to improve and maintain 310 km of feeder roads in the Geita district, a cotton growing area of the Mwanza region and a further 165 km in the Mara region to help the development of the dairy industry in that area. These roads were to be improved to a 14 foot width, all-weather gravel standard which, during implementation, was changed to an 18 foot width. (ii) Supervision, Management and Training 2.14 Funds were provided for consultants to provide experts to plan, lay out, supervise and monitor the betterment works and to train local counter- parts, and two field teams to help manage and organize the operations of the betterment and maintenance units. (c) Technical Assistance 2.15 Funds were provided to top-up the salaries of two experts who were helping the Ministry and who had been recruited and financed under the First Highway Project and whose services were about to terminate; these two experts were to be financed for a further two years. (d) Preinvestment Studies 2.16 Funds were provided for preinvestment studies of about 250 km of primary and secondary roads to prepare for future lending operations support by the Bank Group. The roads were to be selected on the basis of pre-feasibility studies being carried out at the time by the Ministry's Planning Division. - 30 - III. PROJECT IMPLEMENTATION AND COST A. Construction of the Mtwara-Mingoyo-Masasi Road 3.01 The start-up phase of the Mtwara-Mingoyo-Masasi road component took 31 months from signing of the Development Credit Agreement, whereas the original implementation schedule allowed 8 months on the basis that detailed engineering and tender documents were well in hand by the consultants and that the same consultants would be hired to supervise constructior. Contract Awards 3.02 In late 1972 Government approved five of the six firms who had applied for prequalification for bidding for the construction contract, but only four were accepted by IDA as capable of executing the work satis- factorily. Tender documents were completed and approved by IDA in January 1973, following some delays in reaching agreement between Government and IDA on the tender documents. However, tender documents were not released until August 11, 1973 due to delays in complying with the Development Credit Agreement's conditions of appointing consultants for the supervision of this contract, including the bidding process. The delays in appointment of consultants were caused by a breakdown of negotiations between the Government and the consultants, mainly over fees and the question of the consultant staff's liability to pay taxes on earnings under the contract. 3.03 In view of the delay in calling for bids, the Government with IDA agreement, added to the list of prequalified contractors three of four firms who had applied for prequalification after the original closing date for applications. Of the seven firms invited to bid, four made site visits and only two submitted tenders on November 20, 1973. Bids submitted were about twice the appraisal estimate; at about the same time the supervising con- sultants increased their construction cost estimates to about 71% over the appraisal estimate. The increase in bid prices and consultant's estimate over the appraisal estimate were due to the time lapse between the originally estimated bidding time of early 1972 and the actual bid date of late 1973, during which time prices had increased significantly due to oil price increases specifically, and general cost escalation factors. In addition, there was a general lack of interest in bidding, particularly for work in southern Tanzania which was close to, and affected by, the ongoing civil war in Mozambique. One other factor which may have influenced bid prices was the incomplete detailed engineering which could have caused bidders to increase prices to cover themselves against unknown engineering factors in the project. 3.04 Because of the sharp increase in construction cost, the Association reappraised the project road in January 1974 taking into account the lowest bid. This reappraisal showed that the Nanganga-Masasi section had a rate of return of only 2%. The Association then suggested to the Government that the design standard for this section be lowered. However, the Government insisted on retaining the original standards (see Table 2.01), and the Association finally agreed but reduced disbursement percentages from 70% to 32% and on condition that the Government would finance the increased costs. The Govern- ment accepted these conditions and in March 1974 awarded the contract to the lowest bidder for an amount of TSh 82.4 million (US$11.31 million equivalent). - 31 - 3.05 The supervision contract was only signed in November 1973, 27 months after Credit signing. Further delays occurred as a result of the high bid prices received; delays to resolve the issues which arose as a result of the re-evaluation and subsequent dialogue between Government and IDA on the design standards of the Nanganga-Masasi section. The principal delay was caused by the protracted period required to finalize the contract with the consultants for supervision of the construction contract. Although the delays appeared to be due to extended arguments over levels of fees and other terms and conditions the main factor was the slowness of the Ministry in making decisions on financial matters. This resulted from the extra caution exercised by Government engineers following charges, later proven to be incorrect, against a former Director and Senior Engineer of alleged overpayment to contractors. Amendment of Project 3.06 Later in 1974 AfDB agreed to finance the Mingoyo-Masasi section of road and in early 1975 the Bank Group agreed to two changes in its involve- ment with the project. First, the project description was changed to cover only the Mtwara-Mingoyo section of road 1/. Second, the Bank agreed to trans- fer a surplus of US$1.9 million from the Second Highway Project's Loan 586-TA, to help meet the increased foreign exchange cost of the project. At that time, the civil works component consisted of only 82 km of road at an esti- mated total cost of US$6.5 million equivalent, based on the lowest bid. Construction Cost 3.07 The original contract period was 24 months, later extended to 37 months due to the increased scope of work and subsequently increased to 41 months because of an unusually bad rainy season in 1976/77 and a cholera outbreak during the construction period. The contractor eventually took a further 12 months during which time he paid liquidated damages. 3.08 Increased Scope of Work: Between the time the project was formulated in 1970/71 and the time the contract started the road condition deteriorated so that less of the original gravel surfacing was salvageable for incorporation into the sub-base of the new pavement. However, a more significant increase in work was caused by raising the geometric design standards. The existing engineering plans for the gravel road were used by tenderers in preparing their bids but during construction, these plans were virtually ignored by the consultants, who redesigned and reengineered the road to significantly higher standards. The width of the road formation was changed from between 7.8 m and 8.4 m to between 9 m and 11 m, thus increasing the volume of earthworks: the vertical alignment was also significantly improved, further increasing earthworks. Both these changes caused the complete abandonment of any attempt to salvage the existing gravel pavement and accordingly the quantity of sub-base material required, and hence the cost of this element increased significantly. 1/ Because of the Bank Group's involvement in the entire Mtwara-Masasi road from preparation through appraisal, negotiations and the initial implementation, this PCR will cover the entire road and not only the Mtwara-Mingoyo section which was ultimately financed by the Bank/IDA. - 32 - 3.09 Four other factors affectea the rate of progress of the contract. First, the consultant's resident engineer did not have autnority tU make certain decisions without reference to his headquarters, and this delayed works. For example, a decision on the location of a borrow pit took six months. Second, the contractor was not prepared for the scope of work needed for the redesigned road. He was slow in assembling the necessary equipment and in procuring needed supplies. Later in the contract he encountered financial difficulties which made his continuation on the job questionable at one point, although with Government's sympathetic assistance in making advances to overcome his l1quidicy problems at that time, he was able to continue. Third, in the early days of the contract, tight security conditions were imposed on the southern part of the country as a result of the civil war in Mozambique. These conditions made travel and access to the area difficult and impeded communications and supplies as well as supervision by Bank Group staff. Fourth, the Government closed the border with Kenya in early 1977 which adversely affected the contractor's access to supplies of cement and spare parts, particularly in the later stages of the work between Mingoyo and Masasi. The quality of work was good, and the condition of the road is still satisfactory. Construction Supervision 3.10 Apart from the change in design standards and the fact that the Resident Engineer did not have the required delegated authority for efficient supervision, (see paras. 