Morocco - Agricultural Projects

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 12701 PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (LN.1602-MOR) LOUKKOS RURAL DEVELOPMENT PROJECT (LN. 1848-MOR) MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (LN. 2082-MOR) OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (LN. 2217-MOR) JANUARY 25, 1994 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. WEIGHTS AND MEASURES Metric System CURRENCY EQUIVALENTS Currency Unit = Dirham (DH) 1979 US$1 = DH 3.8 1980 US$1 = DH 4.5 1981 US$1 = DH 5.2 1982 US$1 = DH 6.2 1989 US$1 = DH 8.5 1990 US$1 = DH 8.3 1991 US$1 = DH 8.2 1992 US$1 = DH 8.5 ABBREVIATIONS AND ACRONYMS CT Work Center CNCA National Bank for Agricultural Credit DE Livestock Department DEF Department of Water and Forests DERRO Project for Rural Development of the Western Rif DMV Department of (Agricultural) Development DPA Provincial Department of Agriculture DPV Department of Crop Production ED Executive Director ERR Economic Rate of Return FAO Food and Agriculture Organization FERTIMA National Fertilizer Distribution Agency IMF International Monetary Fund INRA National Institute for Agronomic Research MARA Ministry of Agriculture and Agrarian Reform M&E Monitoring and Evaluation MTR Mid-term Review OED Operations Evaluation Department ONEP National Bureau for Potable Water SAP Special Action Program SAR Staff Appraisal Report SONACOS National Agency for Suds' Marketing T&V Training and Visit UNDP United Nations Development Program FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Offloe of Direator-Genel Operations Evaluation January 25, 1994 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT Subject: Performance Audit Report on the Kingdom of Morocco Fes-Karia-Tissa Agriculture Project (Ln. 1602-MOR) Loukkos Rural Development Project (La. 1848-MOR) Middle Atlas Agriculture Development Project (Ln. 2082-MOR) Oulmes-Rommani Agricultural Development Project (Ln 2217-MOR) Attached is the Performance Audit Report on the Kingdom of Morocco - Fes-Karia-Tissa Agriculture Project (Ln. 1602-MOR), Loukkos Rural Development Project (Ln. 1848-MOR), Middle Atlas Agriculture Development Project (Ln. 2082-MOR), and Oulmes-Rommani Agricultural Development Project (Ln 2217-MOR) prepared by the Operations Evaluation Department. These pilot projects promoted rural development in rainfed, upland areas. Each was aimed at a distinct agroecological area, and implemented by a separate directorate of the Ministry of Agriculture and Agrarian Reform. Each included a component of rural roads, and all, except Middle Atlas, supported social infrastructure development. Three projects (Fes-Karia-Tissa, Loukkos, and Oulmes-Rommani) had satisfactory outcomes, and one (Middle Atlas) was unsatisfactory. The PCR review that rated Middle Atlas as satisfactory has been revised downward because the evidence of overall project benefits is minimaL These projects were implemented with less than full Borrower commitment to rainfed area development, at a time when the Borrower was experiencing acute budgetary difficulties. As a result, all projects were starved of counterpart funds, and their size and scope sharply reduced. The counterpart funding problem had been foreseen, but it was ignored. Implementation was overshadowed by the struggle for funds. Help, when it came in the form of a mid-term review and a special action program, came late. Despite these problems, the projects achieved many of their revised physical targets. The satisfactory projects had numerous components which did not perform well and a small number of