Mexico - Rainfed Agricultural Development Project

Document of The World Bank FOR OFFICIAL USE ONLY g8o Report No. PROJECT PERFORMANCE AUDIT REPORT MEXICO RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) JUNE 29, 1990 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official dutles. Its contents may not otherwise be disclosed without World Bank authorization. ABBREVIATIONS BANRURAL National Bank for Rural Credit BTO Back-to-Office (Report) CIMMYT International Maize and Wheat Improvement Center DT Rainfed District FIRA Agriculture Trust Funds of the Bank of Mexico FICART Trust Fund for Credit in Irrigated Areas IBRD International Bank for Reconstruction and Development LAC Latin America and Caribbean LLDP Lilongwe Development Project M&E Monitoring and Evaluation NAFINSA Nacional Financiera, S.A. OED Operations Evaluation Department PCR Project Completion Report PIDER Investment Program for Rural Development PLANAT Natior.al Plan for Rainfed Agriculture Development PPA Project Performance Audit PPAR Prcject Performance Audit Report PROCATI Producer Organization, Training, Extension and Research Project PRODERITH Program for Development of the dumid Tropics PyC Policies & Coordination Directorate SARH Secretaria de Agricultura y de Recursos Hydraulicos (Ministry of Agriculture and Water Resources) T&V Training and Visit FOR OFFICIAL USE ONLY THE WORLD BANK Washington, DC. 20433 U.S. OIke as OectrwWal Operatim Evetuum June 29, 1990 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on Mexico Rainfed Agricultural Development Project (Loan 1945-ME) Attached, for information, is a copy of a report entitled "Project Performance Audit Report on Mexico Rainfed Agricultural Development Project (Loan 1945-ME)' prepared by the Operations Evaluation Department. Attachiuent This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT MEXICO RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) TABLE OF CONTENTS Page No. PREFACE ........................................................... BASIC DATA SHEET .................................................. . 11 EVALUATION SU1MARY .............................................. v PROJECT PERFORMANCE AUDIT ............... ................ 1 I. BACKGROUND ....................... 1 II. DESIGN ........................ . 3 III. IMPLEMENTATION ........................................... 7 IV. IMPACT .......................................... .. 12 V. FINDINGS AND ISSUES .................................... 15 COMMENTS PROM THE BORROER Annex 1 - Comments from the Ministry of Agriculture and Water Resources ................................... 19 Annex 2 - Comments from the Ministry of Agriculture and Water ResourcesNational Institute of Agricultural, Livestock and Forestry Research ...................... 23 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 MEXICO RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) PREFACE 1. This is a Project Performance Audit Report (PPAR) on the Rainfed Agricultural Development Project, generally known by its acronym, PLANAT, involving an IBRD loan in the amount of US$280 million to Nacional Finan- ciera S.A., guaranteed by the United Mexican States. The objective of the project was to increase agricultural production in nine rainfed districts, mainly through adaptive research and extension, as well as laying the basis for a subsequent national rainfed agricultural program in the other 115 rainfed districts. 2. Loan 1945 was approved on January 22, 1981; the Loan, Guarantee, and Project Agreements were signed on March 2, 1981 and became effective on July 1, 1981. US$82.9 million was cancelled on at the request of the Bor- rower. The original closing date of June 30, 1986 was extended to June 30, 1987. Final disbursement was made in April 3, 1986. 3. The PPAR is based on the Project Completion Report (PCR), prepared by the LAC Region and issued in 1990,1/ which was based on a draft comple- tion report prepared by the Borrower, as well as on the Staff Appraisal Report, the President's Report, the loan documents, the transcript of the Executive Director's meetings at which the project was considered, on a study of project files, and on discussions with Bank staff. An OED mission visited Mexico in November 1989, and discussed the effectiveness of the Bank's assistance with the representatives of the Agriculture and Irriga- tion Ministry (Secretaria de Agricultura y de Recursos Hydraulicos), the agricultural research institute (Instituto Nacional de Investigaciones Forestales y Agropecuarias), FIRA (a Government trust fund for agricultural lending). and Nacional Financiera (the borrower). Their kind cooperation and valuable assistance in the preparation of this report is gratefully acknowledged. 4. The PCR provides a satisfactory account and assessment of the project experience, and discusses the performance of the Bank and the project executing agencies. The PPAR elaborates on particular aspects, notably: the project's successes in coordinating many activities at dis- trict level and at obtaining popular participation & maintenance of project civil works; the dubious rationale for inclusion of agricultural credit and for the extraordinary increase in project size during appraisal; and devel- opment options when a project becomes unsustainable because the borrower considers it 'gold-plated.* 5. Following standard OED procedures, copies of the draft PPAR were sent to the Borrower for comments on April 5, 1990. Comments were received from the Ministry of Agriculture and Water Resources. Those from the Guanajuato State Delegate are attached as Annex 1; those from the National Institute for Agricultural, Livestock and Forestry Research are attached as Annex 2. 1/ Project Completion Report, MEXICO - Rainfed Agricultural Development Project (Loan 1945-ME) Report No. 8335, February 2, 1990. - it - PROJECT PERFORMANCE AUDIT REPORT MEXICO RAINFED AGRICULTURAL DEVELOPIUMT PROJECT (LOAN 1945-ME) BASIC DATA SHEET KEY PROJECT DATES Date of Board Approval 01/22/81 Date of Effectiveness 07/01/81 Closing Date 06/30/87 CUMULATIVE DISBURSEMENTS FISCAL YEARS 1981 1982 1983 1984 1985 1986 1987 Appraisal Estimates Annual 18 39 44 43 44 43 49 Cumulative 18 57 101 144 188 231 280 Actual Disbursements 0 34.82 60.63 112.22 180.37 197.12 - Actual as % of Estimate 01 61.1% 60.0% 81.4% 95.9% 85.3% Date of Final Disbursement: 04/03/86 Amount Cancelled : US$82.88 million Repayments : US$85.42 million (as of October 30, 1988) INTERNAL RATES OF RETURN Appraisal Estimate Actual Economic Rate of Return 21% a/ Financial Rate of Return 25%-38% b/ 6.441 c/ a/ The ERR was not calculated in SARH's PCR. In addition, it is no longer Bank practice to calculate ERR for projects with large credit and extension components. b/ Represents the range of expected financial rates of return for the different investment plans evaluated in the SAR. c/ Financial rate of return of the whole project as reported by SARH in its "Informe de Terminacion del Proyecto." The report does not offer a financial rate of return by investment plan or the assumptions upon which the 6.441 is based. However, when adjustments are made to exclude short-term credit from the net cash flows and to include incremental operating costs at a constant rate from 1987 to 2011, a rate of return of 8.691 results. - 111 - PROJECT COSTS (In US$ million) APPRAISAL ESTIMATE ACTUAL COST a/_ Investment Local Foreign Total Local Foreign Total Category Costs Costs Costs Costs Costs Costs Civil Works 79.9 59.1 139.0 39.2 32.1 71.3 Equip. & Veh. 12.8 26.0 38.8 4.1 :).4 7.5 Sal. & Oper. Costs 133.1 8.8 141.9 91.2 74.6 165.8 Training & Studies 9.3 15.1 24.4 6.7 5.4 12.1 Credit 200.0 62.0 262.0 161.9 140.3 302.2 Total Base Line Costs 435.1 171.0 606.1 - Physical Contingencies 26.1 16.0 42.0 - Price Contingencies 104.0 45.0 149.1 -.- -.- -.- TOTAL PROJECT COSTS 565.2 232.0 797.2 303.1 255.8 558.9 ==mz= =m== =a= ==== ==I== 9/ Actual costs as reported in SARH, "Informe de Terminacion del Proyacto," September 1988, in constant 1980 Mex$, converted at average 1980 exchange rate, Mex$23=US$1. NOTEs Figures may not add up due to rounding. - iv - MISSION DATA Date No. of Staff/Days Specialization Performance Types of Lohr) Persons in field represented Rating Problems Identification 08/79 1 6 A n/a a/* Preparation 04/79 5 110 A,S n/ n/O Appraisal 02/90 5 60 A,B,D n/4 n/a Post-Apprleal OS/80 8 15 A,D n/s n/a Supervision 1 08/81 8 88 A,8,D 2 M,T Supervision 2 12/81 4 42 2A,8,D 2 M,T Supervision 8 06/82 1 12 A Supervision 4 09/82 2 21 A,E 2 MY Supervision 6 01/88 2 22 A,C 2 M,T Supervision 6 08/88 2 24 A,B 2 M,T Supervision 7 01/84 1 14 A 2 M,T P/ Supervision S 08/84 1 20 F 2 - Supervision 9 01/85 2 17 2A 2 M,F Supervision 10 11/86 2 18 A,E 2 1 Supervision 11 05/88 1 10 A 8 F Supervision 12 01/87 1 8 A 8 n/a Supervision 18 10/87 2 1 2A 1 n/ if 4 OTHER PROJECT DATA Borrower: NAFINSA Executing Agency: SARH Fiscal Year of Borrower: January 1 - December 81 Name of Currency (Abbreviation): Mexican Pesos (Mext) Currency Exchange Rate: Appraisal Year Average (May 1981): US$1.0 = 28 Intervening Years Average (1982-86) USS1.0 = 180 Lst Year Average: (09/86) project terminated balance of 358 m cancelled US81.0 a 575 A = Agriculturalist; 5 a Economist; C = Irrigation Engineer; D = Financial Analyst; E = Dep. Div. Chief; Fa Roads Specialist. 1 problem-free or minor problems; 2 a moderate problems; 8 = major problems; 4 a major problems, objectives will not be met. F u Financial; M Managerial; T a Technical; P = Political; and 0 = Other overall. 1 Partial in conjunction with other missions. Partial in conjunction with FIRA project. 2 Partial supervision of only roads component. 3/ Special focus on drainage. Partial, only roads. / Report not in Files. STAFF INPUTS (staff weeks) FY79 FY80 FY81 FY82 FY85 FY84 FY86 FY88 FY87 FY88 TOTAL Preappraleal 44.1 4.7 07 Appraisal 79.5 21.1 100.6 Negotiations 11.9 11.9 Supervision 8.4 22.5 24.9 17.8 8.8 9.0 2.0 1.4 89.9 Other .0 .2 .0 .2 Totale 44.1 126.2 56.6 22.5 24.9 17.9 8.8 9.0 2.0 1.4 298.8 * Excludes time spent by young professionals. -v - PROJECT PERFORNANCE AUDIT REPORT MEICO RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) EVALUATION SUMMARY Introduction 1. Seven-eighths of Mexico's farmers farming 3/4 of the cropped land depend on rainfall. Traditionally, the other 1/8, who benefit from irriga- tion, have benefitted most from public expenditures for the sector, getting an estimated 90 percent in the 1945-75 period. In 1979, Mexico discovered that its oil reserves were larger than hitherto realized; it seemed to be rich. Government wished to do more to help the rainfed districts that had been largely ignored and to replace growing food imports. 2. Government's earlier programs to assist rainfed agriculture had been limited to certain pilot areas. The noteworthy precedents for PLANAT (the common name for this project) were Plan Puebla and PIDER. Plan Puebla started in 1967 in Cholula District. It was a joint effort of the Chapingo Postgraduate College, CIMMYT (International Maize and Wheat Improvement Center) and the Puebla State Government. It focussed on applied research, extension, and helping arrange access to existing services, e.g., credit. PIDER started in 1973 and had IBRD support starting in 1975. It was a classical, multi-component, rural development program focussed in selected micro-regions. One result of IBRD involvement was to stress productive investments and high-potential micro-regions. By 1978, when PLANAT was identified, rural-development projects in general and PIDER in particular were being criticized as being too complicated and too multifaceted to manage. 