Tanzania - Second Highway Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 4030 PROJECT PERFORMANCE AUDIT REPORT TANZANIA SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-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 contents may not otherwise be disclosed without World Bank authorization.  FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT TANZANIA: SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-TA) TABLE OF CONTENTS Page No. Preface................. . ..... i Basic Data Sheet.. ...... ......... ........ . ii Righlights...... .... .. . .. ........... .. .. ............ iv PROJECT PERFORMANCE AUDIT MEMORANDUM The Project.... o.,o. .....*........ .... ......... 1 Pavement Design.............. ........oo ... o....o 1 Economic Returns..-.... ... ........................... 3 Recouping Benefits to Transit Traffic......................... 5 Use of Loan Funds............................ . ........... 5 Conclusions...................... ........... 6 Annex: SIDA Comments .................................. 7 PROJECT COMPLETION REPORT Summary and Conclusions............................. 9 I. Introduction..o................ .............. ............ 12 II. Project Preparation and Appraisal........... ......... 13 III. Project Implementation and Cost......... ....... 15 IV. Institutional Development and Performance of Borrower... 21 V. Economic Re-evaluation............................ 22 VI. The Role of the Bank Group.......o........... ..... 31 VII. Conclusions...... ....,. . . . . . . . . . . ...... ... 32 Annex: Sources of Information................. ......... .33 Tables: 1. Actual and Expected Project Implementation......... ... 34 2. Actual and Appraisal Estimates of Project Costs........... 35 3. Detailed Comparison of Different Estimates of Benefits and Costs...................................... 36 4. Comparison of Actual Traffic Volumes, 1970 and 1972 and Appraisal Forecast 1972.......... ... . . 37 5. Comparison of Forecast and Actual Traffic, 1972..oo...... 38 6. Actual and Forecast Traffic, 1967 and 1972....... 39 7. Comparison of Unit Vehicle Operating Costs and Savings Estimates for Different Road Types, Appraisal versus "Actual"'. ....................... ........oooooo 40 8. Schedule of Disbursements as of October 31, 1974.......... 41 Map 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.  PROJECT PERFORMANCE AUDIT REPORT TANZANIA: SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-TA) PREFACE The following is a performance audit on the Second Highway Project in Tanzania for which a combined Bank Loan (US$7.0 million) and IDA Credit (US$8.0 million) were made in February 1969. The financing plan for the project included also a US$15.0 million equivalent credit by the Swedish International Development, Authority (SIDA) and US$8.0 million local funding. Part of the Swedish financing for the project was provided on the understand- ing that once the Second Replenishment of IDA funds became effective, Tanzania would apply for a further IDA credit of US$7.5 million and, if the credit was granted, an equivalent amount of the Swedish credit would be cancelled. In November 1969, the loan/credit documents were amended to reflect that change. The final financing plan for the project included: the Bank loan (US$7.0 million), a US$15.5 million IDA credit, and a US$7.5 million equivalent SIDA credit in addition to the US$8.0 million local funding. The project was fully and satisfactorily implemented with some delay and at costs below appraisal estimates. Final completion of the project was delayed to 1974 due to lengthy proceedings to settle contractor and consultants' claims. The lower costs generated a small surplus (US$1.9 million) in the Loan Account; to use it up, the loan documents were amended in 1975, with the Board's approval, to include a component of the Third Highway Project (Credit 265-TA, 1971), on which heavy cost overruns were incurred. That component, the Mtwara-Masasi road, although strictly part of this project, is not audited here but will be audited as part of the Third Highway Project. However, because of delays in implementing that component, preparation of the PCR for the Second Highway Project was delayed until late 1978 when the project was finally closed. Final disbursement of the loan took place in September, 1979. The attached Project Completion Report (PCR) was prepared by the Eastern Africa Region's Highway Projects Division without the assistance of the Borrower. An Operations Evaluation Department (OED) mission visited Tanzania in July, 1981 to discuss the project with Government officials. The assistance of the Government and the Ministry of Works in particular is gratefully acknowledged. As for all practical purposes the Bank acted on behalf of Sweden in the administration of the project, SIDA was not consulted in preparing this audit. However, a copy of the draft audit was sent for their comments, which are attached as an Annex to the Audit Memorandum and reflected in the text. OED has reviewed the Appraisal and President's Reports, the Minutes of the Executive Directors' Meeting at which the project was considered, Bank records and files, and has found that the PCR covers adequately and comprehen- sively the issues arising from the implementation of this project. However, the benefit of two more years of experience since the PCR was written have allowed OED to emphasize certain salient aspects of the Project in the Audit Memorandum. The draft audit was sent to the borrower for their comments; how- ever, none were received,.  - ii - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET TANZANIA: SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-TA) KEY PROJECT DATA Original Actual or Item Expectation Reestimated Project Costs (US$ million) 38.00/a 35.20 Cost Overrun (%) - - 8 Loan/Credit Amount (US$ million) 22.50 22.50/ Disbursed 22.50 22.50/c Cancelled - 0.00 Repaid to June 30, 1981 0.92 Outstanding to June 30, 1981 - 25.451 Date Physical Components Completed 09/71 05/72 Proportion Completed by Expected Date (%) 100.00 80 Proportion of Time Overrun (%)- 15 Economic Rate of Return (%) 20(15) 16(7)/e Cumulative Estimated and Actual Disbursements (US$ million) FY69 FY70 FY71 FY72 FY73 FY74 Estimated - 5.4 14.0 22.5 - - Actual 1.1 4.6 12.1 18.6 19.8 19.9 Actual/Estimated (Z) - 85 86 83 88 88 FY75 FY76 FY77 FY78 FY79 FY80 Estimated - - - - - - Actual 20.6 20.6 20.6 21.1 21.4 22.5 Actual/Estimated (Z) 91 91 91 94 95 100 /a Financing for the project included the Loan/Credit amount (US$22.5 million), a Swedish International Development Authority Credit (US$7.5 million) and US$8.0 million of local Government funding. /b Loan 586-TA: US$7.0 million; Credit 142-TA: US$15.5 million. /c US$1.9 million from Loan 586 were used, with Board approval, to meet cost overruns on the Third Highway Project (Credit 265-TA, 1971). /d Includes US$3.87 million exchange adjustment. /e Including and excluding benefits to Zambian transit traffic respectively, and considering benefits comparable (in real terms) with those used at appraisal. If an alternative estimate of vehicle operating cost savings (PCR, para. 5.19 et. seq.) is used (expressed in similar real terms) the rates of return of the project would increase to 24% and 14%, with and without Zambian transit traffic, respectively (PPAM, para. 12).  - iii - OTHER PROJECT DATA Original Item Plan Actual First Mention in Files 07/66 Government's Application - 01/68 Appraisal 03/68 04/68 Negotiations 11/68 11/68 Board Approval 02/69 02/18/69 Loan/Credit Agreement Daice 02/69 02/24/69 Effectiveness Date 04/69 04/10/69 Closing Date (Credit) 12/31/72 12/31/72 Closing Date (Loan) 12/31/72 12/31/78 Borrower United Republic of Tanzania Executing Agency Ministry of Works Fiscal Year of Borrower July 1 - June 30 Follow-on Project Name Third Highway Project Credit Number 265-TA Amount (US$ million) 6.5 Credit Agreement Date 08/06/71 MISSION DATA Number of Number of Date of Item Month/Year weeks of Persons Manweeks Report Pre-appraisal 01/68 1.0 1 1.0 02/01/68 Appraisal 04/68 2.0 3 6.0 01/15/69 Supervision I 04/69 3.0 1 3.0 06/05/69 Supervision II 10/69 1.0 1 1.0 12/10/69 Supervision III 07/70 2.5 2 5.0 07/31/70 Supervision IV 11/70 1.0 1 1.0 12/18/70 Supervision V 08/71 2.0 1 2.0 09/24/71 Supervision VI 03/72 1.5 2 3.0 04/25/72 Supervision VII 06/72 2.0 1 2.0 07/28/72 Supervision VIII 11/72 2.0 1 2.0 12/28/72 Supervision IX 03/73 2.0 1 2.0 03/21/73 Supervision X 08/73 3.0 1 3.0 10/12/73 Supervision XI 05/74 2.0 1 2.0 07/11/74 Supervision XII 10/74 3.0 2 6.0 11/27/74 Supervision XIII 11/75 2.0 1 2.0 01/14/76 Supervision XIV 03/76 0.2 1 0.2 05/28/76 Supervision XV 10/77 1.0 2 2.0 12/08/77 Supervision XVI 04/78 0.2 2 0.4 No Report Completion 09/78 2.0 2 4.0 04/79 COUNTRY EXCHANGE RATES Name of Currency (Abbreviation) Tanzania Shilling (TSh) Appraisal Year Average (1968) US$1 = TSh 7.14 Intervening Years Average (1969-1975) US$1 = TSh 7.14 Completion Year Average (1975) US$1 = TSh 7.14 4 - iv - PROJECT PERFORMANCE AUDIT REPORT TANZANIA: SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-TA) HIGHLIGHTS The project assisted the construction of two difficult sections, aggregating 499 km, of the Tanzania to Zambia highway. The goal of providing Zambia with a reliable outlet to the sea was effectively achieved. However, appraisal expectations that the project would be justified by benefits accru- ing to Tanzania alone were not fully met. The project's rate of return to Tanzania was only 7% (PPAM, para. 10)..!/ When benefits accruing to Zambia are included the rate of return of the project increases to an acceptable 16%..2/ This is because, contrary to appraisal expectations, Zambia traffic did not abandon the highway when the TAZARA railway was opened in 1975 (PPAM, para. 11). An important issue on distribution of benefits between benefitting countries is raised by the project. At appraisal it was expected that most benefits (85%) would accrue to Tanzania; on completion it has become evident that an important share of the benefits (45%) has accrued to Zambia (PPAM, para. 13). The PCR suggests that the issue should be reviewed and that Tanzania might consider recouping part of the benefits from Zambia. The audit supports the suggestion,, although it is clear that the reason for the unex- pected distribution of benefits is the failure of Tanzanian traffic to mater- ialize as expected (PPAM, para. 10). Implementation of the project highlights the handling of uncertainty about traffic volumes (PPAM, para. 3). Against the wishes of the Government and the advice of consultants, the Bank initially recommended using a thin asphalt concrete surfacing, risking that faster traffic growth would require earlier than expected strengthening of the road and also risking that the road might fail. In the opinion of the Bank uncertainties about future traffic volumes made it worth running those risks and postponing the additional cost of a stronger surfacing. As a result of protracted discussions between the Government and the Bank, the road was built with a considerably thicker surfacing than initially suggested by the Bank and only slightly thinner than proposed by the consultant. In the event the pavement as built lasted just about as long as the Bank had predicted the pavement with the thinner surfac- ing would last (PPAM, para. 2). However, this was one of the first projects in which risk analysis was applied in the Bank and the PCR (paras. 5.30 to 5.34) contains an interesting discussion of its application, limitations and means to improve the use of the technique. 1/ Based on an estimate of vehicle operating cost savings similar to that used at appraisal (PCR, para. 5.01). An alternative estimate used in the PCR (para. 5.19 et, seq.) produces a rate of return of 14% (PPAM, para. 9). 2/ 24% using the PCR's alternative vehicle operating cost savings estimate (PPAM, para. 9).  - v - The risk analysis concentrated mostly on risks associated with the economic aspects of the project and not those related to the engineering and administrative components of the problem. For instance, no account was taken of the quality of construction, or of the quality of the soil tests used, both of which have an overriding effect on pavement durability. Similarly, the very thin asphaltic concrete surfacing suggested by the Bank was prone to fail early because of poor adhesion to the base, another aspect to which the risk analysis made no reference. Further, no reference was made to the amount, quality and timeliness of maintenance operations or of the effectiveness of vehicle weight controls. There is, however, some question whether the thin surfacing suggested by the Bank was in fact, as understood by the Government and the consultants, asphaltic concrete or some other treatment which may have been less susceptible to failure when used in thin layers (PPAM, paras. 