3.31 to 3.37) the consultant team performed its other technical functions adequately until April 1976 when the team was withdrawn. The consultant contended at that time that his agreement expired at the end of February 1976 and demanded an 86% increase in monthly rates to continue. The Government considered that the supervision contract was valid until completion of construction; in fact, the contract provided for a 13% rate increase after January 1, 1976 which was refused by the consultants. They also refused a compromise offer of about 35% rate increase and withdrew their staff at the end of April 1976. Both the Bank Group and the AfDB concluded that an increase of up to 30-35% may have been justified and that in turning down and not attempting to negotiate the Government's compromise offer, the consultants were at fault. 3.11 The Ministry decided to take over supervision using its own staff and, notwithstanding the fact that both IDA and AfDB staff were skeptical, MOW mobilized its own supervision team which performed quite adequately. The project was completed on August 5, 1978, and the maintenance period expired on February 5, 1979. - 33 - B. Betterment and Maintenance of Feeder Roads Consultants 3.12 At the time of appraisal, Government was prepared to hire a consultant firm for the supervision, field management and training involved in the feeder roads component. Despite the fact that this agreement was also confirmed during negotiations, in early 1972, Government pressed for proceeding without consultants, arguing that they had the capacity to allocate their own staff for this work. At about the same time, the consultants expressed serious reservations about working in Tanzania, citing the diffi- culty in recruiting staff, slow payments by the Government and high withholding taxes as the main reasons. The Bank Group held firm to the agreement that consultants be used to increase Government's capacity and avoid drawing on its scarce resources of trained staff. 3.13 As in the case of the consultants for superv sing the Mtwara- Masasi road, the discussions and negotiations between the Government and the consultants were protracted and after considerable delays in reaching agreement the contract was signed in late 1973. However, due to the problems previously cited by the consultants, they had difficulty in recruiting staff, and the supervision team was not fully assembled until December 1974. From its side, MOW also had difficulty finding qualified counterparts, and those finally assigned were not adequately qualified to be trained to take over the planning, supervision and management of the betterment units. Procurement 3.14 In the meantime, the Ministry, without reference or consul- tation with IDA,had proceeded with preparation of bid documents and specifications and in early January 1972, with the calling of bids for equipment for the betterment and maintenance units. When the Association reviewed the tendering procedure and documents in March 1972, it found that they were not fully consistent, in several respects, with procedures agreed with Government. As the closing date for bids was then imminent, the Association chose to reserve its position until it could review the results of the bid call and evaluation. After numerous discussions, the Association eventually agreed to a compromise arrangement whereby defi- ciencies in the bidding procedures and documents were compensated for to some extent in the "Conditions of Purchase" which formed part of the official purchase contracts negotiated with successful bidders. 3.15 The Association's agreement was given reluctantly and largely to avoid probable price increases and delay to a Norwegian-financed project for which equipment was being procured under the same bid call. However, in agreeing to this compromise, the Association emphasized that Government - 34 - should not consider this as constituting a precedent and that in the future Bank Group procurement guidelines should be adhered to in every respect. I/ 3.16 As a result of the delays in reaching agreement with consultants, the equipment began arriving in Dar es Salaam long before the BMUs were formed and ready to work. The equipment was therefore held at the port where it remained exposed to the elements and where it was subjected to considerable pilferage of parts and tyres. By the time the consultant supervisors were mobilized and the BMUs were ready to receive the equipment, much of it had deteriorated and needed to be repaired and rehabilitated before it could work. Betterment Activities 3.17 Work progressed slowly following delivery of the equipment and commencement of roadworks in April 1974 (see Table 3.01). Problems encoun- tered included: (a) occasional diversion of equipment by the politically powerful regional authorities; (b) interference by the industrial organi- zation (NUTA) and the political party (TANU) in the hiring, firing and management decisions of the units; (c) lack of spare parts and materials caused by Government's t;low procurement procedures and failure of equipment suppliers to provide adequate organizations and maintain a reasonable inven- tory of spare parts in Tanzania; (d) long communication and delivery lines to Dar es Salaam, the main source of supplies; (e) slow decision processes in MOW; (f) unsatisfactory performance of the consultants' resident engineer at Geita; and (g) insufficlently qualified local counterpart staff assigned by the Government. 3.18 The original schedule agreed with the consultants was for completing the 164 km of Mara roads within 24 months and the 303 km of Geita roads 2/ within 30 months. By October 1975, 19 months after field work had started, only about 20% of the work was completed; after 30 months, about 65% of the Mara and 25% of Geita roads were completed. 3.19 Progress remained slow, particularly at Geita, and over the first 18 months relations between the consultants and Government deteriorated. In late 1975 the resident engineer from Geita was withdrawn and the Mara resident engineer took responsibility for both areas. At about that time the consultants also prepared, and the Government accepted, a revised work program and changed schedules for the consultants' staff. The new schedules provided for consultant staff to progressively train and hand over responsibility 1/ Despite this qualification, Government pointed to the Bank Group's willingness to compromise in this instance as a precedent for requesting in a subsequent highway project that the contract for supplying several major items of equip- ment not go to the lowest evaluated bidder; however, in that instance the Bank Group stood fi:rm in its position and refused the request. 2/ Adjusted from Appraisal estimates of 165 and 310 km respectively. - 35 - to Tanzanian counterparts and then leave the site. The revised program provided for completing the Mara roads by January 1977 and the Geita roads by January 1978. The Mara component was completed in March 1977, about 12 months late, compared with the original schedule, while in Geita, at that time, only about 40% was complete. The work at Geita came to a standstill in about November/December 1978 when, following the inadvertent disbursement of all Credit funds (para. 3.28), the Government ceased budgetary allocations to the project except for minor maintenance and for payment of outstanding debts for works carried out before cessation of construction work. 3.20 When work stopped about 70 km of project roads had not been constructed and despite repeated IDA requests to both the Ministry of Works and to Treasury that Government fulfill its obligation under the Development Credit Agreement to complete the project, no action has been taken nor has Government given any firm commitment to complete the project or return the project equipment which has been dispersed to other works. If nothing else, this reflects the relatively low priority given to feeder road improvement projects and casts some doubt on the increasing emphasis being given to these types of roads in agricultural and regional development projects in Tanzania. 