3. Mexico moved rapidly from a country that seemed to be rich early in 1980 to severe financial crisis and retrenchment in 1982, all accom- panied by unprecedented and rising inflation. Objectives 4. PLANAT can be seen as an extension of Plan Puebla, or as a simpli- fication of PIDER. narrowing it to its agricultural components. Its main aim was to increase farm and ranch production in nine high-potential rain- fed districts, mainly through improved extension, applied research, and irrigation. Producers were encouraged to organize into groups, which had a say in productivity-related rural public works financed by the project: principally roads and storehouses; and in the project's investments in afforestation and in soil & water conservation. Credit under the control of PLANAT was added to the design to encourage adoption of the project's technical packages. 5. PLANAT was meant to evolve into a national program for all 124 rainfed districts (DTs) as soon as requisite technical staff could be trained. The project, therefore, financed training and administrative support at the national level. - vi - Emplementation Experience 6. While Government's financial picture was darkening almost from the date of loar effectiveness in mid-1981, PLANAT received reasonable budget allocations for its first two and one-half years (although the 1983 alloca- tion was released very late, disrupting project works). Thereafter, in a climate of increasing budget austerity, PLANAT's allocations were insuffi- cient to carry out the program. By 1985 it was clear that neither Borrower nor Lender considered PLANAT the model for future rainfeJ agricultural development, and by 1986, the Borrower asked to cancel the outstanding balance. As Government's resources for rainfed agricultural development became more constrained and prospects for future expansion of such programs waned, PLANAT came to be regarded as an unwarranted concentration of public resources in a small number of naturally-favored districts. 7. PLANAT's budgetary woes translated into a change in project compo- sition. Salaries and operating costs accounted for 14 to 21 percent of project expenditures in the first two and one-half years, about two-thirds of them in the last two years. Starting in 1 , IBRD disbursements were kept up in the face of waning Borrower budgetary support by increasing the disbursement percentage from 45 percent to 70 percent under the Special Action Program. Budgetary stringency favored the credit component, for which counterpart funds came from the banking system, over the civil works component, whose counterpart funds came from the budget. 8. With negative real interest rates to producers, there was a very strong demand for PLANAT credit. Credit accounted for 56 percent of PLANAT expenditures from mid-1981 through 1982, 64 percent in 1983, and rose to 80 percent in 1984 as the civil works program shrank and the Special Action Program began, all of this despite many months during which agreements not to submit credit disbursement requests were in effect because renk and Borrower had not reached agreement on interest rates as legally required. These were problems affecting all Bank agricultural credit lending to Mexico. PLANAT's credit component had a number of problems of its owns agreement on short-term interest rates, establishing lincrementalityl of short-term credit, absence of required district-level credit committees, submission of disbursement requests for sub-loans made at non-qualifying interest rates, and, most important, disintegration of the linkage between PLANAT investment packages and credit--the justification of inclusion of credit under PLANAT in the first place. Nevertheless, in the event, credit accounted for 72 percent of loan disbursements and 54 percent of PLANAT expenditures. The loan agreements had to be amended to permit this expan- sion of the credit component, which would have expanded further had further amendments permitted; the credit categories were overdrawn long before loan closing. The bulk of the credit-component expansion was for short-term credit, which ended up accounting for 51 percent of the credit and which could not have been closely linked to farmer adoption of PLANAT investment packages. Results 9. PLANAT's national-level records, including those of the M&E unit, were wiped out by the earthquake of September 1985. The Borrower made a major effort to assess PLANAT's results before preparing a PCR, inter alia conducting an exhaustive survey of 824 sample beneficiaries. This scienti- - vii - fically-conducted survey shows that beneficiaries were surprisingly uatis- fied with public services, such as extension, input supplies and credit. The evolution of their farm inputs and outputs is traced. The net incre- mental farm and ranch benefits are then compared with all PLANAT costs of all components at prevailing prices converted to constant Pesos. The resulting internal rate of return is 6 percent. 10. This project-returns' calculation is far better than the usual in PCRs and clearly understates returns in some ways, e.g., afforestation and soil conservatiou costs are taken into account but not the benefits. Howe- ver, benefits are based on the before-after difference, not the with- without one. Casual empiricism indicates that PLANAT districts' agricul- ture did very well during the period in question, but a causality problem remains. How much of the increase was a secular increase in all Mexican rainfed agriculture, how much due to .te selected districts' superior resource endowment, and how much to PLANAT? 11. Beyond what may be, with difficulty, measurable in agricultural production, PLANAT produced some other lasting things of value. Targets were exceeded in minor irrigation and far exceeded in community warehouses. The survey indicates that these are maintained and appropriately used, undoubtedly a result of PLANAT's successes organizing producers' partici- pation. Feeder-road implementation, however, was well short of target, consisted almost exclusively of new roads, not rehabilitation, and has been characterized by poor maintenance. PLANAT supported a great deal of train- ing of lasting value, though more procedural short courses and not as many advanced technical courses for extensionists in Mexico as projected. PLANAT-supported research was important. Forestry and soil-conservation achievements, though well short of objectives, were significant. 12. PLANAT's long-run institutional objectives were not achieved. PLANAT did not become the organizing model for Mexican rainfed agricultural development. What PLANAT did to improve research and extension has not been sustained in the ensuing budgetary retrenchment, except that PROCATI, a largely extension project partly financed by IBRD Loan No. 2859, has tried to build on PLANAT's successes in its four most-successful districts, those in the Central Highlands. In these districts in particular, the producers' organizations that PLANAT helped generate may be a lasting ins- titutional achievement. 13. In general, PLANAT certainly did not fulfill its objectives. When all the uncertainties are taken into account, its investments probably did not yield a return up to the opportunity cost of resources. The fact that honest doubt remains, however, indicates that PLANAT was not a total fai- lure and not too bad an investment. Sustainability 14. As per Para. 12 above, PLANAT has not been sustained as an insti- tutional model for rainfed agricultural development. It led to some sus- tained farm-level changes, whose discounted social value is positive but probably not enough to justify project expenditures (Paras. 9 & 10). It also financed other investments of lasting value, notably minor irrigation, community warehouses, training, research, afforestation and soil conser- vation (Para. 11). -viii - Findings 15. The Audit has five findings. 16. Firstly, under the assumption that resources were plentiful, PLANAT's estimated cost ballooned from US$60 million in the first Project Brief to US$817 million at Bank appraisal, not an expansion in area to be covered but in unit resources per district. The constraint that limited PLANAT to nine of 124 rainfed districts seemed to be trained technical staff, not money. With financial crises and budget austerity, putting 30 percent of the budgeted resources into 7 percent of the districts, the better-endowed ones at that, seemed and had become over-generous, a distor- tion. Cuts were appropriate. The pattern of cuts--a common one-- main- taining salaries and operating costs while cutting the investment program, w&s not good for development. 17. Secondly, PLANAT was justified within the Bank as bringing the T&V extension model, then being transferred to India, to Mexico. This notion cannot be found in the loan agreements and evidently did not go beyond the Bank. PLANAT officers who could be located by the Audit and others felt that Mexico has a perfectly good extension model in Plan Puebla. 18. Thirdly, it was not wise to include credit in PLANAT. One justi- fication, that doing so would help the Bank reform BANRURAL, was not taken seriously within the Bank at the highest levels and was not incorporated into the legal agreements. The principal justification vat that PLANAT control of credit would help establish a link between credit and PLANAT's investment packages with the credit fostering adoption. 19. There is no evidence that banka provided their borrowers with more technical assistance under PLANAT-financed loans than they would have on loans financed from other sources, e.g., IBRD's regular agricultural credit loans. There may, however, have been some investment incentive associated with preferential access to subsidized loans through PLANAT. This incen- tive was attenuated, firstly, by that fact that half of PLANAT's credit was short-term and not for investment and, secondly, by the progressive disap- pearance of PLANAT's investment program. On the whole, the credit compo- nent perverted the project and distracted it from its objectives. 20. Fourthly, PLANAT's efforts to build producers' groups and to include them in planning and decision-making seem to have improved th utility of its public works' program and delivery of services. Its suc- cesses in this regard seem to be influencing its successors. 21. Finally, while PLANAT was a simplification of PIDER, it still involved coordinating twenty-odd government agencies, a situation which development folk wisdom considers unmanageable. But some of PLANAT's dis- trict chiefs did manage this situation very well. Success seems to have been associated with the energy and dedication of the district chiefs, and with the existence of a well-tested technical package for PLANAT to promote, as was tht ease in its Central Highland districts. PROJECT PERFORMANCE AUDIT MEXICO RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) 1. BACKGROUND 1. Until Mexico created its 124 rainfed districts (DTs) by a presi- dential decree in March 1977, the irrigated areas had received the lion's share of technical assistance, Government money for agriculture, and atten- tion. There was no national program to assist rainfed farmers--roughly 7/8 of all farmers and farming about 3/4 of the cropped land. It was estimated that, from 1945 to 1975, only 10 percent of public funds for the sector went to rainfed agriculture. The rainfed farmers who got such public assistance were residents of areas benefitting from special programs, such as Plan Puebla, PIDER and PRODERITH or were getting loans and perhaps technical assistance through the public banking system (FIRA and the private banks working with it or BANRURAL). 2. What became PLANAT, as this project is commonly called, began in 1977 as a plan to coordinate these various activities and to develop an effective national technical assistance program for rainfed agriculture by 1982. PLANAT never reached that goal. However, from 1977 to the reorgani- zations of 1985, and especially while attempts were being made to implement PLANAT in nine pilot districts with Bank financial assistance (1981-6), PLANAT served as a major. perhaps the major blueprint for Mexican rural development. PLANAT was dropped, doubtless never to be revived, in 1986. That does not mean, however, either that it failed or that it failed to influence the future. 3. In what way was PLANAT different from the programs, including those supported by earlier Bank loans, for Mexican rural development? 