components which provided sufficient benefit to earn the satisfactory ratings. Since completion, support for rainfed area development has been modest. Institutional support for agricultural research and extension has replaced area development, building upon lessons learned and capacities built through these projects. Attachment 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 PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (LN. 1602-MOR) LOUKKOS RURAL DEVELOPMENT PROJECT (LN.1848-MOR) MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (LN.2082-MOR) OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (LN. 2217-MOR) TABLE OF CONTENTS Page No. Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i Basic D ata Sheets ..................................................... iii Evaluation Summ ary ................................................... xi I. BACKGROUND... .............................................. 1 II. THE PROJECTS ................................................ 3 Fes-Karia-Tissa Agriculture Project (Ln. 1602-MOR) ....................... 4 Loukkos Rural Development Project (Ln. 1848-MOR) . . 4 Middle Atlas Agriculture Development Project (Ln. 2082-MOR). . ...... .. 6 Oulmes-Rommani Agricultural Development Project (Ln. 2217-MOR) ............. 6 The Main Conclusion of the PCRs ................................... 8 III. THE ORIGINS OF BANK SUPPORT FOR RAINFED AGRICULTURE IN MOROCCO ... 8 The Evolution of the Loukkos Project ................................. 8 The Austerity Program, 1978-80 .................................... 11 IV. IMPLEMENTATION EXPERIENCE .................................... 12 The Effect of Budgetary Shorfalls ................................... 12 The Effect of Inadequate Technology ................................. 15 The Effects of Incomplete Preparation ................................. 16 The Effect of Complex Institutional Arrangements ......................... 16 The Effects of Weak Monitoring and Evaluation .......................... 18 Supervision ................................................. 19 The Mid-Term Review and the Special Action Program ...................... 20 V. PROJECT OUTCOME ......... .... 21 Farm Impacts: Cereals and other Crops ................................ 21 Farm Impacts: Livestock Production .................................. 21 Infrastructural Outcomes ......................................... 22 Economic Rate of Return ......................................... 23 Environmental Effects ........................................... 23 Institutional Development . ......................................... 24 Experience as Pilot Projects ....................................... 24 "This report was prepared by Christopher Gibbs (Task Manager), assisted by Jose Olivares, both of OED, who audited the project in July 1993." 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. VI. FINDINGS AND ISSUES............................................25 Overall Findings...............................................25 Commitment and Political Will......................................25 Implementing Pilot Projects........................................27 Time to Start Implementation Up.....................................27 Coping with Shortage of Counterpart Funding ............................ 27 Complementarity of Components .................................... 27 Beneficiary Participation ......................................... 28 Necessary Conditions for Rural Development Success............... . . . . . .. 28 Bank Performance.................................... . . . . . . . .. 