3.21 The Association reviewed both the consultant's and Government's staffing situation during the course of the project and several times inter- vened to extend the appointments of the expatriates when their local counter- parts were assessed as inadequately qualified and experienced to take over. Consultant's staff were progressively withdrawn except for the Resident Engineer and a Construction Superintendent who remained until March 1978 when their services were terminated concurrently with the posting to the job of an experienced expatriate Executive Engineer seconded on a bilateral arrangement to the Tanzanian Government. From that date progress picked up considerably, largely due to the ability and competence of the expatriate engineer, but came to a standstill in November 1978 when the expatriate engineer was transferred to another project. 3.22 The roads in Mara reverted to the regional authority for maintenance shortly after they were completed in 1977. From that time they have competed with all other roads in the Region for the inadequate road maintenance budget allocated to regional authorities. As a result they have not been maintained properly and have deteriorated almost to their state before the project. 3.23 The completed roads in Geita were maintained by the remnants of the BMU until late in 1979 when they also-reverted to the regional authority. It is anticipated that given the shortage of road maintenance funds in the region, these roads will also deteriorate rapidly to their pre-project condi- tion. The complementary feeder road program being implemented under the Geita Cotton Project, is continuing on other roads in the area. In view of the importance to the cotton industry of the feeder roads improved under the Third Highway Project, it would be desirable if they could be maintained by the Geita Cotton Project road unit. - 36; - C. Preinvestment Studies 3.24 The project as originally appraised included financial provision for feasibility studies and, subject to economic justification, detailed engineering of about 250 km of the country's highest priority roads for future investment projects. The specific roads were to be selected on the basis of preliminary studies by the Government. After these were carried out, the Government decided that there was no priority for further studies of major roads, and requested that the project should instead finance feasi- bility studies of feeder and development roads in two important agricultural areas; the Kilombero area in the south central part of the country and the Kilimanjaro area in the northeast. The Association agreed, after reviewing the request and then helped prepare terms of reference for the two studies. 3.25 Proposals were invited from five firms, and the Government selected separate consultant firms for each area. The studies began in about August/ September 1974, field work was completed in April/May 1975 and the draft final reports received in August/September 1975. The Association provided comments to the Government by November 1975 but the Government delayed until May/June 1976 to give its comments to the consultants. In turn the consultants, who in the meantime had dispersed their study group, took until December 1976 and September 1977 to complete the final reports for Kilimanjaro and Kilombero, respectively. The Consultant for the Kilombero study had also claimed addi- tional payment for extra expenditures caused by the delay. The Government disputed the claim, but: the Association has been unsuccessful, despite repeated requests, in ascertaining the outcome of this matter. 3.26 The studies were competently carried out, although both consultants indicated lack of confidence in the end result because the regional authori- ties, who are responsible for the roads, were not part of the decision to carry out the study and, therefore, did not cooperate with the consultants who needed information which only the regions could provide. The Kilombero study report has since been used as one input for defining an aid program by the EEC and the Swiss Government aid agencies. The Kilimanjaro study report is being used by USAID to help define a feeder roads program in that region. D. Cost Estimates and Disbursements Cost Estimates 3.27 The total project cost was US$25.2 million compared with the appraisal estimate of US$9.5 million (see Table 3.02). Cost increases were particularly severe in the civil works component, because of (a) increases in design standards and deterioration of existing sub-base of the Mtwara-Mingoyo-Masasi road, and (b) for supervision by consultants and Government due to the much longer than anticipated completion time. Delays in setting up of BMUs, in the delivery of equipment, organizational problems and a much slower work rate than anticipated causing labor cost increases, were the main causes of cost increases in the betterment of feeder roads component. - 37 - Disbursements 3.28 Because of the cost increases in the civil works component, the Association first suggested that disbursement percentages for the Mtwara- Masasi road be lowered from 70% to 32%, with Government financing the remain- ing costs. Once the AfDB agreed to assist the Government with financing of the Mingoyo-Masasi section, the Association decided to finance only the Mtwara-Mingoyo section, to transfer a surplus of US$1.9 million from the Second Highway Project to the civil works component and to increase the disbursement percentage for this component to 85%. Disbursements for the feeder road component were made at the original rate of 100% of foreign expenditures and 70% of local expenditures until December 1977. At that time, following a review of project finances and estimates, the Association reduced the rate of disbursement to 40% to ensure financial participation until all roads were complete. However, in March 1978 funds for this component were inadvertently fully disbursed when the reduced disbursement rate was overlooked by the Association. Disbursements were slower than anticipated for all components (see Table 3.03). Credit 265-TA was fully disbursed in March 1978 while the US$1.9 million of Loan 586-TA were fully disbursed in September 1979. E. Contractor's Performance 3.29 The contractor for the civil works component was slow to mobilize; he was not prepared for the scope of work he encountered, as the original bill of quantities provided for only nominal amounts of earthworks and sub-base, whereas the designs as amended by the consultants increased both items to become major components of the works. Accordingly, the contractor's equip- ment was insufficient, and he eventually was forced to buy additional equip- ment. The contractor was also slow in procuring needed supplies such as cement and bitumen. This was partly caused by, the closure of the Kenyan- Tanzanian border and partly by his lack of financial strength. 3.30 At one stage it appeared that the contractor could go bankrupt as he had serious liquidity problems. In order to avert the prospect of such bankruptcy, the Government agreed to make financial advances to the contractor to enable him to buy needed additional equipment and supplies. The financial advances were gradually recouped from subsequent progress payments. This assistance helped the contractor pull out of the crisis, and he recovered quite strongly to the extent that he was later accepted for another major project in Northern Tanzania. The quality of the road as completed by the contractor was acceptable. F. Consultant's Performance Mtwara-Mingoyo-Masasi Road 3.31 The supervising consultant was used for both the detailed engi- neering and the supervision of construction of this road. Engineering carried out for the road to be constructed to engineered gravel standard under the First Highway Project was updated only sufficiently to provide approximate quantities for the tender documents and bill of quantities to be prepared. - 38 - The engineering was not: detailed, as the Government proposed that it would be completed to accurate detail by the supervising consultants during cons- truction. The basic principle on which the designs and estimates were pre- pared was that the standards as well as vertical and horizontal alignment of the road would be the same as the previously engineered gravel road. 3.32 During the course of the new project, the consultants deviated from this principle and redesigned the road, particularly its vertical alignment, thus significantly increasing the quantities of the road's earth- works and sub-base materials, and thus the cost of construction. In addition, the consultants permitted the contractor to construct the road's cuttings and embankments wider than necessary, ostensibly at the contractor's cost (and without additional cost: to Government), in order to facilitate the contractor's later operations on the road. The nature and extent of the advantage to the contractor were not explained. Despite the consultant's assurances that the higher standards were constructed without additional cost to the Government, at least in two respects the Government would have borne part of the extra cost. First, because the vertical alignment was changed, none of the gravel pavement was salvaged and the Government had to pay the cost of the all-new gravel sub-base. Second, the extra earthworks and sub-base quantities took longer to construct than would have been the case if the original standards were retained and, accordingly, the cost of supervision borne by the Government was higher. 