4. Of the many such Bank loans, PLANAT was most similar to and builds most directly on PIDER (Integrated Rural Economic Development Program). Mexico initiated its PIDER in 1973 when integrated rural development was in vogue, and when the Bank's Blue Ribbon Panel on Rural Development, which had a Mexican member, was starting the Bank's formal entry into the field. The Bank supported PIDER with three loans: No. 1110 for US$110 million, approved in May 1975 and disbursing between 1975 and 1980; No. 1462 for US$120 million, approved in June 1977 and disbursing between 1977 and 1981; and No. 2043 for US$175 million, approved in July 1981 and disbursing bet- ween 1982 and 1989. Under PIDER direction, existing line agencies were to implement coordinated multi-sectoral investment packages to benefit the rural poor in selected microregions. PIDER financed investments princi- pally in irrigation, livestock, soil & water conservation, fruit, rural industries, rural roads, rural electrification, primary schools, health centers, drinking water systems, farmer organizations, agricultural exten- sion & applied research, marketing and credit. The Bank's chief contribu- tion was to persuade Mexico to shift PIDER's focus from infrastructure alone to include "productive" investments so that rural development could - 2 - ,pay its way," and to concentrate on microregions that were poor but which had high productive potential. During the early years of Bank involvement, it was the Bank's experimental Rural Development Division in the, then, Central Projects' Staff which dealt with PIDER. The Bank's manpower com- mitment for preparation, appraisal and supervision was exceptionally large. 5. OED has reviewed PIDER I and II, which influenced PLANAT's concep- tion, taking into account findings of many other studies of PIDER expe- rience.1/ PIDER I was audited in depth. There was considerable controversy regarding PIDER's impact in 1983 both within Mexico and within the Bank. It was recognized that the "productive" investments in irrigation, live- stock, soil & water conservation, fruit, bee-keeping, forestry, and fishe- ries (ca. 70 percent of project expenditures), had not, on average, had rates of return up to the opportunity cost of capital, and certainly not high enough to compensate for the supposed low returns to the "nonproduc- tive' investments as had been expected. The PCR estimated the economic rate of return to the *productive" 70 percent at 6 percent but the audit found that 6 percent was too optimistic. (Report No. 4617, Para. 104, p. 36) The crucial controversy boils down to whether the social returns to the 'nonproductive" investments which were not measured by the field impact were enough to justify the poor average economic returns to the 'produc- tive" components. The Borrower feels that they were. As the audit des- cribes its 'Some describe it <PIDER> as a path-breaking effort to introduce socio-economic development to Mexico's poor rural regions and, in the process, initiate improvements in Government agency perfor- mance and obtain knowledge regarding rural development strategy. Others view PIDER I as an inefficient public works program whose 'top-down' efforts to achieve quick resource transfers led to disappointing field impact and too little systematized knowledge or system building. It was probably, perhaps of necessity, some- what of both." (Report No. 4617, Para. 27, p. 8) While agnostic on PIDER I's over-all economic return, the audit argues that PIDER I provides lessons on more efficient and cheaper ways to attack the problem and to realize the project's undeniable benefits. 6. OED did not select PIDER II for audit because, as the PCR Preface states, 'its most important findings and lessons are similar to those noted in the PPAR of the first project." The 1986 Evaluation Summary for PIDER II states what had by then become the principal consensual finding: 'The project was over-ambitious and too dispersed geographically. Coordination between numerous line agencies responsible for project execution proved more difficult than anticipated." 7. PLANAT had another, non-Bank-financed progenitor--Plan Puebla. Starting in 1967, the Graduate College of the National School of Agricul- ture, Chapingo, assisted by CIMMYT (the International Maize and Wheat 1/ Project Performance Audit Report. MeAico - Integrated Rural Development Project (PIDER I) (Loan 1110-ME). OED Report No. 4617, June 30, 1983 and Project Performance Audit Report. Mexico - PIDER II Integrated Rural Development Project (Loan 1462-ME) OED Report No. 6333, June 30, 1986. - 3 - Improvement Center). kicked off its effort to help small farmers in Puebla District. As PIDER would later, Plan Puebla sought to mobilize local resources (credit, fertilizer & other input supply, extension and research) by working through existing agencies. Its focus, though, was very much narrower--agriculture, indeed, principally maize--and it was backed by unlimited access to the agronomic expertise of the Chapingo Graduate Col- lege and CIMMYT. Puebla District had good soils and rainfall, as well as good infrastructure and proximity to markets. This permitted major econo- mies with respect to most of the principal things PIDER spent money on. Plan Puebla was remarkably cheap and successful.21 8. One final word about the environment into which PLANAT was born - the period of oil optimism. Of course, the two dramatic increases in oil prices significantly affected Mexico and Mexico's economic view of itself. Only in the late 1970s, however, when it was discovered that oil reserves were much bigger than had been realized, did Mexican decision-makers think of their country as rich and as having surpluses to dispose of. PLANAT was a child of that period, an effort to see to it that some of the oil wind- fall was used to solve Mexico's recalcitrant problem of rural poverty, and to do so in a way that would be sustainable and productive.3/ 9. As the oil windfall proved to be less large and more temporary than anticipated in the halcyon days, PLANAT came to be regarded as gold- plated, as a distortion channeling excessive resources into a small number of pilot DT. There were signs of this transformation almost as toon as Bank's participation in PLANAT became effective in mid-1981 and the concep- tion had definitely changed by late 1982, the year Mexico brought the debt crisis to the world's attention by defaulting on its debt. Mexican infla- tion, traditionally in the 3-5 percent annual range, rose to the 12-30 percent range between 1973 and 1981 and reached crisis levels in 1982 when Government retrenchment began in a serious way. As we shall see, this caused the Borrower to back away from PLANAT. II. DESIGN 10. To respond to the question of Para. 3 above, PLANAT was different from PIDER in that it responded to the criticisms that PIDER was too com- plicated to manage. PIDER dealt with all rural investments; PLANAT nar- 2/ The best contemporary source on Plan Puebla is The Puebla Project 1967- 69. Mexico: CIMMYT, <n.d.>, 120 pp. For analysis, see Jairo Cano y Donald Winkelman. "Plan Puebla: Analysis de Beneficios y Costos," El Trimestre Economico (1972) XXXIX, #4 or William I. Jones. "Mexico's Puebla Project: Is There Hope for the Minifundistas?," International Development Review (1972) XIV, #2 with Delbert Myren's reply (IDR (1973) #1) and the author's rejoinder (IDR (1973) #2). 31 At the time of PLANAT's approval by the Board, January 22, 1981, the handwriting was already on the wall. Nevertheless, one Directoz queried the loan on grounds of Mexico's comfortable foreign-exchange position and implied that Mexico was a high-income country with a strong balance- of-payments position, a view countered by Mexico's representative who, inter alia, pointed to the US$6 billion 1980 current-account deficit and the substantial deficits projected to 1995. rowed the focus to only those directly related to agriculture. Even so, PLANAT district managers had to coordinate the activities of about 20 government agencies. PLANAT differed from both Plan Puebla and PIDER in that it was intended to be the framework for national intervention to develop rainfed agriculture, coordinating ongoing and future activities, as the Sub-secretary of Agriculture in charge explained to the Bank's identi- fication mission in July-August 1978. 11. The PCR's brief account of project design conveys a faithful gene- ral picture of the process. This audit wishes only to expand on five points: PLANAT's conception as a research-cum-extension project; the high quality of its preparation and appraisal; its extraordinary expansion between identification and appraisal, largely as a result of the addition of a large credit component; the justification of that credit component; and the selection of PLANAT's pilot districts. 12. Firstly, at the time of the above-mentioned identification mis- sion, PLANAT was conceived as a national technical assistance program for rainfed agriculture. As such, it was not expected to be too big. The emphasis, a la Plan Puebla, was on extension and adaptive research. The key worry was that technicians would be scarce and that the DTs would be pinching them from ongoing programs such as PIDER. A good deal of training for technicians was built into the project. The identification mission got high-level assurances that PLANAT would complement, not compete, and that PLANAT's expansion had been phased as a function of manpower availability and planned training. Nevertheless, from the beginning, there was resis- tance from existing organizations. PIDER management was apprehensive. 13. Secondly, PLANAT was prepared and appraised by both the Bank and the Borrower as carefully as a project of this type can be. The prepara- tion mission of May 1979 generated a series of specialized papers on soil & water conservation, agricultural production potential, agricultural train- ing, extension, and applied research. The mission included the former Director General of CIMMYT, one of the original principals of Plan Puebla, and an extension consultant. The Borrower's chief interlocutor with both the preparation and the appraisal mission of January 1980, the Deputy Director of Rural Development, was a former Bank projects' officer and hence had an intimate knowledge of the Bank. Almost all Bank members of both missions had an exceptionally intimate knowledge of Mexican agriculture. The appraisal has been described by eye witnesses as a rolling circus; green buses carried dozens of Mexican officials from the twenty-odd participating agencies, many of whom had not previously met, and the Bank team from district to district, in each of which there were more crowds of officials, effusive welcoming speeches, and the Bank mission leader's exhortation in fractured Spanish with Mexican participants cheerfully volunteering the missing words when he could not find them. The fact is, though, that that rolling circus generated as detailed and carefully thought-out an appraisal report as one could hope for. 14. Thirdly, it is worth elucidating the remarkable growth in project costs, noted by the PCR (Para. 4.2), from US$60 million with a US$6-9 mil- lion Bank loan in the January 1979 Project Brief to US$817 million with a $285 million Bank loan by the Decision Memorandum. The report of the second preparation mission of August 1979 indicates a seven-fold expansion from an extension-research-training project to a much larger and different one. Estimated project cost had grown from US$60 million to US$443 million (without contingencies). Credit, for which US$212 million or 48 percent of the cost of the project as then conceived was earmarked, accounted for most of the expansion. The rest was for infrastructure: roads, storage facili- ties, housing, wells and water storage. According to the contemporary report of the loan officer who usually shadowed these missions, the goal of including credit in PLANAT was "to increase agricultural credit coverage among rainfall dependent farmers," which was consistent with the objective of contemporary Bank lending for agricultural credit nationwide. 