Mexico - Small-scale Agricultural Infrastructure Project

Docunent of The World Bank FOR OFFICIAL USE ONLY Report No. 7253 PROJECT COMPLETION REPORT MEXICO SMALL-SCALE AGRICULTURAL INFRASTRUCTURE PROJECT (LOAN 1643-ME) May 27, 1988 Latin America and the Caribbean Regional Office This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Acronyms BANRURAL National Rural Credit Bank DGCD General Directorate for the Coordination of Delegations DGIR General Directorate for Rural Infrastructure DGID General Directorate of Irrigation and Drainage GOM Government of Mexico ICB International Competitive Bidding IDB Inter-American Development Bank LCB Local Competitive bidding NAFINSA Nacional Financiera, Inc. OHIADR General Directorate of Hydraulic Works and Agricultural Infrastructure for Rural Development PIDER Integral Rural Development Program PNDR National Plan for Small-Scale Agricultural Infrastructure Development SARH Ministry of Agriculture and Hydraulic Resources SPP Ministry of Planning and Budget SUDR General Subdirectorate for Supervision of Rural Development Units B/C ratio Benefit-Cost Ratio ERR Economic Rate of Return M & E Monitoring and Evaluation O & M Operation and Maintenance PCR Project Completion Report  FOR OMCIL. USE ONLY THE WORLD BANK Washington. DC 20413 U.S A OImce d Dvectuw-Col Op.uatm IvakMtrn' May 18, 1988 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report: Mexico Small-Scale Agricultural Infrastructure Project (Loan 1643-ME) Attached, for information. is a copy of a report entitled "Project Completion Report: Mexico - Ssall-Scale Agricultural Infrastructure Project (Loan 1643-ME)" prepared by the Latin America and the Caribbean Regional Office. Under the modified system for project performance auditing further evaluation of this project by the Operations Evaluation Department has not been made. Attachment This document has a restricted distribution and may be used by recipients only in the performanme of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  FOR OMCIAL USE ONLY MEXICO SMALL-SCALE AGRICULTURAL INFRASTRJCTURE PROJECT (LOAN 1643-ME] PROJECT COMPLETION REPORT Table of Contents Page No. Preface . . . . . . . . . . . . . . . . . . . . . i Basic Data Sheet .... . . . . . . . . . . . . . . li Evaluation Summary . .t.i.n.Sw.n. .i.v.. . . . . . . iv I. INTRODUCTION .1. . . . . . . . . . . . . . . . . . 1 II. PROJECT FORMULATION 5 A. Identification, Preparation . . . . . . . . . . 5 B. Appraisal . . . . . . . . . . . . . . . . . . . 5 C. Irrigation and Drainage Development in Mexico . 6 III. IMPLEMENTATION 8 A. General . . . . . . . . . . . . . . . . . 8 B. Project Components . . . . . . . . . . . . . . . 10 C. Credit ............ .* * * 13 D. 0 & M and Cost Recovery . . . . . . . . . 13 E. Monitoring and Evaluation . . . . . . . . . . . 14 F. Disbursements . . . . . . . . . . . . . . 14 G. Procurement . .. . . . . . . . ..- 15 H. Reporting and Auditing . . . . . . . . . . . . 15 I. Adherence to Covenants . . . . . . . . . . . . . 15 IV. ECONOMIC AND AGRICULTURAL IMPACT 16 A. Project Benefits and Incremental Output . . . 16 B. Economic Rate of Return . . . . . . . . . . . . 16 V. INSTITUTIONAL PERFORMANCE . . . . . . . . . . . . . 16 VI. BANK PERFORMANCE . . . . . . . . . . . . . . . . . . 17 VII. CONCLUSIONS AND RECOMMENDATIONS 17 A. Conclusions . . . . . *. . . . . . . . . . . 17 B. Recommendations .... . . . . . .. *.... 19 This document has a restricted distribution and may be used by recipients only-in the perfiormance of their offcial duties. Its contents may not otherwise be disclosed without World Bank authorization.  TABLE OF CONTENTS (Cont'd) PROJECT COM?T,ETION REPORT Page No. List of Annexes I. Project Costs . . . . . . . . . . . . . . . . . . 21 II. Follow-up Projects . . . . . . . . . . . . ... 22 III. Withdrawal of Proceeds of the Loan . . . . . . . . 23 IV. Rate of Return Analysis . . . . . . . . . . . . . 24 V. Assumptions for Benefit Analysis . . . . . . . . . 25 VI. Comments from Borrower . . . . . . . .... 26 Map IBRD 11879R  ~Ico SMALL-SCALE AGRICULTURAL INFRASTIWCTURE PROJECT (LOAN 1643-ME) PROJECT COMPLETION REPORT PREFACE This is the Project Completion Report (PCR) for the Small-Scale Agricultural Infrastructure Project in Mexico, for which Loan 1643-ME in the amount of US$60.0 million was approved on December 12, 1978. The ioan was closed on June 30, 1985, two years behind schedule. It was fully disbursed and the last disbursement was on August 26, 1985. The PCR was prepared by the Latin American and Caribbean Region, Agriculture C Division, after a field visit in June-July 1987. The PCR is based on data prepared by the Borrower since late 1985. Information on the project was incomplete for two reasons: the loss of files as a result of the September 1985 earthquake; and the reorganization of the Ministry of Agriculture and Hydraulic Resources (StRH). The PCR is also based on the Staff Appraisal Report dated December 1, 1978 and the Loan, Guarantee and Project agreements. Supervision reports, correspondence with the Borrower and internal Bank memoranda on project implementation also were reviewed and incorporated in the PCR as appropriate. In accordance with the revised procedures for project performance reporting, this PCR was read by the Operations Evaluation Department (OED) but the project was not audited by OED staff. The draft PCR was sent to the Borrower on February 23, "'988 for comments. Comments received from the Government (Secretaria de Hacienda y Credito Publico and SAhR) have been included and are attached as Annex VI.  - ii - PROJECT COMPLETION REPORT BASIC DATA SHEET MEXICO: SMALL-SCALE AGRICULTURAL INFRASTRUCTURE PROJECT (LOAN 1643-ME) KEY PROJECT DATA Actual as X Appraisal of Appraisal Item Estimate Actual Estimate Project Costs (USS million) 143.7 172.4 1202 Loan Aa-.unt 'US$ million) 60.0 60.0 100Z Date Board Approval 121-178 12112178 Date Effectiveness 05/08179 04/13/79 Date Physical Components Completed 12131/82 06/30/85 Proportion then Completed 57X 1002 Closing Date 06/30/83 06130/85 Econcmic Rate of Returx 23.0 12 Financial Rate of Return n.a. n.a. n.a. Number of Direct Beneficiaries 30,000 22,000 CUMULATIVE DISBURSEMENTS FY79 FY80 FY81 FY82 FY83 FYd4 FY85 Appraisal Estimate (US$ 000) 0.3 21.4 34.6 48.8 60.0 Actual (US$ million) 1.7 4.1 8.2 16.3 23.6 55.9 60.0 Actual as % of Estimate + 19.0 24.0 33.0 39.0 93.0 100.0 Date of Final Disbursement August 26, 1985 STAFF INPUTS (staff-weeks) FY77 FY78 FY79 FY80 FY81 FY82 FY83 FY84 FY85 FY86-87 TOTAL Preparation 5.1 5.1 Appraisal 61.0 16.0 77.0 Neogtiation 12.9 12.9 Supervision 1.8 11.1 c ..3 3.5 13.7 6.0 9.9 69.3 Completion ___ 9.0 9.0 5.1 61.0 30.7 11.1 23.3 3.5 13.7 6.C 9.9 9.0 173.3  - iii - JISS.'ON DATA Date No. Staff- Type & No. of StatusTrend Type of (mo.yr) of days in of of of Problems Item Persons Field Specialists Proi.P o 4 Identification Preparation Appraisal 4178 6 171 a b(2),c d(2) Supervision 1 3/79 3 6 a(2) b 1 - Supervision 2 :2/79 4 12 a(2) d(2) 2 1 F,N Supervision 3 7/80 2 5 d a 2 1 F,M,T Supervision 4 8/80 2 7 a d 2 1 F,M,T Supervision 5 3/81 2 22 a d 2 1 F,M,T Supervision 6 7/81 1 15 a 2 1 Supervision 7 6/82 1 7 a 2 1 F,!M,T Supervision 8 4/83 2 22 a d 2 1 F,M,T Supervision 9 12/83 1 20 d 2 2 F,M,T Supervision 10 1/85 2 12 a d 2 2 M,P,F Supervision 11 5/85 1 6 d 2 2 M,F Total (SPN) 305 OTHER PROJECT DATA Borrower National Finance, Inc (NAPINSA) Executing Agency Ministry of Agriculture and Water Resources (SARH) Fiscal Year January 1 to December 31 COUNTRY EXCHANGE RATES Name of Currency (peso) Currency Exchange Rate: Appraisal Year kverage (1978) US$ 1.00 - 23 Intervening Years Average (1979-84) US$ 1.00 - - Completion Year Average (1985) US$ 1.00 - 250 FOLLOW-ON PROJECT: Not applicable l/ a) Agriculturalist; b) Agricultural economist; c) Financial specialist; d) Irrigation engineer 2/ 1 - Improving; 2 - Stationary; 3 - Deteriorating 3/ 1/ - Problem-free or minor problems; 21 - Moderate problems; 3/ - Major problems 4i F - Financial; M = Management; T = Technical; P - Political  - iv - MEXICO SMALL-SCALE AGRICULTURAL INFRASTRUCTURE PROJECT (LOAN 1643-ME) PROJECT COMPLETION REPORT Evaluation Summary Introduction This was the fifth of six projects designed to promote the development of small-scale agricultural infrastructure in Mexico by means of small-scale irrigation, drainage and livestock schemes. Five of these projects have been financed by loans of the Inter-American Development Bank (IDB), and one by an IBRD loan. The small-scale agricultural infrastructure is s cational program that is being implementid in Mexico since early 1960. Com; -ed to large agricultural infrastructure works this program is spe< .ally attractive in periods of shortage of capital for investments bec,-ase of the lower unit cost, the shorter time required for construction of the small investments, and the rapid realization of benefits. QoD ectives The main project objective was to benefit directly 30,000 rural families by irrigating 56,000 ha of dry lands, mostly (65%) by means of tubewells; draining 36,000 ha of wet lsads; and constructing 50 livestock production schemes. Of