3.33 These matters were raised by the Association following a staff visit to the site in early 1976. The Government followed them up with the consultants, but they were not resolved satisfactorily before the consultants withdrew from the project in April 1976. Following their withdrawal the consultants wrote to Government with some explanations concerning the change in vertical alignment and the increased width of construction and, although the explana- tions were not completely satisfactory the Government did not pursue them further. 3.34 The Government, Association staff and AfDB staff considered that the consultants' performance in these matters was not adequately explained and that some of the problems arose from the fact that the consultants' resident engineer was not competent and experienced enough for the job and also did not have adequate authority delegated from the consultants' headquarters. The resident engineer's lack of authority, coupled with the long communication lines back to headquarters, amongst other things, caused delays in reaching decisions and thus'contributed to the delays in completing the project. 3.35 Apart from doubts remaining over the consultants' performance as indicated above, the circumstances and aftermath of their withdrawal in April 1976 brought criticism from the Government, the Association and the AfDB. Although the original contract between the Government and the consultants was written, for a three-year period to cover the construction period as anticipated at the time, it contained provision for a 13% increase in rates towards the end of the initial contract period. However, the consultants requested an 86% increase and refused the Government's compromise offer of an increase of about 35%. 3.36 When it became obvious to the consultants that the Government would not agree to a larger increase, with little warning to the Government they withdrew their staff, equipment and other facilities at the end of - 39 - April 1976. Both the Association and the AfDB considered that an increaae of between 30-35% may have been justified and that the consultants were at fault in refusing and not attempting to negotiate the Government's compro- mise offer. The Government then faced a serious problem of finding and mobilizing a new supervision team. With the Association's agreement the Government, on a trial basis, put together its own team to supervise the work, on the understanding that if it proved unsuccessful a new team of consultants would be hired. The Government team proved to be quite satis- factory and supervised the project to its completion. However, in order to put such a team together the Government had to withdraw some of its scarce experienced staff from other works. Therefore, although the withdrawal of the supervising consultant did not adversely affect this project, it did adversely affect the implementation of other projects, such as the ongoing IDA Highway Maintenance Project (Credit 507-TA) which suffered continually from lack of experienced local staff. 3.37 During discussions with the supervising consultant, the Government indicated that an unsatisfactory performance in this project could lead the Government to not offer future consultancy tasks to the firm. Although the Government expressed serious dissatisfaction with the firm, shortly after the firm's withdrawal, the Government awarded it another majc;r consultancy task in the power sector. Betterment and Maintenance of Feeder Roads 3.38 Consultants were hired to provide the management and supervision teams for the Mara and Geita BMUs and to train local counterparts. The team provided for Mara was satisfactory, and the work proceeded relatively smoothly, largely due to the ability of the resident engineer to work well with the local political and industrial organizations. The resident engineer's ability to work with and gain the respect of the work team assigned to him and his drive and enthusiasm which led him to substitute labor-inten- sive methods during frequent and extended equipment breakdowns, resulted in the Mara work program being satisfactorily completed. 3.39 The team assigned to the Geita BMU proved less effective than the Mara team largely due to the character and personality of the resident engineer appointed. After about two years of unsuccessfully trying to make this unit work, little work was being achieved, relations deteriorated between the consultants and Government and between the BMU and the local political and industrial organizations until finally the resident engineer was withdrawn in November 1975. At that point the resident engineer at Mara took responsibility for both units and despite the relatively wide separation between the two sites he was able to improve performance at Geita without much adverse effect on performance at Mara. Training 3.40 Although some attempt was made to fulfill the training element of this component, this did not work very effectively due to the conflicts between the need to keep the job moving and the consultants'time that would be needed to train local staff. Other major factors that exacerbated - 40 - this problem were the relatively low calibre of local staff assigned and, in some cases, an unreceptive attitude and a lack of interest on the part of the counterparts which prevented them from accepting training from the consul- tants. Many of the local staff assigned as counterparts to the consultants for training did not have the technical and academic ability needed for them to profit from the training. 3.41 After the Mara component was completed and the remaining consultant staff concentrated on Geita, some deterioration in progress occurred. The resident engineer's enthusiasm waned, the influence of the local political and industrial organizations was less effectively dealt with and the efficiency of the aging equipment was reaching a low level. The last two consultant staff members were terminated in March 1978, with about 134 km of roads uncompleted. The Government posted its own staff with an expatriate resident engineer in charge until November 1978 when all the work came to a standstill. Feasibility Studies 3.42 The two feasibility studies were carried out by the respective consultants quite satisfactorily. G. Progress Reporting 3.43 During the time consultants were supervising the civil works component and the BMU's at Geita and Mara, they were producing adequate regular progress reports on their work. However, despite repeated discus- sions, requests, and letters regarding the overall project, including the Government's own activities, analyses and assessments of progress, the Associa- tion was not successful in achieving its objective of having the Government accept and implement a satisfactory progress reporting procedure. This still remains a major deficiency in project implementation monitoring and attempts to improve the situation are continuing under other ongoing highway sub-sector projects. H. Implementation Schedule 3.44 Table 3.04 gives details of actual and expected schedules of project implementation. -41 - IV. INSTITUTIONAL DEVELOPMENT AND BORROWER PERFORMANCE A. Institutional Development 4.01 The project's main emphasis in this direction was the use of Betterment and Maintenance Units for the Geita and Mara feeder roads. This approach was adopted at the time of appraisal to enable the roads to be improved while at the same time using consultant managers and supervisors who were to train local counterparts in the various functions of such units. 4.02 As mentioned in paragraph 3.40, the training element was not particularly successful because of several factors, of which the most signi- ficant was the conflict between the work output objective and the training and institution building objective. Where work programs are being implemented, the immediacy of site problems, personnel needs and management decisions takes priority and the training and technical assistance program will be relegated to a secondary position for which only residual time will be applied. To better achieve institution building, project technical assistance and training should be provided exclusively for that purpose. 