15. During the appraisal mission, the project base cost expanded by another 37 percent to the US$607 million ($817 million with contingencies) in the February 26, 1980 Issues Paper.4/ The expanded project, as embodied in the staff appraisal report slated spending project funds as follows: 46 percent for credit, 18 percent for research, 8 percent for roads, 8 percent for extension, 7 percent for soil & water conservation, 6 percent for irri- gation, 4 percent for PLANAT overhead expenditures, and 1 percent each for forestry and livestock. 16. Fourthly, the most contentious issue had been inclusion of credit, and for two reasons. Firstly, as the Bank's Loan Committee wanted to know, in view of the fact that the Borrower already had major agricultural credit projects under implementation with Bank financing, why could not PLANAT's credit needs be accommodated from existing sources under existing arrange- ments? The Region responded (April 23, 1980) that: a) with technology and inputs, credit would be part of an integrated package; having it under PLANAT control would assure its availabi- lity; b) in integrating credit with other district programs, PLANAT would increase BANRURAL's 5/ penetration of rainfed areas; and 4/ Although the loan henceforth is projected to be US$285 million, which it was, there continued to be small variations in estimated project costs as the appraisal mission completed its work, e.g. base cost of $567.8 million and total with contingencies of $771.6 million reported by Freeman on April 23rd and base cost of $605.9 million and total with contingencies of $797 million as per the SAR dated August 7th. These differences of ca. 6 percent are not material. 5/ Almost all Bank agricultural credit lending to date had been through FIRA, a central bank trust fund, which rediscounted qualifying loans originated by commercial banks or by BANRURAL. BANRURAL operations as a whole were characterized by extremely high operating costs, poor recoveries, inflexibility in lending norms, et al., but BANRURAL and not the commercial banks was the principal lender to the rural poor. The Borrower proposed to have PLANAT credit channeled not through FIRA but through FICAR (later FICART), another trust fund set up to allow the InterAmerican Development Bank to rediscount loans originated by BANRURAL only, without having to get involved in the creditworthiness of BANRURAL itself as a whole. -6- c) while it would be difficult to achieve the improvements in BANRURAL's lending standards and operations which the Bank sought nationwide, they could be achieved through PLANAT in specific districts. At the Decision Memorandum review in May 1980, the Bank's Senior Vice Pre- sident, prophetically as it turned out, expressed skepticism, particularly about point c): OIt is hard to believe that a section of BANRURAL will work well while the rest continues to flop around." The Staff's arguments were, nonetheless, accepted. 17. Secondly, the appraisal team's credit expert was concerned by: a) BANRURAL's rigid lending practices which used prescribed packages and took no cognizance of borrowers' individual risks and needs; b) BANRURAL's practice of financing 100 percent of investments without requiring any borrower financial participation; c) BANRURAL's lack of rural resource mobilization; and d) BANRURAL's high lending costs, bad-debt losses, operation losses and crop-insurance subsidy amounting to 30 percent of loans at that time.6/ The mission member formulated credit issues and recommendations on February 20th, they were sent to BANRURAL for comment on March 25th with a request for a continued dialogue,7/ and on April 29th formulated proposed *sugges- tions for minimum response levels or fall-back positions' for the May 1980 post-appraisal mission. Whatever happened on that mission,8/ the credit component was in the August 1980 Yellow Cover draft appraisal report, lead- ing Central Projects' Staff to comment (September 22, 1980) that the project "will not significantly alter BANRURAL's overall level of subsidy, and to hope "that this will establish a basis for the region to strengthen its dialogue with BANRURAL." Further Bank staff exchanges occurred over the SAR's not mentioning "that project credit will require a considerable subsidy from the Federal Government" (internal memo of October 15, 1980). 18. During negotiations, the above-mentioned points were not the focus of concern. Rather the Bank was worried as to whether the short-term cre- dit it would be financing would be "incremental" and whether Government 6/ This list does not include interest rates because the problem of negative real lending rates was being dealt with at that time on a country-wide basis. By legal agreement, whatever was agreed between the Borrower and the Bank in the context of the Bank's agricultural credit loans would apply to agricultural lending covered by other loans, e.g. PLANAT. 7/ The Bank's division chief said he looked forward to the Director General's early reply. There is no record of any. 8/ The Audit has not been able to locate a BTO report. - 7 - would raise interest rates on that short-term credit to something closer to long-term rates and to the rate of inflation (26 percent in 1980, 28 percent in 1981 and 59 percent in 1982). The loan agreements contained assurances on both these points, though the Borrower considered the latter such a sensitive aspect of sovereignty that the loan agreements specified only that Bank financing of short-term credit would be subject to Borrower- Bank agreement on interest rates. 19. As regards the concerns of the appraisal mission's credit expert, the agreements required the Borrower's credit rediscounters, whether BANRURAL or FIRA, to adopt for PLANAT purposes the same norms that the Bank and the Borrower had agreed with respect to FIRA in the contemporary agri- cultural credit loans. These norms required a borrower participation rang- ing from 5 percent to 20 percent. They also required that each loan be appraised. Thus they dealt, to some extent, with the appraisal mission's concerns about BANRURAL rigid lending practices and lack of farmer contri- bution (Para. 17, a) & b) above). The other points--BANRURAL's (and FIRA's) failure to mobilize resources, and BANRURAL's deplorable record on administration and loan recovery (Para. 17, c) & d))--were not addressed. 20. Fifthly, PLANAT's nine rainfed districts (DTs) out of Mexico's 124 9/ were presented in files, in the staff's Board presentation, and in the PCR (Summary, Para. 2) as representative of the country's rainfed universe. In fact, the 9 are representative of the best and highest-potential of Mexico's rainfed areas: Cholula, Morelia, Zapopan and Jalpa of the Central Highlands; Martinez de la Torre, Coatzacoalcos, Cardenas, and Emiliano Zapata of the Caribbean Coast; and Chilpancingo in the Pacific Highlands. Cholula is even the same district where Plan Puebla started back in 1967. Part of project design was to concentrate efforts where they were most likely to pay off, just as it had been in PIDER. The Audit commends this strategy. By the same token, however, the nine PLANAT zones are not representative of Mexican rainfed agriculture as a whole, nor could one have expected to generalize results from these favored DTs to the rest. III. IMPLEMENTATION 21. The PCR's brief account of implementation (PCR, Paras. 5.1-5.9, pp. 4-5) is basically correct. Initial Borrower enthusiasm for PLANAT waned and disappeared. PLANAT's budget, initially adequate, was released later in the year, and got progressively smaller until Government signalled that there would not be any funds in 1987. As Borrower commitment waned, Loan 1945 disbursements were kept up by putting the project under the Spe- cial Action Plan and increasing the disbursement percentage from 45 percent to 70 percent, but this did not prevent project slow-down. Civil works slowed down most, while credit, for which the issue of scarcity of counter- part funds did not arise, took up the slack. After re-allocating funds to credit and being unwilling to re-allocate more, Lender and Borrower dis- cussed using the remaining balance for rehabilitation from the September 9/ During project implementation, some larger DTs were sub-divided, so that later PLANAT covered 11 of 139 DTs. - 8 - 1985 earthquake and/or for road-maintenance machinery,10/ but finally agreed to cancel the final 24.1 percent of the loan (US$67.5 million).11/ 22. At the national level, PLANAT worked well for planning & budgeting purposes, but never for technical ones. Technical quality and much else depended on PLANAT leadership in the districts, which varied a lot, or on national organizations that were implementing components of the project. This had implications for extension and for road construction. In the case of the latter, road works were almost exclusively for new roads, not 70 percent for road rehabilitation as envisaged at appraisal, and PLANAT had nothing to do with maintenance. 23. This Audit will expand on the PCR's account of implementation, based mainly on information from the Borrower's PCR and from field inter- views. 24. Borrower Commitment & Provision of Funds. The PCR gives the impression (Para. 5.1, p. 4) that the Borrower was disenchanted with PLANAT almost from the time the Bank's loan became effective in mid-1981. "The most critical variance between planned and actual implementation was the lack of sustained counterpart funds for rainfed district-;.... the new Administration of 1982 became sensitive to PLANAT<'s> channeling large investments to <a> few rainfed districts and was not inclined to further 'distort' the activities of DGDUT." The PCR refers to its Tables 3 and 4 which do not, however, show Borrower investments in PLANAT. 25. A review of supervision reports and of the Borrower's PCR (Table 2 principally, which is in constant Pesos) shows that PLANAT got the budget it needed in 1981 and 1982. It did in 1983 too, but only after a protract- ed fight, and then the money was not released until mid-year, upsetting the construction program. In 1984, 1985 and 1986, PLANAT budget allocations were not enough to fund its planned activities or even to maintain previous levels; and 1984 funds, for example, were released in September. Project expenditures (in constant Pesos) and Loan 1945 disbursements by year are shown below. mid-1981 1983 1984 1985 1986 (Cancelled) TOTAL thru 1982 PLANAT Expenditures 28% 18% 27Z 22Z 5% 1002 Loan 1945 Disbursements 26% 13? 39Z 21% 1% 100% (net of cancellations) Loan 1945 Disbursements 18.5? 92 28Z 15? 0.5? 30% 1002 (of amount approved) Funds were adequate for activities PLANAT was ready to implement in the first 18 months. In 1983 implementation was constrained by late release of 10/ Dropped finally because the Borrower did not agree to purchase it using international competitive bidding. 11/ Another 5.5 pezcent (US$15.4 million) had been cancelled earlier (September 7, 1984). - 9 - funds, by the time needed to prepare more civil-works sub-projects for the pipeline, and perhaps because credit disbursements were blocked for 3.5 of the 12 months by an agreement not to submit disbursement applications. In 1984 and 1985, the considerable drop in effective Government commitment was masked by IBRD's higher disbursement percentage, raised from 45 percent to 70 percent under the Special Action Program. PLANAT dribbled over into 1986. 26. The first evidence of Borrower concern about PLANAT "distortions" and "gold-plating" occur in a letter from NAFINSA's International and Trea- sury Director in September 1984. "Among the 150 rainfed districts that are waiting with similar hopes, approximately 30 percent of budget resources have been assigned to the 11 Loan districts, so that any budgetary realloca- tion that would accelerate disbursements would result in a bigger concentration and, consequently, distortion of DGDUT activities." In view of what the Director calls "the changes in the country's economic reality," he finds even greater concentration of resources on a small num- ber of the best-endowed rainfed districts unacceptable. 