Peru - Sixth Road Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 4252 PROJECT COMPLETION REPORT PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) December 30, 1982 Latin America and the Caribbean Regional Office 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 PERU SIXTH HlGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT TABLE OF CONTENTS Page No. PREFACE . ................................................. i BASIC DATA SHEET ......................................... ii MISSION DATA ............................................. iv HIGHLIGHTS ............................................... v I. INTRODUCTION ............................................. 1 II. PROJECT PREPARATION, APPRAISAL AND NEGOTIATIONS .... ...... 2 III. PROJECT IMPLEMENTATION ................................... 5 IV. COST ESTIMATES AND DISBURSEMENTS ......................... 12 V. ECONOMIC REEVALUATION .................................... 13 VI. INSTITUTIONAL DEVELOPMENT AND PERFORMANCE OF THE BORROWER 15 VII. ROLE OF THE BANK ......................................... 16 VIII. CONCLUSIONS .............................................. 18 ANNEXES 1. Conversion of Current Soles into Constant US$ .... ........ 20 2. Letter of Agreement from the Borrower .... ................ 21 TABLES 1. Chronology of Works Execution, Huanuco-Aguaytia Road ..... 22 2. Huanuco-Aguaytia Road: Estimated and Actual Cost of Civil Works ......... - 23 3. Comparison Between Appraisal Estimates and Actual Project Costs .......................................... 24 4. Scheduled and Actual Disbursements ....................... 25 5. Traffic Estimates: Appraisal Compared with Actual Counts 26 6. Representative Vehicle Operating Costs .... ............... 27 7. Economic Rates of Return: Appraisal Compared with Completion ............................................. 28 MAP IBRD 10610 (PCR) - Sixth Highway Project 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.   PERU PROJECT COMPLETION REPORT SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PREFACE The following is a Project Completion Report on the Sixth Highway Project in Peru for which Loan 1025-PE, in the amount of US$26.0 million, was approved by the Board on April 23, 1974. The Loan was closed on June 30, 1981, and the final disbursement was made on July 29, 1981. This Completion Report was prepared by the Bank's Latin America and the Caribbean Regional Office and is based on information obtained from the LAC Information Center, Appraisal Report No. 302a-PE, Staff Supervision Reports, project progress reports, final construction reports of the Ministerio de Transportes y Comunicaciones (MTC) and information gathered during a mission to Peru in September 1980. In accordance with the revised procedures for project performance audit reporting, this Completion Report was read by the Operations Evaluation Department but the project was not audited by OED staff. The draft Completion Report was sent to the Borrower who, by letter dated November 9, 1982, communi- cated to the Bank its agreement with the concepts expressed in it. The letter has been annexed to the report.   - ii - PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Basic Data Sheet Key Project Data Original Plan Actual Total Project Cost (US$ million) 40.00 49.18 1/ Overrun (%) 72% 2/ Loan Amount (US$ million) 26.00 26.00 Disbursed (US$ million) 26.00 26.00 Repaid (US$ million) 0.61 3/ 0.61 3/ Borrower's Obligations (US$ million) - 24.69 3/ Date Physical Components Completed 01/77 04/80 4/ Time Overrun (%) 134 Economic Rate of Return (%) 24.00 23.2 Institutional Performance - Varied 5/ Other Project Data First Mention in Files - 01/72 Government's Application - 02/73 Negotiations - 03/74 Board Approval - 04/74 Loan Agreement Date - 07/74 Effectiveness Date 08/74 Loan Reformulation Date 03/78 Closing Date 06/30/78 06/30/81 Borrower Republic of Peru Executing Agency Ministerio de Transportes y Comunicaciones Fiscal Year of Borrower January 1-December 31 Follow-on Project: Lima - Amazon Transport Corridor Project Loan Number: 1196-PE Amount : US$76 million Loan Agreement: 05/76 reformulated 07/79 1/ In 1974 dollars, after reformulation of the project. The appraisal cost estimate of the elements in the reformulated project was US$28.5 million. 2/ Up to February 20, 1981, on reformulated project components (from table 3). 3/ As of 11/30/80. 4/ Except feeder road works financed entirely by the Government. 5/ Under difficult circumstances, some of them outside the control of MTC.   PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Basic Data Sheet (Contd) Country Exchange Rate Country Exchange Rate Name of Currency: Soles Exchange Rates: Appraisal Year Average (1974): US$1 = S/-38.7 Intervening and Completion Years: Overall average inappropriate because of very high inflation. Averages for each intervening year were: 1975 US$1 S/-40.80 1976 US$1 = S/-57.43 1977 US$1 = S/-83.81 1978 US$1 = S/-156.34 1979 US$1 = S/-224.55 1980 US$1 = S/-257.29 1/ 1/ First quarter of 1980 only.   - iv - PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Mission Data Month/ No. of No. of Staff Date of Year Weeks Staff Weeks Report 4/ Preparation 02/72 0.5 2 1 02/27/72 Preparation 05-06/72 2 1 2 07/12/72 Preparation 09/72 0.5 1 0.5 09/18/72 Preparation 11/72 0.5 1 0.5 12/07/72 Appraisal 02/73 3 2 6 04/23/73 1/ Post-Appraisal 05/73 0.5 2 1 06/08/73 Post-Appraisal 10/73 - 2/ 1 - 2/ 11/20/73 Totals 6.5 11 Supervision 05/74 3/ 1 2 2 05/28/74 Supervision 06/74 2 1 2 07/15/74 Supervision 07/74 1 1 1 10/11/74 Supervision 01/75 3/ 1 2 2 02/03/75 Supervision 03/75 2 2 4 03/20/75 Supervision 07/75 3/ 2 1 2 08/08/75 Supervision 10/75 3/ 2 1 2 01/07/76 Supervision 04/76 2 1 2 07/12/76 Supervision 07/76 3/ 2 1 2 08/05/76 Supervision 10/76 3/ 0.5 1 0.5 01/05/77 Supervision 02/77 2 3 6 06/02/77 Supervision 06/77 1.5 2 3 07/21/77 Supervision 09-10/77 2 2 4 10/13/77 Supervision 03/78 2 4 8 05/12/78 Supervision 05/78 1 2 2 08/29/78 Supervision 11/78 2 2 4 01/08/79 Supervision 03/79 2.5 2 5 05/18/79 Supervision 07/79 1 3 3 10/01/79 Supervision 11/79 3/ 1 2 2 5/ Supervision 05/80 3/ 1 2 2 5/ Supervision 09/80 3/ 0.5 3 1.5 5/ Totals 32.0 60.0 Total Supervision Time (Includes Mission and Headquarters Time) FY74 - 1.0 Staff-Weeks FY78 - 15.5 Staff-Weeks FY75 - 6.9 Staff-Weeks FY79 - 6.7 Staff-Weeks FY76 - 9.7 Staff-Weeks FY80 - 3.4 Staff-Weeks FY77 - 11.0 Staff-Weeks FY81 - 7.0 Staff-Weeks 6/ 1/ First draft; yellow cover appraisal report produced in September 1973; grey cover in April 1974. 2/ One day only. 3/ Limited supervision only on mission principally devoted to other ongoing projects. 4/ Final report unless otherwise indicated. 5/ No report in files. 6/ Includes preparation of Project Completion Report.   PERU PROJECT COMPLETION REPORT SIXTH HIGHWAY PROJECT (LOAN 1025-PE) HIGHLIGHTS The Sixth Highway Project (Loan 1025-PE) had as a main objective the upgrading of the Huanuco-Aguaytia section of the Carretera Central. Although originally included for financing under Loan 425-PE, its implementation was postponed when cost and time overruns forced the Bank and the Borrower to reduce the scope of that project (paras. 2.01, 2.02). The execution of the Sixth Highway Project encountered basically the same problems that affected the two previous Bank Highway Projects in Peru: (a) time delays due to institutional weaknesses (para. 6.01) and important design changes and additions (paras. 3.06, 3.14), which resulted in the project being completed almost four years later than appraisal estimates; and (b) cost overruns due to high price inflation and currency devaluations, time delays and, most important, substantial underestimation of project costs because of less than adequate detailed engineering (paras. 3.13, 3.14). The final project cost was about 72% in real terms above appraisal estimates, an increase almost entirely due to a real 111% increase in the price of the civil works elements of the project. The Project Completion Report recognizes that, in retrospect, time and cost estimates made during the appraisal were far too optimistic (para. 7.03). Because of insufficient funds to provide for the project as originally envisaged, the project was reformulated in mid-1977 (para. 3.07) and disbursements on civil works after January 1, 1979 were reduced from 65% to 20% to permit continued Bank participation (para. 4.03). Despite only partial completion of the original construction and improvement program, the project met its principal objective of reducing transport costs on roughly four-fifths of the Huanuco-Aguaytia road. For those sections improved under the project, the overall estimated ERR is close to the 24% anticipated at appraisal. This result, in the presence of substantial time and cost overruns, is principally due to considerable increases in vehicle operating costs which occurred between appraisal and the project's first full year of operation (1980) (paras. 