2 and 3). A recent study by the University of Dar Es Salaam and a visual inspection of the road by OED staff confirm that the road needs reinforcing and strengthening soon. To delay the works risks facing a considerably more deteriorated highway on account of its use by overloaded trucks (PPAM, para. 7). Loan/Credit covenants to the effect that axle weight limitations would be strictly enforced were noc complied with, and the aging of the road is mostly due to the overloading of the few trucks that use it (PPAM, para. 6). The audit highlights the fact that the project was expeditiously and well carried out, although it was economically successful only because of fortuitous circumstances (PPAM, para. 15).  - 1 - PROJECT PERFORMANCE AUDIT MEMORANDUM TANZANIA: SECOND HIGHWAY PROJECT (LOAN 586-TA/CREDIT 142-TA) The Project 1. Construction of an improved road link between Tanzania and Zambia was a long-term goal ol the two countries which was reaffirmed when Zambia gained independence in 1964. This long-term aim was turned into an imme- diate objective, however-, by Rhodesia's unilateral declaration of independence in 1965. A series of United Nations resolutions aimed at Rhodesia in 1965, 1966 and 1968 had the effect of placing on landlocked Zambia a successively tighter embargo on shipping through Rhodesia. In an effort to help solve Zambia's transportation dependency problem, the Government of the U.K., USAID and the Bank agreed to help reconstruct the Dar Es Salaam (Tanzania) to Kapiri Mposhi (Zambia) road as follows: the Zambia portions of the highway were to be assisted by a U.K. grant and two Bank loans, and reconstruction of the Tan- zania portions by a USAID and a Bank loan and an IDA credit. The Tanzania portion of the highway (859 km) was divided into four sections: Dar Es Salaam to Morogoro (167 km) supported by USAID, Morogoro to Mahenge (227 km) and Mahenge to Iyayi (272 km) assisted by the Bank and SIDA, and the Iyayi - Tunduma section (193 km) also supported by USAID. The Bank project, conse- quently, assisted the construction of two important, perhaps the most diffi- cult, Tanzanian sections of the highway. The project was expected to generate significant benefits to Zambia but was justified on the basis of vehicle operating and maintenance cost savings accruing to Tanzania alone. There were features of the project, particularly the design standard of certain segments, that lacked full economic justification based on Tanzanian traffic alone. These segments would have been designed to a lower standard had it not been for the Zambia traffic that was expected to use the highway. However, these instances were few and marginal, and consequently accepted by the Bank. Pavement Design 2. Two related issues made it particularly difficult to properly judge the appropriateness of the design standard: one was the uncertainty surround- ing the volume and duration of the Zambian traffic that would use the highway, the second was the design standards proposed by the consultants, in particular the surface thickness. The consultants argued in favor of a 2 to 2-1/2 inch asphalt concrete surfacing to accomodate projected traffic growth without performing major works on the road for 15 years. The Bank, on the other hand, asked the consultant to examine the feasibility of using staged construction with a 3/4 inch thick surface applied initially, to be strengthened by adding a 1-1/2 inch overlay afi:er about 10 years (PCR, para. 2.05). It was unfortu- nate that the presumption was made, by the consultants as well as the Govern- ment, that the 3/4 inch surfacing would be asphaltic concrete. There are bituminous surfacings, for instance surface dressing, which perform well as - 2 - thin layers. But as is known from experience there are grave risks of failure by inadequate adhesion of pre-mixed asphaltic concrete less than 2 inches thick when applied in new construction..! After protracted discussions - involving the Bank, the Government, the consultants and the USAID which was assisting in the finance of the contiguous sections - a compromise was reached on a pavement design that included a 1-1/2 inch asphalt concrete surface, a 4 to 5 inch crushed stone base and a 4 inch sub-base, except on the Mahenge- Iyayi section where no sub-base was provided in order to save in costs (PCR, para. 3.27). However, after failures occurred on the portions of the section opened early to traffic a sub-base was also added to the design of the re- mainder (PCR, para. 3.13). 3. The Bank suggested staged construction approach was based on esti- mates of the likely economic return of different designs under different scenarios, and involved the risk that if traffic increased faster than indica- ted by the best estimate, the road would require strengthening in less than ten years. The suggested design gave clear recognition, however, to uncer- tainty, one of the first occasions in which this was done in the Bank. Uncer- tainty was handled by using risk analysis in assessing the likely rate of return of the project. However, as the methodology was applied, possibly for lack of data, it did not take into account the effect of physical risks such as workmanship falling below normally accepted standards, failure to control vehicle axle weights, maintenance being delayed or not carried out at all, or estimates of soil strength being too high. The PCR (paras. 5.30 to 5.34) contains an interesting discussion on the use and limitations of the tech- nique. Analyses of the type used at the time are seldom used nowadays, and have been replaced in most appraisals by sensitivity tests. 4. Soon after completed portions of the highway were opened to traffic signs of early failure became evident in some places (25 km of the Morogoro- Mahenge section and 6 km of the Mahenge-Iyayi section) (PCR, paras. 3.01 to 3.15). Although these failures were initially blamed on the thin surfacing, subsequent investigations have shown that major contributory factors were (a) inadequate control of the stability of the asphaltic mix on the Morogoro- Mahenge sections, and (b) an inadequate appreciation (by the consultant and Bank engineers) of the effect of moisture on the strength of the subgrade on the Mahenge-lyayi section, which in turn determined that no sub-base was included in the initial design and construction of it. A sub-base was subse- quently added to the unfinished portions of the section. 5. All areas on which failures occurred were repaired, and the two sections of the highway were completed, within cost estimates, by mid 1972. For all practical purposes, therefore, they are now about ten years old, the time estimated they would last before requiring strengthening. During this time the two road sections have held up remarkably well despite the neglect of proper and timely maintenance. More recent research (particularly that of the U.K.'s Transport and Road Research Laboratory) has shown that the design was 1/ Eastern Africa Region staff disagree. In their view, "asphaltic concrete surfacings of 1-1/2 inch thickness have been laid elsewhere with success using appropriate design mix and quality control." - 3 - probably much more sensitive to construction techniques and procedures than to the thickness of the pavement itself. Quality of workmanship by the contractors on both sections was very good, possibly the main reason for the two Bank-assisted sections having held up well. In contrast, the two USAID- supported sections, although built to the same standard, failed much sooner and are now in poor shape for most of their lengths. 6. The two project sections (as well as the rest of the highway) have now reached the end of their design life, not in terms of the number of vehicles that have used zhem, but in terms of the number of equivalent stan- dard axles that have gone over them. An independent study done at the Univer- sity of Dar Es Salaam has concluded that about 825,000 equivalent standard axles have used the road while it was designed for 675,000. This is mainly due to the overloading of about 80% of all loaded trucks (in turn, about 50% of the total number of trucks) that move on the highway, and the exponential damage effect imposed by such overloading. For instance, one axle with double the stapdard weight (8.2 tons) is equivalent to roughly 20 standard weight axles.-' Loan covenants to the effect that vehicle weights would be strict- ly controlled have obviously not been enforced, despite the fact that the main users of the highway are the Government controlled National Transport Company and the Zambia-Tanzania Road Services of which Tanzania owns 35%. 7. A visual inspection of the highway by OED staff showed that both sections exhibit localized surface damage and failures consisting mainly of potholes, surface cracking, edge ravelling, longitudinal and transverse rutting and excessive depressions in some places. The road should, therefore, be resurfaced and strengthened in the near future even though substantial portions of the two sections are still in fairly good condition with only minor rutting and surface cracking. A rapid start on this work is essential and should not be delayed longer than necessary to carry out deterioration measurements. To delay the resurfacing and strengthening would risk having to deal with a significantly more deteriorated road, the reconstruction of which would require major expenditures. In designing the strengthening overlay it would be most appropriate to take into account the lax vehicle weight enforcement in Tanzania.3 Economic Returns 8. The PCR contains a very thorough economic analysis of the project with which the audit fully concurs (PCR, paras. 5.01 to 5.34). It includes a detailed review of the risk analysis method used in the appraisal of this 1/ TRRL, Laboratory Report No. 763, Table 4. 2/ Eastern Africa Regional Office staff have indicated that "a Bank mission in November 1980 discussed and agreed with the Government on the inclu- sion of the rehabilitation/strengthening of parts of the TanZam highway in the proposed Sixth Highway Project and consultants for carrying out the studies have already been selected." 3/ A view supported by SIDA, see Annex. project and explores ways in which it could be improved. However, due to the paucity of data, the PCR, which was completed in 1979, uses an average of 1972 to 1975/76 traffic counts as the most reliable estimate of the 1972 actual traffic volume (PCR, para. 5.08). As far as the audit was able to determine, the reason is that no systematic traffic counting has been done in Tanzania since 1974 when the Ministry of Communications, Transport and Labor was split into the Ministry of Works and the Ministry of Communications and Transport. In addition, information predating the split was lost or has been misplaced. This situation has occurred despite Bank staff consistently and systematically emphasizing the importance of collecting this basic information throughout the implementation of this and three subsequent projects (1971, 1974, 1978). 9. When using vehicle operating costs as estimated by consultants in 1971, which were significantly higher than those used in the appraisal (1969) and which were subsequently used in later project appraisals, the PCR con- cluded (paras. 5.19 to 5.24) that the project had a rate of return of 14.3% without including benefits accruing to Zambian traffic and 23.7% if Zambian benefits were included. 