4.03 Another limiting factor was that the BMUs were directly responsible to MOW which does not have jurisdiction over the feeder roads improved. This meant that as the BMUs completed their work in Mara, for example, they were removed from the area, and the roads reverted to the regional authority for maintenance. Thus the objective of institutional development was largely lost when the collective expertise was dispersed to various MOW projects. 4.04 With regard to the supervision team put together by MOW for the civil works component, the group assembled worked well together and achieved a satisfactory result. However, in this case as well, the group was dispersed when the contract was finished, and although IDA staff suggested to the Government that the unit would be a valuable asset for the staff-deficient Highway Maintenance Project, the suggestion was ignored. 4.05 The project also provided funds for continuing in the MOW for two years, the services of two technical experts previously financed under the First Highway Project. However, shortly after the project was approved the two people left the Ministry. The Association raised the matter of using these funds for other technical assistance and training needs but the funds were never used for that purpose and were eventually disbursed against other project elements. B. Institutional Performance 4.06 Part way through this project the Ministry of Communications, Transport and Labor, which was the original executing agency, was divided into a separate Ministry of Communications and Transport (MCT) and a Ministry of Works (MOW). Responsibility for project execution was allocated to MOW except for the two feasibility studies which remained with MCT until about 1978 when - 42 - they also were taken over by MOW. Both the new Ministries suffered from lack of qualified experienced staff and from weak management and operational systems. This resulted in slow inefficient management and decision making and in an inability to identify the causes of problems and delays in project implementation. This was evident in the civil works component where the problems caused by the weak consultant resident engineer were not perceived until IDA staff pointed them out. Similarly, the problems raised for the BMUs by the politically strong regional authorities (see para. 3.17) could not be handled adequately by MOW staff. Even in the two feasibility studies, the Government ministries could not ensure the necessary involvement of the regional authorities who refused to cooperate with the consultants. 4.07 It became evident that institutional performance was dependent upon the individuals assigned responsibility for the project: where strong, experienced people were in charge, their performance was better and they were receptive to suggestions for improvement and strengthening; where individuals were weak and inadequately skilled, institutional performance was weak, in general obstructive and not receptive to suggestions for improvement. C. Compliance with Covenants 4.08 The main covenants of the Development Credit Agreement (DCA) in relation to project implementation referred to the hiring of consultants for feasibility and engineering studies, supervision of construction and management and supervision of the BMUs; the establishment of BMUs for the feeder roads component; employment of contractors for the civil works component; design standards for the civil works component; provision to the Association of a priority review of its road development program, all to the satisfaction of the Association. In addition the DCA included quite general covenants concerning the maintenance of the highway system, the establishment and enforcing of axle-load regulations, and the collection and recording of technical, economic and financial information and road traffic data. 4.09 Regarding the project specific covenants, compliance was variable as follows: (a) with regard to hiring of consultants for feasibility and engineering studies, compliance was satisfactory; (b) after the Credit was approved the Government was reluctant to hire consultants for supervision as required under the DCA and wanted to use its own staff. However, the Association was adamant in view of the poor staffing situation in the Government and the Government complied with the covenant. On the Mtwara-Mingoyo- Masasi road-, when the Government was left in the lurch following the consultants' departure, the Association agreed to allow an MOW team to supervise the work and this worked out quite satisfactorily; (c) the design standards for the Mtwara-Mingoyo-Masasi road were varied from those agreed. The indications are that this was a consultant's and not a Government decision. - 43 - 4.10 Other covenants relating to the establishment of BMUs,the use of contractors,and the provision of the priority review of the road develop- ment plan were complied with satisfactorily. 4.11 The one outstanding matter is the question of completing the feeder roads at Geita where about 70 km remain incomplete. Despite repeated requests and urging by the Association for the Government to complete these roads, no action has been taken nor is any anticipated. Accordingly, this part of the project remains incomplete. 4.12 The maintenance of the country's highway system and the establish- ment and enforcing of axle load regulations are continuing covenants in later ongoing projects. Progress in improving these matters has been slow but the Association has a continuing dialogue and continuing leverage through other projects, two of which are aimed directly at improving the capacity and capability for trunk road maintenance. - 44 - V. ECONOMIC RE-EVALUATION General 5.01 The two project components re-evaluated are: A. Improvement of the Mtwara-Masasi road; and B. Betterment and maintenance of agricultural and feeder roads in the Geita district and Mara region. A. Mtwara-Masasi Road Component The 1971 Appraisal 5.02 The 1971 Appraisal Report in its economic evaluation of the Mtwara- Masasi road construction component relied heavily on the 1968 feasibility study prepared by consultants. The 200 km road was subdivided into 4 sections and for each section the internal rate of return (IRR) was computed taking into account constructiont and supervision costs and vehicle operating cost (VOC) savings resulting from the road improvement. The IRRs for the 4 sec- tions ranged between 16% and 9% and the IRR for the whole road was 14%. The 1974 Reappraisal 5.03 After considerable delays, two contractors submitted bids in November 1973 for the construction of the 200 km project road. The lowest bid was more than double the cost as estimated in the 1971 appraisal. Because of this drastic increase in costs the Bank then (January 1974) undertook a reappraisal of the project based on the lowest bid cost and updated traffic, road maintenance and VOC savings estimates. The following table presents the IRRs according to the 1971 appraisal and according to the 1974 reappraisal. Estimated Internal Rates of Return (%) Road Section 1971 Appraisal 1974 Reappraisal I Mtwara-Mingoyo 16 10 II Mingoyo-Mtama 15 10 III Mtama-Nanganga 13 8 IV Nanganga-Masasi 9 2 All Sections 14 8 - 45 - 5.04 This reappraisal recommended, inter alia, that the Bank (a) advise Goverment to reduce the design standards of the 58 km Nanganga-Masasi section by omitting base and surfacing and (b) continue to assist financing the work but reduce the disbursement percentage to reflect the increased cost. 5.05 In February 1974 Government indicated that it intended to go ahead with the construction of the whole road as planned (no change in design standards for Section IV) and that it would secure additional financing from sources other than the Bank Group to cover the cost overrun. In addition, Government requested the Bank's urgent approval to award the contract to the lowest bidder, as the expiration date for bid validity was February 28, 1974. The Bank expressed strong reservation about the improvement of Section IV to paved standards but nevertheless approved award of the contract. The Bank moreover indicated that it would reduce its disbursement percentage for the Mtwara-Masasi road and that all cost overruns should be borne by Government. The Government then awarded the contract. 5.06 In August 1974 the Bank informed the Government that the disburse- ment percentage was to be reduced from 70% to 32% reflecting the increase in construction costs. 