5.02, 8.01, 8.02). The Sixth Highway Project included, at the Bank's suggestion, improve- ments to about 200 km of feeder roads connecting to the Huanuco Aguaytia road, but the Borrower attached very low priority to this element of the project and argued for its elimination at the time of project reformulation, three years after loan approval (paras. 3.07, 3.16). The Bank subsequently (1979) agreed   - vi - that improvements be scaled down to the essential minimum, but, despite this concession, MTC continued to delay the works, which were finally completed in 1980. The Project Completion Report critically examines the Bank's performance during project implementation and concludes that: (a) the Bank made a significant contribution to the project by attaching special importance to those elements of the project other than works on the Huanuco-Aguaytia road: the feeder roads, road maintenance (para. 3.21) and technical studies (paras. 3.22, 3.25, 7.01); these elements of the project served to substantially enhance local skills in geotechnical engineering, planning and feasibility studies and led to a number of important technical recommendations of relevance for similar projects in Andean countries (para. 8.03); (b) the Bank was at times ready to respond to the difficulties that arose during project execution; for instance, the Bank gave immediate consent to disburse loan funds without inter- national competitive bidding for emergency landslide clearance (para. 3.02); also, when the Borrower did not give the Bank the opportunity to review contract awards recommendation and draft contracts before their signature, retroactive Bank approval was given (para. 3.05). The PCR recognizes that such Bank responses, if repeated frequently, would undermine the Borrower's sense of obligation to follow procedures and regulations agreed with the Bank (para. 7.02).   PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT I. INTRODUCTION 1.01 Peru's topographical, geological and climatic conditions make the provision of transport facilities extremely difficult. The transport network reflects these difficulties and has shaped the country's geographic development accordingly. Since the nineteenth century, several isolated railways have been built by private interests to connect mining areas in the Sierra with coastal ports. Even today, railways and ports are highly specialized and operate mainly as transport systems for the export of minerals. Only highways and, much more recently, civil aviation have played an important role in integrating the country, but the lack of an adequate transport network remains one of the most difficult problems confronting Peru. 1.02 The highway network totals some 58,500 km, of which about 6,000 km are paved, 12,500 km have gravel surfacing and 40,000 km are earth roads. After the heavy emphasis on highway construction in the mid 1960s, road development has, since 1970, been concentrated on paving and improving some of the more heavily used roads and gradually extending the system. While the whole system grew by 12% during the period 1970-1978, the paved and gravel road networks grew by 25% and 22% respectively. The backbone of the Peruvian road network is the Pan-American Highway, an all-weather asphalt road 2,675 km long, running the entire length of the country along the coast (Map IBRD 10610(PCR)). The highway system is still in a very early stage of development, and Peru's road density is half that of its neighbors, Ecuador and Chile. The few roads connecting the coast with inland areas are generally low standard, unpaved and sometimes, particularly in the inland areas, little more than rudimentary dry weather tracks. In these circumstances, the importance, for the country's economic integration, of the most important major road connection from the coast through the mountains to the inland areas can scarcely be exaggerated. This road, some 800 km in length, is the Carretera Central, which starts from Lima, crosses the highest range of the Andes to La Oroya, continues across high plateau and other mountain ranges to Huanuco and Tingo Maria, and eventually emerges from the Andean foothills to reach Pucallpa on the headwaters of the Amazon river system. The main element of the project reviewed in this report consisted of paving and improving the 219-km section of the Carretera Central between Huanuco and Aguaytia. 1.03 The Bank has been closely involved with the highway subsector in Peru since 1955. Highway development has to date been promoted by eight loans totaling US$278.5 million: Loan 127-PE, US$5 million, 1955; Loan 300-PE, US$10 million, 1961; Loan 271-PE, US$5.5 million, 1960; Loan 425-PE, US$33 million, 1965; Loan 706-PE, US$30 million, 1970; Loan 1025-PE, US$26 million, 1974; Loan 1196-PE, US$76 million, 1976 and Loan 2091-PE, US$93.0 million in 1982. Four of these loans (Loans 271-PE, 425-PE, 1025-PE and 1196-PE) have supported construction and improvement of different sections of the Carretera Central as part of Bank assistance toward the ultimate aim of a paved road between Lima and Pucallpa. All the loans have helped to finance feasibility studies and detailed engineering designs, as well as providing for studies, technical assistance and equipment with the objective of increasing efficiency in the highway subsector.  - 2 - 1.04 This Completion Report is based on information obtained from the LAC Information Center, Appraisal Report No. 302a-PE, staff supervision reports, project progress reports, final construction reports of the Ministerio de Transportes y Comunicaciones (MTC), and information gathered during a mission to Peru in September 1980. II. PROJECT PREPARATION, APPRAISAL AND NEGOTIATIONS 2.01 In September 1965, the Bank made a loan (425-PE, US$33 million) for the improvement of the La Oroya-Aguaytia section of the Carretera Central. The extraordinarily difficult topographic, geological, soils and climatic conditions of the region, however, had been insufficiently considered during the design, and both the cost estimates of the original studies, as well as the bid prices of the contractors, were substantially underestimated. These factors, together with inflation and the Government's inability to pay contrac- tors for extended periods, led to large time and cost overruns. In November 1969, the Borrower and the Bank agreed to reduce the scope of the project as well as eliminating paving and some stabilization and drainage works in order to match the then available loan funds. The improvement and paving of the Huanuco-Aguaytia section were among works which were postponed for a later phase. 2.02 In March 1972, the Government engaged an Italian consultant firm, which had been working for several years on the Carretera Central, to carry out detailed engineering and economic studies for completion of the improve- ments to the La Oroya-Aguaytia road which had been started under 425-PE. This study, financed under 425-PE, was completed in December 1972 and subse- quently updated in April 1973. The study indicated that the improvements contemplated under what was now the second phase of the La Oroya-Aguaytia project were technically and economically feasible. Accordingly, the project was accepted in principle as the Sixth Highway Project, and an appraisal mission visited Peru in March 1973. 2.03 As a result of this appraisal mission, a subsequent visit in May 1973, and discussions with the Government and the consultants, it was agreed that: (a) certain design features included in the plans proposed by the consultants, which were desirable but considered as not technically essential, would be deleted or modified to reduce costs; (b) the Bank would finance the purchase of maintenance equipment for adequate future maintenance of the project road; and (c) up to US$2 million worth of improvements to access roads leading to the La Oroya-Aguaytia section or the Carretera ientral would be included in the Droject. At that time (June 1973), the Bank expressed its concern over the lack of progress in payments to contractors and consultants under Loans 425-PE and 706-PE as well as over the disputes emerging from Loan 271-PE and other  - 3 - foreign-financed loans. It was felt that these delays and disputes could have an adverse impact on the participation and pricing of contractors and consultants in international tenders called by the Peruvian Government. 