Bolivia - Third Livestock Development and Agricultural Credit Projects

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 3996 PROJECT PERFORMANCE AUDIT REPORT BOLIVIA THIRD LIVESTOCK PROJECT (CREDIT 261-BO) & FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) June 23, 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. WEIGHTS AND MEASURES 1 hectare (ha) = 10,000 m2 = 2.47 acres 1 kilometer (km) = 0.62 miles 1 square kilometer (km2) = 0.39 square miles = 100 ha 1,000 kilograms (kg) = 1 metric ton = 0.98 long ton GLOSSARY OF ABBREVIATIONS ALPD - Agricultural and Livestock Project Division BAB - Agricultural Bank of Bolivia BE - State Bank CNECA - National Commission for Sugarcane Studies CB - Central Bank DESEC - Center for Social and Economic Development IDB - Inter-American Development Bank LPD - Livestock Project Division MPAA - Ministry of Peasant and Agricultural Affairs NIS - National Institute of Statistics FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT BOLIVIA: THIRD LIVESTOCK PROJECT (CREDIT 261-BO) AND FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) TABLE OF CONTENTS Page No. Preface ......................................... i Basic Data Sheet .. ... . . . .......... .... .... ........ ii Highlights .... -..... ................. .. . ................... iv PROJECT PERFORMANCE AUDIT MEMORANDUM I. SUMMARY . ................ o ...... ...... 1 . .. Background ..... ........ ......... ......... .......... 1 The Projects ............ . . . ........... 2 Implementation of the Projects ......... ......... 3 Impact of the Projects ......... ........ 5 II. MAIN ISSUES - ........ ........... 6 A. Institution Building .................... ....... 6 B. Indexation of Subloans .............. .. ....... 8 C. The Projects' Achievements ........ . 10 PROJECT COMPLETION REPORT A. Third Livestock Project (Credit 261-BO) ........... 13 I. Project Background ............................ 17 II. Project Formulation ............................... 19 III. Project Implementation ..... o ......... ...... 21 IV. Project Achievements ........................ 23 V. Project Impact .-.................................... 27 B. First Agricultural Credit Project (Credit 561-BO) ....... 33 VI. Project Background ...... . ........ .............. 0 35 VII. Project Formulation - ............ .............. 35 VIII. Project Implementation ...................... 38 IX. Project Achievements ........................... 40 X. Project Impact .... ............. ............. 47 XI. Special Issues ............... ............ . .... . 50 XII. Conclusions ............. ........... ......... .. 56 Annex 1: Tables 1 - 5 .......................... .......... 58 Maps: IBRD 11241 IBRD 2897R1 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 BOLIVIA: THIRD LIVESTOCK PROJECT (CREDIT 261-BO) AND FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) PREFACE This is a performance audit of the Third Livestock Project and the First Agricultural Credit: Project, for which Credits 261-B0 and 561-BO were approved in June 1971 and June 1975 in the sums of US$6.8 million and US$7.5 million respectively. Both projects sought to achieve the same objectives, were administered by the same project unit and were closed at the same time, in December 1980 after cancellation of US$170,000 (Credit 261-BO) and US$60,000 (Credit 561-BO). The audit report consists of an audit memorandum prepared by the Operations Evaluation Department and a Project Completion Report (PCR) dated August 1981. The PCR was prepared by the Latin America and the Caribbean Regional Office on the basis of a country visit in March 1980 and of a Com- pletion Report prepared by the Project Unit within the Agricultural Bank of Bolivia. The audit memorandum is based on a review of the Appraisal Reports (Nos. PA-51a and 695a-BO) dated May 11, 1971 and June 6, 1975, the President's Reports (Nos. P939-BO and 1653-BO) of May 1971 and June 9, 1975, the Project and Credit Agreements dated June 25, 1971 and June 20, 1975, the audit report (OED report no. 537) of the First and Second Livestock Projects, dated Octo- ber 11, 1974, and the PCR; correspondence with the Borrower and internal Bank memoranda on project issues as contained in relevant Bank files have been consulted, and Bank staff associated with the projects have been interviewed. A copy of the draft report was sent to the Borrower on April 22, 1982 for comments. However, none were received. On the basis of this abbre- viated procedure, the audit finds that the PCR covers adequately the project's salient features, and the PPAM agrees with the conclusions. In addition to summarizing the objectives and results of the projects, the PPAM expands upon two main issues--institut:Lon-building and indexation of subloans--because of their importance to this as well as other agricultural credit projects in the same region.  - ii - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET BOLIVIA: THIRD LIVESTOCK PROJECT (CREDIT 261-BO) KEY PROJECT DATA Appraisal Actual or Actual as % of Estimate Estimated Actual Appraisal Estimate Project Costs (US$ million) 11.0 15.9 145 Credit Amount (USS million) 6.8 6.6 97 Cancelled )- 0.2 - Repaid to ) December 31, 1981 - - Outstanding to ) - 6.6 - Date Board Approval 06/01/71 06/01/71 - Date Effectiveness 09/15/71 09/15/71 - Date Physical Component Completed 12/31/76 12/31/79 155/a Proportion then completed (%) 100 100 Closing Date 06/30/77 12/31/80 172a Economic Rate of Return (%) 18 10 - Number of Beneficiaries 400 1,282 320 CUMULATIVE DISBURSEMENTS FY72 FY73 FY74 FY75 FY76 FY77 FY78 FY79 FY80 FY81 Appraisal estimate (US$ million) 0.7 2.4 4.2 5.6 6.4 6.8 - - - Actual (US$ million) 0.3 1.9 2.5 5.4 5.8 5.8 6.0 6.0 6.0 6.6 Actual as % of estimate (US$ million) 43 79 60 96 91 85 88 88 88 97 Date of final disbursement 03/31/81 MISSION DATA Date No. of Manweeks Performar7ce Types of Mission (Month/Year) Persons in Field Ra nb Trend/c Probleas/d Appraisal L1/69 5 15 - - - Supervision I )9/71 2 1 - - - Supervision II )4/72 2 1 - - - Supervision III LO/72 2 1 - - - Supervision IV )2/73 1 2 - - - Supervision V 31/74 1 1 2 1 T, P Supervision VI )6/74 1 1 3 3 M, 0 Supervision VII 12/74 1 2 2 1 T Supervision VIII 02/75 1 2 3 1 F, M Supervision IX L1/75 1 2 2 1 1, F Supervision X 05/76 2 2 2 1 P, M Supervision XI L/76 3 3 2 1 P, M Supervision XII 02/77 2 3 2 1 0, M Supervision XIII 08/77 1 2 2 1 0, M Supervision XIV LO/77 2 4 2 1 M, T Supervision XV 05/78 2 2 2 2 M, T Supervision XVI LO/78 2 2 2 1 P, T Supervision XVII L1/78 1 2 - - - Supervision XVIII 02/79 1 0.5 2 2 P, F Supervision XVIX L0/79 2 2 2 1 P, T Supervision XX 05/80 1 1 2 2 P, T Supervision XXI 12/80 1 0.5 2 2 P Completion 03/80 1 1 - 1 - Total 53 OTHER PROJECT DATA Borrower Republic of Bolivia Executing Agency Banco Agricola de Bolivia (BAB) Fiscal Year of Borrower January 1 - December 31 Name of Currency (Abbreviation) Bolivian Peso ($b) Currency Exchange Rate Appraisal Year Average US$1 - $bll.88 Intervening Year Average US$1 - $b20.0O Completion Year Average US$1 - $b25.00 Follow-on Project Name First Agricultural Credit-Project Credit Number 561-80 Credit Amount (US$ million) 7.5 Credit Agreement Date 06/20/75 /a Calculated from date of Board Apptoval. /b 1 = Problem free or minor problemv; 2 - moderate problems; 3 - major problems. /c 1 = Improving; 2 = Stationary; 3 . Deteriorating. /d F = Financial; M = Managerial; T = Technical; P - Political; 0 = Other.  - iii - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET BOLIVIA: FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) KEY PROJECT DATA Appraisal Actual or Actual as % of Estimate Estimated Actual Appraisal Estimate Project Costs (US$ million) 11.0 11.3 103 Credit Amount (US$ million) 7.5 7.4 99 Cancelled )- 0.1 - Repaid to ) December 31, 1981 - - Outstanding to ) - 7.4 - Date Board Approval 06/19/75 06/19/75 - Date Effectiveness 09/20/75 12/15/75 - Date Physical Component Completed 06/30/79 12/31/80 117 a Proportion then completed (%) 100 100 Closing Date 06/30/79 12/31/80 117,a Economic Rate of Return (%) 22 10-15 - Number of Beneficiaries 4,000 1,139 28 CUMULATIVE DISBURSEMENTS FY76 FY77 FY78 FY79 FY80 FY81 Appraisal estimate (US$ million) 2.1 5.5 7.0 7.5 - - Actual (US$ million) 0.1 1.7 3.2 5.0 5.0 7.4 Actual as % of estimate (US$ million) 5 31 46 66 66 -99 Date of final disbursement 03/31/81 MISSION DATA Date No. of Manweeks Performance Types of Mission (Month/Year) Persons in Field Ratinit- Trend/ ProblemsZd Identification 06/72 4 12 - - Pre-appraisal 02/73 4 12 - - Appraisal 06/74 6 12 - - Supervision I 11/75 1 1.5 1 2 M Supervision II 04/76 2 2 2 1 M Supervision III 11/76 3 3 2 1 M Supervision IV 02/77 2 2 2 2 F, M Supervision V 10/77 2 2 2 2 M, T Supervision VI 04/78 2 3 2 2 M, T Supervision VII 11/78 1 1 2 1 P, T Supervision VIII 02/79 1 /e - - - Supervision IX 10/79 2 2 2 1 P, F Supervision X 06/80 1 1 2 2 P, F Supervision XI 12/80 1 /e 2 2 P Completion 03/81 1 1 - - Total 53.5 OTHER PROJECT DATA Borrower Republic of Bolivia Executing Agency Banco Agricola de Bolivia (BAB) Fiscal Year of Borrower January 1 - December 31 Name of Currency (Abbreviation) Bolivian Peso ($b) Currency Exchange Rate Appraisal Year Average US$1 - $b20.00 Intervening Year Average US$1 - $b25.00 Completion Year Average US$1 - $b25.00 Follow-on Project Name Second Agricultural Credit/f /a Calculated from date of Board Approval. /b 1 = Problem free or minor problems; 2 = moderate problems; 3 = major problems. /c 1 = Improving; 2 = Stationary; 3 = Deteriorating. /d F = Financial; M = Managerial; T = Technical; P = Political; 0 - Other. /e Less than one week. /f At identification stage.  - iv - PROJECT PERFORMANCE AUDIT REPORT BOLIVIA: THIRD LIVESTOCK PROJECT (CREDIT 261-BO) AND FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) HIGHLIGHTS These were the third and the fourth projects assisting the Banco Agricola de Bolivia (BAB) credit operations in the agricultural sector. The Third Livestock Project provided for lending to farmers for the purchase of cattle in the lowlands and sheep in the Altiplano, technical services for livestock development, and construction of new slaughterhouses and cold storage facilities. The First Agricultural Credit Project, which aimed at developing beef and sugarcane in the lowlands, annual crops and grapes in the valleys and sheep in the Altiplano, concentrated on lending to less privileged regions and the rural poor. The projects included an important institution-building component to remedy the main problems which had been identified at completion of the first two projects: inadequate meat pricing and marketing policies and the deteriorating financial situation and poor management of BAB. These problems were expected to be resolved through consultant studies and subsequent imple- mentation of their recommendations. Insufficient interest rates during periods of rapid inflation and subsequent decapitalization of BAB were to be tackled by indexing of loans extended to commercial farmers. At completion of the two projects (three years behind schedule for Credit 261-BO and 1-1/2 years for Credit 561-BO), 2,421 subloans had been made compared with an appraisal estimate of 5,000. Credit allocations were close to appraisal expectations, with beef production receiving about 70% of the total lending, sheep production 18%, grapes 8% and annual crops 4%. Although production impact at farm level is still difficult to determine since a monitoring and evaluation system was only recently established, it can be concluded that the projects had a positive impact on livestock and agricul- tural production. Overall ERR's were re-estimated at 10% for the Third Livestock Project and LO-15% for the First Agricultural Credit Project, compared with appraisal estimates of 18% and 22% respectively. While the two projects were successful in diversifying lending operations and increasing financial assistance to less privileged regions and farmers, their institution-building objective fell short of expectations. The projected studies were carried out but produced no changes or had little effect on the beef industry or on agricultural credit policy and procedures. Continuous government intervention in BAB's operations was the main reason for the failure to improve the institution's efficiency. At project comple- tion, it was found that BAB was in a more difficult situation than at the time of appraisal. - v - The main lessons of these projects are: - studies have a valuable role to play in development but it is unreasonable to expect that the Government and the Bank can commit themselves to implement consultant's recommendations within a relatively short time (PPAM, paras. 24-26, and PCR, paras. 11-16); - a simple agricultural credit project is not the appropriate vehicle to bring about changes of macro-economic policies (PPAM, para. 26 (b) and PCR, paras. 11.02-11.03); - improving performance of a credit institution is difficult when it lacks autonomy and is subject to political interventions (PPAM, para. 26 (c) and PCR, paras. 11.04-11.06); and - indexing is doomed to fail unless it is fully accepted by the Government, understood by borrowers, and applied to all sectors and to all credit resources (PPAM, paras. 27-31). PROJECT PERFORMANCE AUDIT MEMORANDUM BOLIVIA: THIRD LIVESTOCK PROJECT (CREDIT 261-BO) AND FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) I. SUMMARY Background 1. Agriculture produces about 25% of the country's gross Domestic Product (GDP) but employs more than half of the work force. One of the problems affecting the agricultural sector is the imbalance between demo- graphic and natural resources in Bolivia's three main geographic regions: (i) the cold and dry Altiplano, which comprises 44% of the country's culti- vated land and holds 56% of the population, (ii) the temperate and humid Valleys, with 33% of culcivated land and 30% of the population, and (iii) the Tropical Lowlands with 23% of the cultivable land and 14% of the population (see map). 2. Livestock production is the most important agricultural activity; beef production is mostly concentrated in the tropical and little populated lowlands of the east and the Andean foothills, while sheep production is limited to the Altiplano, where it is part of the subsistence small farm sys- tem practiced by the predominantly Indian population. Livestock development suffers from the same general problems which adversely affect Bolivian agri- culture: Use of improved inputs is the lowest in Latin America; technical assistance and credit reach no more than some 10% of all farmers; supporting institutions are weak and suffer from recurrent budgetary deficits. 3. The first two IDA credits for livestock development (Credit 107-BO of May 1977, for US$2.0 million, and Credit 171-BO of January 1970 for US$1.4 million) helped finance a livestock development program in the Department of Beni in the tropical lowlands. This program, administered by a specially created Livestock Project Division (LPD) in the agricultural credit institu- tion (Banco Agricola de Bolivia, BAB), helped finance on-ranch investments such as fencing, corrals and breeding stock and provided short-term credit, especially for the purchase of fattening cattle. Most of the program objec- tives were achieved about. one year earlier than expected. Funds from the two credits were used to make 300 loans to ranchers. The LPD has, on the whole, performed satisfactorily in appraising and supervising ranchers' investment programs, in spite of logistical problems in the Beni and difficulties in obtaining and retaining competent livestock specialists. The BAB provided local funds for the program, as stipulated by the Credit Agreement, undertook changes in its accounting procedures and carried out the land tenure and beef marketing studies required by IDA. However, BAB's financial situation steadily deteriorated during project implementation, largely due to measures taken by the government: imposition of unprofitable lending programs, reduc- tion in the interest rate and devaluation. - 2 - 4. The Project Performance Audit Report (PPAR) issued in October 1976 (OED Report No. 537) concluded that the two projects had a positive impact on cattle management and beef production. The projects' ERR was found similar to the 17-22% estimated at appraisal. The PPAR, however, concluded that the institutional impact of the projects was slight and that BAB's internal efficiency had not improved. In addition, other negative factors to livestock development, such as Government control on the price of beef and export restrictions, as well as marketing and air transport difficulties, remained unresolved. The Projects 5. The projects now under review were the third and the fourth with the Banco Agricola de Bolivia for the purpose of expanding credit operations in the agricultural sector. Although the Third Livestock Project became effective four years before the First Agricultural Credit Project, both sought to achieve the same objectives, were administered by the same project unit, and were closed at the same time (December 1980). 6. In 1969, following completion of the first two credits for livestock development, IDA appraised the Third Livestock Credit Project (Credit 261-BO). The following key aspects distinguished it from its two predecessors: the extension of the beneficiary groups to include small farmers in the Altiplano; the concentration on problems of marketing and processing; and the attempt to charge positive interest rates by indexation of the loan principal during periods of rapid inflation. As approved in June 1971, the project consisted of (i) loans for purchase of cattle and on-farm development to some 250 ranchers in the Beni region; (ii) loans to some 150 sheep owners and small- holder cooperatives in the Altiplano for on-farm development and purchase of improved sheep; (iii) technical services for both sheep and cattle develop- ment; (iv) employment of consultants to make recommendations to the Government on meat pricing and marketing policies, establishment of an adequate meat grading and inspection service, construction of additional slaughterhouses and creation of transport and marketing facilities; and (v) finance for the construction of new slaughterhouses and cold storage facilities. Total project cost was estimated at US$11 million, with IDA contributing US$6.8 million. The project was expected to be disbursed in 5.5 years and to be closed by June 1977. 