5.07 In the meantime, AfDB had shown an interest in assisting in the finance of the project road, and in October 1974 negotiated a US$4.8 million loan with the Government. 5.08 Subsequently it was agreed that the Bank would only finance the Mtwara-Mingoyo section and the AfDB, Mingoyo-Masasi (Sections II, III, and IV). Moreover, in January 1975, the Bank agreed to transfer the undisbursed balance of US$1.9 million from the Second Highway Project (Loan 586-TA) to the Third to assist Government in meeting the cost overrun. The disbursement percentage for the Mtwara-Mingoyo road section was increased from 32% to 85%. Comments on the Reappraisal Memo 5.09 The January 1974 Reappraisal Memo based its calculation of the IRRs of the four project road sections, inter alia, on the lowest amount of the bid submitted in November 1973, i.e. T Sh 86 million. In the economic calculation this amount was reduced by 25% to T Sh 64.5 million apparently to allow for the tax element. No evidence was provided as to the basis for the 25% estimate. In addition, the economic calculation on the cost side ignored any allowance for physical contingencies for construction (the 1971 Appraisal Report allowed for 6%) and it ignored the cost of construction supervision (the 1971 Appraisal Report allowed for supervision costs equivalent to 7% of construction costs plus a corresponding 10% for physical contingencies). 5.10 Correcting the Reappraisal Memo's IRR calculation to allow for a 7.5% tax elementl/ included in the contract price, 6% physical contingency for 1/ Five percent duties and taxes on construction costs and 43% taxes on supervision costs . - 46 - construction, supervision costs estimated at 7% of construction costs and 10% physical contingency for supervision (for details, see Table 5.01) while accepting the benefits as calculated in the Reappraisal Memo, gives the following results: Estimated Internal Rates of Return (%) "Corrected" 1974 1974 Road Section Reappraisal Reappraisal I Mtwara-Mingoyo 10 6 II Mingoyo-Mtama 10 7 III Mtama-Nanganga 8 4 IV Nanganga-Masasi 2 -2 All Sections 8 4 5.11 If the 1974 Reappraisal Memo had included the "correct" IRRs, its recommendation would in all likelihood have been that the design standards for the entire project road (not just one section) be lowered from paved to gravel. A detailed review of the contractor's bid indicates that the correspon- ding reduction in construction costs would have amounted to over 60%. Given this background the subsequent course of events might have been quite dif- ferent from what actually happened as described above (para. 5.05). Re-evaluation 5.12 The following table presents the re-evaluated rates of return of the four road sections of the Mtwara-Masasi Road 1/. For comparison sake the IRRs of the appraisal, reappraisal and "corrected" reappraisal are also included in the table. Estimated Internal Rates of Return (%) 1971 1974 "Corrected" 1974 1980 Re-evaluation Road Section Appr. Reappr. Reappraisal "Actual" I Mtwara-Mingoyo. 16 10 6 4 II Mingoyo-Mtama 15 10 7 7 III Mtama-Nanganga 13 8 4 4 IV Nanganga-Masasi 9 2 -2 1 All Sections 14 8 4 4 1/ Though the last 3 sections (between Mingoyo and Masasi) were ultimately financed by the AfDB, it is nevertheless interesting, given the original scope of the Bank's involvement from preparation, through appraisal and into implementation, to see the actual return of these sections. - 47 - 5.13 Tables 5.01 to 5.04 provide details on the underlying data such as construction and supervision costs, traffic flows and VOC savings and permit their comparison over time. The re-evaluation has been done in terms of constant 1979 prices. 5.14 The re-evaluated or "actual" IRRs closely resemble the "corrected" 1974 reappraisal IRRs. In the re-evaluation the impact of higher construction and supervision costs (see Table 5.01) is offset by higher unit VOC savings (see Table 5.04) while the traffic volumes in both calculations (see Table 5.02) are about the same. The traffic growth rate anticipated at appraisal has not occurred. This is mainly due to the drop in cashew nut production in spite of the Bank Group's project to help increase production. The "actual" overall rate of return of the road is 4% and so is the IRR for the Bank Group financed Mtwara-Mingoyo road section. Conclusions 5.15 Three main points emerge from the foregoing: (a) Though project appraisal reports are closely scrutinized in the Bank and go through several rounds of internal review, the 1974 Reappraisal Memo, while essentially performing the same function, did not go through this process; hence its shortcomings were more likely to go unnoticed: (b) Because of the Reap- praisal Memo's shortcomings, its recommendations formed the wrong basis for subsequent Bank actions and decisions. (c) In spite of an IRR of 2% for the Nanganga-Masasi road section as calculated in the 1974 Reappraisal Memo and Government's refusal to lower the design standard for that road section as recommended in the Reappraisal Memo, IDA in 1974 accepted Government's posi- tion and agreed to assist in financing the construction of this sectionl/. (That financing of this section was subsequently taken over by the AfDB does not affect this point.) 5.16 Two lessons are to be learned from the foregoing. First, the re- appraisal of a project, undertaken because of changed circumstances since the time of appraisal, deserves as much attention and scrutiny as an appraisal de novo. And second, the recommendations and conclusions of such a reappraisal deserve the same respect as those of an appraisal de novo even if they imply a drastic revamping or abandoning of the original project concept. 1/ It is interesting to note in this context that the 1971 Appraisal Report originally did not include this road section as the appraisal team judged the estimated IRR of 8% too low. It was only during credit negotiations that the Bank after considerable discussion--and after increasing the estimated IRR by one percentage point and referring to unquantifiable additional benefits-- included this section in the project at the urgent request of the Tanzanian negotiating team (see also para. 2.05). - 48 - B.1 Geita Feeder Roads Components The 1971 Appraisal 5.17 This component covered betterment of 310 km of feeder roads in Geita district, the economy of which is primarily based on cotton. In its Economic Evaluation the Appraisal Report (page 16) states that "A cotton development project for the Geita district .... is in an advanced stage of preparation with assistance from the PMEAl/ .... Since many of the selected feeder roads are being improved specifically for seed-cotton collection and the distribution of seed, fertilizers and insecticides, (the Geita feeder roads component) cannot be divorced from the overall cotton promotion scheme involving the distribution of essential inputs. Net benefits for the project have been computed after allowance for total additional costs of producing the .... cotton as well as for the road betterment and maintenance. These benefits are expected to yield an overall economic rate of return of about 21%." 5.18 The "cotton development project" referred to above was subsequently appraised by the Bank and approved by the Board in January 1974 (Geita Cotton Project: Credit 454-TA). The Closing Date for this Credit is December 31, 1982. Re-evaluation 5.19 Given the original appraisal of the Geita feeder road component of Highways III as an integral part of the Geita Cotton Project as well as the fact that the Geita Cotton Project will not be completed before the end of 1982, it is logical to undertake the re-evaluation of the feeder roads component within the fcamework of the re-evaluation of the Geita Cotton Project. This has been discussed with the Bank's Agricultural Division con- cerned (EAPSA) and it was agreed that the re-evaluation of the Highways III Geita feeder roads component will be dealt with in the PCR of the Geita Cotton Project. 5.20 In the meantime--pending the preparation of the Geita Cotton Project PCR--the Geita feeder roads component must be judged unsuccessful (see para. 3.23) as some of the improved feeder roads have already reverted to their original condition and others are deteriorating rapidly. In economic terms, the very short life span of the feeder roads would imply a negative rate of return of the investment involved. 5.21 Table 5.05 p:resents the original cost estimates and the actual costs of the Geita feeder roads component. 