10. However, the PCR's analysis demonstrates that, using vehicle operat- ing cost and maintenance cost savings as in the appraisal (properly updated and measured in comparable real terms), the project has a rate of return of only 6.8%. This is less than half the rate of return expected at appraisal (15%) and occurs despite lower than anticipated costs (PCR, Table 2). The main reason for the difference vis-a-vis the appraisal is the appraisal's overly optimistic forecast of Tanzanian traffic. Tanzanian traffic grew at less than half the rate assumed in the appraisal and there is no evidence that construction of the highway has generated any new traffic (PCR, Table 6). This misestimation was recognized by Bank staff soon after the highway opened to traffic (1972), and much lower rates of growth and no generated traffic were considered in subsequent projects (1974, 1978). However, although diffi- cult to assess because of the lack of systematic counts, from a study on axle weights done by the University of Dar Es Salaam in 1979, it would appear that slightly higher growth rates than the 3% assumed in the PCR have occurred. If this represents the true trend and is not just the result of the seasonal- ity of the University's traffic count, the rate of return of the project could be expected to be slightly higher than the 6.8% estimated in the PCR. 11. When benefits accruing to Zambian transit traffic are included, the project's rate of return increases to 15.9% (PCR, para. 5.01). The reason for this significant increase in the return of the project is that while at appraisal it was assumed that Zambian traffic would eventually taper off and disappear as the TAZARA railway became operational, in actual fact it has not, although the railway has been in operation for a number of years. Thus, while according to the PCR, long-run Tanzanian benefits of the project have been about 60% below expectations, Zambian benefits have been about 130% above - 5 - expectations.!L The project has, therefore, failed to realize its objective of being fully justified on the basis of Tanzanian benefits alone (PCR, para. 1.05) but has fully achieved its objective of providing Zambia with a reliable access to the sea and generating substantial benefits for it. 12. It is fortunate from the point of view of the project that, contrary to expectations, Zambia traffic has not shifted to the railway. It is only because of this fortuitous event that the project has a good rate of return. The audit was not able to establish, however, whether this is an optimal long-term transport solution for Zambia. A shift of Zambian traffic to the railway as originally foreseen would decrease the rate of return of the project to a level close to the 6.8% estimated without the Zambian traffic. Recouping Benefits to Transit Traffic 13. The share of benefits between Tanzania and Zambia did not conform with appraisal expectations, instead of splitting roughly 85% for Tanzania and 15% for Zambia they have split about 55% for Tanzania and 45% for Zambia. Given this situation, the PCR raises the interesting question for Tanzania to consider recouping some of this benefit from Zambia, a question that was not raised at appraisal, quite possibly because at the time the estimated share of benefits acccuing to Zambia was considered rather modest. The audit agrees that the issue raised is important and more widely applicable to a number of cases where one country provides transit facilities to others. The equitable sharing of transport cost savings in these cases should be investigated in more depth. Use of Loan Funds 14. Implementation of this project again raises the question of the appropriate time to close a loan/credit account and evaluate the project. In this case, the physical works of the project were completed in 1972, within a year of the schedule foreseen at appraisal. Payments for settlement of claims related to the construction were completed within two more years. However, another five more years elasped before the remaining US$1.9 million (less than 10% of the loan/credit amount) was disbursed, to finance cost overruns of the Third Highway Project, a purpose quite different from that of the original projeci: and with quite a separate implementation schedule. The question is raised whether such use of the funds promotes effective management of Bank Group projects and accords an effective accountability for funds and projects. 1/ It is possible that some of these benefits are recaptured by Tanzania in the form of fuel taxes and other taxes and fees and also, of course, as part owner of Zambia-Tanzania Road Services, one of the principal users of the highway. -6- Conclusions 15. As an incidental consequence of these delays, evaluation of the original project was delayed long after the physical works were completed and long after evaluations are typically undertaken for Bank Group projects. This delay has allowed carrying out the evaluation with nearly a decade of actual experience in the use of the road. Yet, for lack of simple traffic counts it was not possible to measure the impact of the project, and instead the impact had to be estimated on the basis of projections. It is clear, however, that the project had a fundamentally different impact from what was foreseen at appraisal. Tanzanian traffic was less than forecast, reflecting the project's reduced development impact as well as general Tanzanian conditions; Zambian traffic continued at high levels for longer than expected, reflecting a misjudged diversion of traffic to the TAZARA railway and thus a misplaced traffic allocation at appraisal, a detrimental development in an economy short of resources and foreign exchange; and, axle loads were greater than pro- jected, prematurely aging the road and reflecting unrealistic design param- aters, as well as poor law-enforcement -and lack of adherence to Government regulations mostly by the para-statal trucking organizations. The fact that the economic rate of return of the project is comparable to appraisal expecta- tions (when Zambian traffic is included) reflects the concurrence of several factors which have been significantly influenced by chance. -7- ANNEX Datum/Date Dnr/Reg No 1982 04 26 $S0 FIN 11711 9tk VTIfb Bdagor/Encd Industry Division Lars B Lagging/br The World Bank Operations Evaluation Department Att: Mr Shiv S Kapur 1818 H Street, N.W., Washington DC 20433 USA Ref SIDA ref 1.34 Arende/Re 1.TAN.32.8 Project Performance Audit Report Tanzania: Second Highway Project (Loan 586/Credit 142-TA) This is ir reply to your March 29, 1982 letter addressed to the Director General of SIDA, which has been referred to this office for reply. The captioned draft report has been reviewed by our staff who are familiar with the project and I am pleased to in- form you that they think that the report is of great value thanks to the well documented presentation. We do not have any major comments on the report but we have made one observation which may be of interest. The audit report as well as the project completion report point to the fact of unsatisfactory control of truck axle loads and of no or delayed maintenance measures. In our experience these circumstances are not unique for this project but more or less common in development countries. It gives rise to the question whether design criteria should take into account these and other risk factors and whether they should be included in the economic ana- lysis. In the case of the Second Highway Project, presumably the road users savings have not materialized over the whole lifetime but have decreased gradually as surface damages developed., Yours sincerely Action Head of the Industry Division MY 0 3 INZ 0263 Address Office Telephone Telegram Telex Postgiro S-105 25 STOCKHOLM, SWEDEN Birger Jarlsgatan 61 08- 1501 00 sida stockholm 11450 sida sthlm 1 5634- 9  TANZANIA SECOND HIGHWAY PROJECT LOAN 586/CREDIT 142-TA PROJECT COMPLETION REPORT April 30, 1979 Highway Projects Division Eastern Africa Regional Office u - 9 - SUMMARY AND CONCLUSIONS 1. On February 18, 1969, the Board approved a Second Highway Project for Tanzania, authorizing Bank Loan 586-TA for US$7.0 million and IDA Credit 142-TA for US$8.0 million, which, in conjunction with a Credit from the Swedish International Development Authority (SIDA) for US$15.0 million equivalent, helped finance the construction of about 310 miles (496 km) of the Tanzania- Zambia Highway (the Tan-Zam Highway), including consultants' services for construction supervision. 2. Total project cost was estimated at US$38.0 million, and the original financing plan was as follows: US$ Million Bank 7.0 IDA 8.0 SIDA 15.0 30.0 /1 Government 8.0 /2 Total: 38.0 /1 US$25.8 million foreign exchange, and US$4.2 million local costs. /2 Remaining local costs. 3. The original Loan and Credit Agreements, together with an Adminis- tration Letter authorizing IBRD/IDA to act as executing agency for the Swedish Credit, were signed on February 24, 1969. The date of effectiveness was estab- lished as April 10, 1969, and the closing date as December 31, 1972. 4. The legal documents of February 1969 were amended in November 1969 to reflect changes in financing already anticipated at the time of Board approval. The revised financing plan was as follows: US$ Million Bank 7.0 IDA 15.5 SIDA 7.5 30.0 Government 8.0 Total 38.0 5. Construction contracts were awarded in early 1969 after prequalifi- cation and bidding in accordance with Bank Group guidelines. The works were carried out under two separate contracts: (i) the Morogoro-Mahenge road (142 - 10 - miles) and (ii) the Mahenge-Iyayi road (170 miles) Construction supervision was carried out by the consultants who had previously carried out detailed engineering. 6. The Morogoro-Mahenge road was completed in May 1972, five months behind the appraisal schedule, and the Mahenge-Iyayi road was completed on schedule in December 1971. The delay in completion of the former contract was mainly due to slow mobilization by the contractor and to additional works. The quality of workmanship under both contracts was good. The consultants' performance in supervising the contracts was generally adequate, although their detailed engineering of the two road sections (financed partly under an earlier Bank Group project and partly through bilateral assistance) suffered from several shortcomings, including inaccurate ground levels and unsuitable road alignments. These problems were satisfactorily corrected during con- struction. A more serious problem was the failure of pavements, which to some extent was caused by inadequate design standards resulting from a less than thorough design approach. The consultants attributed the failures partly to reductions in the original pavement standards following discussions with Bank Group staff. However, Bank Group involvement came about only because of concern that the consultants' pavement design approach was not sufficiently thorough and analytical and consequently needed improvement. 7. The total project cost was US$35.2 million compared to the appraisal estimate of US$38.0 million. The rate of disbursement was slow, initially but satisfactory overall. The credit of US$15.5 million was fully disbursed in December 1971, and closed as scheduled on December 31, 1972. However loan disbursements continued until October 1974, over two years after project com- pletion, due to the caution exercised by Government engineers in settling contractors' and consultants' claims owing to Government's charges that contractors on other projects had been overpaid by some Government engineers. These charges were later found to be untrue. 8. In October 1974, US$5.1 million of the loan amount of US$7.0 million had been disbursed. With Board approval in January 1975, the remaining US$1.9 million was reallocated to meet cost overruns on the Third Highway Project (Credit 265-TA). The Closing Date for Loan 586-TA, originally December 31, 1972, was extended several times and finally occurred on December 31, 1978. 9. The economic re-evaluation indicates alternative estimates for the rate of return for the whole project of 16% and 24% if the benefits accruing to both Tanzania and Zambia are taken into account. If benefits to Tanzania alone are considered, the corresponding rates are 7% and 14% or on average some 10% as compared with an estimate of 15% at the time of appraisal. This difference between ex ante and re-estimated rates of return can largely be attributed to a considerable overestimation at the time of appraisal of the future volume of Tanzanian traffic using the road and to a minor extent to an underestimation of the economic cost of construction and supervision. 10. As the appraisal report contained the Bank's first serious attempt to deal with risk analysis in project appraisal, this PCR provides specific - 11 - comments on the subject in the context of this project and makes a few suggestions for future appraisal work for consideration and followup by the OPerations Evaluation Department (OED). 