7. The interest rate to be charged on the loans was 12%, with a four year period of grace and repayment over 12 years. Assurances were obtained during negotiations that "if the rate of inflation in Bolivia should rise above 8% per annum during the Credit commitment period, BAB would make no further loan commitments. . .until an agreement with IDA had been reached on arrangements for indexing to be applied in respect of such further loans".1/ 1/ Staff Appraisal Report, para. 3.38. -3 - 8. A Livestock Project was scheduled by the Bank for FY76 and a sepa- rate Agricultural Credit Project for FY74. In 1973, however, IDA decided that for project management simplicity, it would proceed with only one proj- ect. Therefore, the First Agricultural Credit Project (Credit 561-BO) was appraised in June 1974 and approved by the Board in June 1975 in the sum of US$7.5 million. Project objectives were to develop beef and sugarcane in the lowlands, annual crops and grapes in the Valleys and mutton and wool produc- tion in the Altiplano. Total project cost was estimated at US$12 million. The project somewhat differed from its predecessors by increasing its assis- tance to the less privileged regions and the rural poor. Of a total of about 4,000 subloans, 3,600 were expected to be made to farmers belonging to the lowest income stratum of the population. 9. Organization and management were the same as for the ongoing Third Livestock Credit Project, with some expansion of supervision and technical assistance capability to accommodate the wider range of activities and re- gions. In addition, BAB agreed during negotiations to adopt an overall reorganization plan based on the recommendations of consultants. The major financial innovation of the project was a new arrangement for indexing loans to commercial farmers; the system agreed upon was to leave to the sub- borrowers the option of choosing between the two following alternatives: "(i) interest rates of 4% calculated on the outstanding balances adjusted on the basis of the cost of living index compiled by the National Institute of Statistics or (ii) interest rates of 14% calculated on the outstanding bal- ances adjusted on the basis of variations of the rate of exchange between the Bolivian peso and the US dollar".!1 Subloans to subsistence farmers were not subject to indexing and were to bear a nominal interest of 12%. 10. At Board presentation of the Third Livestock Credit Project, the following main issues were discussed: (i) the limited impact of the project on small farmers since most of the loans would be made to large ranchers and (ii) the possible difficulty of introducing an indexing system. At Board presentation of the First Agricultural Credit Project the possible competition between BAB and commercial banks was discussed. Implementation of the Projects 11. The Third Livestock Credit Project started slowly as a result of the then prevailing low beef price and disputes between the Beni ranchers, the Government and IDA on interest rates and indexing (see paras. 27-31). After two years actual disbursements only reached 10% of appraisal estimates. In 1974 as beef prices rose sharply, credit demand increased despite the continuing interest rate problem, but the ranchers refused to make interest payments on the basis of the indexing system. The Bank then requested from Government that no further withdrawal applications be submitted to IDA until 1/ Staff Appraisal Report, para. 3.34. BAB had resolved the application of the indexing formula. The project entered the Bank's problem project list and remained almost continuously a problem project until 1975. In September 1975, the new indexing system proposed under the First Agricultural Credit was finally approved by the ranchers; disburse- ments then increased sharply, and the credit allocation for beef development was fully committed by March 1976. 12. Credit for sheep development was delayed by the lack of land titles, which hampered the farmers' access to medium-term credit. In 1978, BAB changed its lending policy toward small farmers. Land titles as collateral were no longer required and chattel mortgages were accepted as security. As a result, actual disbursements as a percentage of appraisal estimates rose from 44% in 1975 to 87% six months later. 13. Out of four studies financed under the project, three were imple- mented: the meat marketing and production study, the reorganization of BAB, and the study on new meat processing plants for Santa Cruz. The study on changes in meat marketing was taken up by UNDP and is expected to be com- pleted by the end of 1982. Delays in implementing the studies were the main reason for postponing the project closing date several times. The project was finally closed by end December 1980, 3-1/2 years later than expected at appraisal. 14. The First Agricultural Credit Project became effective in December 1975, but disputes over interest rates slowed its initial progress as for the Third Livestock Project. When the new indexing system (para. 11) became effective, the number of loans for beef and grape production rose sharply. No requests were received for sugarcane development because world sugar prices were depressed and because competing lines of credit were available through the Central Bank. Neither were loan applications received from annual crop farmers because of unawareness of credit availability under the Bank project, the existence of alternative sources of credit, and lack of land titles; nor were applications submitted for sheep development, since suffi- cient funds were still available under the Third Livestock Project. In 1978 following BAB's agreement to accept chattel mortgages, subloan commitments for sheep and annual crops increased substantially. The unused allocation for sugarcane was reallocated to support beef and sheep development. By the end of 1979 all project funds had been committed, but disbursements slowed down mostly because of the shortfall in Government counterpart funds. The consul- tants completed their study of BAB and produced detailed operational manuals which were a solid basis for a reorganization plan. A commission was formed to study the report, but its findings are yet to come. The project was fi- nally closed, fully disbursed, in December 1980, 1-1/2 years behind schedule. 15. At completion of the two projects in December 1980, 1,282 subloans had been made under the Third Livestock Project and 1,139 under the First Credit Project, compared with an appraisal estimate of 400 and 4,600 respec- tively. Credit allocations were close to appraisal expectations: beef produc- tion received 70% of the total lending (64% estimated at appraisal), sheep production 18% (20% at appraisal), grapes 8% and annual crops 4% (as projected at appraisal). No subloans were made for sugarcane development. -5- Impact of the Projects 16. Production impact at farm level is difficult to determine since a monitoring and evaluation system was only established at the end of the First Credit Project, and no farm records were kept during the first implementation years. 17. About 45% of the beef loans were used for the purchase of livestock; more funds than expected were allocated for fattening operations; and some funds were used for unspecified and relatively non-productive purposes. Fencing, pasture improvement and machinery also constituted important invest- ments. The economic rate of return (ERR) on beef development is estimated at 9 to 17%, according to farm size and type of investment, compared with an overall 22% expected at appraisal. 18. For the sheep component, the size of sub-borrower farmers proved to be larger, consequently credit needs per borrower greater, and the number of beneficiaries smaller than anticipated. The major investments were purchase of breeding sheep. The recalculated ERR on sheep development varies from 13% to 20% (27% estimated at appraisal). 19. For vineyards, the average size of loans was smaller than antici- pated at appraisal. Most of the loans financed vineyard establishment (172 ha) and rehabilitation (35 ha). The ERR on investments in vineyards is now estimated at 26: to 49% (27% anticipated at appraisal) mostly due to a rise in prices paid to producers as demand for grapes expanded. 20. For annual crops, some 60% of the credit was used for expanding potato production, 20% for purchase of oxen and the remainder for agricultural equipment. Most of the Subloans went to small farmers. No rate of return was recalculated for annual crops. 21. The overall rate of return of the projects is expected to be less than the individual returns indicated above, mostly because the level of arrears remains high (about 20%) and beneficiaries in arrears may be consid- ered as not fully realizing expected profits. In addition, some funds were diverted from intended use. Overall ERRs were re-estimated at 10% for the Third Livestock Project (18% at appraisal) and 10-15% for the First Agricul- tural Credit Project (22% at appraisal).l' 22. The First Agricultural Credit Project succeeded in increasing lend- ing operations in less privileged regions but was less successful in its equity objectives, since the total number of beneficiaries proved to be significantly lower than. anticipated at appraisal. However, the fact that a number of small farmers had access to credit for the first time can be consid- ered an important achievement of the project. 1/ OPS considers that farmers in arrears have received some of the expected benefits and that the ERR should be intermediate between these and the ones presented in the PCR. - 6 - 23. The institution-building objective of the two projects fell far short of expectations (see paras. 24-26). The projected studies were carried out but produced no changes or had little effect on agricultural credit policy and procedures. BAB remained in disarray. With one exception, the attempt to involve commercial banks in agricultural development lending failed, mostly because of the Government's discouraging interest rate policy. MAIN ISSUES A. Institution Building 24. The striking feature of these two projects is the contrast between their relative success in terms of lending operations and increased assistance to less privileged regions and farmers on the one hand, and their failure to substantially improve institutions and sector policies on the other. Develop- ment issues and weaknesses of agricultural credit policies, procedures and institutions were properly identified at the end of the first two livestock projects (para. 3). As a result, the two follow-on projects included an important institution-building component. The main issues identified were the following: (a) inadequate meat pricing policy and absence of differential prices for different grades and qualities of beef; (b) insufficient slaughter- ing and marketing facilities; (c) inadequate interest rates; and (d) deterio- rating financial situation and poor organization and management of BAB. 25. Remedial actions were proposed by the Bank and accepted by the Government at negotiations, and pertinent covenants were included in Credit Agreements. On issues (a) and (b), the Government was to engage consultants to carry out studies and make recommendations within 12 months of the date of Credit effectiveness. Within nine months of receipt of the consultant's report, the Government was expected to adopt a beef price policy acceptable to IDA to provide for differential prices for meat of different grades and qualities. The question of interest rates was tackled by a covenant (in the Third Livestock Project) stipulating that, if the rate of inflation should rise above 8%, BAB would make no further loan commitments until an agreement had been reached with IDA on adopting an indexing system. Finally, BAB's problems were to be reviewed by consultants (under the First Credit Project), whose recommendations were to be implemented within six months after issuance of their report. 26. In sum, with the exception of the indexing problem (see paras. 27-31), all issues identified at appraisal were expected to be resolved through studies and subsequent and rapid implementation of the recommendations made by consultants. In fact, although the studies were carried out, and despite IDA's efforts during supervision of the two projects, little or no action was taken following completion of the studies, and the same main issues are still unresolved. The lessons to be learned from this experience confirm the observations already made in connection with other projects of the same kind: - 7 - (a) Studies have a valuable role to play in development but it is unreasonable to expect that both the Government and Bank can commit themselves to implement consultants' recommendations in a given relatively short time. Past experience shows that conclusions of a study are not necessarily accepted by the parties involved, and often raise important institutional and policy matters which need extended discussions, changes in legislation and involvement of several ministries. By imposing a time frame to implement policy reforms which go beyond the scope of the project and which could not be clearly defined at negotiations, the project covenants proved totally unrealistic. (b) In the audit's view, a simple agricultural credit project is not the appro riate vehicle to bring about changes of macro-economic policies The case of these two projects does not substantially differ from that of a number of other agricultural credit projects audited by OED.2/ The Government's perception of these two agricultural credit projects was essentially that of a lending operation to farmers, aiming at improving agricultural production and rural incomes. The Bank's objective, however, was somewhat different. The Bank wanted to use the project as a means to re- structure the entire beef industry. But many of the issues to be tackled under these projects were simply beyond the control of BAB, the main participating agency. Only strong Government commit- ment would have permitted some progress in resolving sectoral issues, but this commitment was obviously lacking since the very beginning of the projects. (c) Improvement of an agricultural credit institution is not an easy task since it cannot be isolated from the country's political and economic environment. Although the project was successful in intro- ducing some structural changes within BAB, continuous Government intervention in BAB's operations turned out to be one of the reasons for the failure to improve the institution's efficiency; decrees to lower interest rates in 1972 and 1978, uneconomic branches and arrears of other banks imposed on BAB, poor adherence to an indexing clause, insufficient financial contributions, continuous turnover of 1/ It is worth noting that Structural Adjustment Loans (SALs) are now considered by the Bank a more appropriate vehicle to bring about changes of macro economic policies, but SALs did not exist when these two proj- ects were appraised. 2/ Senegal--Second Agricultural Credit Project (0ED Report No. 3514, dated June 25, 1981). Pakistan--Third Credit for the Agricultural Development Bank (OED Report No. 2126, dated June 1978). Costa Rica--Second Agri- cultural Credit Project (OED Report, dated June 1981). - 8 - managers and lack of action regarding issuance of land titles were all Government measures--or lack thereof--which hampered both project implementation and BAB reorganization. At project comple- tion (1980), it was found that BAB was in a more difficult situation than the one prevailing at the time of appraisal of the Third Livestock Project (1971), but most of these shortcomings were due to intervention and the lack of autonomy. B. Indexing of Subloans 27. It was a common practice in the early 1970s to introduce an indexing clause in Loan/Credit Agreements of credit projects, particularly in Latin America where inflation rates were high. The objective of indexing is to sustain a positive rate of interest and prevent decapitalization of lending institutions. Covenants on indexing (see paras. 7 and 9) were included in both the Third Livestock Project and the First Agricultural Credit Project. 28. Soon after start-up of the Third Livestock Project, increasing inflation rates in Bolivia triggered the need to apply the indexing clause, but since indexing mechanisms were not then defined, lengthy discussions took place between Bank staff and Government officials. Agreement on indexing was finally reached in March 1973. The agreement, aiming at a real interest rate of 6% on subloans, provided for a nominal interest rate of 9% and for adjustment of the outstanding principal by a price index!/ less 3 percentage points. However, ranchers strongly opposed any indexing mainly on the grounds that (i) on-going credit for livestock development from the Inter-American Development Bank was not subject to indexing, (ii) the official cost of living index was based on expenditures of the urban population, whose outlook and way of life are different from those of farmers, and (iii) the price of meat was fixed by Government and never related to the La Paz cost of living index. 29. It was not before June 19751/ that the indexing issue was finally resolved, when an option was offered to farmers between (i) 4% interest on loan principal adjusted to the cost of living index or (ii) 14% interest on loan principal adjusted to the peso-dollar rate. With a few exceptions all beef cattle ranchers expressed their preference for the dollar clause option. The agreement, however, was not fully respected either by the Government, which in 1978 unilaterally lowered interest rates, or by farmers, who after two devaluations responded by increased defaults on repayments. 