the total cost of the project, US$143.7, about 77% was going to be invested in irrigation, 12% in drainage and 11% in livestock infrastructure. Beneficiaries were going to participate in financing construction of the works, ead they would also cover the full cost of operating and maintaining the schemes. Implementation Experience Since this was one of a series of already successful projects, the institutional mechanisms were already in place and functioning smoothly. The efficiency and motivation of the agencies involved were important elements in maintaining continuity and progress. Total project cost was 120% of appraisal estimate because of slightly higher infrastructure costs and unexpected purchase of pumping equipment. The expansion of land for irrigation, with 50,000 ha. added was almost 90% of what was targeted at appraisal. Tubewell infrastructure development, which received half (US$87.4 million) of project investment, was completed in 28,000 ha, or 80% of the targeted land. Total investment in irrigation development was US$131.8, with an average cost of US$2,200 per ha. Drainage infrastructure was complcted in 16,000 ha, or 44% of what was targeted at appraisal, at a. cost of US$11.7 million. Drainage development is an excellent investment because of its low cost (US$1,075) per hectare, and its positive impact on production. Its potential, however, is limited to humid areas with good agricultural potential.  The livestock component of the project performed poorly, no surprise since the levels of experience, planning and organization were all deficient; and the market conditions, prices of meat and milk, also were unfavorable for its development. The result for livestock infrastructure were skewed by the addition of extensive sub-projects and thus showed dramatic additions (almost 300%) over what was targeted at appraisal. IBRD's performance during preparation, appraisal and implementation was timely and effective. Bank's staff time devoted to the project was 173.3 staff-weeks of which 305 staff-days were spent in the field, a relatively small amount for a complex national projezt with 351 subprojects. Project management was very good. Results About 22,300 rural families benefitted directly as a result of the project, wi .h indirect employment of another 10,000 expected. The estimated economic rate of return (ERR), excluding the livestock component, was 12%. The result is lower than the ERR of 23% estimated at appraisal because of the follkwing differences: (a) economic prices during the project implementation period were lower than those projected at appraisal; (b) the total area benefitted by investment on irrigation and drainage was smaller--by about 25%-- and the project total cost somewhat higher--by about 202--than those estimated at appraisal; (c) about 15% of tubewell investments were lost due to unproductive wells; and (d) subproject investments, specially storege dams, took longer to const :uct than expected at appraisal. Sustainability The sustainability of the different subprojects varies. The irrigation and drainage subprojects show high sustainability as long as adequate financing is secured to operate and maintain the infrastructure in good working coadition. Beneficiaries participation in financing the investments, as well as in bearing the cost of operating and maintaining agricultural infrastructire, is essential to obtain high sustainability. Findings and Lessons Active involvement of the beneficiaries in the subprojects is essential for success. The most effective way to achieve their participation is to have the beneficiaries bear and finance part of the construction cost through credit made available for this purpose, as has been done with the tubewell schemes. The beneficiaries' assumption of part of the cost of the hydraulic infrastructure, beginning at the outset of the project rather than later, increased the level of commitments of the members of the users' association to use the infrastructure effectively, to recover quickly the money invested in it. The lengthy implementation periods for dam subprojects to irrigate more than 1,000 ha tightened scarce investment funds in a few subprojects. This could have been avoided if the size of the dam subproject were limited to a given amount of area to be irrigated or by setting a ceiling in the cost of dams that would be included in the project. The beneficiaries of  - vi - storage dam schemes also seem to be less eager to optimize water use and crop production. Perhaps they feel less responsible sinc^ they have less invested and committed than the farmers benefitting from the tubowell schemes; the latter pay for electricity and equipment while the former pay nothing. Organizing the farmers is also more difficult in the case of dams, since the average scheme (700 ha) involves the organization of 140 farmers. Furthermore, the water supply of dam schemes is not as reliable as that of tubeaells; this fact became painfully evident in 1987 when a drought hit the Central Region. Drainage development is also an excellent development alternative because of its low investment cost per hectare and its positive impact on production, since it transforms unproductive, waterlogged land into fully productive acreage. Drainage development has excellent potential in the coastal plains of the humid southeastern region, in the states of Tabasco, Veracruz and Chiapas, but is also a possibility in waterlogged depressions around lakes and river in the semi-arid central region. However, beneficiar43s especially in the larger subprojects have been difficult to organize and educate about the need for maintenance, especially of the larger drains which are mechanically maintained. Beneficiaries, organized in users' associations, are responsible for cperating and maintaining the small-scale agricultural infrastructure constructed by the Government. SARH oversees that users' associations fulfill their commitments by periodic supervisions. This set up, in general, works quite well. However, for difficult cases such as new schemes where beneficiaries have little experience with irrigated agriculture or in large schemes requiring more complex organization and management, SARH, due mainly to shortage of funds, has been unable to provide to beneficiaries the extended and intensive training needed to overcome these difficult cases. The relatively minor livestock component fared poorly. The infrastructure provided was not used adequately to increase livestock production as was intended, because of management, credit and marketing problems.  MEXICO SMALL-SCALE AGRICULTURAL INFRASTRUCTURE PROJECT (LOAN 1643-ME) PROJECT COL"FTON REPORT I. INTRODUCTION 1.01 The purpose of this project was to promote the development of small-scale agricultural infrastructure in Mexico using irrigation, also on a small-scale. It was the fifth of six projects designed for this purpose. Five of them have been financed by loans of the Inter-American Development Bank (1DB), and one by an IBRD loan. The small-scale agricultural infrastructure project was a national one, with implications for the entire population, given the nature of its goals and results, but especially for the rural population engaged in agriculture (272). In fact, this project was part of one of the government's most important programs in the agricultural sector, with significant potential for the tuture given the need and success to date. 1.02 Thus, this project fits well into the Government of Mexico's (GOM) priorities for development which are aimed at increasing the production of basic food crops in order to achieve self-sufficiency, increasing employment levels, and generating employment opportunities and improving income among the rural poor. 1.03 The project also formed part of the Bank's overall strategy to assist develop agriculture in Mexico, through water control projects; development of new lands in the humid regions; technical assistance programs; agricultural credit extended on a naticnal, regional or project basis; rural development programs; and conventional itrigation projects. 1.04 This is a Project Completion Report (PCR) for the Small-Scale Agricultural Infrastructure Project (Loan 1643-ME). The Loan Guarantee and Project Agreements are dated February 6, 1979. The total loan amount waq US$60.0 million, approved by the Board on December 12, 1978. The loan was fully disbursed and the last disbursement was on August 26, 1985. 1.05 The PCR was prepared by the Latin America and Caribbean Region, Agricultural Division C, after a field visit in June-July 1987. The PCR is based on data prepared by the Borrower since 1985, that information remains incomplete for two reasons: the reorganization of the Ministry of Agriculture and Hydraulic Resources (SARH); and the loss of files as a result of the September 1985 earthquake. The PCR is also based on the Staff Appraisal Report dated December 1, 1978 and the Loan, Guarantee and Project agreements. Supervision reports, correspondence with the Borrower and internal Bank memoranda on project implementation also were reviewed and incorporated in the PCR as appropriate. 