11. The project objectives, to provide Zambia with access to the sea through Tanzania and to improve sections of the main road network in Tanzania, were fully achieved. The project was prepared and executed satisfactorily by the Government with the assistance of consultants. - 12 - I. INTRODUCTION 1.01 The Bank Group began assisting the Government of Tanzania to develop its highway network in 1964 with the First Highway Project (Cr. 48-TA and Supplementary Cr. 115-TA; US$16.9 million and US$3.0 million, respectively) which comprised detailed engineering and construction of eight main road sections, detailed engineering of feeder roads in the Geita Peninsula, and a 5-year program of staffing and training for the Roads and Aerodromes Division of the Ministry of Communications, Labor and Works. 1.02 The First Highway Project originally included construction of the Morogoro-Iringa road (187 miles), but following Rhodesia's Unilateral Declaration of Independence (UDI) and the resultant UN-imposed sanctions which halted Zambian traffic through that country, construction of the road was dropped from the project pending the results of a study financed by the United States Agency for International Development (USAID) (para 2.01) to find an alternative route to the sea through Tanzania (Map 1). 1.03 The Second Highway Project arose from discussions in 1966 involving the Governments of Tanzania and Zambia, the USAID, and the Bank Group follow- ing UDI. The project was developed to improve two sections of the Tan-Zam Highway, 1/ namely, Morogoro-Mahenge (142 miles) and Mahenge-Iyayi (170 miles). The other sections in Tanzania, Iyayi-Tunduma (147 miles) and Morogoro-Dar es Salaam (120 miles) were slated for construction with assis tance from USAID. Broadly speaking, he project objectives were twofold: to improve part of Tanzania's main road system and to develope an adequate road link to the sea through Tanzania for its landlocked neighbor, Zambia. 1.04 Even before Rhodesia's UDI, Zambia had seen the need to reduce her dependence on Rhodesia for access to the sea, and regarded Tanzania as a natural partner in her plans to accomplish this. After UDI, when Zambia diverted its import and export traffic from the southern route through Rhodesia and Mozambique, the Tan-Zam Highway became its primary alternative route to the coast. Moreover, in view of the political uncertainty elsewhere, adequate upgrading and maintenance of the Tan-Zam Highway would provided Zambia with an alternative route for traffic using other insecure transport links to the sea, such as the Benguela railway to Lobito (Angola) and Zaire's rail-river route to Matadi. 1.05 In terms of economic justification, the project was envisaged of as part of an international road link which would generate substantial bene- fits for Zambia but which would also be justified for Tanzania alone. The appraisal calculations indicated a rate of return of 15% for Tanzanian traffic only, and an overall rate of return of 20% for combined Tanzanian and Zambian traffic. It should be noted, however, that the selected design and construction standards could be justified only on the basis of both countries' traffic. 1/ The Tan-Zam Highway is the main road connection between Tanzania and Zambia linking Dar es Salaam with Lusaka. - 13 - 1.06 A formal request for Bank assistance was made by the Government of Tanzania in January 1968, and the project was appraised in April 1968 (see Appraisal Report TO-690a of January 15, 1969). It was approved by the Board in February 1969, for US$15 million (Bank Loan 586-TA for US$7 million, and IDA Credit 142-TA for US$8 million), with an additional US$15 million of financing from SIDA. The Loan and Credit Agreements were signed a week afterwards and became effective in April 1969. 1.07 The Swedish credit of US$15.0 million equivalent was provided on the understanding that, once the Second IDA Replenishment became effective, Tanzania would apply to the Association for a supplemental credit of US$7.5 million and the equivalent proportion (50%) of the Swedish credit would be cancelled. The Government of Tanzania formally requested the supplemental credit in September 1969, and Amendments to the Loan and Credit Agreements and the Administration Letter were signed on November 24, 1969. They were further amended in early 1975 to reallocate a surplus of US$1.9 million from Loan 586-TA towards cosi: overruns on the Third Highway Project (Credit 265-TA). 1.08 The sources of information on which the PCR is based are listed in Annex A. II. PROJECT PREPARATION AND APPRAISAL Feasibility Study 2.01 The two project road sections (Map ) are part of a highway link between Tanzania and Zambia studied by consultants in 1966 and recommended by them for improvement to two-lane paved standard. The study was financed by USAID, and its objectives were (i) to determine the optimum route to connect the two countries, (ii) to provide Zambia with access to the sea through neighboring Tanzania, and (iii) to study the economic development potential made possible by the construction of such a highway. Detailed Engineering 2.02 In early 1966, as part of the First Highway Project, consultants completed detailed engineering of the Morogoro-Iringa road (187 miles), of which the Morogoro-Mahenge section (142 miles) is a part. At that time, Zambia depended largely on the Rhodesia - Mozambique route to the sea, but when strict UN sanctions were imposed on Rhodesia following UDI in November 1965, Zambian traffic was suddenly diverted to Tanzanian routes, which sharply increased truck traffic on the Tan-Zam Highway. 2.03 The proposed reconstruction of the Morogoro-Iringa road, also under the First Highway Project, was then cancelled because of time needed to review design standards in the light of the increased Zambian traffic and to reduce the scope of that project to avoid cost overruns. Instead, the Association - 14 - agreed that the same consultants should redesign the Morogoro-Mahenge road section based on the findings of the USAID-financed feasibility study (para. 2.01). The redesign was carried out in two phases: Phase I consisted of (i) detailed engineering to estimate quantities and costs for three alternative design standards, and (ii) a traffic evaluation which distinguished between local traffic and Zambian copper and oil traffic. Phase II consisted of final design and documentation on the basis of design standards agreed upon by Government and the Bank Group. Although the designs were accepted by Govern- ment and the Bank Group as generally satisfactory, some changes were later needed (paras. 3.06, 3.21-22) and during construction questions arose as to the adequacy of pavement design (paras. 2.05-2.06, 3.07, 3.23-3.24, 6.01-6.02). 2.04 Detailed engineering of the Mahenge-Iyayi road, financed by USAID, was carried out by consultants under terms of reference similar to those for the other road section. As with the Morogoro-Mahenge sections, the designs were accepted as generally satisfactory but subsequently needed amendment (paras. 3.25-3.26), and the adequacy of the pavement design was also later questioned (paras. 3.11-3.13, 3.27-3.28). Appraisal 2.05 The Bank Group appraisal in April 1968 took place before the detailed engineering of the two roads was completed, and during appraisal much of the technical discussion centered on pavement design. Government and the consul- tants wanted a surface course of 2 to 2-1/2 inches of asphaltic concrete since they believed that it would be impractical to use anything less and since this would conform with the design used for the Iyayi-Tunduma road section being considered for financing by USAID. However, after researching the topic, Bank Group staff questioned the Government's and the consultants' stand and requested the consultants to investigate the feasibility of constructing the surface course in stages, with a minimum thin layer (3/4 inch) asphaltic concrete course applied initially and a 1-1/2 inch overlay after about 10 years. Following further analyses and discussions, a compromise was reached to provide a 1-1/2 inch thick asphaltic concrete surface course. Reports and correspondence in Bank Group project files indicate that the final decision was made on the basis of recommendations by the consultants, and constituted a compromise which the Bank accepted. 2.06 In addition, the Bank staff raised questions about traffic and loading assumptions, pavement design methodology, and the thickness of the base and sub-base, which prompted the consultants to revise the pavement design and to adopt a more thorough and analytical approach than they had previously used (para 6.02). Negotiations and Approval 2.07 Detailed engineering for both project road sections was completed, and bids for the civil works were called before negotiations. The project was negotiated in November 1968 without any significant changes, and approved by the Board in February 1969. No major issues were raised by the Bank and the Government, and Government had no serious reservations on the Credit and Loan covenants. Shortly after negotiations, bids were opened and found in line with the engineers' cost estimates. - 15 - 2.08 The project was expedited as a result of the emergency created by UN-imposed sanctions following UDI and, therefore, was limited to the improve- ment of those parts of the Tan-Zam Highway located in Tanzania. No institu- tional development objectives were envisaged under the project (para 4.01). Project Description 2.09 The overall scope of the project was not modified during implemen- tation. Project revisions were of a technical nature and consisted of alignment changes, the construction of a bypass, and pavement modifications (paras 3.05-3.07). The project as completed comprised the construction to two-lane paved standard of two sections of the Tan-Zam Highway (Morogoro-Mahenge and Mahenge-Iyayi) and consulting services for construction supervision. III. PROJECT IMPLEMENTATION AND COST A. Civil Works 3.01 The civil works comprised the reconstruction to two-lane asphalt standard of two sections (about 310 miles) of the Tan-Zam Highway. The works were completed satisfactorily in May 1972, about five months later than the appraisal schedule (para 3.08 and Table 1). The total construction cost was US$33.9 million, compared to the original contract amounts totalling US$30.8 million and the appraisal estimate of US$31.5 million excluding contingencies. The increase was mainly for additional works, and contingencies provided in the appraisal estimate were more than adequate to cover this (Table 2). 3.02 The works were carried out under two separate contracts by different contractors: (i) the Morogoro-Mahenge road (142 miles) and (ii) the Mahenge- Iyayi road (170 miles). The main problems during construction related to inaccurate ground levels, changes in route alignment, and pavement failures of completed sections (paras 3.05-3.07 and 3.11-3.13). (i) Morogoro-Mahenge Bidding and Award 3.03 The road was divided into two contracts and bidding was allowed for one or both contracts. Bids were invited from prequalified contractors in accordance with ICB conditions as required by Bank Group Guidelines for Procurement. Bids were opened in January 1969, and in April 1969 both con- tracts were awarded to the lowest bidder and administered as a single contract for TSh 119 million (US$16.7 million) (Tables 1 and 2). - 16 - Start-up of Works 3.04 The works were scheduled to commence in May 1969, about a month after contract award as stipulated in the contract. This period was inade- quate for mobilization, which was not completed until September 1969 when the execution of permanent works began. The start-up of works was delayed mainly because of initial inadequate management by the contractor and late deliveries of equipment (para. 3.18). Project Revisions 3.05 During construction, the Government proposed that a bypass be built around the town of Morogoro rather than improving the existing road through the town center. The Bank's economic evaluation, based on the consultants' engineering proposal and cost estimates, proved the bypass to be justified because of the amount of through-traffic on the road. 3.06 A further revision was made by redesigning a 15 mile section near Morogoro which eliminated the need for seven bridges and made the section 1.3 miles shorter than the original design. The two revisions together resulted in a total saving of about four miles in the length of the road (para 3.21). 