27. There was a remarkable gap in Bank supervision between January 1984 and February 1985. While both of these missions rate PLANAT a project with moderate problems, the former does not convey a sense that there are major problems, while the latter does. By the next supervision mission of November 1985, PLANAT was clearly dead and discussion was over how to bury it decently. As the PCR notes (Para. 5.4, p. 5), the administrative reor- ganization of February 1985 effectively made PLANAT irrelevant and made it obvious to all that the Borrower did not consider it the model for future intervention in rainfed agricultural development. Bank staff agreed. 28. Project Composition. The PCR notes (PARA. 5.8, p. 5) that credit was a bigger project expenditure than anticipated--54 percent of actual project expenditures and 72 percent of loan disbursements as vs. 43 percent of both anticipated at appraisal 12/ All of the expansion occurred in short-term credit. Actually, PLANAT's transformation from a research- extension project to a credit one was more substantial than those figures indicate. Credit accounted for 56 percent of project expenditures in the first 18 months, 64 percent in 1983 13/ and 80 percent in 1984, falling to 25 percent in 1985 and nothing in 1986. The 1984 bulge stems from the higher Bank disbursement percentage under the Special Action Program and the counterpart situation. The Borrower's (reduced) counterpart for credit could come from the banking system, while, for civil works and other categories, it had to come from the budget, which was severely constrained. 12/ 43 percent when "unallocated" is excluded, not the PCR's 38.6 percent when it is not. 13/ Despite the fact that credit disbursements were effectively suspended during ca. 10 months of the first 18 and 4 months of the next 12 because the Borrower had agreed not to submit disbursement requests of loans made during these periods because interest rates to sub-borrowers were not consistent with legal obligations under Loan 1945. - 10 - Credit's falling share in 1985 and 1986 stems not from falling demand (the demand for credit at highly negative real interest rates was, not surprisingly, very strong) but first because FIRA was excluded from PLANAT financing in 1985 and 1986, and then because IBRD refused to agree to further reallocations of funds to the credit category, which had been overspent. But for these restrictions, credit would have expanded to account for more than 70 percent of PLANAT expenditures. 29. If PLANAT turned out to be much more of a credit project than intended, even at appraisal, other components lost importance, except for salaries & operating costs. Appraisal Estimate Actual Expenditure Credit 43% 54Z Civil Works 23% 13Z Salaries & Operating Costs 231 30% Equipment & Vehicles 6Z 1% Training & Studies 4% 2% Civil works declined the most, and its composition changed. Minor irriga- tion and community warehouses for fertilizer and other things were far more in demand than anticipated; feeder roads and forest plantings ran far behind projections. Government's reluctance to keep putting its own funds into the project was the major factor in the civil works decline and in that of the equipment & vehicles category. The latter was also affected by the Borrower's refusal to use international competitive bidding, as well as by a sensible reluctance to provide vehicles and equipment to a project that was clearly being phased out in later years. 30. For a project that started to be for extension and research, underspending on training & studies is noteworthy. Moreover, what was spent in this category was disproportionately for procedural short courses and for foreign studies and study tours, for which there was a great rush in the last two project years when it became clear that PLANAT was being phased out. On the other hand, technical training of extensionists was short-changed. As the February 1985 supervision report said, 'The advanced technical training program has made little progress since the project began." 31. A final insight into the erosion of Borrower support for the project emerges when project composition is analyzed over time. From mid- 1981 to end-1983, when PLANAT still had some claim to being the wave of the future, credit made up 56-64 percent of project costs, salaries & operating costs made up 14-21 percent, civil works made up 14-15 percent, and others, 8 percent. In 1985 and 1986, when PLANAT was winding down, salaries & operating costs accounted for 2/3 of project expenditures. As the project was increasingly cut, they became most of what was left. 32. Credit Implementation Problems. To an extent that cannot be appreciated from either the Bank's or the Borrower's PCR, PLANAT credit implementation was fraught with difficulties. Some were just PLANAT's small share of sector-wide tractations between the Borrower and the Bank, - 11 - largely over interest rates.14/ The periodic agreements not to submit dis- bursement requests are an example. Others were more peculiar to PLANAT. 33. One PLANAT-specific problem was that of the district-level credit sub-committees. These were meant to meet the appraisal concern (see Para. 17 a) above) that BANRURAL actually tailor its lending to farmers' needs and that it practice sub-project evaluatioa, things BANRURAL traditionally did not do but which FIRA obliged the commercial banks to do for loans it rediscounted. Getting PLANAT to set these committees up and get them work- ing was a major concern of early supervision missions. Their purpose, howe- ver, was not so much to get BANRURAL to actually respond to farmers' needs, or to use credit to support the technical packages PLANAT was pushing--the justification of the credit component used to Bank management--as to "plan" credit. 34. By credit planning, the parties generally meant allocating credit between BANRURAL and the FIRA-commercial bank complex. At appraisal, it was expected that 80-90 percent of PLANAT's credit would pass through BANRURAL, 10-20 percent through FIRA. During this period, however, commer- cial banks were expanding rapidly and establishing new branches in the higher-potential rainfed districts--the PLANAT districts. They enjoyed a far better reputation for honesty and for responsiveness to clients needs than BANRURAL. The former FIRA Director had become Agriculture Minister at this time. These factors contributed to what several concerned Mexicans called the FIRA "coup." Between loan effectiveness in mid-1981 and the end of 1982, 69 percent of PLANAT credit passed through the commercial banks and FIRA. The district committees' planning generally consisted of cutting down FIRA's share. 35. Other credit implementation problems which supervision had to address repeatedly were the short-term interest rate, applications where sub-borrower interest rate did not conform to legal agreements, and failure to demonstrate the "incrementality" of short-term lending. On the basis of the last of these, FIRA disqualified itself from further short-term lending under PLANAT. Towards the end, all PLANAT lending was limited to BANRURAL. 36. These PLANAT-specific credit implementation problems may be consi- dered details, however, compared to the implementation problem raised by the February 1985 supervision mission. "We are concerned," it said in the supervision letter, "that these funds might not have been used exclusively to supplement the project's infrastructure and extension programs." Since the justification of having Bank credit lending under PLANAT and not just under the Bank's credit loans depended on this supplementation, the concern is a fundamental one. The mission recommended a study by September 1985, but none was carried out. This issue was one focus of the Audit and will be discussed under Impact and under Findings and Issues below. 14/ These are being reviewed in greater depth in a concurrent OED audit of five Mexican agricultural credit loans (Loans 1217, 1569, 1891, 2454 and 2610) which were disbursing between mid-1976 and late 1987. - 12 - IV. IMPACT 37. On Production. It is exceptionally difficult to estimate PLANAT's impact on production. The basis for such estimates was well laid. Prepa- ration and appraisal involved serious attempts to estimate what changes project-sponsored research, extension, irrigation investments, etc. would prompt, how many farmers would change how much, and to quantify all of these effects and aggregate them. Naturally, they were only estimates. PLANAT had a monitoring and evaluation unit to try to measure its impact. Bank supervision missions were routinely critical of the M&E unit, but they always are. Unfortunately, we shall never know how useful the M&E unit's work would have been for it was lost, along with most other PLANAT records, in the September 1985 earthquake. 38. In the absence of PLANAT's records, the Ministry of Agriculture's Policies & Coordination Directorate (PyC) made a major effort to find out what happened, assembling records from the district level and commissioning an in-depth study of 824 PLANAT beneficiaries. For that sample, the survey sought to determine not only the farm outputs and inputs and their values before and after PLANAT, but also the beneficiaries attitude toward public services and what they regarded as their major, unresolved farming problems. The results are presented in the Borrower's PCR and briefly summarized in the Bank's PCR. 39. The Borrower's PCR estimated the value, net of production costs, of incremental agricultural production--maize, beans, rice, wheat, sorghum, cacao, coffee, coconut, fruits, vegetables, meat, milk--and compares it with the entire cost of PLANAT during Loan 1945's disbursement period, using prevailing prices and in constant 1980 Pesos. Under these assump- tions, the internal rate of return is 6 percent and the net present worth is negative at the presumed opportunity cost of capital, 8 to 10 percent (see Borrower's PCR, especially Tables 20 & 21). 40. The Bank's PCR down-plays this effort to ascertain PLANAT's returns to the economys 'real impact cannot be ascertained not only due to data gaps but also problems of causality and correlation. Hence, benefits are at best tangential and impressionistic." (Para. 6.5, p. 6) It also says, Osuch rates of return are no longer required by the Bank for projects with credit and extension." (Para. 6.4, p. 6) The results are therefor presented "merely as points of information., 41. Changes in IBRD requirements and fashions should not be allowed to discourage a Borrower's legitimate desire to find out whether PLANAT was a worthwhile investment or nott So what of PyC's attempt? There are two reasons for believing that the 6 percent finding is conservative. Firstly, it includes only on-farm benefits, thus not the off-farm benefits of the forestry component, soil conservation, warehouses, not to mention the more elusive things like improved competence of extensionists, an expanded banking system, and other such to which both Borrower and Bank evidently attach considerable value. Secondly, the analysis, while global, is conducted in prevailing prices. These may understate the value of agricul- tural produce if the commonly-held view that Mexican agriculture suffers from negative net distortions is correct. For these two reasons, 6 percent may understate PLANAT's true economic rate of return. - 13 - 42. On the other hand, as the Bank's PCR says, there are problems of causality. The net production increments in PyC's IRR are before and after PLANAT, which is not necessarily the same as with and without. This problem is a common one. It recalls IBRD's justifications of the Lilongwe area-development projects in Malawi, first ignoring that similar production experience had occurred in the project district and in neighborIng, non- project districts, later attributing production increases outside the project area to the demonstration effect of the project.15/ Even if PyC had wished to compare with-project agriculture with results outside PLANAT districts the job would not have been easy. PLANAT had picked the best districts and neighboring districts would not be good comparators. 