2.04 Against the background of a series of disputes concerning the nationalization of foreign corporations in Peru, a deteriorating political situation, and a need for substantial updating of the overall list of projects requiring foreign financing, a Country Programs mission visited Peru during the third quarter of 1973. Partly based on the findings of this mission, Bank management decided to reduce lending to Peru in FY1974. Consequently, suffi- cient funding was not available to finance improvements along the whole length of the road between La Oroya and Aguaytia, and, after further discussions with the Government, it was decided to divide the road into two sections: La Oroya-Huanuco and Huanuco-Aguaytia. Although the La Oroya-Huanuco section yielded higher economic benefits because of heavier traffic (over 500 vehicles per day), it was felt by both the Government and the Bank that the Huanuco- Aguaytia section had a higher priority for reconstruction under the proposed project because highly unstable terrain on some sections of the existing road could cause prolonged closure of the only land corridor between Lima and Pucallpa. Most of the La Oroya-Huanuco section passes through more stable terrain, and such mudslides as occur can be cleared relatively rapidly. 2.05 Thus, the works for the completion of the La Oroya-Aguaytia road, which was originally a part of one project (Loan 425-PE), had now been divided into three phases: Phase I: works financed under Loan 425-PE; Phase II: completion of the Huanuco-Aguaytia section to be financed under the Proposed Sixth Highway Project (subsequently Loan 1025-PE); and Phase III: completion of the La Oroya-Huanuco section (subsequently financed under Loan 1196-PE). 2.06 Negotiations were held in Washington, D.C. in March 1974. The loan for US$26.0 million (1025-PE) was approved by the Board of Executive Directors on April 23, 1974, was signed on July 1, 1974 and became effective on August 17, 1974. The original closing date was June 30, 1978.  -4- 2.07 The approved project items and their estimated cost, as shown in the appraisal report, were as follows: Bank Parti- -----US$ Million ------- cipation Local Foreign Total US$ X A. Civil Works (i) Construction and Improvement of Huanuco-Aguaytia Road (219 km) 10.2 10.3 20.5 13.3 65 (ii) Improvement of Feeder Roads (200 km) 1.5 1.0 2.5 1.6 65 Sub-Total A 11.7 11.3 23.0 14.9 65 B. Consulting Services (i) Construction Supervision 0.9 1.4 2.3 1.4 60 (ii) Additional Geotechnic Studies 0.3 0.4 0.7 0.4 60 (iii) Road User Charges and Pre-Investment Studies 0.4 0.7 1.1 0.7 60 Sub-Total B 1.6 2.5 4.1 2.5 60 C. Road Maintenance (i) Equipment - 2.2 2.2 2.2 100 (ii) Resealing of Neshuya-Pucallpa Road 0.4 0.3 0.7 0.4 65 Sub-Total C 0.4 2.5 2.9 2.6 90 Basic Cost Estimate 13.7 16.3 30.0 20.0 67 D. Contingencies (i) Physical (15% of A and C(ii)) 1.8 1.7 3.5 2.1 60 (ii) Price Escalation (22% of Basic Cost Estimate) 3.4 3.1 6.5 3.9 60 Sub-Total D 5.2 4.8 10.0 6.0 60 TOTAL PROJECT COST 18.9 21.1 40.0 26.0 65  III. PROJECT IMPLEMENTATION A. Civil Works (1) Huanuco-Aguaytia Road 3.01 For bidding purposes, the Huanuco-Aguaytia road was divided into five contracts (Lots) as follows: Lot A: Aguaytia-"Tea Gardens" 41.5 km Lot B: "Tea Gardens"-Puente Pumahuasi 38.0 km Lot C: Puente Pumahuasi-Huachipa 47.7 km Lot D: Huachipa-Carpish 45.2 km Lot E: Carpish-Huanuco 46.8 km Tender documents were issued for international competitive bidding with bid opening scheduled for October 1974; contractors were to be post-qualified. Following the usual practice, official cost estimates were published for each Lot, and only bids within 10% of those estimates would be accepted. The official estimates were too low, and the invitation to bid produced virtually no results, with only one bid for Lot E being submitted. In accordance with Peruvian law, the bidding was declared void and the one bid received was returned unopened. 3.02 In the early months of 1974, prior to the initial bidding period, heavy rains caused repeated slides which blocked the Tea Gardens-Puente Pumahuasi section of the road (Lot B). The Bank agreed, in July 1974, to disburse US$400,000 in loan funds, without competitive bidding, to enable slide removal work (to reestablish full roadbed width) to be carried out on an emergency basis by a local contractor already working in the area. In spite of this concession by the Bank and the emergency nature of the work, the slide removal did not begin until November 1974 because of Government administrative delays. 3.03 In March 1975, subsequent to the first bidding, an inspection revealed that Lot B had stabilized reasonably well, but that the section between Huachipa and Carpish (Lot D) had developed some critically unstable zones and was suffering from severe erosion from the rivers, which had swollen to record levels as a result of the unusually heavy rains. Because of the uncertainty of future stability on some sections, it was agreed that this portion of the road should not be paved at this time; it was, therefore, not included in the next round of bidding. It was duly paved as a Government- financed project (para 3.10). 3.04 New bid documents were issued in April 1975. The revised official estimates of construction costs for the five lots were, by then about 50% in excess of the appraisal estimates (US$30.8 million as against the appraisal's US$20.5 million, excluding contingencies), and normal procedures were modified to permit bids within 15% of the official estimate to be accepted. Bids were opened in July 1975, and only one bid was received for each of the four Lots: A, B, C, and E. Price levels were, on average, about 14% above the official estimates.  - 6 - 3.05 In August 1975, separate contracts were awarded to each of the Peruvian firms bidding for the four Lots. The contracts were awarded without the Bank having been informed either of the results of post-qualification or of the evaluation of bids. Nevertheless, in view of the urgency of the proposed works, and the delays which the project had already incurred, the Bank approved the contracts retroactively. Work began on October 10, 1975, with expected completion of Lots A and B in January 1978, Lot C in July 1977, and Lot E in October 1977. 3.06 Subsequent execution and costs of the contracts were substantially affected by design changes and additions (para 3.14). Principal among these were: (a) raising embankments (Lots A and C); (b) widening roadbed (Lot C); (c) changing potentially unstable and failed sideslopes (all lots); (d) adding of protective works against river scour (Lots A and C); (e) additional erosion protection, paved ditches and underdrains (all Lots); (f) substitution of a 400-meter long tunnel for three bridges in the Boqueron Canyon (Lot A); (g) paving of 1.4 km of the main street through the town of Tingo Maria, for which no solution had been included in the original contract (Lot C); (h) increasing pavement width from 6.0 to 6.6 meters in order to reduce raveling of pavement edges (Lots A, B and C); (i) large increases in volumes of landslide removal after severe rains in 1976 and 1977 and addition of protective walls (Lots A, B and C); (j) change in the alignment of the bypass around the town of Huanuco to reduce expropriation of houses (Lot E). These design changes, which were agreed by the Bank, inevitably led to modifi- cations in the construction schedule, which were aggravated by slow and com- plicated administrative procedures for approving contract additions and modifications. Work on Lots C and E was completed by the end of March 1979, Lot B by July 1979, and Lot A by the end of the first quarter of 1980. The final increases in the value of constructed works, excluding cost escalation, were: for Lot A, 97%; Lot B, 32%; Lot C, 71%; and Lot E, 32%. 3.07 By early 1977, it had become evident that there were insufficient funds in the loan to provide for all of the project, and, by letter of April 25, 1977, the Government requested reformulation of the project. The Bank agreed to delete the improvement of Lot D and the resealing of the  - 7 - Neshuya-Pucallpa road section, which had, by this time, deteriorated to a condition which required pavement reconstruction. However, the Government request to also delete feeder road improvements was not accepted by the Bank, and the Government agreed to include, at the Bank's request, detailed engi- neering studies for pavement rehabilitation of the Neshuya-Pucallpa road. The loan agreement was amended accordingly on March 9, 1978. 3.08 In early 1979, an earthquake caused localized damage to a number of paved segments. Although some of the necessary repair work was carried out under the project, there were insufficient funds remaining to provide for all of it. For the remainder, therefore, MTC awarded a separate contract to the contractor still working in the area. There was no Bank participation in the expenditures under these contracts, and they are not included in the cost estimates for this project. 3.09 During several missions, and also by letter, the Bank expressed its concern regarding the lack of effective provisions for the maintenance of the paved road sections completed under this project. MTC had largely depended, for pavement repairs, on the contractor working on Lot D. This solution is no longer acceptable and does not provide for routine mainte- nance. A plan of action to improve MTC's overall maintenance capability was agreed with the government as part of the Eighth Highway Project (Loan 2091-PE). 