1/ Either the index of food prices established by the National Institute of Statistics, or the index of meat prices established by the Ministry of Industry, whichever is lower. 2/ By then the IDB loan for livestock development had been fully disbursed. - 9 - 30. The indexing experience gained under these two projects can be compared with similar experiences in Latin America and other agricultural credit projects audited by OED: (a) In Brazil under the First and Second Livestock Development Projects.! and the Grain Storage Project,2/ indexing arrange- ments were made with the Government after lengthy discussions. The Government, however, was under strong pressure from farmers to introduce softEr lending terms and finally acceded to their demand by establishing new lending programs without monetary correction or Bank participation. (b) The Uruguay Fourth Livestock Development Project-/ introduced an indexing arrangement based on an option between (i) the average increase in beef price during a 12-month period or (ii) the annual rise in the cost of living index. The system was accepted by the Government but was not properly applied since increases were not compounded, as they should have been, but added to calculate the adjustment of the outstanding balance of subloans. The Bank and the Government were aware that the system as applied weakened the effectiveness of indexing under conditions of high inflation (50% at that time). However, since Government policy was to keep beef prices substantially below those which would have prevailed in a free-trade situation, the Bank agreed that the application of the indexing clause would not be changed until Government policy no longer discriminated against the livestock sector. (c) Under the Paraguay Fourth Livestock Credit Project-4 the indexing of subloans was accepted by the Government despite the fact that it was a new practice in Paraguay. There were minor reactions to the system by farmers, who accepted the indexing principle because (i) there were no alternative funds available for medium- and long- term lending for livestock development and (ii) beef prices, which were not controlled by Government, were high during the project period. 31. From these experiences and that of the two projects in Bolivia, some lessons can be learned on indexing of subloans in the agricultural sector. First, there is no doubt that indexing is doomed to fail unless the 1/ OED Report No. 2402, dated April 1979. 2/ OED Report No. 2590, dated June 1979. 3/ OED Report No. 2572, dated June 1979. 4/ OED Report No. 3457, dated May 1981. - 10 - Government fully accepts and understands the monetary correction principle and strongly supports its application during the project life. Second, indexing cannot be applied only in a particular sector or subsector; the Bolivian and Brazilian experiences mentioned above failed partly because by applying indexing to livestock and grain storage the interest rates of these two subsectors soon became the highest in the country except for housing.I. Third, indexing is unanimously rejected by farmers when prices of outputs are controlled by government and generally distorted in favor of urban consumers. Fourth, alternative sources of financing which do not require indexing for similar programs (IDB and USAID in Bolivia and local sources in Brazil) proved to hamper introduction of an indexing system; coordination between the Govern- ment, the Bank and other financing institutions-/ appears to be a prerequi- site for successful introduction of a sound monetary correction system. Finally, moving from zero to full indexing on a gradual basis might be desir- able in countries which have no indexing experience. C. The Project's Achievements 32. When compared with the first two IDA credits for livestock develop- ment, the Third Livestock Credit Project and the First Agriculture Credit Project appear a failure to achieve all their objectives. Such a failure, however, is relative because the objectives of the two series of projects were entirely different. Under the first two IDA credits, project objectives were mainly restricted to lending for livestock development and to commercial farmers. The institutional aspect was limited to the establishment within BAB of the Livestock Project Division, which performed satisfactorily. Although the Bank had recognized the financial and managerial deterioration of BAB, no action was proposed or taken under the first two projects to remedy the situation. When the projects were closed, they were considered successful because disbursements took place faster than expected and the projects' ERRs were satisfactory. 33. By contrast, the two follow-on projects tackled institutional and sectoral issues which were and remained beyond their control. However the failure to restructure a major industry and to significantly improve institutions should not mask some of the projects' achievements like the agreement to accept chattel mortgages for smallholders instead of land titles, the reinforcement of the Agriculture and Livestock Project Division of BAB and the establishment of an evaluation and monitoring system. The projects were also successful in diversifying lending operations from the Beni Province with its large farms to smallholders in the Altiplano and Valleys regions. The 1/ In Bolivia, interest rates for livestock development were higher (in 1978) than for mining or industrial enterprises. 2/ in Bolivia, IMF appears to have disliked indexing in contrast to the Bank's preference for this mechanism. - 11 - project achievements in terms of number of subloans (more than 2,400 compared with 300 during the first two projects) and diversification of beneficiaries are not negligible either. In conclusion, it can be said that the projects were too ambitious in their expectations; but despite their failure to achieve all their objectives their impact on the economy proved to be greater than that of their predecessors.  - 13 - BOLIVIA THIRD LIVESTOCK CREDIT PROJECT (CREDIT 261-BO) AND FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) PROJECT COMPLETION REPORT August 31, 1981  - 15 - BOLIVIA THIRD LIVESTOCK CREDIT PROJECT (CREDIT 261-BO) PROJECT COMPLETION REPORT PART A  - 17 - PART A THIRD LIVESTOCK CREDIT PROJECT (CREDIT 261-BO) PROJECT COMPLETION REPORT I. PROJECT BACKGROUND Sector Setting 1.01 Agriculture is the largest sector in Bolivia accounting for almost half of the value added in commodity production. Despite a decline in the share of agriculture's contribution to GDP, Bolivia's economy remains agrarian in terms of source of personal income, provision of jobs and manner of social organization. Out of the total population of 5.5 million some 52% make their living directly from agriculture. The majority of farm families are of Indian extraction and their economy is based on subsistence- type agriculture. 1.02 Although the growth rate in sectoral output has been some 3.5% per year over the long run, per capita production of food has apparently declined. The value of agricultural imports, mostly cereals, vegetable oils and dairy products, now exceeds the value of sectoral exports, mainly sugar, timber, coffee end cotton. 1.03 Livestock production continues to be the most important activity, contributing 34% to the value of farm output. Beef production accounts for 55% of the value of livestock, with dairying and sheep being 14% and 7%, respectively. In addition to value as a final product, oxen are the dominant draft power inlut in crop husbandry. 1.04 Sheep production is concentrated in the Altiplano where it is an integral part of the subsistence small-farm system practiced by the predominantly Indian population. Dairying is also common near the urban centers of the highlands but is more specialized in the sheltered valleys. The major areas of beef production are the tropical lowlands of the east and the Andean foothills. In the former area, extensive grazing on large properties is the usual form of enterprise management whereas integrated livestock production, using improved pastures and crop residues, is becoming common in the Santa Cruz district. 1.05 The livestock industry suffers from the same general problems of Bolivian agriculture. Use of improved inputs is the lowest in Latin America and among the lowest i. the world. Technical assistance and credit supply reach no more than some 10% of farmers. Supporting institutions are weak and suffer from recurrent budgetary deficiencies. 1.06 In addition, farms in the crowded Altiplano are small and soils are poor with a frequently hostile climate. Although the low standards of livestock husbandry provide opportunities for gains in productivity, these conditions necessarily limit development in the highlands to little more than poverty alleviation. On the other hand, the lowlands offer considerable scope for expansion of farmed area as well as major productivity gains. The binding constraint on lowland development has been the physical difficulty associated - lb - with cattle transport. Large areas of seasonally flooded plains impose unusually high cost on road building and maintenance. Because of low popula- tion density and poor extraction rates, the supply of an adequate road network has not been economically worthwhile. As a result, the movement of finished cattle has been difficult and costly, with high weight loss and mortality. The lack of roads also restricts farm credit supervision and the delivery of technical services. The demand for these services is also weak because of the high incidence of absentee ownership. 1.07 Despite frequent changes in Government, public policy for agricul- ture has remained largely unchanged over the past two decades. Policy objec- tives are the usual intentions to increase exports, decrease imports, and help the rural poor. Policy instruments are the familiar provision of technical services and farm credit as well as price controls, direct state trading and involvement in agricultural development projects. In general, policy management has not been successful because the goals have been too ambitious for the available financial and human resources. For instance, price control manage- ment has not been based on sufficient analytical input and institutional impact has been severely reduced by budgetary difficulties. As a result, Bolivian agriculture is unusually dependent upon external aid flows and on foreign development institutions working within the country. The Project 1.08 The project addressed productivity issues in the livestock subsector by providing technical assistance linked to farm credit. Issues in processing and marketing were dealt with by providing funds for construction and renova- tion of slaughterhouses and for consultant studies of meat marketing. Project objectives were to (a) increase beef exports from the Beni and nearby areas, (b) increase wool production in the Altiplano, and (c) increase mutton produc- tion for local consumption. 1.09 Because of the structure of farming in the two areas, the Altiplano sheep component was essentially to be a small-farm development effort whereas the Beni beef component would serve extensive grazing properties. 1.10 The project was large in size yet small in anticipated coverage. At the time of appraisal, total farm credit available was in the order of US$20 million for 500,000 farmers. The proposed project of US$11 million was expected to reach 250 individual cattle ranchers and 150 sheep properties in cooperatives of some 20 to 50 members, to give a total of around 3,000 beneficiaries. At an overall average investment of only US$2,500 per farm, these figures reflected the inadequacy of credit resources available. 1.11 The credit was the third livestock project supported by IDA in Bolivia. The previous '.wo Credits 107-BO of 1967 for US$2 million and 171-BO of 1970 for US$1.4 million, helped finance a livestock development program in the Department of Beni in the lowlands. This program, administered by a specially created Livestock Project Division (LPD) 1/ in the Bolivian 1/ The Livestock Project Division (LPD) later became the Agricultural and Livestock Project Division (ALPD) to incorporate the First Agricultural Credit Project into the credit program. - 19 - Agricultural Bank (BAB), helped finance on-farm investments such as cattle purchase, fencing and yards, as well as short-term credit. The Project Performance Audit Repori: (SEC M74-710, October 11, 1974) concluded that most of the program's production objectives were achieved, but that institu- tional gains had been slight. The present project, 261-BO, was initiated to continue the program, but with a broader perspective. II. PROJECT FORMULATION Identification and Preparation 2.01 Toward the end of 1968, when lending under the First Livestock Credit (107-BO) was in full operation, it became apparent that funds would be committed faster than anticipated. Bolivia approached IDA for an interim credit of the same type while a more comprehensive project was prepared. The Interim Credit (171-BO) was approved by the IDA Board without an appraisal as the Second Livestock Project. The Project Director then prepared the Third Livestock Credit, which was to become Credit 261-BO. Appraisal and Approval 2.02 In contrast to the speedy passage of the Interim Credit, the project took 18 months to go from appraisal in November 1969 to Board approval in June 1971. The delay in presenting the project was due to country conditions, wh:'ch made it difficult for the Bank to develop a constructive dialog with the Government on difficult policy issues. During one period in 1970, Bolivia had five presidents in four days. The issues to be resolved were: (a) adequate compensation payments for nationalization of assets of GuLf Oil Corporation and two US-owned mining corporations; (b) arrears in Bolivia's US dollar bond payments to the Bank; (c) the budgetary support which the GOB could be expected to give to the BAB; (d) the level of interest rates and the terms and purpose of operating credit to be provided; and (e) the effect of state intervention in marketing particularly through beef export controls and unprofitable charter rates for air transport operators. None of these issues was entirely settled, but sufficient progress was thought to have been made by June 1971 for the project to be approved by the Board, despite a last minute attempt for postponement over the compensation item. - 20 - Project Description 2.03 The project consisted of two subprojects: (a) Beef and Sheep Property Development, providing for: (i) subloans for on-farm development on some 250 Beni area properties; (ii) subloans to some 150 property owners, generally as cooperative groups; and (iii) technical services for both sheep and cattle development; and (b) Cattle Slaughtering and Meat Marketing, providing for: (i) employment of consultants to make detailed recommendations to Government on meat policies, establishment of an adequate meat grading and inspection service, construction of additional slaughterhouses and creation of adequate transport and other marketing facilities; and (ii) finance for construction of new slaughterhouse and cold storage facilities and for renovation of existing ones and for training up to six meat inspectors. 2.04 The interest rate to be charged on the subloans was 12%, with a 12-year repayment period and four years' grace. Assurances were obtained during negotiations that, if the rate of inflation in Bolivia should rise above 8% per year during the life of the credit, the Government would make arrangements with IDA for indexing of future subloans. 2.05 Key aspects of the project distinguished it from its two predecessors. These were (a) the extension of the beneficiary groups to include small farmers in the Altiplano, (b) the concentration on problems of marketing and processing, and (c) the attempt to charge positive interest rates by provision for indexation of loan principal during periods of rapid inflation. For the first time in the program, credit was made available for sheep development. The Credit Agreement required the GOB to make important changes in its beef price policy, meat inspection services and grading system within nine months after the submission of the consultant's study of these problem areas. It further required that the Borrower should abolish beef export restrictions except for temporary emergencies. 2.06 Some 74% of total project c3st was allocated to on-farm investment with a further 12% to provide incremental working capital. Of the balance, some 6% was to be used for technical assistance to farmers, 7% for slaughter- house construction and renovations, with a final 1% for consultant services. Two-thirds of the farm credit was to go to beef enterprises, and one-third to sheep enterprises. The total project cost was estimated at US$11 million, with IDA's contribution of US$6.80 million being the expected foreign exchange component of expenditures under each category. BAB's share was set at US$2.49 million, with the balance to come from beneficiaries' own funds. - 21 - 2.07 Ranch development was to continue the lending program of the IDA's previous projects. Subloans would pay for improved breeding cattle and property improvements such as fences, water facilities, yards, barns, housing, pasture improvement and equipment. Sheep farm development was to follow the same strategy with replacement of "criollo" animals by heavier bodied improved breeds and with replacement of overgrazed native pastures with introduced species. Provision of technical services would be continued through the Livestock Project Division of BAB established under the earlier projects. 2.08 The organization and management system was a continuation of the previous arrangement. The Government was the borrower. All on-lending was to be done through BAB, assisted by a Loan Committee, for the approval of farm plans. A new item was the establishment of a Project Commission for the implementation and supervision of consultant studies and for the preparation of recommendations for expenditures under the processing and marketing components of the project. 