3.07 After the road was completed in 1972, surface failures began to appear, mostly along a 16 mile section in hilly country which had been opened to traffic in late 1970. The failures were due to a separation of the 1-1/2 inch asphalt concrete surfacing from the stabilized gravel base, and occurred at steep gradients and sharp curves. The affected areas were cut out and repaired, and a 1 inch thick asphalt overlay was applied over the entire 16 miles. The additional cost amounted to about TSh 1.9 million (US$0.2 million) (paras 3.23-3.24). Implementation Schedule 3.08 The original contract period of 900 days was extended by 163 days to allow time for additional works, which consisted mainly of increased earthworks involved in the construction of the Morogoro bypass and an align- ment change. With the time extension allowed, the road was scheduled to be completed in February 1972, about two months later than the appraisal schedule of December 31, 1971. However, the road was not completed until May 1972 because of the late start-up of works (Table 1). (ii) Mahenge-Iyayi Bidding and Award 3.09 The road was divided into three contracts and bidding was allowed for one, two or all three contracts. Bids were invited from prequalified contractors under ICB conditions as required by the Bank Group's Guidelines for Procurement. Bids were opened in December 1968, and in January 1969 all three contracts were awarded to the lowest bidder and administered as a single contract in the sum of TSh 101 million (US$14.1 million) (Tables 1 and 2). - 17 - Start-up of Works 3.10 The contract itart-up was quite rapid and problem free. The joint venture contractor had excellent combined resources from the three partici- pating firms, which enabled him to accept and meet a commencement date of only 11 days after contract award, even though the tender documents allowed up to one month for mobilization. Project Revisions 3.11 Early in 1971, some 3 months after traffic was allowed on a completed section of the road (about 70 miles), extensive failures occurred in approximately 4 miles of pavement, and isolated sections in other areas showed signs of breaking up. Temporary repairs were made, followed by per- manent works which consisted of cutting out and repaving the failed parts of the pavement and applying an asphalt overlay on 7 miles of road and a seal coat on the remainder of the completed road. A seal coat was also applied to the shoulders to stop surface water from penetrating into the shoulders and subgrade (paras 3.25-3.26). 3.12 The consultants attributed the failures to excess moisture in the subgrade, vehicle axle overloading, and inadequate pavement design; the original pavement design did not call for a sub-base (paras 3.27-3.28). 3.13 The pavement structure of the rest of the road was modified to incorporate a sub-base of selected granular material, and a seal coat was applied to both shoulders. In addition, steps were taken to stop overloading of vehicle axles, but the enforcement of the regulations governing vehicles loads was not effective (para 4.07). The additional costs of the extra works amounted to about TSh 6.9 million (US$1 million) (paras 3.27 and 3.28). Implementation Schedule 3.14 The original contract period of 27 months was extended by 8 months (from May 1971 to January 1972) mainly to allow for the pavement repairs and other project revisions. Delays due to strikes, late arrival of equipment on site because of port congestion, and inclement weather accounted for part of the time extension granted to the contractor. The road was completed in December 1971 as estimated at appraisal. B. Consulting Services 3.15 The contract for supervision of the Morogoro-Mahenge road section was negotiated with the consultant who had previously undertaken detailed engineering of the road under the First Highway Project (paras 2.02-2.03). A contract for supervision of the Mahenge-Iyayi road section was similarly negotiated. For performance of consultants see paras 3.20-3.28. - 18 - C. Project Cost Estimates and Disbursements Cost Estimates 3.16 Total project cost was US$35.2 million compared to an appraisal estimate, including contingencies, of US$38.0 million, a saving of about 8% (Table 2). Contingencies utilized amounted to about US$2.4 million, compared to a provision of about US$5.3 million in the appraisal estimate. Available data are insufficient to determine how much of this was due to inflation. Actual inflation during the entire construction period is estimated at about 5-1/2% as compared to an appraisal forecast of 10% of the total cost. No figures are available for the actual foreign exchange cost of the project (para 3.30). Disbursements 3.17 There was generally a delay of about three months between the time when expenditures were incurred by Government and the receipt of withdrawal applications by the Bank Group. Disbursements were initially slow (see para 3.18), and only 41% of the appraisal estimate was disbursed at the end of the first nine months. At the end of 21 months, disbursements rose to 66% of the appraisal estimate, and at project completion in 1972 they reached 82% (see Table 8). Disbursements on the project continued until late-1974 due to the caution exercised by Government engineers in settling contractors' and con- sultants' claims following official Government charges (later found untrue) that contractors for other projects had been overpaid by some Government engineers. The Credit Agreement (Cr. 142-TA) was closed as originally sched- uled on December 31, 1972. The closing date of the Loan Agreement (Loan 586-TA) was extended several times and on January 1975 the Board approved the reallocation of undisbursed loan funds (US$1.9 million) to the Third Highway Project (Cr. 265-TA) to meet cost overruns. The closing date for Loan 586-TA was consequently amended to December 31, 1976 and then to December 31, 1978 because of delays in setting the disputes with consultants (para 3.24) and in completion of the Third Highway Project. D. Performance of Contractors Morogoro - Mahenge 3.18 Initially, the contractor had some difficulties in organizing and staffing his management team, contributed to the slow start-up of works. Progress was also affected by the late arrival of some equipment at the site. Both problems were associated with an unexpected delay in the completion of another contract being executed by the same contractor on the Zambian side of the Tan-Zam Highway. In addition, the delivery of some items of equipment was delayed by congestion in Dar es Salaam port. The works were completed in May 1972, three months later than the revised contract period and five months later than the appraisal schedule. The quality of workmanship was good. - 19 - Mahenge - Iyayi 3.19 The performance of the contractor was satisfactory throughout the contract period. There were some signs of lack of experience among equipment operators, but this did not affect progress. The joint venture was responsive to the scheduling changes needed for pavement repairs and pavement design changes. The works were completed in December 1971, within the revised con- tract period and as scheduled at appraisal. The quality of workmanship was good. E. Performance of Consultants Morogoro - Mahenge 3.20 The consultants' performance during construction supervision of the Morogoro-Mahenge road was satisfactory. They provided reliable staff, were able to make important design changes, and incorporated additional works without disturbing the contractor's schedules. 3.21 However, their detailed engineering of the road (financed under the First Highway Project) proved to be inaccurate in its measurement of existing ground levels and failed to identify the new and better alignment for a 15 mile section from Morogoro (para 3.06). The inaccurate ground levels were obtained from aerial surveys in the mountainous regions and were attributed by the consultants to dense vegetation covering the steep areas. It is likely, however, that the errors were due to the use of aerial photographs taken at too high an altitude. 3.22 In order to achieve an economic balance of cut and fill, the ver- tical and horizontal alignments of the mountainous sections (about 25 miles) were redesigned during construction. No delay in construction resulted from these changes, and although no figures are available on changes in earthwork quantities, there seems to have been no significant variation from the original design. 3.23 Government also held that deficiencies in the consultants' design of the road pavement were responsible for the appearance of surface failures after a few months of traffic use (paras 2.05-2.06, 6.02). However, the con- sultants pointed out that surface failures might not have occurred if the legal axle-load limit of 18,000 lbs had been enforced (para 4.06), and that Bank staff had influenced the final decisions or. the thicknesses of the sub-base, base, and surfacing. 3.24 Government also believed that the stability of the asphalt concrete surface course suffered because of the high bitumen content, and suggested that the consultants had not properly controlled the quality of the site mix In reply, the consultants produced test results taken on site to establish that the stability of the mix was not responsible for the failures. The dispute continued for several years, with Government remaining adamant on - 20 - charging the repair costs to the consultants. However, Government recently reported to the Bank that the dispute with the consultants has been settled, but the exact details of the dispute and the terms of the settlement were not disclosed. It is difficult to single out any specific cause for the pavement failures, and it seems reasonable to assume that it was a combination of all the above factors. Mahenge - Iyayi 3.25 The consultants' performance during construction supervision of the Mahenge-lyayi road was satisfactory. However, like the other consultants, their detailed engineering of the road (financed by USAID) showed (i) in- accuracies in measuring existing ground levels which necessitated the redesign of the entire vertical profile of the road during construction, and (ii) in- adequate pavement design, which resulted in pavement failures along a 7-mile section of completed road. The design of the pavement was modified during construction. 3.26 Inaccuracies in the ground levels were due to the use of aerial photographs taken at too high an altitude. The consultants initially planned to fly at 10,000 ft for preliminary design only and to fly at 2,000 ft for the final engineering. However, they later decided to make a single flight at 5,000 ft in order to expedite the work, claiming that in so doing they avoided delaying the project by one year; this, however, is doubtful since the flights and subsequent processing of the information would not have taken much time. 3.27 The consultants attributed the pavement failures to compromises made at the design stage in order to reduce the cost of the pavement, and they pointed out that Bank Group staff had influenced these changes. The pavement was designed as a single-layer base without a sub-base, and constructed in a trench section formed by shoulders and sub-grade of the same material. Fur- thermore, in light of discussions with Bank Group staff, the thickness of the surface course had been reduced from the 2 inch to 2-1/2 inches originally planned to 1-1/2 inches of asphaltic concrete. 3.28 More relevant, however, is the fact that the consultants' design assumed a higher strength sub-grade than the one used. After failures began to occur, the consultants prepared a report on the procedures and criteria used for selecting and testing sub-grade material. These were reviewed by the Government and the Bank Group, who found that the consultants had not taken into account posssible variations in moisture conditions. Accordingly, when the sub-grade became wet with the infiltration through the shoulders of sur- face water, its resulting loss of strength caused it to fail under prevailing traffic conditions and this led to break-up of the whole pavement structure. The consultants did not realize that the sub-grade strength was insufficient until after the road was partly constructed. - 21 - F. Reporting 3.29 Reporting was carried out by each of the consultants responsible for construction supervision. Monthly and quarterly progress reports were submitted regularly and were satisfactory. The consultants were prompt in preparing a report on the pavement failures, together with their recommenda- tions which were accepted by the Bank and Government. 