3.10 With regard to Lot D, detailed engineering was initially delayed by the need for additional geotechnical studies. These studies were financed under the Lima-Amazon Transport Corridor Project (Loan 1196-PE, US$76.5 mil- lion, December 1975). A construction contract was finally awarded in 1979 to the same Peruvian contractor responsible for Lots A and B with total finan- cing by the Government. Work began in the fourth quarter of 1979 and was completed in 1981. 3.11 Table 1 shows the chronology for constructing the various Lots of the Huanuco-Aguaytia Road. Construction on all Lots began in October 1975, some 10 months later than expected, and all work was not completed until April 1980, about 4-1/2 years later. The various Lots were completed from 18 to 27 months later than specified in the contracts and from two-and-half to four years later than estimated at the time of loan negotiations. The main causes of these lengthy delays have been described in the preceding paragraphs. 3.12 Rapid inflation and devaluation of the sol in Peru, especially since 1977, made it difficult to compare appraisal estimates with actual construction costs. The comparison has been made using constant 1974 US dollars, the year of project appraisal with the methodology outlined in Annex 1 used to convert the certified cost of works to constant dollars, using the construction cost index as a basis. Because this index is an annual average, the resulting constant costs represent mid-year values regardless of when works were com- pleted in any given year. In spite of this imprecision, however, the comparison is considered to be a reasonable indication of cost differentials between actual expenditures and appraisal estimates.  - 8 - 3.13 Table 2 shows the appraisal estimates, original contract value and the final cost for each of the four Lots and for the works as a whole. The total cost of the construction was about US$41 million in 1974 values, an increase of 111% over the US$19.4 million appraisal estimate (including physical contingencies) and some 76% more than the original contract value. The increase was particularly pronounced in Lot A, where a tunnel was intro- duced into the project after contract awards. The principal factors causing these cost increases have been described in paragraph 3.06. 3.14 Supervision of the construction work was awarded to a joint venture of the same Italian firm which prepared the detailed engineering (para 2.02) and three Peruvian consultants. Although the work of the joint venture in supervising the construction and carrying out quality control was generally adequate, the quality of its design and redesign work during construction was less than satisfactory in some instances. The numerous major changes and additions required during construction (para 3.06) indicate clearly that the original detailed engineering was inadequate in several respects, particularly with regard to subsurface explorations and interpretation of the results, protection against river action, protection against erosion caused by changing drainage patterns and slope stability investigations. 3.15 Supervision costs for the civil works on the Huanuco-Aguaytia road were about US$3.94 million (in constant 1974 values) compared with an appraisal estimate of approximately US$1.8 million for the four sections of the road on which civil works were completed. This increase (121%) reflects the extended construction period and the redesign work which had to be carried out. Supervision costs were about 10% of construction costs, which is reasonable considering the extended construction periods. (ii) Improvement of Feeder Roads 3.16 The project included improvement of about 200 km of feeder roads connecting to the Huanuco-Aguaytia road, and the Government selected three road sections for this work: Puente Rancho-Panao-Chaglla (70 km); Puente Prado-Monzon (68 km); and Ramal-Tournavista (62 km). However, it soon became apparent that MTC attached a low priority to this element of the project, and, by mid-1977, about three years after loan approval, no design work had started and decisions were still pending on the width of bridges, the method of construction (by contractors or force account), and whether supervision should be by a consultant or MTC's own forces. During discussions on project refor- mulation (para 3.07), MTC personnel argued for the elimination of the feeder roads from the project, but the Bank insisted on their continued inclusion on the grounds that they would substantially extend the area which would benefit from the construction of the Huanuco-Aguaytia road. 3.17 MTC continued to delay the start of design work throughout 1978, and, in early 1979, the Bank agreed with MTC's proposal that improvements be considerably scaled down to the essential minimum--that is, construction of permanent bridges and culverts and spot improvements to road surfaces where necessary. Despite this concessionary agreement by the Bank, it was not until January/February 1980 that construction contracts were awarded, through local competitive bidding, and consultants were contracted for construction  - 9 - supervision. Construction of seven bridges was completed during 1980. MTC's final report on the project (dated December 1980) states that seven bridges have been constructed under the project at a total cost of S/ 223 million (about US$0.8 million equivalent). Improvements to the surface of the Tournavista road have been undertaken by force account, with total financing by the Government, and similar improvements have been started, and are programed to continue, on the Puente Prado-Monzon Road. (iii) Resealing of the Pucallpa-Neshuya Road (61 km) 3.18 Resealing of this road was considered of high priority in order to prevent serious damage to the existing asphalt surface treated pavement, the construction of which was financed under Loan 271-PE in 1967-1968. Despite the urgency of this work, MTC was unable to finalize contract documents until mid-1977, almost three years after the Loan Agreement was signed, because of protracted litigation with contractors who had carried out the improvements in 1967-1968. As a result of this delay, what had been a minor project component in 1974 was, by mid-1977, a major reconstruction task because of predicted disintegration of the neglected pavement. 3.19 In February 1978, the Borrower and the Bank agreed to delete the resealing of this road from the project and to include instead the consulting services for design of the necessary pavement reconstruction. At that time, the Bank emphasized the importance of reinforcing the subbase and base courses prior to paving since the original surfacing had almost totally disappeared on many sections. However, the start of consulting services was delayed until December 1978 because of a contracted, emergency road repair program launched by the Government to offset the discontent that had caused a general strike in the city of Pucallpa. The consultants eventually designed a full reconstruction of the pavement at an estimated cost of about S/. 244 million (US$3 million equivalent), at mid-1977 prices, compared with the appraisal estimates of US$0.7 million equivalent. This satisfied the Government's commitment under the reformulated project, although the design failed to reflect the widely different traffic volumes using different sections of the road. 3.20 Subsequently (late 1979), responsibility for pavement reconstruction on this road was transferred to the newly created Loreto Regional Development Authority (ORDELORETO), under which a special commission for the Pucallpa area had taken charge of starting the improvement of the first 15 km outside of Pucallpa with a S/ 700 million (about US$2.8 million) allocation in the 1980 budget. Agreement was reached between the Bank and ORDELORETO's engineers concerning appropriate pavement design procedures to accommodate varying traffic densities on different sections of the road. Work on the first 5 km started in early 1982. (iv) Procurement of Maintenance Equipment 3.21 During negotiations, agreement was reached with the Government that a maintenance program for the La Oroya-Pucallpa road would be prepared to serve as a basis for procurement of maintenance equipment financed under the project. Such a program was prepared, and, in July 1974, the Bank agreed  - 10 - with a list which included workshop installations, workshop tools, and mobile maintenance trucks. Because of MTC's difficulties in accepting the procurement procedures set out in the Loan Agreement, draft bid documents were prepared and presented to the Bank only in May 