1.06 The Government of Mexico (GOM) prepared, coordinated and executed the project, complemented by limited Bank role in supervision, data analysis, and project formulation. The Bank supported the government's inclusion at minimal cost of a livestock infrastructure component in the project despite the agency inexperience; this was the one component of the project that cannot be called a success, for reasons enplained below. 1.07 SARH implemented the project through its General Directorate of Hydraulic Works and Agricultural Infrastructure (OhIADR) and its General Subdirectorate for Supervision of Rural Development Units (SUDR). The  former was in charge of all infrastructure works, and SUDR provided technical assistance to farmers. At the local level, farmers were organized into User's Associations, responsible for operating and maintaining the irrigation works they used. 1.08 The original project included these sub-project components: tubewell development; storage and diversion dams; pumping plants; drainage; and livestock. (Pumping equipment was added on later.) There were a total of 351 sub-projects, dispersed over 27 of 31 states of Mexico. 1.09 The single most striking fact about the project is the incomplete nature of the data, as mentioned above. The two events, SARH reorganization and the 1985 earthquake, effectively ended any hope of being able to quantify many of the most important results. That lack of data also means that any final judgements on the success or failure of the sub-project components must be tempered by a qualifying note. 1.10 Nevertheless, this small-scale agricultural infrastructure project, which focussed on hydraulic infrastructure, was effective for these reasons: (a) Small-scale irrigation schemes are especially effective because: of a low unit cost, when investment capital is limited; rapid realization of benefits, both agricultural and otherwise; and a shorter construction period. The applicability of this factor extends to many nations in situations like that of Mexico. (b) Since this was one of a series of already successful projects, the institutional mechanisms were functioning smoothly. The cooperation from the GOM and the efficiency and motivation of the agencies and their staff members involved were important elements in maintaining that continuity and progress. (c) The beneficiaries' assumption of part of the cost of the hydraulic infrastructure, beginning at the outset of the project rather than later, increased the level of commitment of the aembers of the Users' Associations to operating and maintaining the infrastructure. The tubewell sub-project in particular was the most successful, in part for this reason. Thus, the concept of a financial partnership even at the individual user's level proved effective. (d) Drainage development is an excellent investment because of its low cost per hectare, and its positive impact on production. Its potential, however, is limited to humid states. 1.11 Conversely, the project's effectiveness has been hampered by certain factors: (a) The lengthy implementation periods for large dam tighted scarce investment funds in few sub-projects. This could be  - 3 - avoided if the historical budgeting allocations are taken into account to set a limit on the dam size and cost that wou'ld be included in the small-scale agricultural infrastructure projects. (b) Unlike participants in the tubewell schemes, storage dam scheme farmers are not well organized and have no financial in- stment from the outset in their operations; such a cost-sharing procedure could help correct the situation. (c) The livestock component of the project performed poorly, no surprise since the levels of experience, planning and organization were all deficient; the market conditions also were unfavorable for its development. (d) Credit problems and associated deficiencies also hampered the development of the livestock sub-project. (e) There was no monitoring and evaluation of the sub-projects, a fact which greatly diminished the amount of data available to properly and fully evaluate the project. Other obstacles also played a role, including the 1985 earthquake, the reorganization o4 SARH, budget cuts and the fact that investments varied and were diffused over 27 states. (f) Inflation made it more difficult to collect users' association water fees to cover part of investment costs, as did the method of cost-sharing, spread over 25 years, with an initial 4-year grace period. 1.12 Certain project achievements should be mentioned: (a) Tubewell infrastructure development, which received half (US$87.4 m) the project investment, was completed in 80% (28,000 ha) of the targeted land. (b) The expansion of land for irrigation was almost what had been t4rgeted, with 89% (50,000 ha) added, while expansion for drainage was 44% (16,000 ha). (c) The results for livestock were skewed by the addition of extensive sub-projects and thus showed dramatic additions (almost 300%). However, in reality, it achieved about what was expected at appraisal. (d) Project costs were 120% of appraisal estimate because of slightly higher infrastructure costs and the unexpected purchase of pumping equipment. (e) Approximately 22,300 additional farm families were employed as a result of the project, with indirect employment of another 10,000 expected. (g) Farm earnings increased on average from about $350.00 to $7,500 per family.  - 4 - (h) The ERR was 12%, at constant 1978 prices, lower than estimated at appraisal for various reasons explained below. (i) Staff time supervising the project was 305 field staff-days, a relatively small amount for a complex national undertaking, with 351 sub-projects. 1.13 Finally, a few recommendations should be noted here: (a) With the planned elimination by 1990 of the subsidy of the cost of energy, there is a greater need to promote cost- sharing -lans more rapidly and more widely. (b) In the future, beneficiaries should contribute toward those costs with special credit from BANRURAL. (c) A mid-implementation evaluation would have helped redirect efforts where necessary, i.e., livestock sub-projects. Other more specific recommendations are found in VII, and suggestions for follow-up projects are found in Annex II.  - 5 - II. PROJECT FORMULATION A. Identification and Preparation 2.01 This project originated as A: logical part of an on-going national program to develop productive, small-scale agricultural Infrastructure, based mainly on small-scale irrigation. The program itself dates back to 1937, when the Directorate for Small-Scale Irrigation was created within the National Cotmmission for Irrigation to benefit small groups of low-income farmers. As a result of the December 1976 merger of the Ministry of Hydraulic Resources and the Ministry of Agriculture and Livestock into the SARH, the scope of the small-scale irrigation program was enlarged to include providing the infrastructure for drainage and livestock development as well. At the same time, the project was renamed the National Plan for Small-Scale Agricultural Infrastructure Development (PNDR). Responsibility for its implementation was given to OHIADR with support from the SUDR to provide technical assistance to the farmers involved in the project. B. Appraisal 2.02 This is the fifth in a series of six small-scale infrastructure projects. All of the other projects, including the sixth, which is still being implemented, have been supported by Inter-American Development Bank (IDB) loans. At appraisal in April 1978, three projects already had been successfully implemented, adding about 300,000 ha to Mexico's irrigable area, which now totals 5,800,000 ha. 2.03 Evaluation data from the third project which were available at appraisal, also showed very positive results. These included low investment cost per hectare and per family; a positive effect on the rural poor; improved and more intensive use of the land; and a shorter period of time needed by sub-projects to achieve full production. 2.04 In addition to these excellent =ouc-ic and social results, the executing agencies already had proved to be efficient, highly motivated organizations whose efforts to a considerabla ettent made the program a success. 2.05 Since OHIADR had been made responsibl_. for the development of livestock infrastructure the GOM proposed the inclusion of a livestock component in the project, to absorb about 52 of total infrastructure costs. The Bank agreed with this proposal, since it woAld develop areas which were generally uncultivated becau.,e of poor rainfall and poor soil as well as the fact that its occupants lacked the necessary resources to obtain credit for improvements and had few, if any alternatives, for development assistance. However, at appraisal, no assessment was made of earlier experience with livestock infrastructure schemes, as had been done with the hydraulic sub-projects. This may have been, because OHIADR's one-year experience in livestock was too short a time on which to base an a3sessment.  - 6 - 2.06 To ensure that the sub-projects selected would meet the necessary standards, it was agreed during negotiations that, they should be approved by the Bank based on feasibility reports, and that the Economic Rate of Return (ERR) for each sub-project would be at least 12%, or that the Benefit-Cost Ratio (B/C ratio), as used by the OHIADR would be higher than 1.0 at a discount rate of 12%. To ensure the project's maximum impact on the rural poor, the average investment per sub-project family would have to be less than US$10,000 (Letter 02/16/79). Beneficiaries would have to agree to take full responsibility for the operation and maintenance (O&M) of infrastructure, and the funding as well, through their water users' association. They would also bear part of the investment cost, which was to be recovered according to normal PNDR procedures. 2.07 During implermentation, the sub-project cost limit per family was increased to $20,000, because costs in general had escalated (Letter of Amendment 05120/81). C. Irrigation and Drainage Development in Mexico 2.08 It is a fact of life that water is a scarce resource in Mexico. About 33% of the land classified is arid, with less than 250 mm of rainfall per year, and 31% semi-arid, with between 250 mm and 500 mm annually. This scarcity of water is compounded further by the fact that about 80% of the population and most cities and industries are located in the arid and semi-arid regions. Taken together, scarcity of water and increasing urban and industrial demands mean increasing competition for limited resources. That this is a serious national predicament is seen in: (a) the paramount importance that Mexican legislation gives to Government intervention for the allocation, use and conservation of water resources; and (b) the amount of public investment devoted to supplying water to cities and towns, and to the development of irrigation. 2.09 Seventy years ago, the Constitution of 1917 established water resources as the property of the state and ruled that concessions for its use could be granted only by the Federal Government , a measure currently authorrized by the Ministry of Agriculture and Water Resources (SARH). The Federal Law provides the legal framework for the allocation, use and conservation of the nation's water resources and designates irrigation districts and irrigation units as the legal entities responsible for irrigation. Irrigation districts are large schemes constructed and owned by the Government, but maintained and operated by SARH staff under guidance from the District Directive Committee, which is composed of government and farmers representatives. These irrigation districts typically involve large dams associated with complex irrigation and drainage systems, as well as roads and other infrastructure. Throughout Mexico, there are 83 large irrigation districts responsible for 3.6 million ha, which benefit about 486,000 farm families, with an average farm size of 7.0 ha. Recently, with SARH's reorganization in 1986, irrigation and rainfed districts were consolidated in 192 Districts for Integral Rural Development, with 7t0 Centros de Apoyo. The Centro de Apoyo is the basic territorial unit for planning and providing agricultural services to farmers. 2.10 The irrigation units, on the other hand, are small schemes, constructed and owned by the Government, but operated and aintained by  members of the unit itself, with periodic supervision from SARH; irrigation units are usually designed and planned at the state level. Currently there are about 26,000 irrigation units in Mexico, totalling 1.3 million ha, which supply 442,000 farm families, with an average farm size of 3.0 ha. Most of the units rely on groundwater. 2.11 Traditionally in Mexico investments in irrigation schemes have represented about 85% of public investment in the agricultural sector. Most investment to date has been to expand the amount of land irrigated, with about 20% invested to improve and rehabilitate irrigation schemes. Since 1960, there has been an impressive expansion of the area irrigated, and of both irrigation districts and irrigation units: Table 1 Irrigation District Irrigation Units Total Area Increased Area Increased Area Increased Period In Period Cumul. in P-riod Cumul. in Period Cumul. (thousand of hectares) Before 1960 - 1,649 - 337 - 1,986 1961-1970 838 2,485 262 689 1,088 3,074 1971-1980 916 8,400 411 1,000 1,,326 4,400 1981-1985 800 3,700 800 1,300 600 6,000 2.12 It is important to note the role of irrigation units in expanding the irrigated area from 1981 to 1985, at the same time that Mexico experienced a period of acute economic difficulty. Table 1 shows that small irrigation schemes can become even more attractive for a nation like Mexico when there is a shortage of capital for investment for several reasons: their lower unit cost; the shorter time required for construction of the units; and the rapid realization of benefits; both agricultural and otherwise. 2.13 Thus, the development of a productive hydraulic infrastructure is one of GOM's main programs to increase agricultural production and, at the same time, improve conditions for the rural population. The projects which GOM devised to meet those objectives can be grouped in the following categories: (i) Productive small-scale hydraulic infrastructure, of which this completed project is a part; (ii) productive, small-scale hydraulic infrascructure combined with productive support works (roads). The Chiapas Project (Loan 2526-ME) falls under this category; (iii) productive small-scale hydraulic infrastructure with both productive support infrastructure (roads and electricity) and  social infrastructure (schools, drinking water, etc.), which is onl- required in the country's poorest subregions. Loans made through Mexico's Investment Program for Rural Development PIDER I (Loan 1110-ME) and II (Loan 1467-ME) provide for these needs; (iv) rehabilitation, modernization and completion of large irrigation systems. The Rio Fuerte/Sinaloa Project (Loan 1706-ME) and the proposed Northwest Irrigation Rehabilitation Project are part of this component; and (v) new irrigation systems which require the development of water supply works (storage dams) as well as distribution works and which are therefore expensive. The Ocoroni Project (Loan 1908-ME) and the Apatzingan Project (Loan 1858-ME) are samples of this. The latter has an important repair/refurnishing component and therefore falls partly under category (iv). 2.14 The program for the development of productive, small-scale hydraulic infrastructure, of which this project is part, is one of GOM's most important programs, and because of its sizeable potential, will remain so in the future. III. IMPLEMENTATION A. General 3.01 The loan was approved in December 1978 and became effective in April 1979. The SAR assumed a fo4r years implementation, with project completion expected by December 31, 1987. rhe actual implementation period was about 6 1/2 years with the project closing on June 30, 1985. The main causes of delay were that, first, at the beginning of the project, the designs for irrigation and drainage sub-projects had not yet been completed and second, that funding was restricted in the later years of the project because of Mexicos's economic difficulties. 3.02 The Bank's supervisory input in the project was relatively limited. Eleven supervision missions visited the country in the 6 1/2 year period, requiring a total of 305 field staff-days. The table belo:r shows projected goals and actual accomplishments, organized by major project components.  - 9 - PROJECTED AND ACHIEVED GOALS Sub-project Physical Beneficiaries Investment No. Units Hectares Families USt million Pro3ected Achieved Pro]. Achlv. Pro]. Achiev. Pro]. Achlev. Tubow*ll (9rn)1 218 (S36)1/ 86,60M 28,000 n.s. 9,700 n.a. 82.6 Storage Diversion Dam 20 25 12,600 17,0W n.e. 4,809 n.e. 89.4 PumDing Plant 18 16 8,000 5.000 n.s. 1.291 n.s. 9.8 Irrigation Subtotal - 259 (636)3/ 56,000 60,000 n.e. 16,800 110.2 181.8 Draingea - 18 36.000 16.000 n.e. 2.174 17.4 11.7 Irrigation and Drainage Subt. - 272 (636)1/ 92,000 66,000 n.a. 17,974 127.6 148.6 Livestock - 79 68,000 172,000 n.e. 4,861 16.1 17.9 Pumping EquiDment - - (428)1/ - - - - 11.0 Total 250 861 (963)1/ 160,000 288,000 80,000 22,836 143.7 172.4 1 ( ) = number of tubewells or pumping plants. 3.03 Total project investment was 20% higher than appraised. Of that total, 51S went into tubewells, 231 to storage and diversion dams, 6% to pumping plant irrigation, 71 to drainage, and 10% to livestock infrastructure. 3.04 The expansion in hectares was slightly below what was planned at appraisal, with 50,000 ha developed for irrigation (89Z) and 16,000 ha for drainage (44Z). Investment in livestock infrastructure was about what was planned at appraisal but, due to the inclusion of more extensive sub-projects, it actually shows far more improvementE, with 172,000 ha developed (2961). 3.05 The increased project costs over those expected are mainly due, first, to the higher cost of the irrigation infrastructure, which averaged US$2,640/ha, or 34% above appraisal estimates and, second, the addition in 1983 of a component to purchase pumping equipment (US$11.0 M). The latter was used to complete sub-projects not included in the project which had funding problems. 