2.09 Financial and economic rates of return for both sheep and cattle development were estimated at around 18%. Incremental production was 2xpected to rise to 10,400 tons of carcass beef, 1,450 tons of carcass mutton and 575 tons of wool worth some US$6 million annually at farmgate prices for an investment: of US$11 million. HII. PROJECT IMPLEMENTATION 3.01 The project started slowly, with actual disbursements reaching only 10% of appraisal estimates after two years. The previously high level of demand for livestock credit had fallen suddenly because of an interest rate dispute between the Ben". ranchers, the GOB and IDA. Funds left from the earlier project proved sufficient to cover disbursements during the first project year. Although these were fully utilized by the second year, the interest rate dispute wi.dened over the indexation issue. 3.02 As export prices rose sharply in 1974 local price increases were permitted by the state. Credit demand then rose despite the continuing interest rate problem. By mid-1974, Beni ranchers had still not begun to make interest payments on the basis of indexed principals in accordance with the Supreme Decree of December 5, 1975, and in compliance with Section 4.03 of the Credit Agreement.. In July 1974, IDA wrote to the Government 1/ and requested that no further withdrawal applications be submitted to IDA under the project until (i) BAB resolves the confusion among beneficiaries regarding the application of the correct indexing formula; and (ii) BAB collects the additional interest payments due from sub-borrowers since November 1, 1973. 1/ Letter from G.K. Wiese to Victor Castillo Suarez, Minister of Finance, July 9, 1974. - 22 - 3.03 This situation continued for more than one year until September 1975 when the indexing system agreed for the new Agricultural Credit Project (561-BO) was applied to the Third Livestock Project. Actual disbursements as a percentage of appraisal estimates rose from 44% in September 1975 to 87% some six months later, by which time almost all credit for on-farm beef development had been committed. A number of extensions of the closing date from June 30, 1977 to December 31, 1980 were granted to allow completion of the sheep development aspects and of the market study. 3.04 Credit for sheep development was finally fully used in 1979 as a result of changes in BAB's lending policy toward small farmers whereby land titles for subloan guarantees were no longer required and as a result of reallocation of part of the sheep credit to beef. In order to close the Credit, the responsibility for a study of an implementation plan to restructure meat marketing was passed to the UNDP as executing agency. By the end of May 1981, the study terms of reference had been written a second time in order to conform to UNDP specifications and invitations to bid had been advertised. Table 3.1 - Loan Allocations by Category Initial Allocation Final Actual (at signature, Allocation Disbursements June 1971) (June 1978) (May 1981) ---------------- US$ millions --------- 1. Beef ranch development 3.300 3.930 3.9 2. Sheep property development 1.500 1.301 1.4 3. Incremental working capital 0.600 0.657 0.7 4. Technical services 0.200 0.210 0.2 5. Consulting services 0.100 0.702 0.5 6. Slaughterhouse construction and renovation 0.400 - - 7. Unallocated 0.700 - - Total 6.800 6.800 6.7 Source: Project files. 3.05 The major movement was the transference of US$500,000 from unallo- cated, Category 7, in 1974 to incremental working (.apital, Category 3, for emergency loans due to flooding in the Beni (Table 3.1). Little demand ever materialized for these funds since other long-term credit was made available to meet this need and, in 1975, the unused portion was moved into beef ranch development credit, Category 1. Similarly, US$200,000 of the funds for smallholder sheep development, Category 2, was moved into beef development in 1978 because of relative demand differences. A supplementary trend was the reallocation of funds toward consultant studies. In 1975, a study of BAB financed by this project, but not foreseen by it, became a condition of - 23 - Board presentation of :he First Agricultural Credit Project, Consequently, almost US$200,000 was cransferred to consultant services, Category 5, from technical services, Cai:egory 4, and from the unallocated funds. In 1978, the total of US$400,000 provided for slaughterhouse construction or renovation was moved into consulting services. At various stages, the GOB requested that this sum be reallocated to beef development but this was refused on the grounds that ample credit for this purpose was available from the Agricultural Credit Project and that restructuring of beef markets was a prime intent of the Livestock Project. 3.06 The balance of US$171,743 remaining after final disbursements was canceled in May 1981 and the project. management unit, the ALPD, was absorbed into BAB's general staff. The First Agricultural Credit Project (561-BO) was also closed at that date. IV. PROJECT ACHIEVEMENTS 4.01 By completion date of December 1980, 641 subloans had been made for a total of US$7.70 million. Of these, 534 were for beef cattle investments, costing US$5.90 million, and 107 were for on-farm sheep developments costing US$1.80 million. Another US$0.20 million was used to support the administration and technical services of the Project Unit. A total of US$0.50 million was spent on consultant studies of BAB and a feasibility study for a new abattoirs at Santa Cruz, as well as two meat industry analyses. On-farm Investments - Beef Production 4.02 Livestock purchases accounted for 45% of investment costs in this category (Table 4.1). Of these, three-quarters were for breeding and one-quarter for finishing. Fencing was second in importance, being 13% of the total. Pastures, stockyards, buildings and machinery accounted for another 20%, with a further 8% divided between provision of watering facilities, roads and airstrip construction and meat refrigeration equipment. Of the balance of 14%, a little over half was used for subsistence loans following serious flooding in the Beni in 1974, a purpose not authorized by the Credit. The rest was lent for miscellaneous items within the project intention except for the amount canceled. - 24 - 4.05 The majority of the 570 subloans were for surprisingly small amounts. This was because of the use of the credit as working capital for steer purchase. Some 68% of subloans were for less than US$10,000, with only 16 being for amounts in the US$50,000 to US$70,000 class, the highest group. The majority of subloans fall into the US$2,500 to US$5,000 class, being 46% of the total. On-farm Investments - Sheep Production 4.06 In contrast to the structure of subloan demand for beef development, the major investment item was farm machinery, reflecting the mixed enterprise type of farming in the Altiplano. Table 4.3 - Sheep Development - Investment Items Item Units No. Units Amount % of Total (US$ '000) Tractors and equipment no. 117 456 23 Sheep head 21,000 336 17 Pasture ha 6,289 339 17 Buildings no. 861 195 10 Fences, watering facilities, etc. no. - 658 33 Total 1,984 100 About 21,000 head of generally improved sheep were bought to improve flocks and some 6,289 ha of pasture were established. The rest of the subloan funds was used for fences, watering facilities, veterinary equipment and minor items. 4.07 An impressive feature of the sheep component was the relative success of group subloans (cooperatives or precooperatives) in channeling farm credit to many small enterprises. Of the 107 subloans made, over half went to groups reaching some 849 beneficiaries through only 61 subloans (Table 4.4). Table 4.4 - Sheep Development - Subloans by Farmer Organization Individuals Groups Total Number of subloans 46 61 107 Number of beneficiaries 46 849 895 Amount (US$ '000) 749 1,235 1,984 % of total 38 62 100 4.08 Some 84% of holdings classified as large (more than 500 breeders). were cooperatively managed for credit operations. Most of the individual farmers obtaining subloans were classified as small (less than 100 breeders). Subloans were concentrated in the Department of Oruro (87%), with lesser participation in the La Paz (13%) and Cochabamba (7%) areas. Much of this distribution is due to presence of competing credit lines. - 25 - Results Compared with Appraisal Estimates 4.09 The 45% cost overrun for the project as a whole is largely explained by the strong demand for credit for cattle ranch development in the Beni. The number of subloans granted almost doubled the 250 expected at appraisal (Table 4.5). This, combined with the increase in average subloan amount for beef from US$15,300 to US$22,800, led to an increase in the cost of this component, from US$5,700 million to US$7,027 million. Despite the rise in the number of subprojects, the total area reached no more than 1 million ha projected because the average property size was almost one-half that expected. Table 4.5 - Comparison of Expected with Actual Achievements Beef Sheep Expected Actual Expected Actual Enterprise size (ha) 4,000 2,856 2,000 345 Number of subloans 250 459 150 107 Number of families 250 459 3,790 895 Area (ha) 1,000,000 992,000 140,000 105,000 Total investment (US$ '000) 5,700 7,027 2,500 1,984 Average subloan (US$) 22,800 15,300 16,600 18,000 Source: Completion Report of Project Unit. 4.10 On the other hand, all the coefficients for the sheep component turned out to be much less than originally thought. For example, the number of families reached through group loans totaled only 895, which is scarcely 25% of the 3,790 anticipated. Similarly, the estimates for average enterprise size, the number of subloans and, hence, the total area to be covered, proved overoptimistic and were not achieved. However, it must be stressed that such numerical goals can be no more than illustrative, particularly where census data are inadequate. 4.11 The compositi.on of investments actually financed also differed markedly from expectations. The major change was a shift away from long-term development items such as breeding cows, fences, pastures and water facilities and a move to the purchase of store cattle for fattening to take advantage of seasonal surplus of feed. At appraisal, it was thought that some 22,500 breeding cows would be bought. Instead, property owners only invested in 9,770 of these, preferring to substitute an almost equal number of unfinishe( cattle for fattening. The "one time" nature of the fattening operation, as compared to the long-run gains from building up breeding herds, accounts for much of the shortfall in reaching production targets. The major change in sheep investments was the movement away from pastures into animals. The 6,000 ha of pasture actually planted was one-quarter of that hoped for to relieve the chronic overgrazing problems of the Altiplano. The 21,000 sheep financed was almost double the number expected. Technical Services 4.12 Corresponding with the expansion of the Project Unit's credit activity, the number of technical personnel rose to 42. These specialists assisted farmers under both 261-B0 and 561-BO in drawing up farm plans, making subloan applications and implementing projects. The project supplied the technicians, as well as the office support staff, with six vehicles, five motorcycles, five radio communication sets and office equipment, as well as veterinary materials. Two courses were held with participation of foreign experts in training of farm extension personnel of the Project Unit. Consulting Services 4.13 Four studies were financed by the project: (a) Meat Marketing and Production Policy; (b) Reorganization of BAB; (c) New Meatworks for Santa Cruz; and (d) Implementation of Changes in Meat Marketing. In 1974, an international consulting firm working with a local firm produced a comprehensive and exhaustive diagnosis of all aspects of the Bolivian meat industry. The work was of excellent quality and provided a sound basis for authorities to plan a rational meat industry policy. The total cost was US$156,000, against which US$130,000 was disbursed. The major recommendations were: (a) establishment of a beef industry regulation commission, to decide policy details; (b) establishment of a technical secretariat to provide analytical input for commission decisions; (c) preparation of a plan for a meat inspection system; (d) formulation of a grading system and a price policy; and (e) initiation of a specific investment program for construction and renovation of slaughterhouse facilities. 4.14 A study of BAB was undertaken in 1976 by an international consulting firm associated with a local company, for a cost of US$90,000, of which US$76,000 was disbursed by IDA. The study analyzed and made recommendations on the finencial position, organization and management, accounting system and staff development of BAB. Thirty detailed manuals on organization and proce- dures were produced and accepted by BAB management. BAB implemented many of the recommended reorganizations, which were designed to alleviate structural deficiencies, define and decentralize authority, integrate functions, and transfer the responsibilities of day-to-day operations from the Genaral Manager to the operational departments and the regional offices. - 27 - 4.15 The funds from Category 6, US$400,000 for slaughterhouse construc- tion and renovation, were reallocated to Category 5, consulting services, in June 1978. This was to enable the National Meat Commission to define a meat production and marketing policy for Bolivia and to finance feasibility studies for new meatworks. EvEntuaLy, interest was shown *i studies fr- meatworks at Santa Cruz and at Trinidad. rhe Trinidad study was finally financed by the Argentinian Government. 4.16 In 1980, a local consulting group produced a two volume report recom- mending construction of new meatworks at Santa Cruz. The report recommended against rehabilitation of the present facility, based on market projections. The study projected a 27% rate of return to a meatworks costing some US$4.5 million, but pointed out the need for changes in Government policy for the meat industry. The consultants further recommended that, under the present system of marketing without proper refrigeration and hygiene facilities, as well as arbitrary price controls, no system of differentiation of meat by class, type or cut should be introduced at the abattoir level. The study cost US$72,250, against which IDA disbursed US$68,000. 4.17 Following from the recommendations of the 1974 meat industry study, the Government established an "Interinstitutional Commission" in 1975 to examine the proposals put forward. No changes were made, as proposed, until the establishment of a Technical Secretariat (Comite Nacional de Carnes) in March 1977. The Comite was to be funded by the Ministry of Finance (Article 5 of Supreme Decree 14415) and was charged with drawing up classifi- cation norms, quality control standards and inspection systems, particularly with a view to expanding meat exports. However, the Comite de Carnes still has no operating budger, travel funds, vehicles or equipment and is grossly understaffed. As a result of the 1978 reallocation of funds from Category 6, the Comite entered into an agreement in 1980 with the Andean Group (Junta del Acuerdo de Cartagena) to finally execute the meat study. The study was to be in two parts: (a) an update of the 1974 study but adding to it a series of specific and implementable recommendations; and (b) a retail trial marketing of precut meats differentiated by type of cut and quality. However, following political differences with the new Government of Bolivia in July 1980, the Andean Group withdrew their support before the work had begun. 4.18 In February L981, the UNDP became executing agency for the study, to cost US$250,000. Terms of reference were drawn up and invitations to bid were advertised in May 1981. The study is expected to be completed in late 1982, some 11 years after the credit was signed. V. PROJECT IMPACT Production Effects - Beef Components 5.01 Twenty-one thousand head of cattle were purchased with project finances. Of these, almost 10,000 were store cattle purchased from neighbors for fattening. For this group, the incremental production of meat is the difference between the weight of the finished cattle at slaughter and the - 28 - weight of those same cattle if slaughtered before fattening. Accordingly, the production impact from this item is small at around 100 tons of meat per year out of a total annual production of 70,000 tons. Of more importance is the contribution of the 9,000 improved breeding cows, which were puichased for reproduction instead of slaughtered. The slaughter of females of breeding age has been identified as a major impediment to development of the industry. Similarly, the contribution to improved productivity of the 2,500 improved bulls, mostly Zebu Nellore from Brazil, raised performance coefficients when the use of improved sires was combined with complementary measures, such as parasite and disease control. Almost 200 dip and spray facilities were established for control of ectoparasites. 