1975. However, once started, procurement proceeded smoothly, and, by mid-1977, a total of 15 separate contracts had been awarded for the supply of 19 items of equipment by 11 suppliers. The value of the contracts amounted to about US$1.95 million, leaving a balance of almost US$250,000 in the Loan account in this category. MTC's request to use this balance of funds for additional maintenance equipment from the same suppliers was accepted by the Bank, and all procurement was completed in January 1978 at a total cost of US$2.20 million. The equipment thus acquired was used to maintain the unpaved portions of the road, including Lot D. B. Geotechnical Studies (i) Studies by Consultants 3.22 The principal objective of these studies was to investigate and recommend methods for improving the stability of road sections through unstable mountain areas or in zones of "huaycos" (mud flows). The start of this work was delayed extensively. Several months were required for MTC and the Bank to agree on the scope of work and detailed terms of reference, and this delay was followed by protracted and unsuccessful negotiations, first between MTC and a local geophysical institute, and then with a French Government agency. It was finally agreed that the studies would be done by the Special Studies Department of MTC with assistance from selected individual consultants. Work began in January 1979 and was carried out in two phases: Phase I during the period January-July 1979; and Phase II, starting in April 1980, was not fully completed and no report was issued. 3.23 The Phase I studies were carried out successfully, and resulted in a report which presented specific engineering recommendations to deal with various types of instability and slides. However, these recommendations, including pre- ventive measures (tree planting, ditch lining, increased drainage, etc.), were not utilized fully by the MTC Design Department and their consultants who still rely too much on a "structural" approach with the emphasis on rigid retaining walls and tunnels. Phase II of the studies were eventually reduced to conducting limited monitoring of stability of several landslides in the Tea Garden area. The work was constrained by inadequate budgets and MTC eventually lost all their qualified geotechnical staff. 3.24 Since the study was carried out by MTC personnel with some supple- mentary budget allocations, the portion of the cost of the studies financed under the project was only US$33,000 for the services of a foreign expert, which is extremely low compared with the appraisal report estimate of US$700,000. (ii) Study of Road User Charges 3.25 A contract for this study was signed in late 1976 with a joint venture of two consulting firms, one US and the other Peruvian. At the time of contract award, the Bank questioned the qualifications of some members of the proposed team, and, although some changes were made in the team composition,  - 11 - the results of the study reflected certain of the staff's shortcomings. The work, which required numerous reviews and changes, progressed slowly, with completion in April 1978, some nine months later than the July 1977 date specified in the contract. Some basic faults, such as the analysis of rail and shipping costs being based on average rather than marginal costs, remained even in the final report. Nevertheless, the main goal of the study - the determination of road user costs and allocations - was on the whole achieved, and the methodology and traffic assignment models developed were utilized in later studies. The consultants were willing and cooperative in undertaking the considerable amount of additional work necessary to complete their assign- ment satisfactorily. The cost of the study was S/. 15.2 million (US$236,000 approximately). There was no separate estimate of the cost of this item in the appraisal report (para 3.28). (iii) Feasibility Study of the Tarma-San Ramon-Oxapampa Road 3.26 The road was separated into two sections for study purposes: Tarma-San Ramon and San Ramon-Oxapampa. For the first section, MTC awarded the contract to a Peruvian consultancy with limited experience, without taking into account the Bank's comments on the proposal and recommendation that two experienced expatriate experts be incorporated into the study team. However, the Bank subsequently approved the contract, and work began in April 1976. The study was completed satisfactorily, and a draft final report, including engineer- ing plans, was submitted to MTC in March 1978. The engineering solutions proposed for several critical points on the winding Palca-San Ramon section have been further evaluated, discussed with MTC, and modified in the road design which has been included in the Eighth Highway Project. 3.27 The study for the second section was awarded to the same Peruvian 'consultancy in October 1979. Initiation of this work had been delayed pending the results of a general transport study of the area by MTC to determine if the proposed branch road to Oxapampa constituted the most appropriate alter- native. For this section, the consultant's draft final report included only preliminary engineering. Final engineering is eligible for financing under the Eighth Highway project. The study was completed satisfactorily in July 1980 with recommendations for road improvements to the level of paved road to Puente Paucartambo. The consultant's recommendation for a surface treatment of the existing branch road to Oxapampa (3.5 to 5.0 meters wide), however, has not been accepted by MTC and the Bank, and further study of an alternative alignment is being undertaken under the Eighth Highway Project. 3.28 The total cost of the feasibility studies for the Tarma-Oxapampa road and the study of road user charges was approximately US$436,000 equiv- alent compared with an appraisal report estimate of US$1.1 million. The major reasons for this much lower cost were the extensive use of Peruvian consultants and the assistance provided by MTC's planning service. As a result of the austerity in Peru in this period, salaries paid for these services have been very low by international standards.  - 12 - IV. COST ESTIMATES AND DISBURSEMENTS 4.01 A summary comparison between appraisal estimates and actual costs for the various project elements is shown in Table 3. The final total project cost was US$49.18 million 1/, about 72% over the appraisal estimate of US$28.53 million for the same project elements. The increase in costs was entirely due to the 111% increase in the amount of civil works. In 1974 prices, major increases in contract amounts resulted from the following addi- tional work: Lot A Boqueron Tunnel and approaches US$3.79 million (equivalent) Slides and rain damage (1976 and 1977) US$2.65 million (equivalent) Lot B Rain damage (1976 and 1977) US$1.33 million (equivalent) River damage and protective works US$0.27 million (equivalent) Lot C Additional bridge works US$0.67 million (equivalent) Tingo Maria street (addition) US$0.85 million (equivalent) Lot E Additional slope protection works US$0.10 million (equivalent) Additional maintenance US$0.10 million (equivalent) 4.02 Loan funds were reallocated twice, as follows: Category Items Loan Agreement Reallocated --------US$ Millions--------- Original 5/79 6/81 I Civil works 15.3 21.3 21.2 II Maintenance equipment 2.2 2.2 2.2 III Consulting services 2.5 2.5 2.6 IV Unallocated 6.0 0.0 0.0 Total 26.0 26.0 26.0 This reallocation between broad loan categories simply consisted of the absorption of the contingency items into civil works expenditures, masking some important changes, other than those introduced by project reformulation, within the different categories. These were: 1/ In 1974 dollars for the reformulated project.  - 13 - (a) the reduction in cost of the feeder roads improvement (para 3.17) and the absorption of this reduction into the construction costs of the principal highway; and (b) the substantially higher cost of construction supervision within consulting services, partially offset by reductions in the cost of the geotechnical studies and the feasibility studies of the Tarma- Oxapampa road. 4.03 As shown in Table 4, disbursements lagged substantially behind appraisal estimates from the start. In the early years of the project, this resulted mainly from extensive delays by MTC in the bidding process and in awarding contracts. In spite of a substantial acceleration in project implementation in 1977 and 1978, the initial delays in disbursements were never overcome. By agreement between the Bank and MTC, disbursements on civil works after January 1, 1979 were reduced from 65% to 20% to permit continued Bank participation. The loan was fully disbursed on July 29, 1981. 