1/ Permanent Mission in East Africa; since then the name has been changed to Resident Mission in East Africa (RMEA). - 49 - B.2 Mars Feeder Roads Component The 1971 Appraisal 5.22 This component covered betterment of 165 km of feeder roads in Mara region. In its Economic Evaluation the Appraisal Report (pages 16 and 17) states that this component "relates specifically to a Government-spon- sored dairy development scheme which is supported by Danish technical and financial assistance. The roads in question are essential to the collection of milk for two milk processing plants now under construction and virtually all traffic on the roads will relate to the dairy scheme. It is therefore impossible to separate the benefits accruing to the road and milk collection and processing elements. Justification for both the road betterment program and the dairy scheme is based on the net benefits to be obtained from induced milk production, after allowance for all costs of milk processing and of road betterment and maintenance .... The combined dairy scheme and road betterment program is expected to yield an economic return of 28%." Re-evaluation 5.23 As in the case of the Geita feeder roads, the Mara feeder roads improved under this project have deteriorated and are deteriorating quickly-- because of lack of subsequent maintenance--to the point where they have reverted to their original poor condition (see para. 3.22). Hence, given the very short life span of the feeder roads, and as in the case of Geita (see para.. 5.20), the economic return on the investment made in improving these roads is negative. 5.24 Table 5.05 presents the original cost estimates and the actual cost of the Mars feeder roads component. - 50 - VI. THE ROLE OF THE BANK GROUP 6.01 Despite the Bank Group's usual commitment to financing only proven economically viable projects, it yielded to Government pressure to include in the project a section of road which had been shown to be uneconomic at the design standards proposed. The re-evaluation has shown that the ex-ante economics of the whole civil works component were not properly calculated when the component was reappraised using actual bid prices. The Bank Group's rather cursory review of the reappraisal calculations and their subsequent acceptance, resulted in a questionable decision to proceed with the project as appraised. 6.02 Another factor which further adversely affected the economics of the civil works component was the fact that the engineering for this component was not fully detailed but was accepted by the Bank Group as sufficiently accurate for the project to be appraised and approved. In addition, it was only in late 1975--when much of the Bank Group-financed Mtwara-Mingoyo road was completed--that Bank Group supervision missions became aware that actual construction was being completed at considerably higher standards than those agreed at appraisal. Earlier and more close attention by the Bank Group to the detailed engineering being prepared by the consultants during construction could have reduced the cost and time of construction and hence improved the economics of the road. 6.03 As mentioned above and elsewhere in this report, on a number of occasions during the preparation, appraisal and implementation of this project, the Bank Group faced and yielded to pressures from Government to retreat from its usual policies, guidelines and procedures. First, it agreed to include for appraisal a road component when there were good reasons to anticipate that there was insufficient time to prepare needed economic data, detailed engi- neering and accurate cost estimates. In hindsight, the Bank Group's assess- ment was right and the data was not ready for appraisal. Second, after deleting an uneconomic section of road from the project during appraisal, the Bank Group during Credit negotiations, agreed to its re-inclusion on the basis of a one percentage point higher rate of return calculation and a reference to additional unquantifiable benefits. Third, after the project was appraised and approved, contract bids were called and a contract was awarded on the basis of incomplete detailed engineering. Fourth, after bids came in much higher than expected the recalculated economic rate of return for one section of road was reduced to 2%. Instead of revising the project by either reducing standards--and therefore costs--or by deleting known uneconomic components, the Bank Group agreed to proceed with the project as it was and also to transfer an unexpended Loan Balance from another project to help meet the increased costs. Each of the above Bank Group deci- sions, to yield to Government requests, progressively moved this project from a position where it was designed to help the economic development of the country to the end result as reported herein, of a project with negligible if any positive economic impact. - 51 - 6.04 The Bank Group insisted on the use of consultants for supervision of both the civil works and the feeder road components. In view of the serious lack of trained experienced staff in the Ministry concerned, this stance proved to be fully justified and correct, as verified by the fact that the Ministry had great difficulty even mobilizing suitable counterparts for training under the feeder roads component. Also, when the Ministry was forced to use some of its own staff to supervise the civil works element when the consultants left, this exacerbated the local staff shortage in other Bank Group projects in the country. On this point, the Bank Group intervened to extend the appointments of expatriates on several occasions when the Ministry planned to release some of the consultant team for the feeder roads element and substitute its own inadequately qualified and experienced staff. VII. CONCLUSIONS 7.01 The project's objectives in terms of physical outcome were mostly achieved in that the civil works component was completed and most of the feeder roads were improved. However, the objectives in terms of economic outcome were not achieved as the re-evaluation of the main investments shows low or negative rates of return. This occurred because of: (a) the failure to pay as much attention to the reappraisal of the civil works component as is normally given to the initial appraisal; (b) the unusually "soft" position taken by the Bank Group to the low rate of return of part of the civil works component; and (c) the lack of clear arrangements for the continued maintenance of the feeder roads following completion of the project. 7.02 The Fifth Highway Project contains a component for a Pilot Regional Roads program under which an attempt will be made to find a more satisfactory answer to the problem of improving and maintaining feeder roads, taking into account the institutional arrangements and deficiencies in the country. - 52 - ANNEX TANZANIA PROJECT COMPLETION REPORT THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Sources of Information 1. Bank Project Files - First Highway Project (Credit 48-TA/Supp. Cr.115-TA) 2. Bank Project Files - Second Highway Project (Loan 586-TA/Credit 142-TA) 3. Bank Project Files - Third Highway Project (Credit 265-TA/Loan 586-TA) 4. Contract Documents - Mtwara-Mingoyo-Masasi Road 5. Design and detailed engineering reports for Mtwara-Mingoyo-Masasi road (1970-1971) 6. Monthly and quarterly progress reports on Mtwara-Mingoyo-Masasi road 1974-1978 7. A Study of Feeder Road Development in Six Areas in Tanzania by United Research Company (1969) 8. Consultant Study of Feeder Road Development in Mara and Mbulu Areas (1970) 9. Monthly and quarterly progress reports on Geita and Mara feeder roads work 1974-1977 10. Traffic Counts - Ministry of Communications and Transport (1979) 11. Cost Information from Ministry of Works 1979 and 1980 12. Appraisal Report and Credit and Loan Agreements 13. Staff Appraisal Report (1978), Tanzania Fifth Highway Project June 1980 - 53 - TABLE 2.01 TANZANIA THIRD HIGHWAY PROJECT - CREDIT 265-TA/LOAN 586-TA PROJECT COMPLETION REPORT Appraisal Design Standards for Mtwara-Masasi Road Design speed 80 km/h Minimum radius 300 m Maximum gradient 6% Pavement width 6 m Maximum shoulder width in cut 0.9 m Minimum shoulder width on fill 1.2 m Width of structures 1/ 2 lanes Pavement design axle load 8 m tons 1/ Three existing one-lane bridges, in good condition, would be retained. Source: IBRD, "Appraisal of Third Highway Project, Tanzania, July 1971, Table 4. June 1980 TANZANIA PROJECT COMPLETION REPORT CREDIT 265-TA/LOAN 586-TA - THIRD HIGHWAY PROJECT Feeder Road Betterment Program Geita district Completed Beginning of Work Completion of work Road Name/Spine Roads Length to 1/79 Appraisal Act. 