5.02 The production impact at the farm level was difficult to establish for two reasons. First, because of the age of the project (appraised in 1969), no system of farm record-keeping or surveying was set up to monitor the project impact. Second, over the project life the number and composition of herds belonging to beneficiaries fluctuated with the incidence of serious flooding and changes of ownership. The monitoring unit set up under the later Credit Project (561-BO) is currently attempting to quantify production effects. If it can be validly assumed that the productivity gains under both projects for similar investments are of the same order, then incremental rates of return to individual borrowers who followed through with the development proposed would have reached close to the 18% projected. Production Effects - Sheep Components 5.03 About 21,000 sheep were purchased with project funds. With a national inventory that grew from 7 million to 8.2 million during the project period, the overall production impact was small. The project monitoring group of the Agricultural Credit Project (561-BO) estimated rates of return in the order of the projected 17% to 20% for this component under the same assumptions about productivity gains as for beef. Income Effects 5.04 Although individual farm models showed 18% rates of return to investment in cattle activities, the models were made under the assumption of (a) a constant relation between the prices of factors and those of products; and (b) an uninterrupted series of average seasons. Neither of these assumptions proved valid and the discrepancy became the basis of protracted disputes between IDA, BAB, the Government and producers. The 18 months' delay in effectiveness was partly due to the dispute over interest rates in which the producers claimed that the incremental income could not cover the interest payments proposed by IDA because of the above problems. The passage of water draining from the Eastern Andes across the flat terrain of northern Bolivia to the Amazon renders the countryside susceptible to some degree of flooding every year. From time to time, such as in 1974 and 1979, more serious damage occurred than usual covering valuable pasture for long periods, washing away fencing and roads and bringing movement to a halt. - 29 - 5.05 Producers' complaints about changing price relationships seem to have some substance, as shown in Table 5.1, used to indicate the broad direc- tion of general trends only. Table 5.1 - RelatLonship Between Beef Prices and General Price Level kear CPI 1/ Beef Price Index 2/ Ratio 1973 100 100 1.00 1974 162 133 0.82 1975 176 133 0.76 1976 183 134 0.73 1977 198 140 0.71 1978 219 156 0.71 1/ Consummer Price Index, La Paz. 2/ Official prize at meatworks. Source: Computed. If the change in the CPI is used as an indicator of the movement in the general level of prices in commercial centers, then clearly the terms of trade for beef producers deteriorated over the period shown - even allowing a wide margin for error. On the other hand, such a trend is not limited to Bolivia but is worldwide. Many Beni producers remain caught in a profit squeeze because they have been unable to make productivity gains faster than the deterioration in their terms of trade. The position has also been aggra- vated by the inability of the state to institute the market reforms envisaged by the project. 5.06 Perhaps the major achievement of the project was its success in changing BAB's requirements for lending to small sheep producers. Prior to February 1978, sheep subloans were moving slowly because of BAB's insistence on land mortgages as guarantees. Many small farmers wished to take out group subloans for items to be shared. However, many groups had no legal standing, many farmers had no land titles and titles, under the Land Reform Program could not be mortgaged. In 1978, BAB changed its security requirements to acceptance of chattel mortgage (prendaria) which freed up this type of lending for the Agricultural Credit Project. The sheep component took seven years to disburse rather than the three years anticipated. Table 5.2 indicates the slow progress of subloan approvals. - 30 - Table 5.2 - Subloan Approval, Sheep Component Number of Approvals Cumulative Percent of Funds 1972 23 14 1973 16 28 1974 16 45 1975 14 65 1976 16 79 1977 14 95 1978 8 100 Total 107 100 The group loan method was successful in channeling credit to a large number of small farmers with relatively few transactions and at lower cost. The 107 subloans made benefited 895 individual families. Because of the experience with this project, many small farmers in the Altiplano received their first opportunity to increase their meager incomes through the use of farm credit at reasonable cost for long-term development. Production Support 5.07 A key inclusion in this project that differentiated it from its predecessors was the large measure of production support. Experiences in the previous First Livestock Project and Second (Interim) Livestock Project had pointed to the need for changes in beef industry conditions to be made in conjunction with provision of direct production credit. Support provisions fell into four categories: (a) policy changes; (b) market studies; (c) market facilities; and (d) animal health. 5.08 Policy changes to be made resulting from agreements reached during negotiations were: (a) maintenance of adequate rates for air freight of beef (Section 4.01 of Credit Agreement); (b) maintenance of an adequate beef price policy following a beef industry analysis (Sec. 3.02); (c) removal of export restrictions on beef trade; and (d) establishment of an effective procedure for clarification of land titles (Sec. 4.05). Government price fixing was said to be adversely affecting incentives to develop and low prices were claimed to be the major cause of arrears. As shown in Table 5.1, changes in producer prices lagged behind changes in the general price level and this continues to be a source of friction. Adjustments in consumer and producer prices remain compromises between consumer and producer goals. The consultants study was accepted by the Government in 1974 and provided a solid basis for intro- ducing price differentials to provide producer incenLives. The Government was unwilling or unable to act and is still in default of the covenant in which it agreed, "On a date not later than nine months after the date of the said submission report (consultants study) . . . the Borrower shall adopt a beef price policy . . . establishing price differentials between different grades and qualities of meat." (Section 3.02). However, the issue is not the overall price level but the relative prices between different cuts and grades. No comparative analysis has ever been done to provide a factual basis on the equity effect of administered beef prices versus the cyclical movements of free market levels. Air freight prices ceased to be a source of complaint and are assumed to be satisfactory. - 31 - 5.09 Since construction and/or renovation of slaughterhouses were contingent on introduction of a meat inspection and grading system (Section 3.03) following the marketing studies, no money for this purpose (Part D) was used. Instead, the US$0.4 million was reallocated in June 1978 to enable the Comite de Carnes (a) to study ways to implement the recommendations of the 1974 meat industry study; and (b) to prepare feasibility studies for specific meatworks proposals. The Comite de Carnes remains functionally inoperative. The above study, now under UNDP execution, has not yet started and a feasibility study for a new abattoir at Santa Cruz has been completed, but, to date, no follow-up action has been taken. 5.10 Despite assurances that prompt clarification of land titles (Section 4.05(i)) would enable rapid subloan disbursements, the problem of documentation of land titles to satisfy BAB requirements for mortgage guarantee was still the major reason for rejection of credit application as late as 1978, some seven years after effectiveness. The change of BAB's requirement from land mortgage to chattel mortgage in 1978 overcame this difficulty, rather than compliance with the covenant. 5.11 Foot-and-mouth disease was responsible for significant cattle losses and poor performance. Such viral disease proved a serious impediment to beef development in extensive grazing areas, such as the Beni where animal control is minimal. An agreement was signed (Section 4.05(ii)) under which the Borrower undertook to establish and maintain an effective vaccina- tion program. In 1974, an Aftosa (foot-and-mouth disease) program was established under a project of the Inter-American Development Bank and conti- nues to expand the "clean" areas free from the disease or under treatment. 5.12 Because of the presence of foot-and-mount disease, the unsani- tary state of the meatworks and the lack of meat cut standardization and grading systems, Bolivian exporters eventually lost the small trade, begin- ning with Peru and northern Chile. At the time of appraisal, it was con- sidered that quantitative restrictions were the factor limiting exports and, under Section 4.06 of the Development Credit Agreement, the Borrower undertook to remove such restrictions and not reimpose them except for "temporary" and "emergency" conditions. The Government replaced the export ban with a licensing and taxing system but, because of the unsatisfactory marketing conditions, such controls became irrelevant. Unofficial export of live cattle from border areas to neighboring countries is common because of ease of access, relative transport costs, exchange rates and market distortions. 5.13 In general, the production support activities of the project have nt had a great direct impact on beef development. Except for the BID foot-and-mouth disease control program, these components provided only a sound basis on which to restructure beef marketing. Because of lack of commitment or ability, the failure to follow through has left beef industry conditions little better than at the start of the project.  - 33 - BOLIVIA FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-B0) PROJECT COMPLETION REPORT PART B  3.53 - PART B FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-BO) VI. PROJECT BACKGROUND Sector Developments 6.01 During the mid-1960s and early 1970s, at the time the Third Livestock Credit was being processed, rapid developments in commercial agriculture were taking place in line with similar developments worldwide. During the periods of 1963-65 and 1970-72, the annual average area under cotton rose from 3,000 ha to 23,000 ha and sugar area increased from 24,000 ha to 36,000 ha. The value of cotton exports rose from zero to US$7.5 million and that for sugar from US$0.4 million to US$1.0 million. In addition, poultry numbers increased from 4.6 million head to over 6 million and whole milk production expanded from 25 million liters to 35 million liters. The most rapid increases in poultry and milk production were in the Santa Cruz and Cochabamba areas following the growth in commercial feed availability, principally oilseeds and wheat by-products. The Project 6.02 The Banco Agzicola de Bolivia carried out several studies of these developments and proposed to IDA that it approach the Cooperative Program to put together an agricultural credit project focused on these activities, more as a companion project to the Livestock Credit than as a follow-on. In 1973, a Livestock IV Project was scheduled for FY76 and a separate Agricul- tural Credit Project for FY74. In January 1974, IDA decided that, for project management simplicity, it would proceed with only one project. Accordingly, beef and sheep components were added to the proposed Agricultural Credit and poultry production was dropped due to declining profitability as feed prices rapidly increased. Because of the need to develop sound husbandry practices in the newly cleared areas north of Santa Cruz, a small component for research was added.. At appraisal, in 1974, the project provided US$11.00 million in credit for on-farm developments for beef, sheep, sugarcane, annual crops, and grapes. A f:urther US$0.14 million was included for support to agricultural research and US$0.86 million for project administration. VII. PROJECT FORMULATION Identification and Preparation 7.01 The Cooperative Program (FAO/IBRD/CP) identified and prepared the project between June and October 1972 with assistance from the Project Unit of the ongoing Livestock Credit Projects. At this time, considerable concern was being felt in IDA over these projects because of "(a) . . . deterioration - 36 - in the management of BAB due to recent changes in senior positions; and (b) the question of interest rate policy." 1/ Because a decree had reduced interest rates on subloans under these credits to 10%, whereas the agreed rate was 12%, IDA was initially reluctant to proceed past this stage. At the time of the interest rate reduction, the rate of inflation had reached the point where indexing of subloans should have been begun in accordance with the conditions agreed for the Third Livestock Credit. Section 4.03 of the Credit Agreement required indexing if the domestic inflation rate were to exceed 8% per year. On October 27, 1972, the Bolivian peso was devalued by 67% (from $b ll.88/US$1 to $b 20.OO/US$l), making BAB's repayment of US dollar obligations, relent in pesos, more difficult. In addition to increasing arrears, the Government transferred uncollectable farm loan portfolios to BAB from other institutions. 7.02 However, some progress was being made and a preappraisal mission went to Bolivia in February 1973. The March 1973 supervision mission reported that ranchers were progressively reducing arrears because BAB was now branding cattle purchased with the subloans, giving itself prior claim to subsequent sale proceeds. An agreement on indexing was reached between IDA, the Government and ranchers, providing for a 9% interest rate, with periodic adjustments of outstanding principal, based on the movement of the lower of two indices (consumer price index and beef or mutton price index), minus 3%. On loans obtained prior to November 1, 1972, borrowers could choose to retain the existing 12% rate without adjustments of principal (March 16, 1973). This system was calculated to give a real rate of interest of 6% and was to be applied to the new project. Appraisal, Negotiations and Approval 7.03 Nevertheless, IDA informed the Minister of Finance in May 1973 that it would not appraise the project until ". . . we have been notified that the Government has adopted such a plan of action (to reorganize BAB) and taken the first measures to carry it out ....." 2/ The Government initiated the required reforms in mid-1973 but, in September, was advised of IDA's concern at the slow progress of the reorganization committee's work and the continued deterioration of BAB's financial condition. At this time, IDA instructed the Bank mission to Bolivia to examine the possibility of using the Central Bank through its Agricultural Development Fund as an alternative channel for the proposed agricultural credit project. This course was rejected on the grounds that a principal project objective was the institution building of BAB. Prior to the early 1970s, BAB had a reasonably good operating record and a developmental perspective not shared by the banks participating in the Agricultural Development Fund. 1/ Letter of November 15, 1972 from G. F. Darnell, IDA, to J..P. Huyser, FAO/CP. 2/ Letter from G.K. Weise to L. Bedregal, May 30, 1973. - 37 - 7.04 Upon completion of the reorganization committee's work, BAB's management introduced changes recommended in a draft Presidential Decree of December 1973. The decree contained virtually all IDA's recommendations. In the same month, a fi:aal decree was promulgated on indexing applied retroactively to November 1972. The level of arrears on interest payments dropped from 40% to 5% between 1973 to 1974, but it is not clear how much of the fall was due to rescheduling. In any case, sufficient progress was thought to have been made that IDA proceeded with appraisal in June 1974. 7.05 However, at this time, the ranchers persuaded the Government to press for repeal of the indexing decree because: (a) there were flooding losses in the Beni; (b) a 60% adjustment had to be made to principal in some cases because of the retroactive clause; (c) indexing had not yet been intro- duced in the Santa Cruz area. Although flooding had been severe, the project manager reported that only one-third of the ranching area was affected and relatively few participating ranchers suffered in a serious way. The appraisal mission was sent, but was instructed to advise Govermment that IDA management would not be in a position to recommend to the Executive Directors further lending to Beni ranchers until these issues were settled. 7.06 Between appraisal in June L974 and negotiations in May 1975, several missions worked to break the impasse. By negotiations, seven issues remained namely: (a) Government contributions to BAB equity, counterpart funds, project administration and agricultnral research; (b) initiation of a consul- tant study on BAB reorganization and finances; (c) implementation of an agri- cultural research and land titling program; (d) staffing of the Project Unit; (e) establishing a project monitoring unit; (f) agreement on indexation; and (g) participation of commercial banks in the project and the agreement of the Government to make up any shortfall in counterpart funds expected from these banks. Agreement was reached on these points and the project was presented to the Board under Special Procedure on June 19, 1975. Effective- ness was delayed by three months over meeting the conditions with regard to Government contributions to BAB and appointment of a Bolivian national as project director. Project Description 7.07 Project objectives were to (a) expand production for export (live- stock and sugarcane) arid domestic consumption (wool, mutton, wheat, potatoes, oats and grapes); (b) improve the economic welfare of some 3,600 families of rural poor; and (c) improve regional distribution of economic development. Beef would be financed in the Beni and Santa Cruz areas, sugarcane in Santa Cruz, annual crops in the Cochabamba region, and grapes in the Tarija valley and sheep in the Altiplano. 7.08 Of a total of 4,600 subloans, some 3,600 were targeted to the small farmer poverty group. Beef ranchers were expected to be in predominantly medium to upper income groups. Farmers growing sugarcane on farms in excess of 50 ha and grapes on farms greater than 4 ha would not qualify for a subloan. In addition to on-farm investments, provision was made for an expanded Project Unit and some support to research in crop rotation trials near Santa Cruz. - 38 - The total project cost was estimated at US$12 million, of which US$11 million was for farm credit, US$0.86 million for project administration, and US$0.14 million for research support. The shares in total costs were as follows: Source Share of Total (%) IDA 62 Government 20 Commercial banks 7 Beneficiaries 11 Total 100 The Government was the borrower and the Banco Agricola de Bolivia, through the ALPD, the executing agency, with both BAB and commercial banks as parti- cipating institutions. It was hoped that 10 commercial banks, as well as one cooperative society, DESEC (Centro para el Desarrollo Social y Economico), would be involved. 