4.04 The loan closing date was extended twice: first, in February 1978, from the original date of June 30, 1978 to June 30, 1980, and then, in August 1980, to June 30, 1981. V. ECONOMIC REEVALUATION 5.01 The Sixth Highway Project was justified at appraisal on the basis of quantifiable benefits accruing to road users, mainly from reduced vehicle operating costs and travel time which would result from improvements in road design and surface characteristics. In addition to road user savings, changes which were expected to occur in maintenance cost as a result of the proposed improvements, including the provision of additional maintenance equipment for the Huanuco-Aguaytia section, were included in the evaluation. A separate economic analysis of the maintenance program for the remainder of the La Oroya-Pucallpa road was also carried out. 5.02 The appraisal estimate of the economic rate of return (ERR), not including time savings, for the improvements to the Huanuco-Aguaytia road was 24%, with an average traffic growth rate of 7% per year from 1977 to 1996. Reevaluation, based on the methodology discussed in the following paragraphs, resulted in an overall estimated ERR, for the four road sections improved under the project, of 23.2% with an average traffic growth of 4.1% per year from 1980 to 1999 (para 5.04). For truck traffic only, the average estimated ERR is 17.4%. Only savings in vehicle operating costs were considered in the reevaluation analysis. These highly satisfactory rates of return were obtained despite sub- stantial construction delays and cost overruns, and with a lower assumed annual growth of traffic than at appraisal. This result is principally due to the considerable increases which occurred in vehicle operating costs between the date of appraisal and the project's first full year of operation (1980). With  - 14 - the high proportion of trucks on the project roads, averaging 41% of total traffic (Table 5), and the competitive nature of the trucking industry, it is likely that project benefits related to freight movements are being reasonably distributed among truckers and shippers by limiting freight tariff increases. The road improvement is a vital element in reducing transport costs and the risk of closures on the Carretera Central, which is the only land corridor between Lima, the agriculture development areas beyond Tingo Maria and the river port of Pucallpa, on a tributary of the Amazon. 5.03 Base year traffic estimates rely on counts conducted in 1979 by MTC. From Table 5 it can be seen that the appraisal projection of 7% annual traffic growth was very close to the actual rate (7.2%) for the period 1972-1979 on the road as a whole. There were, however, substantial differences between the rates of growth experienced on different sections, with those beyond Huachipa showing considerably faster growth than the area between Huanuco and Huachipa. This is in large part explained by the previously poor condition of the road in these sections, especially in the Carpish mountain range, and the interrup- tion to the flow of traffic caused by landslides. The high rates of growth elsewhere on the road reflect the importance of locally generated traffic in the rapidly developing area beyond Tingo Maria. 5.04 On the basis of annual traffic counts, MTC derived a trend line, using the least squares method, based on data for the 1965-1979 period. This trend line has been used for projections of traffic from 1980 onward and results in an average annual growth of 4.1% per year during the period to 1999. On the basis of past experience on the road as a whole, this can reasonably be considered as conservative. 5.05 For purposes of the economic reevaluation, all costs and benefits were expressed in constant soles (February 1980 prices). Vehicle operating costs used in the original feasibility study and appraisal were updated to February 1980 prices using data from the "Estudio de Transporte para la Selva Central," prepared by the Economic Studies Office of MTC, and the general consumer price index for items other than food, housing and clothing. The updated vehicle operating costs are shown in Table 6 . Project bene- fits were calculated on the basis of the traffic mix identified from MTC's counts in 1979, taking into account the improvements actually carried out on each section of the road. For example, in the case of Lot B (Puente Pumahuasi- Tea Gardens, part of the road was not paved (because of continuing stability problems), but was, rather, improved to good gravel standards (para 3.03), and this is reflected in the calculation of the benefits accruing to this section. Construction, supervision, and maintenance costs were converted to constant February 1980 prices, using the construction cost index. 5.06 The estimated ERRs for each of the four road sections, and the improve- ment program as a whole, are shown in Table 7. For the two sections between Huanuco and Puente Pumahuasi (Lots E and C), the reevaluated ERRs are very close to those of the appraisal. In the case of Lot B (Puente Pumahuasi- Tea Gardens), the ERR is substantially higher than at appraisal, mainly because of a higher rate of traffic growth between 1972 and 1979 than predicted for this section (13.8% per year as against the appraisal forecast of 7% per year). The much lower ERR on Lot A is due to the very high actual construction  - 15 - costs incurred. However, despite an average construction cost in this section of US$440,000 per km (in constant 1974 dollars), the estimated ERR is still satisfactory and underscores the high justification for road paving projects once traffic reaches levels in excess of 500 ADT. Furthermore, the rate of return understates the total benefits of the project in that it reflects only savings in vehicle operating costs. VI. INSTITUTIONAL DEVELOPMENT AND PERFORMANCE OF THE BORROWER 6.01 Even before implementation of the Sixth Highway Project, MTC suf- fered numerous institutional and administrative weaknesses which frequently impeded the achievement of its objectives in the transport sector. Within MTC, planning, construction, and maintenance of the national road network is the responsibility of the Direccion General de Transporte Terrestre (DGTT). In January 1973, DGTT was reorganized and divided into six directorates. As was foreseen in the appraisal report, this new organization did not have a significant beneficial impact on MTC's management performance. The substantial administrative delays experienced under previous highway loans were equally prevalent in the case of this project. As noted in Chapter III, virtually no element in the project started without delays of at least 12 months and, in some cases, of more than three years. These delays undoubtedly contributed to the substantial cost overruns which were incurred. Furthermore, in at least three individual instances, MTC did not comply with procedures set out in the Loan Agreement with regard to the preparation and review of bidding, procure- ment and contract documents. 6.02 Both this project and Loan 706-PE (US$30 million, September 1970) provided for reinforcement of the road maintenance system, which remains the most critical area of road transport infrastructure. Technical assistance under Loan 706-PE was instrumental in developing a maintenance management system which has since been progressively introduced as the basis for planning and programing maintenance in the departments. However, the partial intro- duction of the system has not led to any significant improvement in the road maintenance situation because of insufficient funding, lack of working equip- ment, and institutional difficulties including frequent organizational changes. The acquisition of small lots of maintenance equipment was financed under this project, but this has not resolved the underlying institutional problems. In October 1980, the Bank expressed its concern that MTC was still largely dependent on the contractor working in the area for pavement repairs and, in June 1981, that routine maintenance of the improved sections of the Huanuco- Aguaytia road was not being performed. The implementation of a cohesive maintenance program for the La Oroya-Pucallpa road, which was a specific undertaking by the Government expressed in a covenant of the Sixth Highway Project Loan Agreement, has at best been erratic. 