3.06 Of t:he five sub-project categories completed thus far, the tubewell sub-projects have been the most successful, particularly in length of time for implementation, farmers' response and cost-sharing. Drainage and pumping plants were the next most successful, but had less satisfactory cost-sharing arrangements; they were followed by storage dams, which had less satisfactory implementation periods, farmers' response and cost-  - 10 - sharing. Nevertheless, all of these sub-projects created viable development components, and important and tangible result. However, the majority of the livestock sub-projects have been disappointing in terms of farmers' response and cost-sharing arrangements. These factors, coupled with stagnant meat and dairy prices in recent years, have made the livestock component less than marginally viable. B. Proiect Components 3.07 Tubewells. More than half of the project investmenc was for tubewell development (US$87.4 M). The appraisal objective of equipping 36,000 ha has been substantially achieved (80%), with the development of infrastructure on 28,000 ha. 3.08 In general, the tubewell projects were well prepared, as expected of OHIADR, given its years of experience. Its designs provided the minimum infrastructure necessary for adequate operation and maintenance. However, the impression remains that OHIADR was rushed into producing sub-projects without having had sufficient time to study the hydrogeology of the wells since, at the beginning of the project implementation, the preparation of hydraulic sub-projects had not been completed. The number of wells in the initial proposals totalled 998, spread over 360 sub-projects. In the end, 535 wells in 218 sub-projects were constructed. In the state of Zacatecas in particular the attrition rate was high (56%) because groundwater potential did not meet expectations. Lost investment as a result of preliminary studies, geophysical exploration, test wells and production wells, the latter mainly due to Insufficient tubewell capacity amounted to 15% of the cost of viable tubewells. In Mexico, where exploitable aquifers are of the alluvial type and therefore rather predictable, losses could have been reduced if more studies at the pre-feasibility level has been made of the hydrogeological conditions in the project areas. 3.09 The quality of the groundwater has been good in general and no salinity problems have been reported. Only two tubewells had to be abandoned because of saline water (Oaxaca - BM 145 and Michoacan - BM 267). 3.10 It was not possible for the farmers to monitor the water-use because flow-meters were not provided. Although they had been included in the PNDR's first projects, they were omitted in later ones, since farmers showed not interested, probably due to a lack of extension service efforts to train them in the benefits of good water management. 3.11 Storage and Diversion Dams. The development of small dam irrigation constituted the second most important financial component of the project, after tubewells, at 23% of total project investment. There were 25 schemes serving an average of 700 ha each, with the largest one irrigating 3,665 ha. Two-thirds of the operations were storage dams, and one-third diversion dam. 3.12 Larger sub-projects were hurt by long implementation periods (5-7 years). In fact, one sub-project in Jalisco has not yet been completed; the problem is typical for large projects in Mexico in general. Due to political necessity scarce investment funds have to be spread-out each year over as many sub-projects as possible, making it difficult to concentrate  - 11 - funding on a few large sub-projects. For future small-scale infrastructure projects, therefore, the size of storage dams that would be included should be decided by considering the availability of funds--budgetary allocations- -to avoid tightening scarce investment funds in few large sub-projects. 3.13 The beneficiaries of storage dam schemes also seem to be less eager to optimize water-use and crop production. Pert aps they feel less responsible since they have less invested and committed than the farmers benefitting from the tubewell schemes; the latter pay for electricity and equipment which the former pay nothing. Organizing the farmers is also more difficult in the case of dams, since the average scheme (700 ha) involves the organization of approximately 140 farmerr. Furthermore, the water supply of dam schemes is not as reliable as that of a tubewell; this fact became painfully evident !.his year (1987) when a drought hit the Central Region. 3.14 One remedy for the lack of farmers' involvement is to institute cost-sharing by requiring their participation in financing part of the project cost at the outset instead of waiting until afterwards, when water charges would be spread over a 25-year period. 3.15 The problem of the variability of the water supply could be mitigated by incorporating tubewells in the sub-projects. This appears to be a particularly appropriate solutior. for the Joia de Mulas storage dam scheme in Guanajato. In this case a partly filled reservoir, a result of the 1987 drought, and a dispute over water rights on a canal shared with a neighboring scheme have resu.lted in very poor utilization of the land. 3.16 PumpinR Plants. The development of the pumping plant component constituted a valuable project component, even though it used only 7% of total project investment. Since pumping plants depend for their water supply on rivers or lakes, in general already nearly exploited to full capacity, the potential for further pumping plant development is relatively limited. 3.17 Drainage. Drainage development, although also a minor project component, with only 7% of total project investment, is also an excellent development alternative because of its low investment cost per hectare (US$700/ha), and its positive impact on production, since it transforms unproductive, waterlogged land into fully productive acreage. Thus, drainage is an excellent way for the GOM to raise agricultural production and improve the living conditions of its beneficiaries while keeping the costs low. Drainage development has excellent potcntial in the coastal plains of the humid southeastern region, in the states of Tabasco, Vera Cruz and Chiapas, but is also a possibility in waterlogged depressions around lakes and rivers in the semi-arid central region. However, beneficiaries especially in the larger sub-projects (the largest is 4,500 ha), have been difficult to organize and educate about the need for maintenance, especially of the larger drains which are mechanically maintained. 3.18 The dam scheme cost-sharing plan, which levies annual water charges for 25 years, is not very effective in giving farmers a sense of being owners with full responsibility for maintenance and operation of the  - 12 - dams. For future projects, the GOM should consider alternative arrangeiments, wa.'reby the beneficiaries finance part of the project works themselves with appropriate types of credit beginning at the outset of the projects. 3.19 Livestock. This relatively minor project component, with 10X of investment, fared poorly. The infrastructure provided was not used adequately to increase livestock production as was incended, because of management, credit and marketing problems. 3.20 The main impediments to a more successful livestock component have been: (a) OHIADR's lack of experience in the sector led to sub-projects which were overdesigned and too costly. At appraisal it was just one year since OHIADR had been made responsible for the development of small-scale infrastructure for livestock. 'hus, an accurate assessment of previous problems and successes in developing livestock infrastructure enterprises was not possible then given OHIADR's inexperience. (b) The beneficiaries' experience with intensive livestock raising methods and their lack of skills and motivation needed for collective management meant a gap in execution of the project. (c) The large additional investment to be provided by the farmers through medium- and short-term credit for iutputs such as improved livestock, feed and veterinary services was significant, compared to the relatively small input to be provided by their existing resources in cattle, land, crop production and labor. (d) Since the end of 1985, the prices of meat and dairy products, which are not subsidized by GOM, heve not kept pace with the price increases in feed and other inputs. 3.21 The most successful livestock sub-projects were those with a relatively simple infrastructure, a low level of additional input requirements, and technology with which the beneficiaries were already, to a large extent, acquainted. Typical of such operations are the livestock fattening sub-projects in Tabasco, in the humid Southeast region, where thexe is no need to purchase supplemental feed because the rainfall is adequate for year-round growing of crops and pasture. 