7.09 The organization and management was essentially the same as for the ongoing livestock project, with some expansion of supervision and technical assistance capability to accommodate the wider range of activities and regions. The financial rates of return expected varied from 13% (potatoes) to 27% (sheep and grapes). No overall economic rate of return was given, but economic rates of return for the various components were expected to be similar to those at financial prices. 7.10 The major financial innovation was the settling of the interest rate dispute. The system agreed upon was to leave to the sub-borrowers the option of choosing between (a) an interest rate of 4% on outstanding balances adjusted by the cost of living index for La Paz; or (b) an interest rate of 14% on the outstanding balances adjusted for variations of the rate of exchange between the Bolivian peso and the US dollar. Subloans to sub- sistence (sheep and potatos) farmers would bear a 12% nominal interest rate. Grace periods were to be four years (cattle, sheep and grapes), three years (sugarcane) and two years (potatoes). Subloan periods were to be 12 years for sheep and cattle, six years for sugarcane, eight years for grapes, and four years for potatoes. VIII. PROJECT IMPLEMENTATION 8.01 Disputes over interest rates slowed the initial progress of the Agricultural Credit Project as happened under its predecessor, the Third Livestock Credit Project. After 18 months, only US$1.70 million had been disbursed compared with an appraisal estimate of US$5.50 million. Although the Government allocated part of its agreed equity contributions, BAB's position continued to deteriorate. Despite earlier promise, arrears once more became critical. The consultants completed their studies of BAB and produced detailed operational manuals which were a solid basis for restructur- ing the bank and BAB, once more, was reported to be in process of implementing - 39 - a reorganization plan. During this period, no applications were received for sheep subloans because sufficient funds were left over from the livestock project to cover demand; no requests were received for sugarcane development because the world sugar industry was in a cyclical downswing in profitability and subloans were available to growers through a Central Bank line with BAB; and no demand was forthcoming from annual crop farmers because of lack of knowledge of the availability of funds. The April 7, 1977 Supervision Report drew attention to the point that Government policy discouraged sugarcane plantations of less than 100 ha whereas the Credit Agreement (Schedule 3) restricted subloans to farms under 50 ha. Although the number of subloans to beef and grape producers were substantially on target, the average investment size was some three times that envisaged, partly because of inflation, but mostly because of underestimation of investment needs per unit of an activity. Beef property sizes were substantially in excess of those envisaged. Only one private bank had made any subloans even though 10 signed subsidiary loan agreements. The cooperative participant, DESEC, had decided not to participate. The rate of granting land titles even slowed down, in spite of undertakings to the contrary. 8.02 However, in 1978, following BAB's agreement to accept chattel mortgages as security Eor smallholders' loans, subloan commitments for sheep and annual crops rose sharply to 186. Demand for beef subloans increased and, early in 1979, new commitments for grape development ceased because all funds had been allocated for that purpose. In April 1978, the unused alloca- tion for sugarcane development was reallocated to beef and sheep development (Table 8.1). At the same time, the financing proportion for the farm equip- ment for agricultural experimentation was changed from 100% of foreign expen- ditures to 50% of total cost to facilitate the administrative process of bank disbursements for these items since they would be purchased locally. Also, the proportion of the local ALPD Director's salary was increased from the 24% of the total agreed to in an amendment in December 1975 to 37.5%. The first amendment was made because no expatriate staff were hired by the Project Unit, making the proposed IDA contribution to foreign exchange costs redundant. In order to keep quality staff, IDA agreed to contribute the difference between a BAB salary and an international level salary for the Project Director. The second amendment was to maintain a real salary value as BAB salaries eroded in real terms. No changes were made in the actual dollar amounts allocated to project administration and agricultural experimentation. Table 8.1 - Allocation of Loan Proceeds Original Allocation of Reallocation of Final Allocation Contingencies Sugarcane Funds Allocation % --------------------------US$ '000------------------------ Beef 3,170 1,138 0,443 4,750 67 Grapes 0,580 0,396 - 0,976 14 Annual crops 0,470 - - 0,426 6 Sheep 0:570 0,217 0,117 0,948 13 Sugarcane 0:560 - - - Contingencies 1.,750 - - Total 7,100 1,750 0,560 7,100 100 - 40 - 8.03 By the end of 1979, some 67% of IDA funds had been disbursed and virtually all project funds had been committed. New subloan disbursements by the Project Unit slowed down from June 1979 until completion for several reasons: first, because of the shortfall in the Government's contributions of counterpart funds, the Project Unit was forced to wait for subloan recuper- ations to use instead; second, project management failed to process the necessary claims for disbursement in a timely manner; and third, uncertainty surrounded the devaluation in November 1979. The closing date was extended six months three different times from June 30, 1979 to December 31, 1980. 8.04 Following the second devaluation, the interest rate issue was raised again. Growers petitioned the Government against application of the principal adjustment clause. In addition, hail damage in Tarija had reduced grape yields in 1978 and 1979 from the long-run average of around 15,000 kg/ha to 5,000 and 9,000 kg/ha in the two years, respectively. Temporary moratoria on debt repayment were applied in the Beni and Tarija Departments. These were rescinded in 1980 and the indexing mechanism applied. The credit was closed on December 31, 1980 and disbursements were made until March 31, 1981. The balance of US$64,000 remaining for project administration was cancelled. IX. PROJECT ACHIEVEMENTS Farm Credit 9.01 By the completion date in December 1980, some 1,139 subloans had been made, for a total of US$9,726,000. Of these, 43% went to beef invest- ments, 20% to sheep, 15% to vineyards and 22% to annual crops. Another US$200,000 went to support agricultural experimentation and US$600,000 to project administration. Most of the on-farm credit advanced (67%) was used for beef development, with the balance going to investments in vineyards, 14%; sheep grazing, 13%; and annual cropping, 6% (Table 9.1). 9.02 The beef areas of the northern lowlands absorbed 49% of the farm credit advanced for 23% of the total subloans made. Although about the same number of subloans were made in the Altiplano for sheep (20%), the smaller size of investments required less than one-third (13%) of the amount spent on beef. However, the greatest concentration of subloans was in the Cochabamba valleys for annual cropping, but reached only 6% of the credit amount because of the smallness of individual subloans. Table 9.1 - Distribution of Subloans by Region and Type Regional Agency Total Beef Sheep Grapes Annual Crops No. Loans % US$ m % No. Loans US$ m No. Loans US$ m No. Loans US$ m No. Loans US$ m Triniad 220 20 4.386 45 220 4.368 - - - - - Riberalta 8 1 0.153 2 8 0.153 - - - - - Cubija 25 2 0.150 2 25 0.150 - - - - - Santa Cruz 221 20 1.751 18 221 1.751 - - - - - La Paz 82 7 0.488 5 5 0.016 77 0.471 - - - Oruro 151 13 0.827 8 - - 151 0.827 - - - - Cochabamba 274 23 0.737 7 6 0.068 - - 13 0.085 255 0.583 Tarija 158 14 1.252 13 -- - 158 1.252 - - Subtotal 1,139 100 9.744 100 485 6.507 228 1.298 171 1.337 255 0.583 Z of Total 100 - 100 - 43 67 20 13 15 14 22 6 Source: Completion Report of Project Unit. - 42 - On-farm Investments - Beef 9.03 The pattern of investments made was consistent with what was to be expected for development of extensive grazing systems. Investments in order of importance were fencing and yards (38%), breeding cattle (36%), pastures and watering points (14%), and miscellaneous (12%) (housing, airstrips, machinery and equipment). 9.04 The weighted average property size proved to be considerably greater than that assumed at appraisal, leading to a reduction in the number of subloans that could be made with a given credit total. Even after adjust- ment for considerable local inflation, the realized investment per family was almost three times that anticipated. However, it is interesting to note that investment per hectare was almost half that expected. In other words, Bolivian beef properties proved to be quite large, with high credit requirements overall, but low on a per hectare basis. This is consistent with the asset structure of lowland ranches where the land has scarcely any value. Table 9.2 - Beef - Actual Compared with Estimated Achievements Appraisal Actual Percent Item Estimate Actual of Estimate Number of subloans 930 482 52 Average property size (ha) 500 2,500 500 Investment per family (US$) 5,250 15,000 286 Investment per hectare (US$) 10.5 6 57 9.05 The distribution of subloans by beneficiary farm size shows a concentration of number in the small-size farm class and a concentration of funds in the medium-size farm class. About 56% of subloans went to those classified as small, 31% to medium, and 13% to large, respectively. Of the credit total, 42% went to medium, with 36% to small and 22% to large properties. However, these figures need to be interpreted with care since the size by which farms are classified refers to enterprise size and not farm size. For example, a sugarcane grower with 500 ha of cane is a large farmer by income definitions. If he has 50 head of cattle and applies for a beef development loan, his herd size classifies him as "pequeno". A similar situation commonly occurs where absentee owners not dependent on their beef holdings for their principal income source maintain understocked properties. Annex, Table 3 classifies subloans according to size in detail. 9.06 Repeat loans with IDA accounted for 26% of all subloans, with 74% being IDA clients for the first time. Diversion of funds was significant. Some 109 beneficiaries (24% of the total) used farm credit received for other purposes, representing 7% of the total amount (US$430,100). The greatest diversion occurred among the large property class with most being in the Beni. Diversion of funds was of three types, namely: (i) non- compliance with the agreed farm investment plan but use of the funds for other productive on-farm investments; (ii) compliance with agreed investment plans - 43 - but at a lower cost for inferior materials, with surplus funds being retained for unspecified purposes;; and (iii) non-compliance with the agreed farm plan and use of the funds outside the agricultural sector. The Completion Report of the ALPD identified this non-compliance with agreed investments as a principal cause of disappointing financial results on the properties concerned and as a major contribu:or to the overall arrears level. No action was taken by the ALPD because of insufficient support from BAB. No action was taken by IDA because the problem was not identified until after the project was closed. A further type of diversion was the failure of beneficiaries to contribute the agreed share of the investment cost from their own resources. Only 47% of the beenficiaries' agreed contribution was actually made. On-farm Investments - Shaeep 9.07 As for the beef component, property sizes proved larger, credit needs per subproject greater and, hence, number of beneficiaries considerably less than anticipated. Table 9.3 - Sheep - Actual Compared with Estimated Achievements Appraisal Actual Percent Item Estimate Actual of Estimate Number of subloans 1,600 228 14 Number of beneficiaries N.A. 531 - Average farm size (ha) 100 430 430 Investment per subproject (US$) 560 2,445 436 Investment per ha (US$) 5.6 5.6 100 As shown in Table 9.3, the weighted average farm size proved to be over four times the appraisal estimate; the average investment, over four times that anticipated, even with adjustment for inflation; and the total number of subloans, less than one-fifth of that expected. Because 88 of the subloans were to groups, the number of beneficiaries was over twice the number of subloans. However, since no applications were received from production cooperatives, no subloans were made to such group. It had been thought that cooperatives with an average of 20 family members would seek help. In fact, although farmer groups averaged four beneficiaries per group, many beneficiary pairs were spouses, reducing the number of beneficiary families. 9.08 Most subloans (76%) and credit supplied (57%) went to individuals or groups with less than 100 breeders. Of the balance, only one subloan for 1% of the funds was advanced to large farmers having more than 500 breeder sheep. 9.09 The major investments were in purchase of breeding sheep (33%); other livestock, such as dairy cattle, work oxen, and llamas (29%); and dips, yards, and other structures (22%). Some 24,000 head of sheep,364 dairy cows, 580 oxen, and 2,086 llamas were financed. Minor items included 727 ha of pasture establishment and 33 irrigation pumps (Table 9.4). The level of diversion of farm credit proceeds to purposes not authorized by the agreed - 44 - farm plans was 12% of the total funds, which is almost twice that for the beef component. The major non-compliance was for construction and reached 25% of the subloans for this item. Table 9.4 - Sheep Component Items Financed Item Number Amount % US$ Sheep 24,000 425,000 33 Cattle, llamas 3,944 377,000 29 Dips, yards, structures 800 292,000 22 Pasture (ha) 727 62,000 5 Veterinary equipment tools - 54,000 4 Irrigation pumps 33 43,000 3 Infrastructure roads 96 23,000 2 Forage, seeds, other - 22,000 2 Total , 1,298,000 100 On-farm Investments - Vineyards 9.10 Most of the 171 subloans for grapes were made in Tarija where production for wine making is an important industry. Scarcely 8% of bene- ficiaries were located around Cochabamba where generally grapes are grown only for table use. Some 172 ha of new grapevines were planted with 76% of the credit advanced and an extra 35 ha of vineyards were rehabilitated with another 9%. The balance of 15% was almost equally divided between nets for protecting the crop from hail damage (5%), irrigation pumps and tractors (4%); and miscellaneous items, such as land levelling and fencing (Table 9.5). Table 9.5 - Grape Component - Items Financed Item Number Amount % US$ Vineyard establishment (ha) 172 1,013,000 76 Vineyard rehabilitation (ha) 35 122,000 9 Anti-hail nets (ha) 15 72,000 5 Irrigation pumps, tractors 19 54,000 4 Land levelling, fences, other - 76,000 6 Total 1,337,000 100 9.11 The number of beneficiaries reached 201, which is 30 greater than the number of subloans because some credit went to cooperatives (Annex, Table 4). In contrast to the cattle and sheep components, the average sub- project size was considerably smaller than anticipated at appraisal. However, the 62% higher cost of investments led to shortfalls in the number of subloans and in the area planted. Only 39% of the area intended for grape planting was reached (Table 9.6). Deviation of funds for purposes not authorized was quite small at some 2%. Most of the subloans (78%) went to small growers having up to 8 ha with irrigation, not necessarily all for grapes. - 45 - Table 9.6 - Grape Component - Basic Indicators Appraisal Actual Percent Estimate Actual of Estimate Number of subloans 220 171 78 Number of beneficiaries 220 201 91 Area of subproject (ha) 2 0.85 42 Total area (ha) 440 172 39 Average subloan amount (US$) 4,090 6,652 162 Investment cost/ha vineyard (US$) 4'435 5,900 133 On-farm Investments - Aanual Crops 9.12 One of the key intentions of the project was to bring in Bolivian institutions other than BAB to participate in farmer credit programs. Not only was the attempt to interest commercial banks in development lending for beef unsuccessful, but, also, the involvement of the DESEC cooperative movement failed to materialize. At appraisal, DESEC was to be the major intermediary for channeling funds for improvement of annual cropping among small farmers in the Cochabamba valleys. DESEC established and continues to operate an inte- grated system of marketing, technical assistance and planning among small scale farmer associations. Under the project, this organization planned to assign nine staff to assist project beneficiaries on the basis of a commission from the participating credit institutions. However, a misunderstanding arose (or existed from the beginning) between DESEC and the ALPD over the roles of each, the method of sharing the financial costs and benefits involved, and the amount of off-farm investment for cooperative groups. After a six-months' delay, the ALPD managed this credit line without DESEC's involvement. 