29 Figure 4.1: Chronology of Significant Events in Morocco and in the Projects under Audit . . . . . . . . 14 Figure 4.2: Organizational Arrangements for Implementation: Membership of National, Provincial and Local Coordinating Mechanisms ......................... 17 Figure 4.3: Project Supervision Time (Staff Weeks) ............................... 19 Table 2.1: Cost Structure of the Projects, at Appraisal (US$ millions)............ . . . . . . . 3 Table 4.1: Loans, Disbursements and Cancellations (US$ millions) ...................... 12 Table 6.1: Project Performance Summary ...................................... 25 MAPS IBRD 13377R IBRD 14879R IBRD 15866RI IBRD 15865RI IBRD 15898R IBRD 16473 IBRD 16474 PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (LN. 1602-MOR) LOUKKOS RURAL DEVELOPMENT PROJECT (LN.1848-MOR) MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (LN.2082-MOR) OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (LN. 2217-MOR) PREFACE Between 1978 and 1982, the World Bank supported four rainfed rural development projects in Morocco. The projects aimed to increase the profitability of agriculture in upper watershed areas, conserve soil to protect impoundments, and relieve poverty. These were pilot actions intended to build experience which could be extended to other comparable areas in Morocco. One earlier project (Cr. 0555-MOR) was audited by OED in 1986.11 Bank loans to these projects totaled $158.0 million, of which $72.4 million (45.8 percent) was disbursed and $85.6 million (54.2 percent) canceled. Board approval dates ranged from 1978 and 1982 and closing dates from 1988 to 1992: Project Loan No. US$ Approval Closing Date Final (m) Date Disbursement Fes-Karia Tissa AP 1602 65.0 6/78 6/90 11/90 Loukkos RDP 1848 34.0 5/80 12/88 6/89 Middle Atlas ADP 2082 29.0 1/82 12/90 4/91 Oulmes-Rommani ADP 2217 30.0 12/82 12/92 5/93 This PAR is based on the Project Completion Reports2/, Staff Appraisal Reports (SARs), loan documents, project files, discussions with active and retired Bank staff involved in project design and implementation, and an audit mission to Morocco in July 1993. The audit mission visited all the project areas and discussed the projects with officials of the Ministry of Agriculture and Agrarian Reform (MARA) in Rabat and in the provinces, and with farmers in the project areas. Their kind cooperation and assistance is gratefully acknowledged. The draft PAR was sent to the Borrower for comments but none was received. Meknes Agricultural Development Project, OED Report No. 6167. The PCR for Oulmes-Rommani had not been received by the time the draft PAR was completed, but this project was included in the audit for completeness, at the request of the Region.  - Ill - PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (Loan 1602-MOR) BASIC DATA SHEET Key Project Data Actual % of Appraisal Actual Appraisal Estimate Estimate Total Project Cost 161.5 86.8 56% Loan Amount (US$m) 65.00 Disbursed (US$ million) 36.25 Cancelled (US$ million) 28.75 Economic Rate of Return 21% 9% Institutional Performance Poor Cumulative Estimated and Actual Disbursement (in US$ m) FY80 FY81 FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 Appraisal Estimate 3.00 10.00 22.00 36.00 48.00 58.00 65.00 - - - - - Actual 1.63 2.31 5.47 9.63 12.02 14.75 18.60 22.64 26.19 30.56 34.14 36.25 Actual as % of Appraisal 54 23 25 27 25 25 29 35 40 47 53 56 Date Final Disbursement November 30, 1990 Project Dates Original Revised Actual Identification May 1976 Preparation October 1976 May 1977 May 1977 Appraisal October 1977 - October 1977 Negotiations May 1978 - May 1978 