'Exp. Act. Exp. ------------------------------ km------------------------ Buchenzi-Kamanga 62.0 62.0 68 75 ) 10/76 ) Sengerena-Katunguru 16.7 16.7 17 74 ) 7/76 ) Busisi-Buyagu 17.6) 17.6) 76 ) 9/72 77 ) 7/74 Buyagu-Karumwa 55.7) 55.7) 88.0 76 ) 78 ) Busolwa-Kasamwa 32.9) 19.5) 77 ) n.y.c. ) Total spine roads 184.9 171.5 173.0 Minor feeder roads 124.6 58.4 137.0 Mara Region: Road Name/Spine Roads Kitarene Rodi) - - 65.0 Rodi-Shirati) 1/ Rodi-Lyagoro 16.6 16.6 17.0 75 ) mid-76 ) Ochuna-Nyamonga 54.0 54.0 55.0 74 ) late-75) Maji Moto-Mara River 28.2 28.2 28.0 75 ) 7/72 77 ) 12/73 Kiagata-Simba 29.5 29.5 - 75 ) mid-76 ) Simba-Maji Moto 18.0 18.0 - 75) 77 ) Total spine roads 146.3 146.3 165.0 Link roads Kiagata-Mugumu-Tarime 7.5 7.5 76 77 Nyibako to Borenga 4.0 4.0 - 76 n.a. 77 n.a. 157.8 157.8 165.0 o I/ Improved by COMWORKS in 1973-74,therefore deleted from Program. Source: Progress Reports. June 1980 Table 3.02 - 55 - TANZANIA PROJECT COMPLETION REPORT CREDIT 265-TA/LOAN 586-TA - THIRD HIGHWAY PROJECT Actual and Appraisal Estimates of Project Costs (US$m) Project Actual Cost/ Appraisal Estimate Actual cost as % component Local Foreign Total Local Foreign Total of Appraisal Est. 1. Road Construction: Mtwara-Mingoyo 2,087 4,435 6,522 2/ )1,793 3,810 5,603 316 Mingoyo-Masasi 3,575 7,598 11,173 2/ ) 2. Betterment of Feeder Roads: (i) Equipment and spares 410 930 1,340 - 791 791 169 (ii) Culvert pipes and fuel 393 917 1,310 93 372 465 282 (iii) Labor 1,700 - 1,700 520 - 520 327 3. Preinvestment Studies: (i) Mtwara-Masasi ) 39 145 184) (ii) Masasi-Songea ) 50 160 210 ) 98 (iii) Mara & Mbulu ) ) Feeder Roads ) 12 18 30) (iv) Kilimanjaro Kilombero studies 190 580 770 157 593 750 103 4. Supervision and Training: (i) Mtwara-Mingoyo 514 183 697 3/ ) Mingoyo-Masasi 528 124 652 ) 120 280 400 337 (ii) Feeder Roads 186 626 812 289 431 720 113 5. Staffing and Training Program: 0 0 0 50 50 Total 9,633 15,553 2 3,023 6,490 9,513 265 1/ Rate of exchange used TSh 7.7 to US$1 2/ Foreign cost element estimated to be similar to that used in Appraisal Report. 3/ Includes supervision by Government upon termination of consultant contract. Source: MOW reports 1976-1980 June 1980 - 56 - Table 3.03 TANZANIA PROJECT COMPLETION REPORT CREDIT 265-TA/LOAN 586-TA - THIRD HIGHWAY PROJECT Schedule of Disbursements as of June 1980 IBRD Accumulated Disbursements (US$ '000) Actual Disbursement Fiscal Year Actual Total Appraisal as % of Appraisal & Quarter Disbursements Estimate Estimates 1972 1st - 2nd 150 3rd 170 4th 360 1973 lst 113 1,230 10 2nd 113 1,920 6 3rd 113 2,480 5 4th 161 3,110 5 1974 1st 174 4,030 4 2nd 174 5,030 3 3rd 174 5,810 3 4th 174 6,300 3 1975 lst 770 6,400 12 2nd 1,735 6,500 27 3rd 1,735 27 4th 2,290 35 1976 1st 2,966 46 2nd 3,454 53 3rd 3,924 60 4th 3,936 61 1977 1st 4,865 75 2nd 5,400 83 3rd 6,300 97 4th 6,300 97 1978 1st 6,400 98 2nd 6,500 100 1/ 3rd 7,000 4th 7,000 1979 1st 7,000 2nd 7,000 3rd 7,000 4th 7,300 1980 1st 8,400 2/ 1/ Credit 265-TA fully disbursed. 2/ Surplus of US$1.9 million transferred from Ln. 586-TA fully disbursed. Source: IBRD Controllers Department. June 1980 TANZANIA PROJECT COMPLETION REPORT CREDIT 265-TA/LOAN 586-TA - THIRD HIGHWAY PROJECT Actual and Expected Project Implementation Percent the work- Dates of comp ietLe Completion by expetzet Project Bid receipt Contract award Beginning of work of work completion component Actual Expected Actual Expected Actual Expected Actual Expected date Construction Mtwara-Masasi 11/73 late 71 3/74 3/72 10/74 4/72 8/18 12/73 0 Betterment Feeder Roads 1/ n.a. n.a. late 73 6/72 12/74 7/72 11/78 2/ 73-74 0 Preinvestment studies: Kilimanjaro 6/74 8/74 9/74 12/76 3/ 11/75 Kilombero 6/74 n.a. 8/74 n.a. 10/74 n.a. 9/77 3/ 11/75 90% Supervision and Training: Mtwara-Masasi n.a. 72 11/73 3/72 11/73 4/72 4/76 4/ 12/73 0 Feeder Roads n.a. 72 12/73 6/72 12/74 7/72 3/78 5/ 7/74 0 1/ See Table 3.01 for detailed list of road betterment program. 2/ All work was completed at Mara in March 1977 and roads transferred to Regional Authorities for maintenance. Major portion of construction works completed at Geita by late 1978 but connecting feeder roads not yet completed. 3/ Draft final reports submitted August 1975. C 4/ Date when consultants left and Government took over supervision. 5/ Date when last member of consultant's team left Geita district. Sources: IBRD Projec1 Files. IBRD, 'Appraisal of Third Highway Project, Tanzania", July 1971. - 58 - Table 5.01 TANZANIA PROJECT COMPLETION REPORT THIRD HIGRWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Mtwara-Masasi Road Construction and Supervision Costs (Current T Sh Million) "Corrected" 1980 1971 1974 1974 Reevaluation Appraisal Reappraisal Reappraisal Actual 1/ 2/ 3/ 4/ Construction: - Costs 34.5 64.5 81.7 - - Physical Contingencies (6%) 2.1 - 4.9 - - Total 36.6 64.5 86.6 129.0 Supervision: - Costs 2.4 - 3.3 - - Physical Contingencies (10%) .2 - .3 - - Total 2.6 - 3.6 7.0 Grand Total Construction and Supervision: 39.2 64.5 90.2 126.0 Index - 1971 Appraisal - 100 100 165 230 347 - 1974 Reappraisal - 100 100 140 210 - 1974 "Corrected". Reappraisal - 100 100 151 1/ Includes taxes 2/ Excludes 25% taxes on construction costs 3/ Excludes 5% taxes on construction costs and 43% taxes on supervision costs 4/ Excludes5.4% taxes on construction costs and 34% taxes on supervision costs Sources: IBRD, "Appraisal of Third Highway Project, Tanzania", July 1971, page 11; Memo from Mr. Reid and Mr. Kirmani, "Mtwara-Masasi Road Paving, Economic Reappraisal", January 17, 1974 and information from Ministry of Works, Tanzania. June 1980 - 59 - Table 5.02 TANZANIA PROJECT COMPLETION REPORT THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Mtwara-Masasi Road 1970/78 Actual Traffic and Forecast for 1978 1978 Reapprai- Actual Forecast For sal Forecast 1970 1978 1978 1978 As % of Road Section (Appraisal) (MCT) (Appraisal) (Reappraisal) 1978 Actual ---------------vehicles per day---------------- ------%------- I. Mtwara-Mingoyo 165 133 234 150 113 II. Mingoyo-Mtama 180 165 210 143 87 III. Mtama-Nanganga 120 133 169 134 100 IV. Nanganga-Masasi 85 98 117 80 82 Simple Average 130 132 183 127 96 Note: For purpose of the re-evaluation, future traffic growth has been estimated at 3% per year for all traffic. This compares with a forecast at appraisal of 3% per year for cars and 5% per year for trucks and buses, and with a forecast at reappraisal (based on actual 1973 traffic counts) of 3% per year for cars, 5% per year for buses and no increase for trucks through 1979. Sources; For 1970 and forecast 1978 traffic, see IBRD, "Appraisal of the Third High- way Project, Tanzania", July 1971, page 15, and "Mtwara-Masasi Road Paving, Economic Reappraisal", January 17, 1974, page 2 and Table,1. For actual 1978 data, Ministry of Communications and Transport (MCT), 1979. June 1980 - 60 - Table 5.03 TANZANIA PROJECT COMPLETION REPORT THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Detailed Actual Traffic, 1978/79 Mtwara-Masasi Road Pickups/ Trucks Buses Road Section Cars Landrovers (9-ton) (50-passenger) Total -------------------vehicles per day--------------------- I Mtwara-Mingoyo 9 46 68 10 133 II Mingoyo-Mtama 4 57 86 18 165 III Mtama-Nanganga 2 53 69 9 133 IV Nanganga-Masasi 3 40 43 12 98 Source: Ministry of Communications and Transport, 1979. June 1980 - 61 - Table 5.04 TANZANIA PROJECT COMPLETION REPORT THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Vehicle Operating Costs Mtwara-Masasi Road (T Sh/100 veh-km) Type and Condition Pickups/ Trucks Buses of Road Cars Landrovers (9-ton) (50-passengers) 1. From the January 1974 Reappraisal Report 1/: Bitumen, good 2/ 23 34 102 119 Gravel, good 2/ 29 44 158 182 2. Used in the Re-evaluation of the Project 3/: A. Flat to Rolling Terrain Bitumen, good 68 77 198 214 Gravel, fair 104 118 332 363 B. Rolling to Hilly Terrain Bitumen, good 70 79 200 216 Gravel, fair 106 121 357 369 1/ In December 1973 prices. 2/ If surface condition is poor, operating costs are assumed to be 15% (bitumen) and 20% (gravel) higher. 3/ In January 1979 prices; source: IBRD, "Tanzania, Fifth Highway Project,"Staff Appraisal Report, December 1978, page 37. June 1980 - 62 - Table 5.05 TANZANIA PROJECT COMPLETION REPORT THIRD HIGHWAY PROJECT (CREDIT 265-TA/LOAN 586-TA) Betterment and Supervision Cost Geita and Mara Feeder Roads 1980 Re-evaluation 1971 Appraisal 1/ Actual Geita Mara Total Geita Mara Total -------T Sh Millions in current prices------- Betterment: - Equipment and spares 5.13 - Culverts pipes, fuel 2.56 - Labor 2.86 Subtotal 10.55 - Physical con- 1.08 tingencies Total 7.56 4.07 11.63 23.23 10.27 33.50 Supervision: - Consultants 4.28 - Physical con- tingencies .43 Total 3.06 1.65 4.71 3.76 2.50 6.26 Grand Total 10.62 5.72 16.34 26.99 12.77 39.76 1/ The Appraisal Report only gives the combined cost estimates for Geita and Mara; this table breaks these costs down in proportion to the mileage of feeder roads covered in each area to allow a comparison with the actual costs by area. Source: IBRD "Appraisal of Third Highway Project, Tanzania" July 1971, page 11 and Table 5 and Ministry of Works, Progress Reports. June 1980