43. At the end of all this methodological discussion, we are left with two basic facts. Firsly, crop and livestock production did grow a lot in PLANAT districts over the period in question, probably more than would have been expected without-project given the economic and policy environment. This probably has something to do with PLANAT extension and other services, the 824-farmer survey's having revealed that farmers were remarkably satis- fied with those services. Secondly, the indeterminacy remains. As the supervision letter following the February 1985 mission said, "Despite improvements in yields and areas under crops in the PLANAT districts, the mission was unable to determine how much this was due to extension services and how much to non-project factors." PLANAT was certainly not a clear success as an investment project, but neither was it a clear failure either. 44. On Institutions. PLANAT disappeared without a trace. The Audit had quite a time tracking down PLANAT veterans, now scattered in many agen- cies. The administrative reorganization of February 1985 made PLANAT's national level obsolete. After PLANAT's demise, progress at the local level in research and extension was strangled by budgetary retrenchment. So one would expect that there was no institutional impact, in any case not a sustained one. 45. Well, not quite. Firstly, as PyC's survey shows, except for roads, PLANAT-built civil works get maintained and taken care of, much more so than is usually the case for Government-sponsored rural public works. That is probably because there really was popular consultation and partici- pation in establishing PLANAT's works program, a contrast with PIDER, where such programs were externally decided. 15/ Credits 113, 244, and 550, disbursing between 1968 and 1979, and reviewed by OED audits Nos. 751, 1597 and 3414 of 1975, 1977 and 1981 respectively. The PPAM of Lilongwe I is a classic discussion of the difficulties and pitfalls of determining the with-without net benefits of area-development projects, which, however, shies away from the accumulating evidence that there were not any. Despite that fact that, as the audit of Phase III said, 'Probably more has been written about the impact of LLDP <Lilongwe> since its inception than any other Bank project,' and despite Lilongwe's elaborate M&E efforts, the positive return found by the 3rd PCR for all three phases is suspicious. Most benefits were from increased production of tobacco, a crop not promoted by LLDP; there was little or no incremental production of the crops that were promoted, maize and groundnuts. - 14 - 46. That is the second point. As a former Director explained, PLANAT was the pioneer of producers, participation which is law today. The Direc- tor also insisted on the distinction between the dis-concentration previously practiced in Mexico, in which local or district representatives are subordinate branches of the center, and the truly autonomous decentra- lization of which PLANAT was the pioneer. This, of course, is the same phenomenon that troubled supervision missions as "lack of technical coordi- nation" and which led to the wide divergences in performance between PLANAT districts. 47. Thirdly, both PCRs emphasize the importance of human resource development under PLANAT. 48. Finally, PLANAT's successes in its most successful districts, those in the Central Highlands, including the technical packages it devel- oped, were picked up in PROCATI, largely an extension project partly financed by Loan 2859-ME. 49. It is important to remember some important institutional impacts that PLANAT did not have too, mainly in the context of arguments that were made to justify it in internal IBRD discussions. 50. Firstly, PLANAT did not convert Mexican agricultural extension to the Indian T&V model. Mexicans the Audit interviewed on this point seemed unaware that any such conversion had been contemplated or even what was being talked about. In their minds, Plan Puebla was the model and the staff of the Chapingo Graduate School and of CIMMYT were the gurus. They, after all, handled extension training for PLANAT. Some Mexicans did visit India for training under PLANAT, but they went to see Anand cooperative dairy development, not T&V extension. 51. Secondly, there is no evidence that PLANAT provided an entering wedge for reform of BANRURAL, an illusion cherished by some Bank staff but not by the Senior Vice President or the Borrower's representatives. 52. And thirdly, the existence of a link between PLANAT's credit and PLANAT's technical packages--used by the Bank to justify inclusion of credit in the loan--is highly suspect. Some Mexican observers argued that credit is fungible and that FIRA, at least, simply submitted loans in PLANAT districts under Loan 1945, freeing more funds to submit under IBRD's straight agricultural credit zones in other districts. That may be an overstatement. One key figure in the program convinced the Audit that there initially was linkage between credit and PLANAT packages, but that the linkage disappeared as the investment program disappeared. At that time of enormous inflation, there was an enormous demand for (highly subsi- dized) credit. Until it was prevented from doing so, FIRA just stepped in and used PLANAT as an additional source of funds for short-term credit, which certainly had little to do with stimulating investment and up-take of PLANAT's technical packages. As regards medium-term credit, FIRA and asso- ciated commercial banks offered some technical assistance with their credit, whether associated with PLANAT or not, and BANRURAL did not, whether associated with PLANAT or not. Nevertheless, PLANAT undoubtedly had some effect in accelerating the expansion of bank lending in its selected rainfed area-, both on the part of the commercial banks and of BANRURAL. - 15 - 53. Obviously, PLANAT fell far short of achieving its objectives. As an entity and as a concept for developing Mexican rainfed agriculture, it has been discarded. Its initial concept became grossly distorted by enlarging the financial size of the project, mainly by adding agricultural credit. Little of what it did has been sustained. But PLANAT was not a disaster. Subject to a number of the usual caveats, the productive invest- ments it fostered probably are having a return to the economy modestly below or about equal to the opportunity cost of resources. Most of the public works built under PLANAT are appreciated by their intended benefi- ciaries, are maintained, and are used for the purpose intended. V. FINDINGS AND ISSUES 54. The Audit wishes to highlight five findings. 55. Are the accusations of *gold-plating" and "distortion', directed against PLANAT ex post facto, justified? Well, yes, even though what hap- pened is very understandable. As a former Vice-Minister of Agriculture intimately involved with PLANAT tells it, in the philosophy of the late 1970s and very early 1980s, the country was rich. Government wanted self- sufficie.cy and wanted to do more for rainfed districts that had been largely ignored. Since these produced mainly maize and beans and other staples, the two wants fit together. With each revision, PLANAT got more generous. It was limited to 9 "pilot" districts not for want of funds or even because the concept was being tested (it was regarded as a refinement of what had already been learned through Plan Puebla and PIDER) but because there were not enough trained agricultural change agents and it would take time to train them. Suddenly, in 1981, the Government was poor. It was worse in 1982, and in 1983, and in 1984.... 56. We have seen what happened to PLANAT under these circumstances. Firstly, under assumptions of plentiful public resources, Borrower and would-be Lender allowed the project to grow very large indeed. Then, when public resources became constrained, PLANAT's generosity to the most- favored rainfed districts--30 percent of budgeted resources for the best- off 7 percent of the districts as per the Director's letter cited above (Para. 26)--really was a distortion. The opportunity cost of public funds for rainfed agriculture in a seemingly rich Mexico seemed rather low in 1979 and 1980; a year later, that opportunity cost was much higher and the same investments did not look so attractive to planners. 57. In retrospect, it is hard to believe that PLANAT's preparers, both Mexican and Bank, did not see the economic danger signals, but they did not, not having the gift of foresight. When retrenchment came, cutting PLANAT back was appropriate. And it did happen, at Borrower initiative, not at the initiative of the Bank. It happened as these things usually do in such circumstances, which is a way not conducive to development pur- poses. PLANAT's investment program was cut, so that salaries and other operating costs, and subsidized credit (the local part of which could be financed from the banking system) became almost all of the program. 58. Is there any lesson or advice in this experience? Only that for the Bank in a BTO from the November 1985 supervision mission: "we should be sure to do no more than can be fully funded by Government." - 16 - 59. Did PLANAT achieve technology transfer by introducing the T&V extension model to Mexico? No, it did not. This notion seems to have been a self-delusion within the Bank. It was discussed once in Washington between a high-level Bank official and a high-level Mexican one, but after Board approval the notion is never mentioned in files. There was a PLANAT training mission to India, but not having to do with extension. Mexican officials the Audit contacted were unaware of T&V and of any intent to introduce it through PLANAT. They all considered that Mexico's own Plan Puebla experience provides a perfectly good model for the obviously much- needed improvement of Mexican extension. The Audit agrees with them. The Audit also feels that cross-fertilization can be a good thing and that exposing Mexican extensionists to other traditions of successful extension would be a good thing. 60. Was inclusion of credit in PLANAT sensible? No, it was not. One justification for the inclusion was introducing reform into BANRURAL (see Para. 16 above), but this was not taken seriously at higher levels within the Bank and was not reflected in any way in the loan agreements. There- fore it deserves to be regarded as a pious hope. The serious justification was the argument that credit could be used to foster adoption of PLANAT's technical packages if it were under project control (see Para. 16 above). The argument itself presumes either that there are imperfections in the credit market which will be overcome in this way or that credit is going to be used in a way that the market would not dictate to make it a carrot for making a particular investment. Of course, Mexico's rural credit markets were and are severely distorted, a situation which the Bank and the Bor- rower have been seeking to rectify through a series of large agricultural credit loans since 1965. Therefore, preferential access to credit in the distorted market was indeed an inducement to adoption of PLANAT packages. The assurance that it would not induce investments that would have been economically and financially stupid in the absence of the credit lay in the technical assistance that was to accompany it, from PLANAT and from the banks. 61. Even if package-adopters got credit from banks not under PLANAT control, they would get technical assistance from PLANAT. Therefore, the argument for having credit as part of PLANAT hinges on getting the partici- pating banks to provide more technical assistance than they would have normally. There is no evidence that they did. FIRA and associated commer- cial banks provided some technical assistance through normal Bank-financed agricultural credit loans. There is no evidence that they provided more when financed through PLANAT. BANRURAL provides little or no technical assistance to its borrowers as a general rule. This does not seem to have changed under PLANAT. 