Board Approval June 1978 - June 1978 Loan Signing September 1978 - September 1978 Loan Effectiveness December 1978 - February 1979 Loan Closing June 1986 June 1990 June 1990 Completion Date June 1992 Indicative only. - iv - Staff Inputs (Staff weeks) TO FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 FY92 Total Preparation 42.6 42.6 Appraisal 128.6 128.6 Negotiation 30.0 30.0 Supervision 42.2 15.5 13.7 11.2 19.1 4.8 5.4 5.0 3.6 17.6 2.0 140.1 Other 0.8 - 0.4 Total 244.2 15.5 13.7 11.2 19.5 4.8 5.4 5.0 3.6 17.6 2.0 342.5 Mission Data Date No. of Staff Days Performance Month/Year Persons in Field Specializationk' Rating& Identification& 5/76 n/a Preparation 5/77 n/a Appraisal 10/77 7 21 E(2),IEn,A,Con(2) n/a Appraisal Follow-up 2/78 5 12 E(2),A,En,Lo n/a Supervision 1 10/78 2 8 A,Le 2/F Supervision 2 4/79 4 8 n/a n/a Supervision 3 10/79 1 22 A 1/- Supervision 4 6/80 1 13 A 1/F Supervision 5 11/80 2 17 A,C 2/M,F Supervision 6 6/81 3 20 C,A 1/rn Supervision 7 1/82 2 7 A,E 1/m Supervision 8 7/82 3 17 A,E,ME 1/m Supervision 9 1/83 2 16 A,E 2/m Supervision 10 10/83 2 6 A,E 2/F,M Supervision 11 7/84 2 5 A,E 2/F Supervision 12 10/84 1 5 A n/a Supervision 1311 7/85 5 6 A,L,E 2/F Supervision 14 2/86 1 3 A 2/F Supervision 15 11/87 1 3 A 2/F Supervision l6/ 2/88 1 2 I n/a Supervision 17 4/88 1 n/a I n/a Supervision 18 10/88 1 1 A 2/F Supervision 19 5/89 1 1 A 2 Supervision 20 6/89 1 2 I n/a Supervision 21 12/89 1 1 A 2 Supervision 22 6/90 1 1 A n/a Supervision 23 2/91 2 7 A,E n/a b/ A = Agricultural; E = Economist; C = Credit Specialist; Con = Consultants; IEn = Irrigation Engineer; L - Livestock Specialist; Le = Legal Specialist; Lo = Loan Officer; ME = Monitoring and Evaluation Specialist; En = Engineer. V 1 = Improving; 2 = Stationary; 3 = Deterioration; M = Managerial; F = Financial. Project preparation was essentially undertaken by Moroccan staff with the assistance of FAO/CP. Several FAO/CP missions took place between 5/76 and 5/77 for reconnaissance, pre-identification and preparation. Manpower details are not available. Combined mid-term review for three projects: Fes-Karia-Tissa, Loukkos, and Middle Atlas. Partial Supervision of Rural Watersupply Schemes only. Some staff paid a visit in July 1988. - PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO LOUKKOS RURAL DEVELOPMENT PROJECT (Loan 1848-MOR) BASIC DATA SHEET Key Project Data Appraisal Actual % of Estimate Actual Appraisal Estimate Total Project Cost (US$ m) 67.6 n.a. 42% Loan Amount (US$ m) 34.0 Disbursed (US$ m) 14.38 Cancelled (US$ m) 19.62 Economic Rate of Return 21.0 15.01, Institutional Performance Fair Cumulative Estimated and Actual Disbursement (in US$ m) FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 Appraisal Estimate 9.40 16.80 22.80 28.00 32.20 34.00 - - - Actual 1.28 3.52 6.15 7.04 8.47 10.02 13.39 14.13 14.38 Actual as % of Appraisal 14 21 27 25 26 29 39 42 42 Date Final Disbursement July 31, 1989 Project Dates Original Revised Actual Identification - Feb. and May 1975 Preparation 1 - October 1975 Preparation 2 - April 1977 through July 1978 Pre-appraisal July 1979 - July 1979 Appraisal September 1979 - September 1979 Negotiations April 1980 - April 1980 Board Approval May 1980 - May 1980 Loan Signing December 1980 - December 1980 Loan Effectiveness March 1981 June 1981 September 1981 Loan Closing June 1987 December 1988 December 1988 Completion - - June 1988 Indicative only. - Vi - Staff Inputs (Staff weeks) TO FY80 FY81 FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 Total Preappraisal 48.3 48.3 Appraisal 50.3 50.3 Negotiation 3.7 3.7 Supervision - 9.5 21.2 24.4 