6.03 In general, these difficulties stem from four basic characteristics of MTC's recent administrative structure: (a) partial decentralization without corresponding clear delineation of responsibilities regarding execution and supervision of central- ized and decentralized functions at different levels (departmental and regional);  - 16 - (b) lack of specific focus on road infrastructure within DGTT (which was also responsible for railways) and a similar lack of focus on road maintenance in the department and regions; (c) insufficient coordination among the different activities carried out within the DGTT and between DGTT and other MTC directorates; and (d) depletion of technical and managerial expertise due to low compensation levels and uncertain career prospects. Measures aimed at improvement of these weaknesses are currently being intro- duced by the new MTC management; however, time will be required to effect the changes since it is impossible to separate the solution to some of these problems from the broader aspects of Government policy concerning issues such as decentralization and the salaries of public employees. In the case of this project, it should be added that the execution of the project coincided with a particularly difficult and turbulent period for the Peruvian economy as a whole, which undoubtedly contributed to the budgetary and administrative problems faced by MTC. Under the subsequent project, Loan 1196-PE, six road contracts are presently being executed. Initiation of these works also sustained considerable delay, but, since the date of contract signing, the works have remained generally close to schedule, with only one contract indicating troubles of the nature experienced under the project. The new MTC administration is also taking measures to restore DGTT to highway-department status 1/ and the field-engineering departments to being exclusively maintenance organizations. 6.04 Despite the problems discussed in the preceding paragraphs, there is no doubt that the project did have a number of positive results from the point of view of the Borrower's institutional development. The project supported broad based studies related to transport planning and the prepara- tion of future investment programs as well as specific road feasibility and engineering studies. These studies have led to the identification and subsequent preparation of high priority programs and to improvements in planning. The transport planning studies were largely carried out "in-house" while specific road feasibility and engineering studies were contracted to local consulting firms. Furthermore, the geotechnical studies prepared by MTC's Special Studies Department, supported by Bank-financed expatriate experts, contained high quality work providing valuable recommendations on more appropriate measures to deal with various types of road instability and landslides common in the Andes (para 3.23), and serving as a basis for upgrad- ing highway engineering in Peru by including more attention to geotechnical, drainage and soils problems. VII. ROLE OF THE BANK 7.01 The Bank contributed significantly to project identification and preparation via the preceding highway projects on the Carretera Central (especially Loan 425-PE; US$33 million; September 1965) and during the execu- tion of the project provided essential continuity and advice in a difficult 1/ In 1981 DGTT became DGC (General Roads Directorate) being concerned only with highway and rail infrastructure directorates.  - 17 - period for the Peruvian economy and for MTC. The number and timing of super- vision missions were appropriate for the circumstances during the execution of this difficult project. A particularly significant contribution of the Bank was the importance which it attached throughout to those elements in the project other than the construction and improvement of the Huanuco-Aguaytia road itself: the feeder roads, road maintenance, and technical studies. On several occasions during execution, the Borrower tended to attach a low priority to these elements and even, in some cases, wished to eliminate them and reallocate the funds to civil works construction or supervision. When these difficulties arose, the Bank was always swift to point out to the Borrower the benefits of the studies, not merely from the technical point of view, but also from that of building local knowledge and institutional develop- ment. It was thanks to Bank insistence that feeder road improvements took place at all under this project and that the geotechnical studies were finally carried out satisfactorily. In the case of the maintenance program, problems associated with lack of adequate funding and institutional weaknesses within MTC limited the achievements that could take place during the period of the project's execution. The Bank's missions have, however, served to continue focusing MTC's management attention on maintenance and rehabilitation and suggesting ways in which some of the institutional difficulties might be overcome. This is being done under the reformulated Loan 1196-PE and the Eighth Highway Project. 7.02 The Bank was always ready, perhaps at times a little too ready, to respond to the continual difficulties which arose, particularly during the early stages of project implementation. When the Borrower requested a rapid disbursement of US$400,000 in order to contract extremely urgent emer- gency repairs without international competitive bidding (para 3.02), consent was given immediately, although several months were to elapse before these funds were put to use. When the Borrower did not comply with proper procedures, for example by failing to allow the Bank the opportunity to review contract documents before they were signed, retroactive Bank approval was nearly always given within a short space of time. While this was done with the objective of avoiding still further delays in project implementation, there must at least be a possibility that such a response, if repeated too frequently, will undermine the Borrower's sense of obligation to follow strictly the procedures laid down in the loan agreement and in the general regulations of the Bank. 7.03 The estimates made at appraisal of the time which would be required to e,eecute the project and disburse the loan were, in retrospect, far too optimistic (para 4.04). In view of the fact that, first, very long delays and difficulties were experienced on previous highway projects in Peru and, second, that considerable skepticism was rightly expressed in the appraisal report concerning the efficacy of administrative reforms carried out at MTC in 1973, part of these delays should have been foreseen and allowed for at the project preparation stage. Furthermore, especially in the difficult terrain and climatic conditions of the project area, there is a natural tendency to underestimate construction costs at the final design stage. The consultants' estimates were indeed much too low, and were reduced even further, at the Bank's suggestion, before incorporation into the appraisal report. As a consequence, the initial invitations to bid produced virtually no response,  - 18 - and cost increases of 50% were included in the re-issued invitations. On future occasions, especially under conditions such as those experienced on the Carretera Central, construction cost estimates produced by the Borrower or his consultant should be examined more critically by the Bank, not just to see whether they can be reduced, but also to judge whether they might not be too low. In addition, the Bank should continue to support the recommendations arising from the geotechnical studies financed under this project concerning more innovative methods of dealing with stability problems. VIII. CONCLUSIONS 8.01 Despite only partial completion of the construction and improvement program originally envisaged at the appraisal stage, the project represents an important contribution to the betterment of the vital land transport link between Lima and the headwaters of the Amazon at Pucallpa. On roughly four- fifths of the Huanuco-Aguaytia road, the project met its principal objective of upgrading the highway and reducing transport costs; work on the remaining portion of the road was completed in 1981, with financing wholly by the Borrower. This has been achieved despite institutional difficulties and physical conditions which challenge the current state of the art in highway engineering. 8.02 For those sections improved under the project, the overall estimated ERR is close to the 24% anticipated at appraisal despite the extended construction time and heavy cost overruns. This result emphasizes the critical importance of the highway to the development of the economy in general and of the Selva Central area in particular. Furthermore, the ERR reflects only the savings in vehicle operating costs. In addition, the project will play an important role in the integration of the area east of Huanuco into the wider economy and the opening up of markets for the agricultural produce of a number of fertile valleys. Previously, the growth of agriculture in these valleys had been impeded by their inaccessibility and the high cost of transport to the urban markets, especially Lima. The impact of the project on these areas will be enhanced by the improvement of feeder roads leading into the principal highway. 