9.13 This component deviated the most from project intentions. No off-farm items such as warehouses, trucks and other group marketing facilities were purchased and the cooperative aspect was considerably diminished. Eventually, 22% of beneficiary farmers received subloans in groups. It is interesting to note that farmers preferred to make relatively small invest- ments in items shared by a few, such as teams of working bullocks, than in large cooperative facilities shared by many, such as the warehouses and trucks envisaged As shown in Table 9.7, the considerable underestimation of per hectare costs by four-fifths resulted in the achievement of only 15% of the area and 18% of the number of subloans proposed. Table 9.7 - Annual Crops Component - Basic Indicators Appraisal Actual Percent Item Estimate Actual of Estimate Number of subloans 1,800 327 18 Average area of subproject (ha) 2 1.6 80 Total area (ha) 3,600 531 15 Average subloan amount (US$) 406 1,782 440 Investment cost/ha (US$) 202 1,100 545 - 46 - There are two reasons for this underestimation. One is the fall in popularity of oats and wheat and the rise in demand for potatoes in response to relative profit shifts. The cost of potato growing is much higher, resulting in no demand at all for grain investments. Second is the unexpected demand for farm oxen. Some 69% of the farm credit was used for potato growing (Table 9.8) and 20% for purchase of oxen, used mostly in land preparation for potatoes. The other 11% bought spraying equipment, tractors, trucks, mules and buildings. Most of these were for use in potato growing. Of the total of subloans given, 88% went to small farmers having less than 6 ha under irrigation or 12 ha rainfed for a total of 76% of the funds. Diversion of funds reached 16% of the total. Table 9.8 - Annual Crops, Items Financed Number of Subloans Amount % (US$) Potato cultivation (ha) 531 403,000 69 Working oxen 458 117,000 20 Spraying equipment 105 15,000 3 Tractors 2 20,000 4 Trucks 1 10,000 1.5 Mules 63 14,000 2 Buildings 7 3,000 0.5 Technical Assistance 9.14 Of the Project Unit's 90 member staff, 42 were classified as techni- cians being veterinarians, agronomists or others. Because of the inadequacy of Bolivia's extension service, something less than 10% of farmers ever receive technical assistance. For most project beneficiaries, ALPD technical staff provided the only extension they were receiving. Because of the number of beneficiaries, their geographical dispersion, the range of crops covered and the need to be involved in project administration, services were thinly spread. Project Monitoring 9.15 The Project Unit was set up under Credit 107-BO, First Livestock Project, and continued to manage the Second and Third Livestock Projects. Because of the need for tighter subloan supervision and better analytical preparation for follow-on projects, the Unit was expanded and strengthened. A focal point was the inclusion of a project monitoring group for systenatic documentation and field evaluation of the program. One headquarters' techni- cian worked fulltime on monitoring, with parttime assistance from field staff. For 1979 and 1980 farms were sampled to measure project impact. Much of the data in this report is a result of the monitoring group work and, on completion of their analyses in 1981, a firm basis for future credit projects will be available. -47 - Agricultural Experimentation 9.16 The sum of USS100,000 was provided to support the work of CIAT (Centro de Investigation Agricola Tropical) at Saavedra, north of Santa Cruz. The objective was to assist in research, experimentation and extension in pasture and livestock management, especially in the newly cleared areas. The use of these funds was approved by IDA on March 31, 1977, after receipt of CIAT's research prog-cam and evidence of a capital contribution of US$100,000 by the Bolivian Government to the Center. Because of problems with local procure- ment procedures necessitating amendment of the Credit Agreement, disburse- ments were held up until the end of 1978. Most of the funds were used for farm machinery, which, Dn inspection, was gainfully used and well maintained. However, no defined plan was ever drawn up by IDA or CIAT as to how the results of the investment were to be evaluated. IDA's contribution can be seen only as a part of the total support coming to CIAT from the Government and other agencies, such as the British mission. The results of CIAT's work have been generally less than expected because of lack of local support, and this is the most that can be said for this component. Land Titling 9.17 Following the recommendations of the consultant study of land tenure, financed under the First Livestock Project and costing US$133,000, assurances were obtained from the Government of budgetary support for a high priority to the land titling program. In the northern Santa Cruz area and the Altiplano, the lack of titles was a serious impediment to obtaining development credit. However, the support which the Government was able to give was insufficient and the increase in the number of titles has been only 27,000 during the project life, from 566,000 to 593,000. This amounts to an annual increase of 1.5%. During 1980, the titling process almost ceased. Project Auditing 9.18 In general, the audits of the project accounts were satisfactory. However, they suffered from a lack of continuity due to the Bolivian law requiring a different external audit firm each year. Retention of the same firm for more than one year at a time would have improved the overall result. X. PROJECT IMPACT Production Effects 10.01 The rates of return on individual farms that complied with farm plans were generally acceptable, but less than anticipated. Table 10.1 lists rates calculated from actual farm data sampled by the project monitoring unit. Grape returns of 26% to 49% exceeded the 27% anticipated due to the steep rise in prices paid to producers as demand from the wine industry expanded. Small and medium beef producers realized 9% to 17% compared with the 22% expected, whereas the larger properties seemed to give a slightly better result, but still somewhat lower than the appraisal value of 22%. Returns on sheep activities also at 13% to 18% were below the expected 27%. Although adverse changes in cost-price relationships may be claimed for the shottfall, the more plausible reason is that the original estimates were a little overoptimistic since the actual returns achieved are in the range normally encountered elsewhere. - 48 - Table 10.1 - Rates of Return Based on Actual Farm Data Component Appraisal Estimate Actual Recalculation Grapes 27 26-49 Beef 22 small size 9.4-16 medium size 9-17 large size 15-17 Sheep 27 13-18 Annual crops 13 N.A. 10.2 The overall rate of return for the project as a whole, which is the weighted average of all investments properly made, plus others not complying with plans, is not yet available. The monitoring unit is processing the sample survey results and a report is due in late 1981. However, it is expected to be less than the individual returns above indicated for three reasons. First, beneficiaries in arrears may be considered as not realizing expected profits because of a variety of reasons; second, some funds were diverted from intended use; and, third, the rate of investment was slightly slower. For instance, grape growers in localities hit by the hail damage in 1978 and 1980 would not have obtained these rates of return whereas those not so afflicted would receive high rates inflated by the shortage. Also, the appraisal expected a faster rate of investment not reached. Year 1 2 3 4 5 Percent expected 42 46 8 3 1 Percent realized 37 28 24 7 4 10.03 No significant technological changes were introduced by the project. The support to experimentation produced little that could be of immediate use to the beneficiaries involved. Income Effects 10.04 The total number of beneficiaries the project hoped to reach was around 4,850. Those actually participating totaled 1,560, receiving 1,139 subloans, some as groups. Because many members of groups were married to each other, the number of beneficiary families is probably little are than one-quarter anticipated. From these may be subtracted those considerably in arrears, who, for various reasons, do not consider the benefits received worth paying for. For instance, samples taken from farm supervision reports occasionally showed herd numbers being less after four years of project participation and the property owner in arrears. On the other hand, on-farm inspections showed ample evidence of the benefits received by those complying with the agreed investment plans. In particular, the access to credit which small farmers were able to get for the first time was an important equity gain. - 49 - 10.05 The income characteristics of the beneficiary farmers in different farm activities are interesting and worth noting. Of the eight grape growers whose files were sampled at random, four were lawyers, of whom two owned hotels, two owned factories and two were fulltime farmers with no off-farm income stated (Table 10.2). Of the nine beef property owners sampled, six were absentee owners with off-farm income greater than farm incomes. In contrast, none of the sheep growers or annual crops beneficiaries sampled had assets or income other than those related to their farms. In addition, the grape grower receiving the largest loan among the sample was not eligible for a loan at all since he possessed 10 ha planted in grapes. The Credit Agreement states: "No farm loan will be made to any grape grower with a farm exceeding 4 ha of area," (Schedule 3, Part B, 1(ii)). It must be stressed that this 3% sample may well not prolerly reflect the characteristics of the beneficiary population as a whole, but it tends to support a perspective built up from other experience. Table 10.2 - Some Characteristics of a Small Sample Farm Type Net Farm Income Average Loan Amount Fulltime Farmers (US$ (US$) Grapes 0-36,000 4,000-46,000 2 out of 8 Beef 3,000-15,000 12,000-36,000 3 out of 9 Sheep 100-2,000 500-4,000 all Annual crops 800-6,000 700-7,000 all 10.06 Not only are the efficiency and equity effects smaller than anticipated, but the lUck of follow-on credit will not sustain most of the gains realized. At project closure, considerable leakage from the credit program was occurring due to: (a) non-collection of arrears; (b) failure of the Government to supply its counterpart funds, as agreed; and (c) executive orders from BAB to ALPI) directing US$610,000 equivalent from recuperations to purposes contrary to the intent of the project. At the same time, genuine farm credit needs could not be met. Under Section 3.07 of the Credit Agreement, Article III, such recuperations are required to be paid to the Central Bank, which, after deducting the amounts needed by the Borrower to service the credit, is to use the balance for further lending for the same purpose under teras agreed to with IDA. No debt service is yet due from the Borrower to IDA. - 50 - XI. SPECIAL ISSUES 11.01 Both projects experienced many of the difficulties common to credit projects, such as cost increases, land titling deficiencies and institutional weaknesses. However, four problem areas are noted here as special issues because they have existed since the beginning of IDA's involvement in Bolivian agriculture in 1965 and are still germane after four projects and 15 years of effort to solve them. In addition, the three ongoing rural development projects share the same impediments and the processing of a follow-on credit project is stalled for similar reasons. These issues concern: (a) the commitment of the Government and its ability to comply with covenants agreed to; (b) BAB and its competence as a development bank; (c) the Project Unit, ALPD, and its appropriate role; and (d) IDA and the choice of development strategy to pursue in Bolivian agriculture. Undoubtedly, the major issue has been institutional. Government Commitment - Counterpart Contributions 11.02 During negotiations of both projects, the Government agreed to a series of covenants in which it undertook to provide counterpart funds, restructure BAB and agricultural lending, restructure the beef industry and accelerate its land titling program. Despite partial compliance, the objectives of the covenants were not reached and the deficiencies remain. Part of the difficulty may well be that the covenants required much that the Government was unable to do. 11.03 The Project Performance Audit Report for the First and Second Livestock Projects (October 11, 1974) had mentioned delays in provision of local funds to these projects. Although a general covenant obliging the Government to make counterpart funds available (Section 3.01(a)) was consequently included in the Credit Agreement for the Third Livestock Project, shortage of funds continued to be identified as the main problem for the project (July 9, 1974). IDA wrote to the Government on July 19, 1974 and requested it not to submit withdrawal applications until US$1.0 million had been transferred to BAB and the indexing question was settled. The transfer was eventually made but BAB's ability to provide local funds continued as a problem. In June 1975, the Credit Agreement of the First Agricultural Credit Project included specific provisions requiring capital subscription of US$4.3 million (3.01(c)) plus US$0.56 for subloan counterpart contributions (3.03(ii)). At completion, the Government had transferred US$4.3 million in total, leaving a shortfall of US$0.56. In addition, the Government agreed (3.04) to make up.the short- fall in counterpart funds expected from the private banks. Since such - 51 - participation was small and restricted to one bank, the required Government contribution was US$822,000. No payments were made on this account. During the period 1978-1980, the Project Unit was so short of counterpart funds it had to wait for recuperations before making disbursements, even though this was in contravention of Section 3.07 of the Credit Agreement (561-BQ) requiring recuperations to be repaid to the Central Bank. Government Commitment - Restructuring BAB and Agricultural Lending 11.04 One of the first questions concerning the preparation of the Agricultural Credit Prcject was whether or not BAB was an appropriate lending channel, given IDA's experience with the Livestock Projects. In 1968, a a consultant firm. studied BAB's accounting system and proposed changes. The supervision report of February 24, 1969 of the ongoing Livestock Project related that the reorganization of the accounting system had started. A report of October 2, 1973 concerning a Bolivian delegation's meeting with Mr. Alter (RVP) stated that the reorganization would now take much longer than anticipated. The delegation inquired if IDA would consider implementation of the proposed Agricultural Project by the Agricultural Development Fund of the Central Bank. This course was rejected on the grounds that a major objective of the project was the long-term improvemnt of BAB. An earlier proposal to split BAB into two regional banks had also been considered (261-BO Supervision Report, January 3, 1973). 11.05 A decree reorganizing BAB was promulgated on July 26, 1974, but was not implemented to any substantial degree. As detailed in para 4.14, IDA agreed to finance under the Livestock Project another study for reorganization of BAB as a condition ':or presentation of the Agricultural Credit Project. This study was done and the recommendations were approved in 1976. By late 1978, although a "partial" reorganization had been implemented, BAB's situation continued to worsen. 11.06 In 1979, USAED, as a condition to a proposed grant and loan to BAB, requested another reorganization study. This study was only partially completed. In 1980, IDA requested yet another study of BAB as a condition of processing a follow-on Agricultural Credit Project. The study was completed in 1980 and a commission was formed to study the report. Its findings are yet to come. Agricultural Lending Policy 11.07 The reason for BAB's recurrent difficulties has been the lack of a sound agricultural credit policy on the part of the state. The interest rates fixed by the Government for subloans are not sufficient to cover BAB's cost of borrowing and administration. In addition, liberal rescheduling and subloan writeoffs constantly erode BAB's capital. Moreover, most of BAB's borrowings are in US dollars, whereas it relends in Bolivian pesos. As a result, the devaluations of 1972 and 1979 increased the difference between BAB's obligations in US dollar amounts and the dollar equivalent of pesos - 52 - owed to it at the new exchange rate. Since it is a matter of public policy that BAB should absorb the losses, IDA insisted on indexation of outstanding subloan balances for beef and grape subloans. 11.08 Borrowers were given the choice of (a) adjustment by the ratio of the new to the old exchange rates on the occasion of a devaluation; or (b) adjustment by the annual change in the La Paz Consumer Price Index each year. All chose the "dollar clause" option. Following the two devaluations, the adjustments were made after protracted disputes. However, many beneficia- ries simply responded by defaulting on repayments. Table 11.1 is a summary of BAB's financial performance during the period 1978-1980. In 1978, interest income covered less than one-half that year's expenses, with similar losses in the following years. In summary, expenses are running from 17% to 22% of the loan portfolio whereas income is only some 10% to 11%. Table 11.1 - Financial Performance of BAB 1978 1979 . 1980 (Est.) --------- ($b millions) ----------- Loan Portfolio 1,114 1,598 1,957 Income 127 160 209 Expense: Administration 101 105 124 Interest 100 85 111 Loan writeoffs 53 43 72 Other 1 45 27 Total 255 278 334 Operating loss 128 118 125 Ratios: income/loan portfolio 0.11 0.10 0.11 expense/loan portfolio 0.22 0.17 0.17 Source: Appraisal of Santa Cruz Agricultural Development Project. 