3.22 The project's livestock development component can be said to have had the same results as of a similar livestock sub-project component in the PIDER II project. However, in addition, this livestock component was hurt by domestic meat and dairy marketing conditions, which were deteriorating since the end of 1985. (The problems and results of livestock development in the PIDER II Project (Loan 1464-ME) were extensively analyzed in its completion report of June 1986). 3.23 In the PIDER II project PCR, it was estimated that about 501 of the livestock enterprise did not operate at all or was operated poorly.  - 13 - The rate of return for the total investment in livestock development was calculated at approximately 9%. C. Credit 3.24 The project did not include a credit component because it was assumed by GOM that existing facilities and institutions had adequate resources to meet incremental demand. However, crucial to the project's success was the medium-term credit required for financing the purchase of tubewell equipment and livestock; this was to be provided by the state's National Rural Credit Bank (BANRURAL). Sufficient credit was provided to buy the tubewell equipment although, in some cases, the credit was released rather slowly, plagued by a large service debt in users' associations, and by an inadequate budget in field branches because of the difficult economic circumstances at that time. No data have been made available on the total credit amount but, based on tubewells installed, it is the approximate equivalent of US$15 million, whereas the amount of medium-term credit for the purchase of livestock has been minor by comparisons. Since the end of 1985, when meat and dairy prices deteriorated and profits from livestock enterprises seemed unlikely, BANRURAL has almost completely cut its credit line dedicated to livestock purchases. 3.25 The amount of incremental demand for short-term credit generated by the project is also unknown. Based on an area of 66,000 ha added by the project, 't is estimated that the equivalent of US$10-15 million would have been needed to make an adequate productive use of constructed irrigation and drainage infrastructure. D. O&M and Cost Recovery 3.26 In general, O&M work was satisfactorily performed and was paid for by the beneficiaries who are organized in water users' associations for the purposes of the irrigation and drainage sub-projects. 3.27 These associations &ppear to be best organized and most active in using the tubewell schemes. This can be explained by the clear need for prompt repair of the tubewells' mechanical and electrical equipment in case of breakdown and by the need for prompt payment of electricity bills. The cost of energy is about Mex$20 per kwh, of which 2/3 is subsidized by GOM. However, the subsidy is being reduced, with the goal of eliminating it by the end of 1990. 3.28 With average energy requirements of 1,000 kwh per hectare, the current average annual payment of the equivalent of US$6.5/ha would be increased to US$20.00/ha by at the end of 1990 if the GOM subsidy were eliminated. All other O&M charges or costs are paid for by contributions, either in-kind, (mostly labor), or b; annual quotas, with major repairs financed through ad hoc levies. The tasks of operating the tubewells and distributing the water are often performed by the beneficiaries themselves on a rotating basis. 3.29 In all hydraulic schemes, the total cost of the required O&M work, including labor and energy, is the equivalent of approximately US$10/ha/year. In storage dam schemes where there is no pumping, the  - 14 - savings on energy costs are offset by higher O&M costs for dams and larger irrigation networks. However, since the project works are new, the need for maintenance is minimal; therefore, farmers thus far have been required to contribute much less than they will in the future. 3.30 The beneficiaries began their ma4.ntenance of the drains by cutting weeds growing there. Since the drains are new, there has been no need to rent equipment (backhoes) from SARH to clean the drains and remove silt deposits. Thus, there have been no ptoblems ti date. 3.31 Cost-sharing by t}e beneficiaries has been very satisfactorily arranged in the case of the tubewell sub-project. The farmers finance about 95% of its construction cost by paying for the mechanical and electrical equipment (such as the transformer) through BANRURAL credit. In the case of the other sub-projects, cost recovery will be achieved through annual charges, paid over a 21-year period after an initial grace period of 4 years. However, each year the amount has to be renegotiated because of inflation. Unfortunately, the annual charges make the beneficiaries feel they are renters rather than owners; naturally, then, they do not feel responsible for the conservatio'n of the project works. Thus, recovery of project costs through annual :harges is a less than satisfactory procedure. In the future, the possibility should be considered for all sub-project categories that beneficiaries contribute from the outset, using a special line of credit from the BANRURAL, as is done with tubewells costs. E. Monitoring and Evaluation 3.32 Due to SARH's decreasing operational budgets monitoring and evaluation of the implementation of the sub-projects was inadequate. Furthermore, the failure to maintain a continual account of the physical progress and the effects of the sub-projects on agricultural development, summarized in semi-annual progress reports, was the main reasor for the poor information available to prepare the PCR. OHIADR's reports focused mainly on the progress of construction despite the fact that the Bank continually asked that this lack of appropriate information be redressed. However, it did not ask the cooperation of BANRURAL for data on agricultural production. Overall, the monitoring and evaluation (M&E) system faced certain obstacles: (a) the inherent difficulty and high cost of collecting technical, financial, socioeconomic and production data from a large variety of different types of investments scattered over a wide geographical area in 27 Mexican states; (b) the fact that SARH field visits were curtailed due to budget restrictions, which also involved a reduction of personnel; (c) the reorganization and decentralization of SARH, which included the reassignment and dispersion of staff members who were familiar with the project; and (d) the fact that important data were lost in the September 1985 earthquake in Mexico City. F. Disbursements 3.33 Disbursements initially were three years behind schedule because of a slow project start. Delays were reduced to approximately one year by 1984, with the replacement of cancelled sub-projects and the procurement of additional equipment at the end of 1983, but rose again to roughly two years by the time of the final disbursement.  - 15 - Fiscal Year Appraisal Estimate Actual Total Actual as x of Appraisal 1979 0.3 1.70 0.56 1980 21.4 4.10 0.13 1981 34.6 8.20 0.23 1982 48.8 16.28 0.33 1983 60.0 23.63 0.40 1984 - 55.98 0.93 1985 60.00 100.00 Procurement 3.34 No problems worth mentioning have arisen from procurement. International Competitive Bidding (ICB) has been limited to the purchase of construction equipment. The loan financed equipment as was planned at appraisal, for a total of US$10.3 million. All equipment was bought before the end of 1983 and therefore was not affected by restrictions imposed later by the Ministry of Planning and Budget (SPP) on the import of equipment for which there are comparable products manufactured in Mexico. Civil works for the project have been awarded through Local Competitive Bidding (LCB) and local shopping for small works, and force accounts. Reporting and Auditing 3.35 Annual reports were submitted to the Bank with some delays. The reports focused mainly on the progress of construction based on financial disbursements, which in most cases did not contain information on the physical progress of the project necessary to gauge efficiency and unit cost. A review of the supervision missions' reports indicates that in general auditing was not a problem during project implementation, and that separate accounts were maintained for each sub-project, as agreed in the loan document. Adherence to Covenants 3.36 In general, compliance with covenants was satisfactory except for two important operational aspects: First, the recovery of part of the investment costs through the collection of water charges has not yet been possible. The collection is still in its initial stages, because of the four-year grace period, the recovery, even if successful will take some time to show results. Second, payment of 0 & M costs was difficult, despite the fact that SARH had determined the appropriate charge. This Was because the users' associations had problems collecting those fees because of rampant inflation and because the farmers themselves resisted paying water charges that covered the full cost of operating and maintaining the system.  - 16 - IV. ECOvM.TC AND AGRICULTURAL IMPACT A. Project Benefits and Incremental Output 4.01 Benefits from the irrigation and drainage component include the employment of about 18,000 additional farm families, thereby increasing employment opportunities across the country. At full development, the .ncremental production of principal products, mainly corn, sorghum, wheat, beans, oil seeds, vegetables and fruits, would amount to a coLal market value of about 45,000 million Mexican pesos in 1987 prices (about US$33.0 million). Eventually the project will provide indirect employment for about 10,000 additional persons, helping to stem migration from the project areas to Mexico's already overpopulated cities. As a result, too, the generation of employment will help improve income distribution. Indeed, average farm earnings for direct project beneficiaries have already increased from about $350.0 to about US$7,500 per family at full development of sub-project, an impressive accomplishment. B. Economic Rate of Return 4.02 The estimated economic rate of return (ERR), excluding the livestock component, was 12% at a constant peso flow of 1978 prices. The ERR is roughly the opportunity cost of capital. For details see Annexes IV and V. The result is lower than that estimated at appraisal (23%) because of these differences: (a) economic prices during the project implementation period were lower than those projected at appraisal; (b) the total area benefitted by investment on irrigation and drainage was smaller - by about 28% - than estimated at appraisal. At the same time, total investments in constant terms are somewhat higher than those calculated at appraisal; (c) about 15% of tubewell investments were lost in sub-projects cancelled due to unproductive wells; and (d) sub-project investments took longer than expected at appraisal. 4.03 In general, each sub-project was implemented according to a methodology agreed upon during appraisal, whereby SARH would furnish the Bank with a "summary sheet' containing technical information, as well as the costs and benefits for each sub-project, including information about the sub-project's ERR. This measure not only provided flexibility in investment decisions but also an adequate aggregated ERR for the entire project. V. INSTITUTIONAL PERFORMANCE 5.01 The PNDR has been successful in achieving the main objective of the project, which is to develop small-scale irrigation and drainage infrastructure, by operating through its executing agencies, OHIADR and SUDR. It was less successful in developing viable livestock enterprises, as is explained above (see paragraph 3.20). 5.02 The reorganization of the SARH occurred toward the end of the project implementation period. As a result, responsibility for larger sub-  - 17 - projects (storage and diversion dams, pumping plants and drainage) was given to the General Directorate of Irrigation and Drainage (DGID) while the General Directorate for Rural Infrastructure (DGIR) was charged with the smaller sub-projects (tubewells and livestock). Technical assistance was transferred to the General Directorate for the Coordination of Delegations (DGDC) but had no appreciable effect on project implementation. However, the reorganization meant valuable information was lost because of staff transfers. Thus, it is very di,ficult to make a quantitatively accurate evaluation of the institutional performance for the purposes of the PCR. Although unified project management by OHIADR, in close cooperation with the SUDR for technical assistance, is clearly preferred for efficient implementation of the project, it is estimated that the new arrangement can provide acceptable management. One advantage is that project coordination was shifted from Mexico City to regional offices which are better acquainted with local conditions and problems. VI. BANK PERFORMANCE 6.01 The staff time spent in supervising the project amounted to 305 field man-days, which is comparatively little for a complex project, ultimately including 351 sub-projects. It is no surprise that reporting on project progress was not fully satisfactory, especially after the reorganization of the SARH. The Bank missions have not been able to obtain accurate information on the project's readjusted objectives nor on its physical, financial and production achievements, a fact whi^h has resulted in considerable variation in quantitative results reported from one supervision mission to the next. The results from monitoring project Implementation could have been considerably improved if a mid- implementation evaluation had been made. That would have provided the necessary information for GOM and the Bank to consider what actions were needed to improve implementation, especially concerning components which were not doing well, such as the livestock sub-projects. VII. CONCLUSIONS AND RECOMMENDATIONS A. Conclusions 7.01 This project is part of a nationwide program for developing an agricultural infrastructure. It was initially conceived as a means to improve employment and raise the income of the poorest of the rural population. The scope of the first projects was limited to irrigation development and projects were justified for social reasons rather than economic ones. Since then, more attention has been given to their economic justification. 7.02 This project, which includes drainage and livestock infrastructure, is the fifth (Quinta Etapa) in a series of six. However, it was the first in which the Bank participated although the first four projects, as well as the sixth, which is still being implemented, have been supported by IDB loans.  - 18 - 7.03 In the field of small-scale infrastructure development the Bank also has participated in two integrated development projects, PIDER I and II which, in addition to components for productive infrastructure, have components for technical assistance, productive support infrastructure (roads and electrification) and social infrastructure (water supply and schools). The second project, which mainly focussed on livestock development, was terminated at the beginning of 1983. It was not very successful, perhaps due to its complexity and to the inexperience of its collective managers. 7.01- The Bank is also financing the recently initiated (1985) Chiapas Rural Development Project which is aimed at boosting production of rain-fed agriculture through improved soil conservation, and drainage and flood control, along with the development of productive support infrastructure, such as feeder roads, and extension services. 7.05 The small-scale infrastructure project has had an overall positive result owing to successful implementation of its irrigation and drainage components, consisting of tubevells, small storage and diversion dams, pumping plants and drainage schemes, and to the positive response of the beneficiaries in making productive use of the infrastructure. The largest schemes (more than US$2.5 million) took more time to construct (5 to 7 years) than planned, due to funding restraints. Monitoring of water is still less than perfect for two reasons: most of the schemes have not been equipped with measuring devices, and technical assistance does not educate farmers on the importance of monitoring water use for optimizing crop production. 7.06 The livestock infrastructure development, which was a relatively minor component, using only 10% of total project cost, was not so successful. The majority of the sub-projects either did not become operative or were operated poorly, because of a lack of motivation and organization among the beneficiaries, problems to obtain credit and the recent deterioration of marketing conditions for meat and dairy products. 7.07 Responsibility for the organization and management of the project implementation initially belonged to one SARH department, the OHIADR, was supported by SUDR, which gave technical assistance to the farmers. OHIADR had extensive experience with preparing and implementing small-scale hydraulic infrastructure sub-projects and was successful in executing them. It was less successful with livestock infrastructure sub-projects because of its inexperience. However, schemes selected at appraisal depended too much on substantial amounts of credit and on good management by the beneficiaries' collectives. 