15.0 14.3 16.6 4.8 3.4 5.1 9.1 0.3 123.7 Other -LU - - 1LZ Total 120.9 9.5 21.2 24.4 15.0 14.3 16.7 4.8 3.4 5.1 9.1 0.3 244.7 Mission Data Date No. of Staff Days Performance Month/Year Persons in Field Specalizationy RatingS Preparationk' Pro-Appraisal 07/79 2 10 Ec, LS Appraisal 09/79 4 21 Ec, Ag, LS, FS Appraisal Follow-up 01/80 2 5 Ec, PA Supervision 1s, 09/80 2 - IS, FA Supervision 2 05/81 2 6 AgC, Ec 1 Supervision 3 10/81 4 5 Ec, FA, HEA, TB 1 Supervision 4 11/81 1 6 FS 1 Supervision 5 02/82 4 5 Ec(2), Ag, LS 2 Supervision 6 06/82 1 6 FS I Supervision 7 11/82 3 7 Ec, Ag, LS 2 Supervision 8 04/83 2 8 Ag, FS 3 Supervision 9 02/84 2 7 Ag, FS 2 Supervision 10 10/84 2 7 Ag, LS 2 Supervision 11 06/85 4 22 Ag, LS, Ec - Supervision 12 02/86 1 7 Ag 2 Supervision 13 02/87 2 7 Ag(2) 2 Supervision 14 11/87 1 5 Ag 2 Supervision 15 10/88 1 2 Ag 2 PCR preparationt' 04/89 1 10 Ec - There were about 3 FAO/CP reconnaissence/pre-identification missions between February 1974 and May 1975, an FAO/CP identification mission in October 1975, and a Bank identification mission in August 1976. In addition, there were about six FAO/CP preparation missions between December 1977 and July 1978. Manpower details are not available. Supervision mission to review progress towards effectiveness. Combined with Vegetable Production Marketing Project Supervision. 2 weeks altogether. Combined Mid-Term Review for three rural development projects Fes-Karia Tissa, Luokkos, and Middle Atlas. 1 Economist from FAO/CP to assist with the economic re-evaluation to be included in borrower's contribution to PCR. f/ Ec = Economist, LS = Livestock Specialist, Ag - Agronomist, FS = Forestry Specialist, FA = Financial Analyst, IS = Irrigation Specialist, AgC = Agriculture Credit Specialist, HEA = Highway Engineer Analyst, and TB = Transpost Economist. W Types of Problems: 1 = Problem free or minor problem, 2 = Moderate problem, and 3 = Major problem.  Vii - PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (Loan 2082-MOR) BASIC DATA SHEET Key Project Data Appraisal Actual % of Estimate Actual Appraisal Estimate Total Project Cost 92.2 15.9 31% Loan Amount (US$ m) 29.0 Disbursed (US$ m) 9.00 Cancelled (US$ m) 20.00 Economic Rate of Return 21% n.a. Institutional Performance poor Cumulative Estimated and Actual Disbursement (in USS m) FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 Appraisal Estimate 5.00 11.00 19.00 26.00 29.00 - - - Actual 1.13 1.77 3.22 4.29 5.29 7.54 8.67 9.00 Actual as % of Appraisal 23 16 17 17 18 26 30 31 Date Final Disbursement April 30, 1991 Project Dates Original Revised Actual Identification - May 1979 Preparation -- July 1980 Pre-appraisal November 1980 - November 1980 Appraisal April 1981 - October 1981 Negotiations December 1981 - December 1981 Board Approval January 1982 - January 1982 Loan Signing March 1982 - March 1982 Loan Effectiveness September 1982 November 1982 Loan Closing June 1988 June 1989-December 1990 December 1990 Completion June 1987 April 1991 Staff Inputs (Staff weeks) TO FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 FY92 FY93 Total Preappraisal 216.6 216.6 Appraisal 58.7 34.8 93.5 Negotiation 1.7 1.7 Supervision 12.5 18.1 15.2 10.9 22.0 5.6 6.7 8.4 3.5 1.3 2.6 106.8 Other 83.9_ 5 8 3. . 