62. The inappro?riateness of credit under PLANAT increased as time passed. As the investment program disappeared, so did the link between investment packages and credit (see above, Para. 52). Moreover, 51 percent of PLANAT credit cannot be regarded as linked to or fostering investment packages in any case. It was short-term. The banking system treated PLANAT as one of several sources of fungible funds. Thus inclusion of credit under PLANAT did not significantly contribute to achievement of its objectives but perverted the project and distracted it from its objectives. - 17 - 63. Was popular participation important to PLANAT? Apparently so. Popular participation is as popular in principle as motherhood used to be, but there are important reasons why it rarely gets far, the most important being that central governments that are putting up most of the resources want a decisive say in how they are used. Mexico's centralizing tradition was a strong as most countries' in this regard. PLANAT did initiate a change here. Producers' groups did have a say in PLANAT's investment program. PLANAT leaders fought to have their civil works program kept separate from the regular programs of the line departments,161 which did not involve popular participation. 64. According to the survey conducted for the Borrower's PCR, and to casual empiricism and interviews, PLANAT-financed civil works are more likely to be used and maintained than other rural public works. According to the survey, PLANAT beneficiaries are more likely to think that they are well served by extension, credit, input-delivery and the like than the average Mexican in rainfed districts. It seems that Government, and the Bank, accept this evidence and have drawn the consequences from it in designing subsequent interventions. 65. Was multiple-component PLANAT manageable? PLANAT was a simplified version of PIDER, but its district chiefs still had twenty-odd public agen- cies to coordinate. Development folk wisdom deems this to be impossible. But it was not. PLANAT's stronger district chiefs seem to have implemented this complicated project better than might have been expected, really as well as budget limitations permitted. Since there really was decentrali- zation, more than Bank supervision missions appreciated (see Para. 46 above), PLANAT district chiefs had the freedom not to do well too, and some of them used it. Decentralization has its risks as well as its benefits. 'Where multi-component management was successful, having the manager at the district level, closer to the problems and opportunities, seems to have helped. 66. There were two other factors associated with successful management in PLANAT's good districts. One was the personality of the district chief, his energy and dedication. The other was the existence of a proven tech- nical package. There is consensus on which of PLANAT's districts were successful, and they are all those in the Central Highlands, where the Chapingo Graduate School and CIMMYT had been doing most of their work for several decades and where they had been trying to promote adoption of improved practices through Plan Puebla for more than a decade. The confidence level in what PLANAT was pushing farmers to do was much higher in the successful districts. That seems to have helped make management successful. 16/ To the point of an argument with the state governor in Tabasco where PLANAT's district chief, as a result of his own position within the district's normal civil service, wanted to use PLANAT funds to finance the normal road-building program. PLANAT managed to avoid this. COMMENTS FROM THE BORROWER - 19 - Annex 1 Page 1 of 4 INTBAFRAD/Washington, USA Attentions Mr. Graham Donalson Chief, Agriculture, Infrastructure and Human Resources Division Operations Evaluation Department Dear 14r. Donalsonz Here are a few brief comments on the Mexico/Rainfed Agricultural Development Project PPAR of March 1990 (Loan 1945-ME) for your information. First of all, I want you to know that I did not get the report until April 27 in an envelope received by the Bank office on April 16, which prevented my consulting certain people and my own files, to compare notes and study the ideas expressed in the PPAR. Secondly, I want to congratulate you for your obvious effort part to investigate the successes and failures of PLANAT in a spirit of objectivity. Unfortunately, I feel you were not able to be adequately objective having failed, among other things, to focus on the country's changing economic circumstances and government structure and on the numerous changes in officials with decisionmaking responsibility in regard to project implementation. Before continuing with my comments, I would like to point out that I am the former Director General of DGDUT, the former World Bank staff member referred to on page 4, paragraph 13, of the PPAR, which fact could well disqualify me as an objective commentator on the report's criticisms of PLANAT. You can judge this for yourself. I want to call your attention to the fact that PLANAT was designed for a specific SARK (Agriculture and Irrigation Ministry), banking system, agricultural policy and national economy existing in 1979/1980. The PLANET evaluated by the OED is the product of an entirely different SARH, a totally different, nationalized banking system, a different agricultural policy and a diametrically different national economy from that of the past. I should add that there was also an entirely different government resource allocation system prior to the crisis, which could have easily affected the PLANAT evaluated in the report. You do make some mention of these points, including the fact that the earthquake destroyed PLANAT files and documents which might have shed some light on its successes and failures. However, I feel that the possible effects on PLANAT of the changing climate within the country were not covered in adequate detail. Let me be more specifics 1. SARH underwent its first structural change in 1980/1981 when an unexpected decision eliminated the Agricultural Extension Directorate (Direcci6n General de Extensi6n Agricola) and transfered its functions to the DGDUT, changing the whole concept of extension. This may have been the root of the project's failure to successfully introduce the T&V extension method. 2. FICAR was reorganized in 1980/1981, resulting in the emergence of - 20 - Annex 1 Page 2 of 4 FICART, whose role in the intermediation of BANRURAL loans is, incidentally, never mentioned in the PPAR. 3. The commercial banks were nationalized on September 1, 1982 and the rules of the agricultural credit game changed to the advantage of the formerly forgotten rainfed areas. 4. In May of 1982, the country was declared in crisis, with the economy and especially inflation showing an upward trend, disrupting the government budgeting system and thereby affecting the allocation of counterpart funds. 5. In 1982/1983, we saw a change in the administration and, hence, new administration officials who viewed things differently. The definition of the "FIRA coup" in paragraph 34 of the PPAR is noteworthy. 6. In 1983/1984, under SAP, the Borrower proposed that, to dispel the image of PLANAT as a "gold-plated project", the Bank make loan proceeds available to a larger number of districts, as had previously been done in PIDER III, which would have allowed for the use of 40 detailed studies of an equal number of districts and, thus, for a better distribution of funds originally allocated to a mere nine districts. However, alluding to technicalities in the Project Agreement, the Bank opposed this change which had been so positive in PIDER III. 7. In a climate of budget austerity, it was only natural that the Borrower would attempt to disburse funds as quickly as possible using the most convenient items. It is through no fault of the project design that credit disbursements turned out to be most convenient. I refer you to the case of the Uruguayan loans for agricultural development where, as a result of the crisis which hit the country in 1973, credit allocations remained virtually unused while the allocation for fertilizer imports was invariably exhausted, requiring reallocations of credit funds. This was circumstantial. In Mexico, the opposite happened, and this, too, was circumstantial. 8. In an economy with high inflation, the propensity to make long-term investments tends to decline, which explains the abuse of short-term credit. In our case, we should measure the extent to which nationalization of the commercial banks affected the demand for credit funds in rainfed areas. The PPAR has very little to say on this subject. It merely states that it was a bad idea to link credit and development under PLANAT and that credit perverted the program. 9. I feel that BANRURAL's role as far as PLANAT is concerned has been satanized, with no mention whatsoever made of that of FICART, particularly of its efforts to evaluate producer subloans and instill discipline in BANRURAL system banks. 10. In 1984 a new Minister of Agriculture arrived on the scene, with numerous changes in government officials and, hence, in agricultural policy. One such change was the unification of irrigated and rainfed districts into what were referred to as "rural development districts" making the PLANAT approach totally obsolete. Incidentally, in 1989, the irrigated districts - 21 - Annex 1 Page 3 of 4 seceded from the rural development districts and the Agriculture and Irrigation Ministry was restructured. Obviously, district decentralization and deconcentration policy is different now. 11. I do think that the PPAR makes an interesting study of the background of PLANAT. However, it is worth noting that PLANAT was actually designed to coordinate the activities of FIRA, BANRURAL and SARH and not of twenty-odd government agencies as mentioned in the report. It was to have coordinated the operations of the research, infrastructure and extension arms of SARH, but this should not have been difficult considering that there was a Minister of Agriculture facilitating such task. This is quite different from the multisector PIDER project requiring coordination of the Agriculture, Fisheries, Agrarian Reform, Health, Education, Transportation and other Ministers under the Ministry of Planning and Budget. 12. I feel that too much emphasis is being placed on the fact that PLANAT is an extension of Plan Puebla, which is not entirely true in that Plan Puebla only involved maize, was not delimited by a region or district and did not include infrastructure. PIDER, on the other hand, though it overemphasized public works, had no project appraisals and tended to neglect productive activities, did use the microregion as its physical frame of reference. 13. PLANAT clearly had a dual focus. Locally, it focused on nine districts. At the nationwide level, it focused on laying the groundwork for the inclusion of another 40 or 50 districts under a second loan. It initially attempted to include a minimum of 20 districts in its local focus. However, district subproject preparation activities resulted in the completion of subprojects for only nine districts, all of them well endowed, which was precisely the intention, the idea being to sell a second project based on good results. The nationwide focus was on studies of the other districts and on improving research, extension, training and, of course, coordination. This ensured a more successful second stage. 