3.30 One consultant also submitted a satisfactory completion report in final form, whereas the other submitted only a brief draft interim completion report. A significant omission in the reports of both consultants was the determination of actual foreign exchange costs, taxes and duties paid by the contractor. Despite requests by the Bank, Government did not submit a project completion report. IV. INSTITUTIONAL DEVELOPMENT AND PERFORMANCE OF BORROWER Institutional Development 4.01 The objectives of the project were limited to improving the Tan-Zam Highway (para 2.08). Institutional development objectives were included in subsequent highway projects (Credits 265-TA in 1971, 507-TA in 1974 and 876-TA approved by the Board in January 1979). Project Performance 4.02 Despite shortcomings in organization, staffing and management, the performance of the Borrower in the execution of the project was satisfactory. Significant delays occurred only in the settlement of claims from contractors and consultants (paras 3.17 and 3.24). The satisfactory performance was clearly due to the employment of consultants to supervise construction and good contractors to execute the roadworks. Covenants 4.03 The Borrower complied satisfactorily with all covenants except for those relating to maintenance of the highway network and to control of axle loads. (Section 4.06 of the Credit Agreement and Section 5.07 of the Loan Agreement). 4.04 Recurrent expenditures on trunk roads and bridges actually dropped during the execution of the project from TSh 73.7 million in 1970/71 to TSh 45.1 million in 1971/72. Since then, however, expenditures more than doubled to TSh 107.7 million in 1973/74, increased to TSh 111.8 million in 1975/76, and have since increased further to reasonable levels. The alloca- tion for 1977/78 was TSh 220.6 million, but the funds were not fully utilized because of the country'S low capacity and capability to carry out maintenance works. - 22 - 4.05 Under the Fourth Highway Project (Cr. 507-TA), the Association is providing assistance to increase capacity and capability for maintenance of the trunk roads in ten regions. Maintenance is improving, and the recently approved Fifth Highway Project will extend this work to the other regions of the country. It will, however, be some time before the entire trunk road system is well maintained. 4.06 Vehicle axle overloading, particularly from Zambian traffic, was a major problem on the project road during construction, and the supervising consultants attributed damage which occurred soon after opening of the road sections to axle loads in excess of the design load of 18,000 lbs. One type of truck using the road carried up to 13,000 lbs in excess of the legal limit; the consultants estimated that "five repetitions a day of a 30,000 pound axle load could destroy the pavement in six months". 4.07 The control of truck axle-loads was unsatisfactory because of (i) a lack of weighing equipment and trained staff, and (ii) ineffective regulations. Weighbridges have been installed at three locations on the Tan-Zam Highway and a program is now underway to install weighbridges at ten other locations around the country. Furthermore, the 1973 Road Traffic Act makes adequate provision for controlling and enforcing axle loads as well as the gross weight of a vehicle (including trailer where applicable). The Government began enforcing the regulations on the Tan-Zam Highway in late-1976, but enforcement still suffers somewhat from lack of cooperation from the police who are responsible fur arrebtiig and prosecuting offenders. 4.08 Vehicle operators on the Tan-Zam Highway, including those from Zambia, are now pressing the Government to relax enforcement of the regula- tions, particularly those relating to gross vehicle weights. The Government has retained consultants to carry out a study to assess the feasibility of this change. V. ECONOMIC RE-EVALUATION Summary of Results 5.01 The following table summarizes by road section the rates of return of the project (i) according to the appraisal report; (ii) as recalculated on the basis of the detailed information provided in the appraisal report (A Estimate); and (iii) based on a re-evaluation of the project with the benefit of current information and foresight (two alternatives: B and C Estimates). The rates of return including Zambian transit traffic reflect the economic return of the project regardless of which country reaps the benefits. The rates of return excluding Zambian transit traffic ignore the benefits accruing to Zambia and reflect the economic return of the project to Tanzania. - 23 - Estimated Rates of Return Project Road Sections (%) Incl. Zambian Transit Traffic Excl. Zambian Transit Traffic Appraisal A B C Appraisal A B C Road Section Mitimate Est. Est. Est. Estimate Est. Est. Est. 1. Morogoro - 38 miles West 22.6 19.2 10.4 18.5 19.1 17.4 5.3 10.8 2. Morogoro 38 miles-Mikumi 23.0 21.5 6.9 13.3 18.2 18.8 2.8 10.8 3. Mikumi-Mahenge 20.1 18.0 12.6 23.4 14.4 15.1 5.6 13.0 4. Kitonga Gorge 11.8 13.3 5.9 12.9 9.7 12.0 1.4 7.0 5. Mahenge-Iringa 13.8 15.0 7.7 15.1 11.2 13.4 2.8 8.5 6. Iringa-Sao Hill 18.9 22.1 16.8 28.2 14.1 18.4 10.0 17.8 7. Sao Hill-Makumbako 23.6 24.7 22.9 37.9 16.3 19.0 13.6 23.7 S. Makumbako-Iyayi 22.7 24.6 24.3 40.0 13.9 17.8 13.1 22.8 All Road Sections Simple Average 19.5 .19.0 15.9 23.7 14.6 16.5 6.8 14.3 A Estimate: RecalculatLd rate of return based on appraisal report information. B Estimate: Re-evaluated rate of return accepting unit vehicle operating cost savings as estimated in the appraisal report. C Estimate: Alternative re-evaluated rate of return including alternative unit vehicle operating cost savings. The A Estimate 5.02 The A Estimate is the rate of return recalculated on the basis of information provided in the appraisal report with minor assumptions made in cases where the report did not provide sufficient information. This recal- culation serves three purposes: (i) to confirm that the original calculation was correct; (ii) to ensure that in the re-evaluation the comparison is between like and like; and (iii) to provide a more detailed basis for analysis going beyond a simple comparison of rates of return (see Table 3, "Detailed Comparison of Different Estimates of Benefits and Costs"). - 24 - 5.03 As para 5.01 shows, although there are some differences, the A Estimates of the rate of return reflect the same orders of magnitude as the rates calculated in the appraisal report. The minor differences are probably caused by the few assumptions which had to be made in the absence of full information in the appraisal report. In the following, the A Estimate will be used as the basis for comparison with the re-evaluated rates of return: the B and C Estimates. The B Estimate 5.04 The B Estimate differs from the A Estimate, as follows: Item B Estimate A Estimate (i) Construction and supervision costs Actual costs Appraisal estimate (ii) Road length after construction Actual length Appraisal estimate (iii) Tanzanian traffic using the Actual 1972 Actual 1967 data plus project road data based on 6% (cars) and 8% traffic counts (buses and trucks) plus 3% annual annual growth; growth for all generated traffic traffic; no assuming demand generated elasticity of 1.0 traffic (iv) Zambian traffic using the project 75 thirty-ton 75 thirty-ton truck road truck trailers trailers per day per day during until 1975 and 50% life of project annual reduction thereafter (v) Unit vehicle operating cost As in appraisal As in appraisal savings (to be modified in the C Esti- mate) (vi) Road maintenance cost savings Based on Based on formula formula as in as in appraisal appraisal In the following (paras 5.05-5.12), each of the above items is briefly discussed. 5.05 With regard to construction and supervision costs, the appraisal report's "best estimate" of the total economic cost of the project is TSh 203 million (US$28.5 million). This is 18% lower than the appraisal estimate of financial costs, including physical but excluding price contingencies of TSh 247 million (US$34.6 million). The appraisal report allows "about 5 to 8% of the total costs for duties and taxes on fuel, equipment and company - 25 - taxes" and excludes the 5% physical contingency allowance, thus leaving about 5 to 8 percentage points of the 18% difference between financial and economic costs for other "adjustments by the (appraisal) mission". 5.06 In the re-evaluation, the total economic costs were estimated at TSh 244 million (US$34.1 million), compared with the actual total financial cost of TSh 251 million :US$35.2 million). This allows for a 3% component for duties, taxes, etc. Actual economic cost, thus, was about TSh 41 million (US$5.7 million) or 20% higher than the corresponding appraisal estimate. 5.07 With regard to road length after construction, the Morogoro - 38 miles West section was shortened by 4 miles compared wtih the appraisal estimate (see para 3.06). For all other road sections the actual length after construction was as estimated at the time of appraisal. 5.08 With regard to Tanzanian traffic using the project road actual traffic counts were collected from the Ministry of Communications and Transport, by road section, for the period 1972 to 1975 or 1976. Unfor- tunately, more up-to-date data are not available. To provide a sound basis for determining the 1972 actual traffic volume and reduce possible error, the available traffic count information for the 3 to 4 years was averaged by road section. These averages were used as the most reliable indication of the 1972 actual traffic volume using the project road. Traffic count data collected from various sources covering the period 1968-1971 provide a very similar result (see Table 4). The data were later used as a basis for the appraisal of the Fourth Highway Project (1974). 5.09 The annual growth rate of traffic on the project road from 1972 on is estimated at 3% as compared to the appraisal estimate of 6% for cars and 8% for trucks and buses. The 3% rate is identical to the growth rate assumed for the Tan-Zam Highway traffic in the Fourth Highway Project (1974) and to the growth rate assumed for traffic on the trunk road system in the northern half of the country in the Fifth Highway Project (1978). Moreover, neither of these projects assumed additional traffic generation and the present re-evaluation is also based on the assumption that no traffic generation has taken or will take place on the project road. 5.10 With regard to Zambian traffic using the project road, although the TAZARA railway had commenced operation after the project was completed, Zambian road traffic is likely to continue at a level of about 75 thirty-ton truck trailers a day. A recent Bank mission studied the Tan-Zam corridor transport problems in detail and concurs with this assumption. 5.11 With regard to unit vehicle operating cost savings, estimates of this kind are notoriously uncertain and contain a large margin of error and subjective judgment. Because of this and because of the relative importance of this type of benefit, the unit operating cost savings as estimated in the appraisal report were maintained in the B Estimate; the impact of alternative unit operating cost savings is analyzed separately in the C Estimate. - 26 - ,.12 With regard to road maintenance cost savings, the road maintenance cost formula used in the appraisal 1/ was checked against data in the 1971 Economic and Engineering Study-Tanzania Highway and found to yield estimates of the right order of magnitude. These estimates were also found reasonable when compared with other scattered information. The road main- tenance cost formula used in the appraisal was therefore accepted in the re-evaluation as a basis for determining "actual" road maintenance cost savings. Actual maintenance cost data on the Tan-Zam Highway were requested from the Ministry of Works but this information has not been forthcoming. Comparison of B with A Estimate 5.13 As shown in para 5.01, the rate of return for all road sections combined and including Zambian traffic drops from 19.5% in the A Estimate to 15.9% in the B Estimate. The following table provides some details on the underlying differences. Comparison Between B and A Estimates: All Project Road Sections Combined, Present Value B Estimate Compared A Estimate B Estimate with A Estimate ------TSh million------ (% difference) 1. Costs - Construction and Supervision 182 213 + 17 2. Gross Benefits - Vehicle Operating Cost Savings Tanzanian Traffic 290 125 - 57 Zambian Traffic 56 131 +134 - Road Maintenance Cost Savings 66 39 -41 Total Gross Benefits 412 295 - 28 Total Gross Benefits (excl. Zambian Traffic) 356 164 - 54 3. Total Net Benefits 229 82 - 64 Total Net Benefits (excl. Zambian Traffic) 173 - 49 -128 Source: Table 3. 