11.09 A special feature of the project was the setting up of a project management unit to provide, among other things, close supervision of subloans. Table 11.2 shows that, at the project closing date, some 20% of subloans were overdue and, of these, 19% are in the process of legal action for recovery. The figures range from 13% for the Agricultural Credit Project to 30% for the Third Livestock Project. The difference is said to be due to better selection of clients. - 53 - Table 11.2 - Status of Subloans December 31, 1980 Classification of Portfolio by Percent Total Normal Overdue Under One Year In Legal Process 561-BO 100 86 1 13 261-BO 100 69 1 30 107/171-BO 100 78 5 17 Total 100 80 1 19 Of the first two livestock projects, all the overdues in process for recovery are for beef. In the Third Livestock Project, the beef overdues total 31% of the portfolio and sheep, 24%. A distinct improvement is shown in the Agricul- tural Credit Project with improved management performance of the Project Unit. Beef overdues fell to 13%; sheep, to 4%; and annual crops to 9%. Grape defaults were the highest in this credit, at 19%. The only class of subloans for the whole program with repayments that can be considered satisfactory are those for sheep in the Agricultural Credit Project. In reality, the arrears situation is probably worse because of rescheduling unpaid debt. BAB and Alternate Lending Channels 11.10 From the history of attempts to reform BAB, it is clear that struc- tural reorganization without basic changes in public policy for agricultural lending has brought no more than marginal gains. Until the Government allows BAB to: (a) set interest rates that cover the cost of borrowing; (b) take effective legal action to recover overdues; (c) discriminate in its choice of clients; and (d) rationalize its numbers and use of staff, efforts to improve BAB's performance at the level of field supervision and accounting systems are bound to have limited impact on the institution's overall worth as the major channel for farm development credit. Over the years, it appears that the many governments have been unwilling or unable to make the required changes. Besides debilitating BAB, the interest rate policy has led to disinterest on the part of commercial banks to be involved in agricultural development lending, thereby reducing the flow of funds to the sector. Under the Agricul- tural Credit Project, a serious attempt was made to involve these banks and 10 signed Supplemental Agreements to become participating banks. However, only one commercial bank actually participated and granted just 19 subloans out of a total of 1,139. This bank had no problems with arrears and, although it did not supply detailed da:a as requested for the Completion Reports, its executives expressed complete satisfaction with the program and wished to be involved in future operations. The key difference was that this bank was able to choose its clients primarily on the basis of creditworthiness and the cost of subloan administration. As a result, the average subloan size made was US$30,000 compared with the average figure of US$13,000 made by BAB to this category for beef deveLopment. - 54 - The Project Unit, ALPD and its Role 11.11 The Project Unit was established in 1968 to manage the first livestock project between IDA and BAB. It was subsequently expanded to cope with the changing scope of the follow-on projects (para 9.14). The initial directors were expatriates but were replaced by local staff when it proved impossible to find a suitable international recruit for the start of the First Agricultural Credit Project in 1975. 11.12 The performance of the ALPD has been mixed. On the one hand, the unit has not been able to bring down the level of arrears to an acceptable figure nor is the technical assistance it gives adequate. On the other hand, its achievements are considerable in bringing the concept of supervised farm credit to such a difficult area as the Beni and extending it to small sheep farmers in the Altiplano whose attractiveness to credit institutions is negligible. The consensus of the 30 or more supervision reports written since the program began with the First Livestock Project is that the deficiencies of the ALPD's performance are due to a lack of training and exposure rather than to any shortage of dedication and ability. It is noteworthy that the perfor- mance of the Project Unit seems to have been the same under local directors as under expatriate directors. The critical factor in determining the unit's performance is that the task set has been too much for the resources provided. First, in the absence of a functioning national extension and research apparatus, the Project Unit could not be expected to fill the vacuum across the range of activities eventually financed and across wide geographical diapersion of beneficiaries. Second, the ALPD, as a department within BAB, has been subject to the same general policies and problems with which that institution is beset. Any real progress in the ALPD's performance will depend on either the progress of the institution-building of BAB as a whole or on placing the ALPD in another institution altogether, such as the Central Bank. 11.13 In contrast to BAB's overall losses, the ALPD was able to show a modest profit of 5% on its operations with the Agricultural Credit Project. This was achieved in spite of BAB's insistence that the ALPD absorb the 50% to 100% of the cost of various personnel whose participation in the project was minor. At the time of appraisal of the Agricultural Credit Project, it was intended that the ALPD should be absorbed into BAB's general operations within 18 months. This was not done because of the failure of BAB to reach the minimum institutional development goals required to be able to absorb the ALPD without a negative impact on project performance. IDA Performance and Project Concept 11.14 After the relative success of the first two livestock projects, which were simple in concept and design, it became clear that factors other than simply shortage of credit were responsible for the slow development of the beef industry. At about this time, IDA's own ideas about income distribution effects of projects were rapidly finding expression in project design. As a result, the follow-on projects also sought to restructure the marketing system for beef in the hope that changes would be a stimulus to production. At the same time, more effort was put into a reorganization of BAB and its interest rate policies in the hope that its improved ability - 55 - would ensure a reliable delivery of credit to farmers. In addition, the inclusion of areas other than the Beni, the extension of coverage from beef to small farmer activities such as sheep, annual crops and grapes, and the focus on the farm size of intended beneficiaries sought to redress imbalances in regional development and income distribution. 11.15 The concept was logical but the question remains as to whether it was overly ambitious, given the limited capacity of the institutions concerned and the primitive state of developent of the Beni. Also, these secondary objectives were important enough of themselves to warrant another means of achievement than as side conditions in credit projects. For instance, the treatment of beef profitability as a function of state producer price intervention to be resolved by a study to set the "right" price is questionable. At no stage was it shown that producer prices were "wrong" with respect to public policy objectives. For instance, a comparison of the domestic producer price index for the period 1973 to 1978, given in Annex, Table 5, with the world price index, shows that adoption of world parity prices would have resulted in instability. The view that Bolivian producer prices were unusually low was gained because world prices were climbing during an upswing of the cyclical trend at the period of preparation of the Third Livestock Project. This view prevailed even during the dramatic fall in world prices that resulted from the oversupply because world prices were not "right" for long-run equilibrium. Second, the changes proposed in the studies involved added costs. No analysis was made as to who would bear these extra costs among producers, consumers and Government, how they were to be financed, and what the welfare implications would be. In other words, a more comprehensive treatment was required and it is not altogether surprising that the Government did not act within the required nine months after receipt of the consultant's study. The supervision reports did not consider these aspects but took a somewhat antagonistic view toward the Borrower's failure to act, an attitude that was not justified, given the uncertainties. 11.16 The four credit projects financed six studies costing almost US$1 million. In general, they generated a lot of data and helped quantify situations known to ex2ist, but the effective results have been few. No changes were made as a result of the land titling study, recommendations of the study of BAB were only partly implemented and the meat industry studies produced no changes. From these experiences, it seems clear that IDA's expectations of what a study can achieve were unrealistic. However, the studies were neither unjustified nor failures. They provided factual basis on which to structure subsequent discussions. Since their recommendations could only be partial, the real failure was the lack of subsequent dialog between IDA and the Borrower to modify those recommendations to fit the general system. The distinct impression gained from the correspondence aad reports is that neither IDA nor the Borrower was prepared to do this, but reached "stand-offs" over consultant's recommendations that were not imple- mentable, as given. For instance, nine months after the receipt of the meat marketing study in 1974 was not an opportune time to change the Borrower's price policy (i.e., increase prices) to stimulate exports as required by the covenants. The price of beef in world trade was at an all-time low and an actual producer price reduction was warranted in Bolivis for World parity. Subsequent events showed that it was not prices but the inability to reach - 56 - international trade standards that caused the loss of export markets to Peru and Chile. 1/ Producer price increases would have been regressive income transfers and would not have overcome the export problem. 11.17 In conclusion, the lesson learned is that studies have a valuable role to play in development but care needs to be taken as to when studies are to be used, what they can achieve, and how the recommendations are to be applied. It is also questionable as to whether a production credit project is the appropriate vehicle to which to attach a requirement for restructuring a major industry with all the implications for required changes in macro- economic policy. Targets 11.18 The achievements of both projects in terms of reaching numerical targets differed markedly from those expected at appraisal. This was not due to any deficiencies in design but to the nature of credit projects that cover many enterprises and regions. The period 1968-1980 covered by the projects was unprecedented in international price instability and local price inflation. Accordingly, the structure of demand, the cost of invest- ments financed, and the number of subloans adjusted to current conditions. For example, at appraisal of the Agricultural Credit, sugar prices were high but, by the time subloans began, the sugar industry was in a major slump and no subloan applications were received. Because of these factors, the appraisal and supervision missions concentrated on the policies, procedures and financial performance of the institutions involved rather than field-level development. Because of this, there has been a tendency to look at project problems as a function of the institutions rather than as farm problems caused by conflict between what the project was trying to do and what made sense to the sub- borrower. The sole criticism of IDA's performance made in the Completion Report of the Project Unit was that supervision missions tended to give little weight to the technical aspects of the difficulties encountered by the program. XII. CONCLUSIONS 12.01 Although implementation of both projects encountered numerous difficulties, the problems arose precisely because they tackled basic development issues in Bolivia. In addition to providing funds for farm investment, they were used as a means to change agricultural credit policies and procedures. The Livestock Project also sought to restructure the entire beef industry and the agricultural Project extended beneficiary coverage and focused on poverty alleviation. Because of the failure to attain specified targets, it would be easy to take an overly negative view of the results of 1/ FAO/CP Preparation Report No. 18/78 BOL.10 for a Meat Industry Project in Bolivia, 1978. - 57 - the program that is not warranted. The acceptance of indexation, the agreement to accept chattel mortgages and the building of the ALPD were important gains. Although certain features of the projects, as appraised, could be brought into q,aestion, the view is held that the results would have been much the same under any system of project design. The limiting factor in the failure to achieve the institution-building and beef industry restructuring goals was the Government's unwillingness or inability to follow through. The studies, in general, gave a good analytical basis for needed reforms that is still valid. 12.02 The Government's failure to act may have been due to the partial treatment given by the recommendations. For example, the consumption implications of meat producer price rises were not followed through in the studies. Given the best intentions in the world, it is difficult to see where the state should begin to tackle macro-economic reforms based on partial micro-economic analysis. Although projects are appropriate vehicles for financing studies, the implementation of study results should not be a condition of the same project. Implementation of recommendations of studies with wide implications are probably best left to the new Structural Adjustment Loans. 12.03 The results of the four agricultural sector credit projects suggest that future projects should be limited in the number of objectives each can pursue. The closure of the Third Livestock Credit Project was delayed by three years because of the beef studies, which are still not completed 1/ nor are likely to produce changes. Second, the infrastructure deficiencies in Bolivia impose special burdens on supervision. This has become a chronic problem in the Beni where floods, lack of roads and absentee ownership make regular and effective supervision unusually difficult and costly. More emphasis would be placed on initial creditworthiness, previous credit history and the pursuit of a more vigorous policy on collection of payments due. The experience with the one commercial bank that cooperated in the First Agricultural Credit Project indicates that Beni beef development would be better handled as commercial credit. Third, the lack of a functioning research and extension system necessarily requires modest goals in improving farm technology for a supervised credit project covering diverse enterprises, regions and target groups. The alternative is either to bear the high cost needed for effective results, which would probably be best left for a specific extension project, or to reduce the coverage. In summary, the experiences suggest that the major objective of a follow-on project should be to develop BAB into a reasonably strong and financially autonomous credit institution in order to ensure a continued flow of credit to the agricultural sector of Bolivia. 1/ Execution of the Eecond marketing study to cost US$250,000 was passed on to the UNDP to allow closure of 261-BO.  ANNEX 1 Table 1 BOLIVIA THIRD LIVESTOCK CREDIT PROJECT (CREDIT 261-B0) Beef - Subloan Classification by Size Classification Number of Loans % Amount % (US$) 0-2,500 41 7 51,789 1 2,500-5,000 254 45 966,023 15 5,000-10,000 99 17 889,700 14 10,000-15,000 45 8 573,216 9 15,000-25,000 74 13 1,361,170 21 25,000-50,000 41 7 1,392,300 22 50,000-75,000 16 3 1,143,800 18 Total 570 100 6,377,998 100 Source: PCR of ALPD. Table 2 Sheep- Subloan Classification by Size Classification Number of Loans % Amount % (US$) 0-2,500 8 .7 16,740 1 2,500-5,000 14 13 47,630 3 5,000-10,000 21 20 190,639 11 10,000-15,000 23 21 326,667 18 15,000-25,000 20 19 474,607 26 25,000-50,000 19 18 617,977 35 50,000-75,000 2 2 111,395 6 Total 107 100 1,785,655 100 Source: PCR of ALPD. - 59 - ANNEX 1 Table 3 BOLIVIA FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-B0) Beef - Subloan Classification by Size Classification Number of Loans % Amount % (US$) 0-2,500 22 5 57,157 1 2,500-5,000 144 31 608,505 10 5,000-10,000 .70 15 499,548 9 10,000-15,000 79 17 1,002,323 17 15,000-25,000 97 21 1,858,309 31 25,000-50,000 54 11 1,896,023 32 Total 466 100 5,921,865 100 Source: PCR of ALPD. Table 4 Grapes - Beneficiaries' Grape Area Before Subloan Total Tarija Cochabamba Number of Area in Hectares Beneficiaries % Number % Number % 0 116 68 106 62 10 6 1 35 20 33 19 2 1 2 16 4 6 4 - - 3 2 1 2 1 - - 4 4 2 4 2 - - 5, 6 2 1 2 1 - - Greater than 7 6 4 5 3 1 1 Total 171 100 158 92 13 8 Source: PCR of ALPD. - 60 - ANNEX 1 Table 5 BOLIVIA FIRST AGRICULTURAL CREDIT PROJECT (CREDIT 561-B0) Index of Meat Prices (1973 = 100) 1/ 2/ World Price- Bolivian Price- 1968 43 1969 41 1970 47 1971 57 1972 68 1973 100 100 1974 83 133 1975 45 133 1976 53 134 1977 56 140 1978 64 156 1/ Composite World Indicator Price Index. 2/ Index computed from official producer price series. 4 1 BRD 1241 SEPTEMBER 1974 FIRST AGRICULTURAL CREDIT PROJECT Mafn roads .:-- Radroads .~~~~R -v s n - Rers SA- -N - Provincial boundaries S-- I .nternational boundaries ~U0, It - --B R A Z 1 L Santa Ana L \~ 0 omaereo Reyes TRINIDAD L A0 OSan Bortax- P A Z PAToos Santoso Vc C OC hBAM B A ~ cauz 7SanJosé SUCREZ - -< -Puo 71m-N (~$P A. R A G U A y - H UJ J U i CcA.O HI'UI.~ jAMERICA \--P O T< 5iSO TAR iJ A T 4.-.Manor -ocean - - .0.  BOLIVIA THIRD LIVESTOCK DEVELOPMENT PROJECT PROJECT AREAS NbT ANOETTASA RT A N B R A[MAUUAILEA OILIUNCIA SUCR ER .RA0NMLATOS JYU CHABBA OF ICE QUECHN SLA --- + - JALOAS MAY 197MAINSDROADS 2 TARIJA...... ...PROV.C.AL.BONDAR..E CL T l A7Q NKUIM TE2E0 ý0~~ *,UNI ARC R 8 1\i MAY 191 IBRD2897R