232. Total 283.9 53.9 18.1 15.2 10.9 22.1 5.7 6.7 8.4 3.5 1.3 2.6 432.3 - Viii - Mission Data Date No. of Performance Month/Year Persons Specialization Ratin' Identification 05/79 Preparation1 07/80 Pre-Appraisal 11/80 Appraisal 04/81 5 Ec, RE, Ag, LS, FS Supervision 1 05/82 4 Ec(2), En, Agn 1 Supervision 2 06/82 1 F 1 Supervision 3 12/82 3 Agn, LS, FS 2 Supervision 4 05/83 1 Agn 3 Supervision 5 11/83 1 Agn 3 Review 02/84 1 FS - Supervision 6 05/84 1 Agn 3 Supervision 7 07/84 1 FS - Supervision Sh/ 07/85 5 Agn(2), LS, Ec, Agn(Cons.) 3 Supervision 9 10/85 2 FS - Supervision 10 05/86 1 Agn, LS 3 Supervision 11V 12/86 3 Agn 3 Supervision 12 n.a/87 7 Agn, LS, Agn n.a. Supervision 13 07/88 2 n.a n.a Supervision 14 11/88 1 Agn, LS n.a. Supervision 15 05/89 1 Agn, LS 2 Supervision 16 11/89 Agn 2 Supervision 17 06/90 Agn n.a. al Project preparation was essentially undertaken by Moroccan staff with the assistance of FAO/CP. Several FAO/CP mission took place between 5/79 and 7/80 for reconnaissance, pre-identification and preparation. Staff details are not available. Combined mid-term review: Fes-Karia Tissa, Louddos and Oulmes-Rommani. & Supervision reports for period 12/86 to 10/88 (not available) in project files. At least two supervisions took place. Reports and details are not available. 21 Ec = Economist, LS = Livestock Specialist, Ag = Agriculturalist, Agn= Agronomist, F = Forester, FS = Forestry Specialist, RE = Rural Engineer, and En = Engineer. f/ Types of Problems: 1 = Problem free or minor problem, 2 = Moderate problem, and 3 = Major problem. - ix - PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (Loan 2217-MOR) BASIC DATA SHEET Key Project Data Appraisal Actral % of Estimate Actual Appraisal Estimate Total Project Cost (US$ m) 105.0 n.a. 45% Loan Amount (US$ m) 30.0 Disbursed (US$ m) 13.50 Cancelled (US$ m) 16.50 Economic Rate of Return 14.0% 14.0% Institutional Performance poor Cumulative Estimated and Actual Disbursement (in USS m) FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 FY92 FY93 Appraisal Estimate 1.60 4.10 9.00 14.80 24.10 29.00 30.00 - - - Actual 0.44 1.06 1.21 3.07 5.81 5.92 6.98 8.87 10.45 13.50 Actual as % of Appraisal 28 27 13 21 24 20 23 30 35 45 Date Final Disbursement May 17, 1993 Project Dates Planned Revised Actual Appraisal March 1982 March 1982 Negotiations October 1982 October 1982 Board Approval December 1982 - December 1982 Loan Signing May 1983 May 1983 Loan Effectiveness August 1983 November 1983 March 1984 Febnuary 1984 March 1984 Loan Closing December 1989 June 1991 December 1992 June 1992 December 1992 Completion Date -n/a A/ Indicative only. - x - Staff Inputs (Staff weeks) TO FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 FY92 FY93 Total Preappraisal 65.9 65.9 Appraisal 77.5 9.4 86.9 Negotiation 2.9 2.9 Supervision 1.9 8.9 5.0 9.5 6.9 3.3 10.4 7.5 4.1 3.7 9.2 70.4 Other 3--. _.5 Total 144.7 21.4 8.9 5.0 9.5 6.9 3.3 10.4 7.5 4.1 3.7 9.2 234.6 Mission Data Date No. of Days in Performance Ratings1 Overall Month/Year Persons Field Specializationh' Avail Funds Status Preparation* Appraisal 03/82 5 105 A, E, I, FO, L Supervision 1 10/83 2 6 A, FO 2 Supervision 2 05/84 2 6 A, FO 2 Supervision 3 10/84 1 5 A 2 Supervision 4 10/85 2 8 A, FO n/s# Supervision 5 07/86 1 6 A 2 Supervision 6 02/87 2 14 A, A 3 2 2 Supervision 7 11/87 1 3 A 3 1 2 Supervision 8 10/88 1 3 A 3 1 3 Supervision 9 04/89 1 4 A 3 2 3 Supervision 10 05/90 2 8 A, L 3 1 2 Supervision 11 12/90 1 2 A 3 1 2 Supervision 12 06/91 1 4 A 2 1 2 Supervision 13 05/92 1 5 A 2 1 2 hI A = Agriculturist; E - Economist; FO = Forestry Specialist; I - Irrigation Engineer; and L - Livestock Specialist. V' 1 - Problem free; 2 - Moderate problems; and 3 - Major problems. Il The project was prepared between 1979-81 by the Government with the assistance of the CP and there are no sufficient details on staff inputs. 