1/ This formula relates road maintenance costs to road condition and traffic volume and composition. - 27 - 5.14 Net benefits, including Zambian transit traffic benefits, decrease by 64% from the appraisal estimate. The principal reason for this drop is the lower level of benefits from Tanzanian traffic (-57%) caused by the lower initial (1972) traffic base, the lower (3%) annual growth rate assumed for later years and the absence of generated traffic. An important offsetting factor is the increase in benefits for Zambian transit traffic caused by the assumption that this traffic will not taper off after 1975 but will continue to use the road. The construction cost increase (17%) and the decrease in road maintenance cost savings (-41%), caused by the lower initial (1972) traffic volumes and the lower growth rate thereafter, have a relatively minor effect. 5.15 When the Zambian transit traffic benefits are excluded, the rate of return drops from 16.5% in the A Estimate to 6.8% in the B Estimate (see para 5.01) and net benefits become negative and drop by 128% (see para 5.13). 5.16 Turning now to the individual road sections (see para 5.01), and taking into account the benefits accruing to Zambian transit traffic, three road sections show a rate of return of less than 10%: Kitonga Gorge (rate of return 5.9%), Morogoro +38 miles West-Mikumi (rate of return 6.9%) and Mahenge-Iringa (rate of return 7.7%). The section Morogoro -38 miles West becomes marginal (rate of return 10.4%) and would have have a rate of return below 10% were it not for the reduction in actual road length (4 miles) which was decided upon after appraisal (see para 5.07). 5.17 A combination of factors explains these results. Comparing the appraisal forecast for the 1972 traffic with the actual 1972 traffic (see Table 5), the four road sections mentioned in the previous paragraph show the largest differences rangiLng between +69% and +120%. (The remaining road sec- tions in the project show differences of between +4% and +52%). In addition, two of the road sections (Kitonga Gorge and Mahenge-Iringa) already had the lowest rate of return of all road sections in the original appraisal and, therefore, had the smallest safety margin to start with. Finally, a contri- buting factor in the case of the Morogoro +38 miles West-Mikumi road section is that whereas the traffic composition for all road sections shows a lower percentage of heavy vehicles in the actual 1972 traffic data than in the appraisal forecast for 1972, the difference in this case is the largest (see Table 6), adversely affecting total benefits from vehicle operating cost savings. 5.18 Considering the individual road sections but disregarding the benefits associated wth the Zambian transit traffic, 6 of the sections have a rate of return of 10% or below. Only the last two sections, Sao Hill- Makumbako and Makumbako-Iyayi have a satisfactory rate of return. This mainly reflects the fact that the appraisal forecast of the 1972 traffic shows only a small difference with the actual 1972 traffic: +18% and +4% respectively (see Table 5). - 28 - The C Estimate 5.19 This estimate is based on the same data as the B Estimate (see para 5.04) except for the unit vehicle operating costs savings i.e. vehicle operat- ing cost savings per vehicle-mile. As mentioned before (para 5.11) estimates of vehicle operating costs savings are uncertain and contain a considerable margin of error and subjective judgment. Nevertheless it is useful to see what the economic results of the project is if an alternative set of vehicle operating cost data is used. In light of the above the E Estimate, therefore, should be considered as an alternative estimate of the actual return of the project and not necessarily a better estimate. 5.20 Unit vehicle operating costs and savings were estimated in 1971. 1/ These estimates, with minor modifications, were later used in the Bank's appraisal of the Fourth Highway Project (1974) 2/ and subsequently, an updated version, in the appraisal of the Fifth Highway Project, 1978 and have been adopted, suitably adjusted to fit the specific conditions of the project road sections, to calculate the C Estimate. 5.21 Table 7 compares the unit vehicle operating costs and savings of the appraisal report with the modified estimates. The principal differences are: (i) whereas the appraisal estimate for truck operating costs was based on a 7-ton truck, a 9-ton truck was considered typical; (ii) operating costs on bitumen roads are lower than costs used in the appraisal for cars, slightly higher for pick-ups and considerably higher for the other vehicle types. (The Economic & Engineering Study report does not contain an estimate for operating costs of 30-ton truck trailers. These have been estimated as 2.2 times the cost of a 9-ton truck). The percentage difference between operating costs on different road types is about the same for both the appraisal estimates and the Economic & Engineering Study estimates. 5.22 The impact of the generally higher unit vehicle operati g cost savings are reflected in the C Estimate. As a result, the C Estimate average rate of return for all project road sections combined, and including the benefits of the Zambian transit traffic, increases to 23.7% compared with 15.9% for the corresponding B Estimate. In addition, under the C Estimate, no individual road section has a return of less than 12% when benefits to Zambian transit traffic are included. 5.23 Comparing the C and B Estimates but excluding the Zambian transit traffic benefits, the average rate of return increases to 14.3% for the C Estimate compared with 6.8% for the B Estimate. However, two road sections (Kitonga Gorge and Mahenge-Iringa) have rates of return of less than 9% while 1/ Economic and Engineering Study Tanzania Highway, 1972, Vol. II, page 7-120. 2/ IBRD Appraisal of a Highway Maintenance Project, Tanzania, July 19, 1974, Table 10. -. 29 - two other sections (Morogoro -38 miles West and Morogoro +38 miles West-Mikumi) show rates of just under 11%. These are the same four road sections which in the B Estimate, including the benefits from Zambian transit traffic, showed a marginal or below 10% rate of return (see para 5.16). 5.24 For a detailed comparison of the C and B Estimates, see para 5.01 and Table 3. Conclusions The Economics of the Project 5.25 Looking at the project as a whole and considering benefits accruing to both Tanzania and Zambia, the overall conclusion is that the project was an economic success. Irrespective of whether one prefers the B or the C Estimates, the return is some 20% or about the same as the appraisal report estimate (19.5%). This similarity is not due to the accuracy of the original estimates at the time of appraisal with respect to costs and benefits, but to a combination of factors as analyzed in some detail in the foregoing. 5.26 Focusing on the individual road sections within the project, some of them have a marginal rate of return. 5.27 Looking at the project as a whole from a Tanzanian point of view, that is, ignoring benefits accruing to Zambia, the alternative rates and return are about 7% (B Estimate) and 14% (C Estimate) or an avevage rate of return of some 10%. ICn this case, some of the individual road sections have an inadequate rate of return. 5.28 Given this marginal overall rate of return of the project from a Tanzanian point of view it is interesting to compare the gross benefits of the project accruing to Tanzania with those accruing to Zambia. % of Total Gross Benefits Accruing to: Tanzania Zambia Vehicle Road Vehicle Operating Maintenance Operating Cost Savings Cost Savings Total Cost Savings Appraisal Estimate A Estimate 70 16 86 14 Re-evaluation B Estimate 42 13 56 44 C Estimate 49 7 56 44 - 30 - Whereas the appraisal expected only 14% of the benefits to accrue to Zambia, Zambia's share in the re-evaluated benefits is some 44% or three times as high. 5.29 While the appraisal report in its economic evaluation did distin- guish between benefits accruing to Tanzania and those accruing to Zambia, there was no explicit discussion about possible cost sharing of road con- truction and maintenance costs between the countries and/or possibilities for Tanzania to recoup some or all of the benefits accruing to Zambia. The appraisal report probably ignored this issue at the time because the estimated return to Tanzania was considered more than adequate and the share of benefits accruing to Zambia rather modest. Risk Analysis 5.30 The appraisal report of the Second Highway Project in Tanzania contained the Bank's first attempt at a comprehensive, systematic, detailed and quantitative treatment in probability terms of the risk aspect in project appraisal. Against the background of the present re-evaluation, what obser- vations can we now make? 5.31 The final result of the risk analysis in the appraisal report is presented in the form of different rates of return with the corresponding probabilities for the project as a whole, i.e. for all road sections combined. As the re-evaluation shows, though the overall rate of return of the project may be satisfactory, the return on some road sections may not be. The risk analysis, therefore, should have focused on individual road sections and highlighted the risks by section. 5.32 Some 95% of the overall rate of return curve in the appraisal report (excluding Zambian transit traffic) falls within plus or minus 5 percentage points of the "best" estimate of 15% and indicates a very slight chance (some 2%) of a return below 10%. In hindsight this would seem to result from overly cautious estimates of the probability distributions of important variables entering into the computation, i.e. a too narrow spread (or no spread at all) is assumed around the "best estimate" of the variables. A second factor may possibly have been the method of combining probability distributions of indi- vidual variables in the overall calculation through a random process whereby pluses and minuses tend to offset each other. The appraisal report, in this context, mentions that "care was taken to correlate interdependent variables" but does not indicate which variables were considered interdependent. 5.33 Examples of overly cautious probability estimates in the appraisal report are: (i) construction and supervision costs are assumed to have a uniform probability distribution between the "best" estimate and the "best" estimate plus 15%; (ii) though projected "generated traffic" accounts for about one-third of total projected traffic (see Table 6), and hence for about one-fifth of total vehicle operating cost savings, the elasticity of traffic demand was not con- sidered sensitive and received a probability of 10%; - 31 - (iii) the traffic growth rate is assumed to vary between 4% and 8% per year for light vehicles, and between 6% and 10% for heavy vehicles, with equal probability for each percentage within each range; and (iv) estimates of vehicle operating costs and savings are assumed to vary between -12% and +15% of the "best" estimate. 