8.03 In addition to the construction and improvement of the main highway and feeder roads, the project has had a highly beneficial impact on transport planning and the identification and preparation of future investment programs. There is no doubt that, as a result of the project, local skills in geotechnical engineering, planning and feasibility studies have been substantially enhanced. This work has led, not only to the development of high priority programs, but also to a number of important technical recommendations of relevance for other similar projects in Andean countries. 8.04 In the light of the problems encountered during the implementation of this project, future Bank involvement in the Peruvian highway system should stress:  - 19 - (a) the need to establish a proper balance between maintenance and rehabilitation on the one hand and new construction on the other (a rehabilitation program is being prepared under the reformulated Loan 1196-PE); (b) continued support of carefully selected studies leading to identification of investment programs, better transport planning, and improved engineering practices; (c) the need for a more realistic assessment of MTC's capacity, and for a time schedule which fully recognizes local practices and conditions; and (d) the need to strengthen and reorganize the services responsible for road infrastructure, and for stricter adherence to the Bank's procurement procedures. The three latter areas are addressed in the Eighth Highway Project.   -20 - ANNEX 1 PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Conversion of Current Soles into Constant US$ 1. The rapid rate of inflation and devaluation of the Sol in Peru, especially since 1976, makes a comparison of the appraisal estimates and the actual final costs difficult. To minimize the problem, the Peruvian construc- tion cost index has been used to convert the actual final costs to costs in 1974, when the appraisal estimates were made. In Peru, the construction cost index is divided into two indices, one for labor and the other for materials. For purposes of this exercise, these two indices have been combined into a single average index. For the conversion of current soles into constant US$, the following procedure was used: (a) divide the construction cost index for the base year by the index for the year in which construction costs were incurred (this refers to total construction costs, including compensations for inflation, paid in that year); (b) multiply the construction costs by the fraction obtained from (a) above; (c) divide the result of (b) by the exchange rate in the base year (throughout 1974, the exchange rate was S/. 38.70 per US$); and (d) add the results of (a) through (c) for all years in which construction costs were incurred to obtain the total construction cost in 1974 US dollars. 2. The construction cost index used is the following: Yearly Average Index (1960 = 100) 1974 399 1975 474 1976 668 1977 989 1978 1,494 1979 2,523  - 21 - ANNEX 2 TRANSLATION National Planning Institute Mr. Shiv S. Kapur Director Operations Evaluation Department World Bank Washington, D.C. Dear Mr. Kapur: I write you in relation to the Project Completion Report on the Sixth Highway Project (Loan 1025-PE) which you sent for my comments. I am pleased to let you know that I have no observations to make to the PCR; I find it comprehensive in the analysis and of fairly balanced conclusions. I would like to thank you for sending the report which I consider very useful to my Directorate, I would also like to apologize for the delay in sending these comments. Sincerely, Ing. Wenceslao Urbina Director General Programming & Investments  - 22 - TABLE 1 PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Chronology of Works Execution Huanuco- Aguaytia Road Date of Commencement Date of Completion Appraisal Appraisal11 Contract 2/ Lot Estimate Actual Estimate - Contract Extensions Actual- A 12/74 10/75 3/77 01/78 08/79 04/80 B 12/74 10/75 3/77 01/78 09/78 07/79 C 12/74 10/75 9/76 07/77 11/78 03/79 E 12/74 10/75 12/76 10/77 10/78 03/79 1/ Individual Lot data not available. Based on contract duration used at time of bidding. 2/ Includes works which were authorized for construction under the respective contracts but finished during the guarantee period.  - 23 - TABLE 2 PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Huanuco - Aguaytia Road Estimated and Actual Cost of Civil Works (US$ million Equivalent-/) Appraisal Original Final % Difference - - - - Lot Estimate2/ Contract Cost Final/Appraisal Final/Original A 5.66 6.85 15.35 171.2 124.1 B 4.72 5.62 8.10 71.6 44.1 C 4.01 4.76 8.47 111.2 77.9 E 4.96 6.01 8.98 81.0 49.4 Total 19.35 23.24 40.90 111.4 76.0 1/ Expressed in constant 1974 US dollars. 2/ The appraisal report did not provide a breakdown of the costs by Lot. The total estimate in the appraisal report (US$20.5 million excluding contingencies), which included Lot D has been divided proportionally, in accordance with MTC's original division into Lots A-E in October 1974. The appraisal estimates in this table include physical contingencies.  - 24 - TABLE 3 PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Comparison Between Appraisal Estimates and Actual Project Costs (Expressed in Constant 1974 US$ millions) Appraisal Actual Estimate 1/ Cost A. Civil Works (i) Construction and improvement of 2/ 2/ Huanuco-Aguaytia road 19.35 - 40.90 - (ii) Improvement of feeder roads 2.88 1.34 - B. Consulting Services (i) Construction supervision 2.30 4.27 4/ (ii) Geotechnical studies 0.70 0.03 5/ (iii) Road user charges and pre- investment studies 1.10 0.44 C. Road Maintenance (i) Equipment 2.20 2.20 TOTAL 28.53 49.18 6/ 1/ Appraisal estimate includes physical contingencies but not price escalation contingencies as both totals are expressed in 1974 US dollars. 2/ Cost of Lots A, B, C and E only; Lot D was deleted from the project. (See Table 2) 3/ Bank-financed contracts only 4/ Supervision costs for the four sections of the Huanuco-Aguaytia road and feeder roads improved under the project. 5/ Bank-financed costs only. 6/ Partial total only.  - 25 - TABLE 4 PERU SIXTH HIGHWAY PROJECT PROJECT COMPLETION REPORT Scheduled and Actual Disbursements (US$ millions) Appraisal EstimateL/ Actual_!/ Actual as % of IBRD FY Amount % of Total Amount % of Total Appraisal Estimate 1975 4.1 15.6 0.4 1.5 9.8 1976 12.3 47.3 3.5 13.5 28.5 1977 20.3 78.1 10.9 41.9 53.7 1978 26.0 100.0 19.2 73.8 73.8 1979 23.8 91.5 91.5 1980 25.6 98.5 98.5 1981 26.0 100.0 lQO.O 1/ Cumulative totals  PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Traffic Estimates: Appraisal Compared with Actual Counts Appfaisal - Estimated Actual Actual Section ADT 1972 - 1979 % Trucks Annual Growth % ADT 1979 % Trucks Annual Growth % Huanuco-Carpish 530 851 45.8 7.0 644 41.9 2.8 Carpish-Huachipa 397 637 71.1 7.0 490 50.6 3.1 Huachipa-Pte. Pumahuasi 527 846 43.8 7.0 931 32.5 8.5 Pte. Pumahuasi-Tea Gardens 212 340 64.5 7.0 524 42.4 13.8 Tea Gardens-Aguaytia 200 321 80.0 7.0 524 42.4 14.8 Total Average 388 623 60.0 7.0 633 41.8 7.2 1/ The appraisal report did not break the road down into the sections shown in the table which correspond to the contracted lots. The 1972 ADT figures therefore represent an average, weighted by distance, calculated from the different sectional breakdown given in the appraisal report. Sources: MTC Final Report, appraisal report.  PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Representative Vehicle Operating Costs -/ Expressed in Soles of February, 1980 Soles per km. 2/ Type of Road Surface Savings per km Poor Gravel/ Poor Gravel/Earth Terrain Paved Gravel Earth Gravel to Paved to Gravel MEDIUM TRUCK Flat 164.4 257.6 419.5 93.2 161.9 Undulating 195.6 343.3 559.0 147.7 215.7 Mountainous 228.3 463.7 765.5 235.4 301.8 CARS AND LIGHT VEHICLES Flat 52.3 69.2 88.2 16.9 19.0 Undulating 60.8 92.4 130.1 31.6 37.7 Mountainous 72.9 111.6 169.4 38.7 57.8 BUSES Flat 185.6 283.2 493.2 97.6 210.0 Undulating 224.1 379.1 666.5 155.0 287.4 Mountainous 269.0 531.8 923.-f 262.8 391.0 1/ Economic costs. 2/ Rate of exchange, February 1980: S/257.29 = US$1. Sources: MTC Office of Economic Studies, "Estudio de Transportes para la Selva Central," original feasibility study, appraisal report.  -28 - TABLE 7 PERU SIXTH HIGHWAY PROJECT (LOAN 1025-PE) PROJECT COMPLETION REPORT Economic Rates of Return: Appraisal Estimates Compared with Completion Estimates Appraisal Estimate Completion Estimate Road Section Rate of Return (%) Rate of Return (Z) Huanuco - Carpish (Lot E) 26.0 26.5 Huachipa - Puente Pumahuasi (Lot C) 28.0 28.7 Puente Pumahuasi - Tea Garden (Lot B) 22.0 28.2 Tea Garden - Aguaytia (Lot A) 22.0 12.2 Total Average 24.0 23.2 1/ See footnote 2/ to Table 2. Source: See text.   J p -.~~~~~~~~~. L.12 PT. PS ;e7 7. \'PERU N)|. t SOUTH< / , C 0 0 M B A E c UA DR I/ El AltL 3 ~~~/ 7 a. ALchsoM I' 2--2\ Joen4A<,+gwo Yambe>oboubu vf2 5 P- S4t 0 A<.. '-- -/ GOA ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~' 0i ".- . ',j- I ffc GULF OF 4 // , ;;; I 8' c~~~- SAIN ROd < '' 12@ 1'@