7.08 O&M and cost recovery objectives of the project, in general, have been adequately met but, except for the tubewell sub-projects arrangements, are not very satisfactory. The beneficiaries are organized in users' associations which are responsible for operating and maintaining the project infrastructure. Members of the associations perform part of the O&M work, while annual charges pay for the remainder of it. GOM recovered part of the project investment cost for the tubewell schemes by having the farmers pay for pumping and electrical equipment (about 25% of sub-project  - 19 - construction cost). Cost recovery for the other types of sub-projects has been initiated recently with annual water charges levies. The cost will be recovered, according to PNDR guidelines, over a 25-year period, including a 4-year grace period. However, that collection process is cumbersome; furthermore the need to adjust cost charges annually for inflation has created a climate of uncertainty and friction between SARH and farmers. B. Recommendations 7.09 A number of lessons can be drawn from this project which are pertinent to future small-scale infrastructure development projects: (a) Future projects should include only the development of tubewells, storage and diversion dams, pumping plants and drainage schemes. The inclusion of livestock schemes should be decided on consideration of prices and marketing conditions for meat and dairy products. When their inclusion is justified, credit institutions (BANRURAL) should play a more prominent role in the preparation of livestock sub-projects in order to ensure that they are viable and not too dependent on credit or unrealistically high levels of managerial skill among members of the beneficiaries' collectives. (b) Active involvement of the beneficiaries in the sub-projects is essential for success. The most effective way to achieve their participation is to have the beneficiaries bear and finance part of the construction costs through credit made available for this purpose, as has been done with the tubewell schemes. As a result, crops in the tubewell areas are the best cared for and farmers the most motivated. The opportunity to increase production through the use of the infrastructure in the other sub-projects appears to be less well-used. Cost-recovery through annual water charges perpetuates the sense among farmers that the government and not they owns the scheme. Furthermore, the need to annually increase the water charges because of inflation makes it a built-in source of uncertainty and friction. Therefore, GOM should promote beneficiaries cost-sharing from the outset for all sub-project categories in future projects. (c) Monitoring water use in irrigation sub-projects is essential for good management of this scarce resource. Flow meters, a relatively minor cost-item, should be provided as part of the sub-project for future tubewell and pumping plant schemes. Extension services should educate the farmers about the importance of the meters in optimizing crop production. In the larger surface water schemes, measuring devices also should be provided as a matter of course at all distribution structures, including even the smallest farm turnouts (approx. 50 ha), as has been done with the Texalcame scheme. (d) To ensure that project funds can be dispersed equitably over all sub-projects, large ones (more than US$2.0 million)  - 20 - should be excluded in future projects; their exclusion would eliminate the diversion of a disproportional amount of project funds during construction for only one project, and would mean that construction would take less time. Reasonable sub-project averages might be 1,000 ha for storage dam schemes and 2,000 ha for drainage schemes, with upper limits approximately double that. (e) Future projects should include a mandatory mid-implementation evaluation of the project.  - 21 - ANNEX I MEXICO Prolect Completion ROport Small-Scale Aaricultural Infrastructure Project Project Costs (1978 MexI Millions) 1979 1980 1981 1982 1983 1984 1995 TOTAL A. Engineering / 9.80 40.4 81.21 19.81 15.60 17.88 7.a8 142.07 S. Investaent V Tubewelis 187.85 677.10 408.02 156.29 108.40 165.62 62.15 1699.86 Dan. 14.18 18.94 121.38 184.87 204.76 204.85 109.29 882.21 Pump Station 18.04 14.98 59.68 46.58 87.e0 44.81 11.02 224.17 Drainage 0 7.52 102.27 103.88 81.88 28.26 o 288.17 Failed !I 88.09 88.89 80.88 8. o. 16.28 16.28 266.64 Total Invoestmnt 241.66 698.58 771.38 489.54 885.64 448.71 182.46 3209.92 Admin. J 86.85 149.79 115.70 78.48 657.8 66.11 27.84 628.60 TOTAL COST 287.91 888.77 918.24 682.78 459.10 524.65 216.70 8878.49 i/ About 3.4x of execution costs. V Civil work and equipment. 8/ Tubewell lnv-stments that were cancelled. / About 15X of Investment cost. September 11, 1987  - 22 - ANNEX II Follow-up ProJects 1. For Mexico's agricultural development it is important that GOM continue to finance productive small-scale agricultur-l infrastructure. There is still great potential for the development of tubewells, small- scale storage or diversion dams, pumping plants, and drainage schemes. Especially when compared with other infrastructure development programs, the advantages of small-scale agricultural infrastructure are clear: their low investment cost per hectare, the quick results on agricultural production per family; the positive response from farmers; and, for the tubewells schemes in particular the successful cost-sharing method. The size of dams, pumping plants and drainage schemes should be kept small-scale in order to have the best chance for success. 2. If the need exists for GOM to develop larger schemes, it should group 5 to 10 of them in one medium-scale hydraulic infrastructure project. 3. Future cost-sharing of project investment costs by beneficiaries for the dam, pumping plant and drainage schemes should be arranged as in the case with tubewell schemes, that is, with farmers paying 25-30% of sub-project costs at the outset. Cost recovery after the fact through annual water charges should be discontinued because it discourages beneficiaries from assuming full responsibility for O&M and causes friction because of the annual readjustments for inflation. 4. Drainage schemes with have particularly low investments per hectare and per family have the greatest potential in humid areas, such as the coastal plains of Tabasco and Chiapas and in the depressions around the rivers and lakes of the Central region (Queretaro-Guanajato-!kichoacan- Jalisco). They should be exploited in these regions to their utmast. 5. Drainage schemes in Tabasco and Chiapas would have the additional advantage, if land is available, of providing possibilities for settling indigenous farmers from the mountain areas and slowing the rate of deforestation in these areas. 6. A reassessment should be made of livestock infrastructure development. Because of the need for substantial credit for livestock schemes, the BANRURAL should have a major role in the preparation of any livestock sub-projects.  - 23 - ANNEX III MEXICO Loan 1643-ME Small-Scale Agricultural Infrastructure Prolect Project Completion Report Withdrawal of Proceeds of the Loan Amounts of the Amounts Actually Actual Dis- Loan Allocated Disbursed bursement US$ US$ Amount All Civil Works 38,500 49,700 129 Equipment 10,300 10,300 100 Unallocated 112200 TOTAL 60,000 60,000  - 24 - ANNEX IV MEXICO Small-Scale Azricultural Infrastructure Proiect Proiect Completion Report Rate of Return Analysis Constant 1978 Prices (MEX$ Million) Total Incremental Total Incremental Net Incremental Investments 11 Benefits Benefits 2J 1979 193.4 12.6 (180.8) 1980 556.0 66.4 (489.6) 1981 617.0 113.2 (503.8) 1982 391.6 146.8 (244.8) 1983 308.6 214.8 (93.8) 1984 352.0 288.4 (63.6) 1985 146.0 304.3 158.3 1986 0 317.9 317.9 1987 321.1 321.1 1988-2005 321.1 321.1 ERR = 12% 11 Adjusted for unskilled labor on civil works and installations. 2/ Benefits from tubewells, dams, pumping and drainage; net of O&M costs.  - 25 - ANNEX V MEXICO Small-Scale Agricultural Infrastructure Prolect Project Completion Report Assumptions for Benefit Analysis 1. The economic rate of return for this project; estimated at 12%, was derived under the following assumptions: (a) because the implementation of the sub-project was a national operation, regional unit cost per ha and unit benefits per ha were estimated to determine a net flow of benefits which, in turn, were based on the cumulative annual area under production as a result of project investments in tubewells, dams, pumping stations and drainage; (b) for each specific type of investment, the average benefit was based on a representative mix of the prevailing production pattern and prices, which were then adjusted to reflect economic farm-gate prices; (c) production costs were adjusted for taxes and other transfer costs; (d) the opportunity cost of unskilled labor was adjusted to reflect the regional unemployment rates in Mexico. On average it was estimated at 70% of its market price; (e) calculations of the economic rate of return were indexed at 1978 constant Mexican pesos in order to compare the results with those estimated at appraisal; (f) costs for purposes of economic analysis are all physical infrastructure works related to the project, including equipment and operation costs; (g) the cattle development program was not included in the analysis because of the lack of adequate information. As this component represents about 10 percent of total investment, iA-ts exclusion does not affect the economic analysis results significantly; and (h) benefits of the project to employment, family income and value of production were estimated from the aggregated data on benefits and sub-projects summary sheets.  - 26 - ANNEX VI Page 1 of 3 Comments Received from Borrower 88E0601 April 12, 1988 Spanish (Mexico) OEDD1 RP:me General Directorate of Public Credit Directorate of International Financial Organizations Subdirectorate of Projects, Agricultural Zector Mexico, DF. March 29, 1988 Mr. Graham Donaldson Division Chief, Agriculture, Infrastructure and Human Resources OED World Bank Washington, DC. Communication No. 305.III.3.-1108 I refer to your commun