5.34 An alternative to the risk analysis as presented in the appraisal report is the sensitivity analysis currently used in most appraisal reports. The rate of return is then calculated on the basis of a combination of generally rather arbitrary "pessimistic" estimates of some or all variables: costs +x% and benefits -y%. It might be useful if OED, on the basis of an analysis of past project completion reports, probably by sector or type of project, could provide some guidance in this respect: for which variables "pessimistic" estimates should be made and how much the adjustments should be as compared with the "best" estimates. In addition, it would be useful if OED, on the basis of its accumulated experience, could provide some indication (or guidance) as to the magnitude of the probability that a rate of return calculated on the basis of these "pessimistic" estimates will in fact be realized. VI. THE ROLE OF THE BANK GROUP 6.01 The Government and the Bank Group were in agreement on the limited project objectives, and there was general accord on the project content and the implementation schedule. The project was initiated, studied, and appraised within about two years. The Bank staff was closely associated with the Gov- ernment and consultants in the preparation of the project, and the only major issue during preparation was the design of the pavement, which was not resolved until after appraisal. 6.02 Bank Group staff believed that the consultants had not been suffi- ciently thorough and analytical in their approach to the design of the road pavement. They thought it necessary to carefully review the designs which had been presented as fait accompli, to raise questions on the designs, and to work closely with the consultants in order to agree on methodology, traffic and loading assumptions, thickness of base and sub-base and type and thickness of the surface course. During this review, the surface course design proposed by the consultants was questioned and the pavement standards were reduced. Although it is debatable whether this change caused the subsequent pavement failures, it does suggest that design criteria and choices of methods and technology should be analyzed and agreed upon before detailed engineering proceeds and that Bank staff should not need to be involved in issues such as these late in the project cycle. - 32 - VII. CONCLUSIONS 7.01 The project's objectives, to provide Zambia with road access to the sea and to improve sections of the main road system in Tanzania, have been achieved. However, maximum benefits can only be derived if the Government regularly maintains the road and enforces vehicle axle-load regulations. 7.02 Road maintenance in the area of the project road is currently being improved under the Fourth Highway Project (Cr. 507-TA). The problems of staffing and training of road maintenance personnel are expected to be alle- viated under that project, as well as by the recently approved Fifth Highway Project (Cr. 876-TA). Vehicle axle-load control has improved with the opera- tion of weighbridges at three locations along the Tan-Zam Highway, although vehicle operators are now pressing Government to relax the enforcement of vehicle weight regulations. Government is now studying the possible effects of this change on the economic life of the road pavement before reaching a decision. - 33 - ANNEX TANZANIA PROJECT COMPLETION REPORT SECOND HIGHWAY PROJECT (Loan 586-TA/Credit 142-1-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. Appraisal Report and Credit and Loan Agreements, 1969. 4. Tanzania-Zambia Highway Study Vol. X by Stanford Research Institute, 1966. 5. Contract Documents Morogoro-Mahenge Road and Iyayi-Mahenge Road. 6. Comparative Costs of One-Inch and Two-Inch Asphalt Road Surfaces, Iyayi - Mahenge Road Project, 1969. 7. Analysis of Pavement Failures and Recommended Solutions, Iyayi-Mahenge Road Project, 1971. 8. Design Report on Section I, Morogoro-Mahenge Road Project, 1968. 9. Monthly and Quarterly Progress Reports on Morogoro-Mahenge Road. 10. Monthly and Quarterly Progress Reports on Iyayi-Mahenge Road. 11. The TAN-ZAM Highway Project, 1972. 12. Final Report on the Iyayi-Mahenge Road Project, 1973. 13. Draft Interim Completion Report on the Morogoro-Mahenge Road Project, 1972. TANZANIA PROJECT COMPLETION REPORT SECOND HIGHWAY PROJECT (Loan 586-TA/Credit 142-TA) Actual and Expected Project Implementation % of the vork- Completion completed by Bid Receipt Contract Award Beginning ot Work of Work expected Project Component Actual Expected Actual Expected Actual Expected Actual Expect,a 1/coFplet ion daze A. Civil Works (i) Morogoro-Mahenge Road 1/20/69 n.a 4/2/69 n.a. 4/30/69 5/2/69 5/29/72 9/19/71 80 (ii) Mahenge-Iyayi Road 12/2/68 n.a. 1/20/69 n.a. 2/1/69 2//0/69 12/22/71 5/1/71 80 B. Consultants' Services (i) Supervision of A(i) - - 2/26/68 n.a. 2/26/68 n.a. 5 n.a. 9/19/71 (ii) Supervision of A(ii) - - 2/1/68 n.a. 2/1/68. n.a. n.a. 5/1/71 1/ Refers to completion date agreed upon in the original contract, and does not include any subsequent time extensionk' ). Appraisal schedule completion date for construction: December 31, 1971. April 1979 - 35 - TABLE 2 TANZANIA PROJECT COMPLETION REPORT SECOND HIGHWAY PROJECT (Loan 586-TA/Credit 142-TA) Actual and Appraisal Estimates of Project Costs Appraisal Estimate Contract Actual Actual Costs as % of Project (incl. cont.) Amount Costs Appraisal Contract Component US$ ('000) US$ ('000) US$ ('000) Amount Amount A. Construction (a) Morogoro-Mahenge 19,500 16,668 17,537 90 105 (b) Mahenge-Iyayi 1.7,100 14,139 16,383 96 116 36,600 30,807 33,920 93 110 B. Consulting Services (a) Moregoro-Mahenge 700 425 569 81 134 (b) Mahenge-Iyayi 800 660 695 87 105 1,500 1,085 1,264 84 116 TOTAL 38,100 31,892 35,184 92 110 April 1979 - 36 - TABLF 3 TA"M ANIA PROJECT COMPlTION HFPORT SECOND HTCHWAY PROJECT (Lonn 586-TA/Credit 142-TA) Detailed Comparison of Different Estimates of Benefits and Costs Present Value Gross Benefits Benefit/Cost Ratio Net Present Value Vehil.e Operating Road Present including Excluding including Excluding Cost Savings Maintenance Total Value Length Zambian Zambian Zambian Zambian Tanzanian Zambian Cost Gross Construction Road Section (miles) Traffic Traffic Traffic Traffic Traffic Traffic Savings Benefits Costs 1. Morogoro - 38 miles Vest 34 T Sb Million- A E3timate 2.23 2.06 24.0 20.7 30.9 3.3 9.4 43.6 19.6 B Estimate 1.03 .61 .6 -9.0 10.3 9.7 3.8 23.8 23.0 C Estimate 1.84 1.07 19.4 1.7 20.8 17.7 3.8 42.3 23.0 2. Morogoro + 38 miles-Mikumi 37 A Estimate 2.58 2.32 36.3 30.4 41.9 5.9 11.5 59.3 23.0 B Estimate .74 .45 -7.1 -15.0 7.9 7.9 4.3 20.1 27.3 C Estimate 1.30 1.07 8.2 1.7 16.7 14.6 4.3 35.6 27.3 .3. Mikumi-Mahenge 65 A Estimate 1.97 1.68 45.1 31.5 63.6 13.5 14.8 91.9 46.3 B Estimate 1.23 .63 12.7 -20.1 26.9 32.6 7.3 66.8 54.4 C Estimate 2.39 1.29 75.9 15.6 62.6 60.3 7.3 130.2 54.4 4. Kitonga Corge 5 A Estimate 1.38 1.24 2.8 1.8 7.9 1.0 1.3 10.2 7.4 B Estimate .66 .36 -2.9 -5.5 2.6 2.6 .4 5.6 8.6 C Estimate .1.28 .73 2.4 -2.3 5.7 4.7 .4 10.8 8.6 5. Mahenge-Iringa 40 A Estimate 1.59 1.43 12.5 9.0 23.1 3.5 7.2 33.8 21.1 B Estimate .79 .44 -5.0 -13.6 7.7 8.6 3.1 19.4 . 24.3 C Estimate 1.51 .86 12.2 -3.5 17.7 15.7 3.1 36.- 24.2 6. Iringa-Sao Hill 49 A Estimate 2.56 2.22 44.7 35.1 52.6 9.6 11.4 73.6 28.8 B Estimate 1.68 .99 23.2 -.3 24.6 23.4 9.2 57.2 33.9 C Tt rte '.12 1.85 71.8 29.0 53.9 42.9 9.2 106 0 33.9 7. Sano 'ill-MIakumbako 40 A Estimate 2.83 2.32 42.2 30.5 46.2 11.7 7.4 65.3 23.1 P. K stImate 2.43 1.37 38.5 10.0 29.2 28.5 7.9 65.6 26.9 C Lt..aLe 4.,1 2.66 97.2 44.7 63 7 52.5 7.0 12A.1 26.9 8. Makumbako-lyayi 27 A Estimate 2.72 2.13 21.6 14.1 24.0 7.4 2.6 34.0 12.5 B Estimate 2.58 1.31 22.5 4.4 15.5 18.1 3.3 3t 9 14.2 C Estimate 4.85 2.51 54.7 21.5 32.3 33,3 3.3 68.9 14.2 All Road Sections (1 to 8) 306 A Estimate 2.26 1.95 229.3 173.4 290.2 55.9 65.6 411.7 181.8 B Estimate 1.39 .77 82.4 -49.0 124.7 131.4 39.3 295.4 212.5 C Estimate 2.61 1.47 341.8 100.1 273.4 241.7 39.3 554.4 212.5 April 1979 - 3- TABLE 4 TANZANIA PROJECT COMPLETION REPORT SCOND ]IiCHWAY PROJECT (Loan 586-TA/Credit 142-TA) Comparison of Actual Traffic Volumes, 1970 and 1972 and Appraisal Forecast 1972 (vehicles per day) Actual Road Section/Length 1970 1972 Appraisal Forecast 1972 (miles) 1. Morogoro - 38 miles West (34) 275 301 572 2. Morogoro + 38 miles-Mikuni (37) 275 301 457 3. Mikumi-Mahenge (65) 275 260 338 4. Kitonga Gorge (5) 275 260 453 5. Mahenge-Iringa (40) 275 260 427 6. Tringa-Sao Hill (49) 250 292 406 7. Sao Hill-Makumbako (49) 250 292 330 8. Makumbako-Iyayi (27) 275 252 259 (306) Source: (i) For actual 1970 data, see Economic and Engineering Study, Tanzania Highway, Vol. III, tables following pages 9-34 and for identical (1973) data, see IBRD, Appraisal of a Highway Maintenance Project, Tanzania, July 19, 1974, Table 9; the actual 1970 data are based on an exhaustive review of all available traffic data for the years 1968 through 1971. (ii) For actual 1972 traffic data, see Table 6. (iii) For appraisal forecast 1972, see Table 6. April 1979 - 38 - TANZANIA TABLE 5 PROJECT COMPLETION RE-< " SECOND HIGHWAY PROJECT (Loan 586-TA/Credit 142-TA) Connr:sar- ofFocs:- Actual Traffic, 1972 Traffic Volume, 1972 Actual Passenger Total Excl. Passenger cars Forecast total or Cars plus Zambia and pick-ups as % of Forecast Pick-ups Transit as % of total Actual total Road Section No. 1 ('Iorogroro - 36 miles west) Actual 116 226 51 Forecast 163 407 33 220 Road Section No. 2 (38 milos west of Morogoro-Mikumi) Actual 116 226 51 Forecast 95 382 25 169 Road Section No. 3 ('ik-i-Yhence) Actual 66 185 36 Forecast 65 263 25 142 Road Section No. 4 (Kitonga Gorge) Actual 66 185 36 Forecast 105 378 28 201- Road Section No. 5 (Mahenge-Iringa) Actual 66 185 36 Forecast 95 352 27 190 Road Section No. 6 (Iringa-SaU hill) Actual 89 217 41 Forecast 116 331 35 152 Road Section lo. 7 (Sao Hill-Makumbako) Actual 89 217 41 118 Forecast 69 255 27 Road Section No. 6 (Makumbako-Iyayi) Actual 77 177 214 Forecast 46 184 25 104 All Road Sections Combined, Simple Average Actual 86 202 43 Vorecast 06 343 28 170 April 1979 &-―・―・”--―『――「―-―「―--&―「―「，-―「――「‘---&-：、、・・ (&1：ま・：：・：益・1：・1§・：ま・1ミ・：三・華1 (：・（：・―証・1：・1§・（ま・1ミ・（三・（-1( 11m ー6Eー 40 - TABLE 7 TANZANIA vPn,TVr.T rOMPLETION REPORT SEC04D HIGHWAY PROJECT (Loan 586-TA/Credit 142-TA) Comparison of Unit Vehicle Operatins Costs and Savings Estimates for Different Road TIP089 Appraisal versus "Actual" 50-Passenger 30-too Cars Pick-ups 9-ton Trucks BU"s Truck-TrailM 2r. Act- Appr- Act. A r. t 22_ AL.A. Appr. Art. Appr. Act. Ap V ------------------------------------- T Sh/100 veh. mile ------------------------------- Costs A. Flat to Rolling Terrain - Bitumen 45.8 36.9 49.8 54.1 65.7 162.8 93.B 191.0 181.5 334.2 - Gravel 58.5 46.8 63.6 69.9 99.6 252.9 143.6 290.8 298.4 556.4 - Earth 80.5 63.7 92.9 99.1 163.1 443.3 227.8 487.5 510.5 975.3 SavinU Gravel to Bitumen 12.7 9.9 13.8 15.8 33.9 90.1 49,8 99.8 116.9 1".2, Ditto % savings 22 21 22 23 34 36 35 34 39 36 Earth to Bitumen 34,7 26.8 43.1 45.0 97.4 280.5 134.0 296.5 329.0 617.1 Ditto % savings 43 42 46 45 60 63 59 61 64 64 Costs B. Rolling to Hilly Terrain - Bitumen 47,5 37.5 52.9 55.5 69.6 164.3 103.5 193.'0 191.4 361.5 - Gravel 61.3 47.5 65.0 71.4 107.5 263.7 151.4 294.8 310.5 M*.l - Earth 84.3 65.0 99.2 104.6 17.8.6 "6.5 245.8 500.0 547.3 10"14 Savings Gravel to Bitumen 13.8 10.0 12.1 15.9 37.9 99.4 47.9 101.8 119.1 218.6 Ditto % savings 23 21 0 22 35 38 32 35 U 36 Earth to Bitumen 36.8 27.5 46.3 49.1 109.0 302.2 142.3 307.0 355.9 6". 8 Ditto % savings 43 42 47 47 61 65 58 61 65 65- Ratio of savings: Appr. vs Act. 1 76 1 1.09 1 2.78 1 2.15 1 1'. Note: Exchange rate used US$1.00-T Sh 7.14 1 Vit 1Q69 Appraisal Report refers to 7-ton trucks hut indicates some uncertainty regardingttrurk size (see Annex B. page 2) 2' T,sLi:iiz1tes; calCul,-ILed on basts of 2.2 times costs -.nd savinj;s of 9-ton trucks. So-,irce: 1Ul), Tanzania Appraisal ot the Second Higliway Project, January 15, ' 1969, Table 4 IBRD, Appraisal of a Highway Maintenance Project, Tanzania, July 19, 1974, Table 10 April 1979 - 41 - TABLE 8 TANZANIA PROJECT COMPLETION REPORT SECOND HIGHWAY PROJECT (Loan 586-TA/Credit 142-TA) Schedule of Disbursements as of October 31, 1974 ACCUMULATED DISBURSEMENTS IN THOUSANDS ACTUAL DISBURSEMENTS AS OF U.S. DOLLARS EOUIVALENT A PERCENTAGE OF APPRAISAL IBRD/IDA ESTIMATE (THROUGH FIRST FISCAL YEAR ACTUAL TOTAL APPRAISAL QUARTER 74/75) AND QUARTER DISBURSEMENTS ESTIMATE 1 2-a/ 3 1968/69 4th 1,070 1969/70 1st 1,240 2nd 2,190 5,400 41% 3rd 3,650 4th 4,570 1970/71 1st 7,010 2nd 9,280 14,000 66% 3rd 10,650 4th 12,110 1971/72 1st 14,300 2nd 15,500 22,500 69% 3rd 16,830 22,500 75% 4th 18,620 22,500 83% 1972/73 1st 18,750 22,500 83% 2nd 19,520 22,500 87% 3rd 19,520) 87% 4th 19,830 88% 1973/74 1st 19,830 88% 2nd 19,900 88% 3rd 19,920 88% 4th 19,920 88% 1974/75 1st 20,623 92-- a/ Appraisal report did not contain quarterly figures. b/ Remaining $1.9 million re-allocated to the Third Highway Project (Credit 265) April 1979 n y0 -LU Z o