14. The credit component became increasingly complicated with the pasage of time, not for project-related reasons but, rather, due to national concerns. Inflation was definitely a factor in the dispute between the Borrower and the Bank over interest rates. The fact that it stopped making credit disbursements did not make PLANAT any worse than other credit projects which were also forced to stop disbursing temporarily. Prior to PLANAT, the entire banking system discriminated against farmers in rainfed areas. Now we see less of this sort of discrimination. This could be counted as a plus and could, in part, justify inclusion of the credit component. According to a field survey, 90% of productive infrastructure subprojects under PIDER were inoperative due to a lack of credit. I believe this point was made clear in the PPAR for PIDER I. In an inflationary economy, I would not recommend including credit for this type of project. 15. Unfortunately, I do not have access to the ten original briefs for the proposed agricultural projects to comment on the alleged cost inflation between the time of project preparation and appraisal. It may have been due, certainly, to the credit component. But it may also have been due to the advanced stage of irrigation and drainage studies supplying better cost data. - 22 - Annex 1 Page 4 of 4 16. I agree that the concept of incrementality caused disagreements between the Borrower and the Bank and was misunderstood by certain Mexican officials. However, it sought only to prevent duplicate disbursement of the project account. Perhaps if another simpler concept had been used there would have been less diisension. 17. I have no idea whether the introduction of T&V in Mexico has been used to justify PLANAT within the Bank. I do feel that paragraph 17 on page viii of the PPAR exaggerates the importance of an extension method. On the other hand, field auditors may have failed to contact the proper Mexican officials. DGDUT organized several seminars with TAHAL of Israel at which the T&V system was examined, laying the groundwork for its introduction in the nine PLANAT districes. What happened is that, with the crisis, budget appropriations for investments with foreign exchange components were cut back by SPP (the Ministry of Planning and Budget). Moreover, with changes in DGDUT officials, this item disappeared altogether. It's true that the introduction of T&V was not recognized in the legal documents. These comments cover most of the points made in the PPAR. However, let me make one final comment on the findings presented by the report, namely that, although they appear to be correct as far as I am concerned, their validity should be considered from a circumstantial viewpoint. Take, for example, the finding to the effect that PLANAT failed to attain its objectives. If this is the way of thinking for rainfed area development, then this raises a question. If irrigated districts are in crisis, shall we be once more forced to justify abandoning rainfed areas? If PLANAT failed as a result of special circumstances, was it improperly designed? I believe it would be worthwhile to meet with other former officials and with more farmers in rainfed areas. Referring to the BANRURAL and ANAGSA crisis, from this crisis emerged the concept of high-risk areas in which BANRURAL will no longer operate. It will now be up to SPP, in a gesture of solidarity, to channel scarce funds into producer goods on farms abandoned by BANRURAL for nonpayment. All this leads me to conclude that residents of rainfed areas will continue to suffer from their lack of a means of demanding recognition of their public and private investment needs. I am going to try and study the PPAR more calmly and compare notes with my records in Mexico City. If I can, I will try and formulate recommendations, although I know it will be too late for them to be incorporated into the final version of the report. In any event, it was a pleasure for me to reexamine PLANAT as a hope, though failed, of channeling more funds into the country's poorest farming areas. Sincerely, /S J. JESUS ROMERO CHAVEZ - 23 - COMMENTS FROM THE BORROWER Annex 2 Page 1 of 5 SECRETARIA DE AGRICULTURA Y RECURSOS IDRAVULCOS JMSTITUTO NACIONAL DE INVESTIGACIONES FORKSTALES Y AGROPECUARIAS Mexico, D.F. June 8, 1990 Dr. Graham Donaldson Chief, Agriculture, Infrastructure and Human Resources Division Operations Evaluation Department World Bank Dear Dr. Donaldson: Thank you for your letter of April 5, 1990, accompanied by the preliminary version of the PPAR on the Rainfed Agricultural Development Project (Loan 1945-ME). As you requested, I am enclosing a memorandum setting out some observations on it, and ask you to accept my apologies for not letting you have them by the deadline date. Yours, etc. /s/ Luz Maria Bassoco Deputy Director for Special Research and Development Projects - 24 - Annex 2 Page 2 of 5 COMMENTS ON DRAFT VERSION OF PPAR ON RAINFED AGRICULTURAL DEVELOPMENT PROJECT (LOAN 1945-ME) This project, generally known by its acronym PLANAT, which was originally designed to increase on-farm production in nine high-potential rainfed districts through improvements in extension services, applied research and irrigation, is presented in this PPAR in a manner which sets out the circumstances and limitations that were met with during the execution process and explains the transformation of the original plan into a scheme for the development of rainfed districts through agricultural credit. According to the original design, 43% of project spending was allocated to agricultural credit; however, by 1984 this figure had risen to 80% of total spending, a change that affected other project expenditure categories, but mainly the investment effort. Although the PPAR draws attention to the effects of the economic crisis and consequent budgetary restrictions on the Borrower's ability to finance the counterpart share of project investment expenditure, the magnitude of the problem created by Mexico's economic crisis is not quantified. It is therefore suggested that some quantitative indicators of that problem be included. For instance, between 1981 and 1987, the years scheduled for commencement and closure of the project, overall supply, at constant 1980 prices, shrank by an annual average of 1.3%, while GDP declined by 0.2% annually and imports by 11% annually. As to overall demand, consumption also shrank, at annual rates of 0.6% in the private sector and 2.2% in the government sector. Gross fixed capital formation declined as well, by 8.2% annually, while exports over the same period increased at an annual rate of 8%. (Refer to attached table: Overall Supply and Demand in Mexico, 1981-87.) Economic evaluation of the project is incomplete, and assessment of the resulting internal rate of return at 6% -- a calculation based on comparison of producers' net income flows with the cost of all PLANAT components -- is an underestimate. This fact is referred to in para. 10, page vii, of the PPAR, which notes that the benefits of soil and water conservation works were not taken into account, whereas costs in this category, and others, were. Some of the major benefits from development of the rainfed districts in fact came from soil and water conservation works, and from cropgrowing techniques (no tilling, minimum tilling, crop rotations, etc.) aimed at reducing soil erosion and the losses in soil productivity it causes. On this score, it is suggested that the PPAR take a wider view that allows the economic benefits derived from PLANAT expenditure in these arenas to be taken into account. Methodologies for evaluation of the benefits and costs of farming practices and techniques that focused on soil conservation and preventing soil deterioration are examined extensively in various publications of the World - 25 - Annex 2 Page 3 of 5 Resources Institute (e.g. "The Economics of Sustainable Agriculture," Paul Feath ane Robert Repetto, World Resources Institute, May 3, 1990). The "Objectives" section of the PPAR points out that the credit under the control of PLANAT was incorporated into the project to encourage adoption of its technical packages, while the "Implementation Experience" section notes the very heavy demand for PLANAT credit and the problems that arose in connection with this component. There is no explicit indication in the PPAR of the existence of any financial mechanism for sharing with PLANAT producers the risks they ran by adopting the technical packages approved for the project districts. It is worth mentioning, however, that one of the major factors deterring producers from adopting modern production techniques, such as the use of improved seed and chemical inputs, is aversion to the risk of losing what they invest in them owing to crop failure and other causes beyond their control. In actual fact, rainfed crops are at high risk of failure. Over the 10-year period 1978-87, an average of 11.2 million ha in the country's rainfed areas was sown with the major crops (rice, beans, maize, wheat, safflower, sorghum, soybeans, cottonseed and barley), although an average of only 9.1 million ha was harvested. This indicates a failure rate averaging 18.8% of surface areas sown during that period. The average failure rate for the States in which the PLANAT districts are located is estimated at 13.6% over the same period (refer to the accompanying table). The greatest differences between surface areas sown and harvested were seen in the States of Puebla (17.5%) and Zacatecas (20.3%). One of the programs run by FIRCO (a Government trust fund for risk-sharing purposes) came into effect in 1981 and was designed principally to share with producers in high-potential rainfed districts the risk inherent in their adopting technical innovations. It would be desirable to indicate in the PPAR whether any relationship existed between PLANAT and FIRCO and what type of relationship it was. The PPAR makes some very important judgements on what are considered to be successes in execution of the project, namely: the coordination of many activities at district level, and the eliciting of community participation in and maintenance of project civil works. Equally important are the observations made regarding the addition of a large credit component to the project and the increase in project size. The findings of this Performance Audit and the pinpointing of the limitations and problems that had to be faced during project execution will undoubtedly serve as feedback whenever other development projects with possibly similar focus are being designed and implemented. - 26 - Annex 2 Page 4 of 5 MEXIC OVERALL SUPPLY AND DEMAND AT 1980 PRICES (in billions of pesos) Concept 1981 1987 Average Annual Growth Rate Supply 5,544.9 5,138.5 - 1.26 GDP 4,862.2 4,802.4 - 0.21 Imports of goods and services 682.7 336.1 - 11.14 Demand 5,544.9 5,138.5 - 1.26 Private-sector consumption 3,123.2 3,014.1 - 0.59 Government-sector consumption 494.8 565.3 - 2.24 Gross fixed capital formation 1,286.4 772.3 - 8.15 Inventory changes 106.6 - 58.5 - 9.52 Exports of goods and services 533.9 845.3 7.96 Source: Statistical Annex to First Government Report of Salinas de Gortari Administration. AREAS SOWN AND HARVESTED: MAJOR CROPS IN STATES WHERE THE MINE RAINFED PLANAT DISTRICTS WERE LOCATED (Averages 1978-1987, in hectares) State Areas sown Rice Maize Beans Wheat Sesame Cotton Safflower Soybeans Barley Sorghtes Totals Ratio of and seed area harvested harvested to ares somn Jeliscol sown 145 811196 82973 15234 2414 926 43 3943 176470 1093344 90.3 Harvested 103 738076 65650 12536 2141 830 43 3109 164434 986922 Mlichoacan2 Sown 87 412239 21479 12717 34373 2432 25 598 112072 596022 86.1 Harvested 87 355238 17704 11079 29716 2206 9 428 96489 512956 Puebta3 Sown 550101 59242 16052 4041 43960 8403 681799 82.5 Harvested 448704 47113 13517 3837 42149 6913 562233 Veracruz4 Som 16740 537461 55063 3197 740 504 904 6177 2643 5982 629411 87.2 Harvested 15132 470899 47831 2000 673 486 614 4096 2068 5005 548804 Guerrero5 Sown 718 412462 12825 35954 27 52 3228 456246 94.0 Harvested 680 387934 11580 34486 26 49 2720 437475 Tabasco Sown 9667 51764 5977 115000 8 1884 184300 94.5 Harvested 6925 45679 5064 115000 1486 174156 Zacetecas7 Sown 438843 550100 13331 12184 £067 1018489 79.7 Harvested 341322 449848 9142 8401 333M 812048 Totat for Sown 4668611 86.4 the seven Harvested 4034594 states Zapopan M artinez de ta Torre Coatacoatcos & Morel fa / Chi pencingo 2/ Chotula Cardenas E. Zapata Jalpa do 0 D