1" Form 590 was not completed during this supervision. RELATED BANK LOANS AND CREDITS Loan/Credit No. Project Name Loan Amount Year of Year of OED PAR (US$millon) Approval Closlnt Year No. Cr.0555-MOR Meknes Agricultural Development Project 14.0 1975 1984 1986 6167 Ln.1704-MOR Fourth Agricultural Credit Project 70.0 1979 1983 1984 5392 Ln.2110-MOR Forestry Project 27.5 1982 1987 1992 10508 La.2367-MOR Fifth Agricultural Credit Project 115.4 1984 1987 1992 11491 Ln.2590-MOR Agricultural Sector Adjustment Loan 100.0 1985 1987 1989 7868 Ln.2731-MOR Sixth Agricultural Credit Project 100.0 1987 1990 1992 11492 Ln.2885-MOR Second Agricultural Sector Adjustment Loan 225.0 1988 1992 - - Ln.3036-MOR Research and Extension Project 28.0 1989 1995 - - Ln.3156-MOR Second Forestry Project 49.0 1990 1996 - - PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (LN. 1602-MOR) LOUKKOS RURAL DEVELOPMENT PROJECT (LN.1848-MOR) MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (LN.2082-MOR) OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (LN. 2217-MOR) EVALUATION SUMMARY Introduction about the use of Bank loan funds for rainfed development returned during project preparation, 1. Agricultural development in Morocco and in 1978 when a three-year austerity budget has been dominated by investment in irrigation was introduced. for more than 70 years. The vagaries of the Mediterranean climate were to be overcome by 5. Despite budget austerity, Fes-Karia- irrigation of a million hectares in the plains. A Tissa was approved in 1978 and Loukkos in concerted attack on the problems of rainfed 1980. At Loukkos' approval, the Bank's Board areas was mounted only in the 1970s. supported Morocco's rainfed strategy but indicated concern that counterpart funding could 2. By 1977, almost all of Morocco's 1.2 be a problem, and questioned its replicability million farmers were rainfed farmers, but because unit costs were high. The Board was rainfed agriculture attracted little Government assured by staff that Government had pledged investment. In this context, the Bank and FAO the necessary support, and that costs would made a concerted effort to raise the profile of come down as experience was gained. rainfed agriculture. The audit examines the four projects which resulted. The projects were 6. Despite counterpart funding problems in approved between 1978 and 1982, with a total Fes-Karia-Tissa and Loukkos, two more projects loan commitment of US$158 million. were approved in 1982. Ultimately, only US$72.4 million (45.8 percent)' of the loan funds 3. The Bank initiated a dialogue with the approved were disbursed, and US$85.6 million Government on rainfed agriculture in 1972. The (54.2 percent) was canceled. All projects were initial focus was on the Loukkos River reduced in scope and each required two or three watershed where the Bank was concerned that extensions before closing. Implementation was further investment in irrigation would be beset throughout by counterpart funding infeasible without profitable development shortages, a risk which had not been properly upstream. Over time, the dialogue grew to assessed at appraisal. include watershed management and poverty alleviation. 7. The Borrower's inability to providelocal funds, reflected both budget austerity and weak 4. The Government was not initially commitment to rainfed agricultural development. enthusiastic about this strategy, but it appears to Whereas Bank-supported irrigation and have been persuaded of its soundness by 1976, agricultural credit projects were more or less when preparation of the first of the projects fully disbursed, rainfed agriculture and forestry under audit began. But Government uncertainty were allowed to lag. - xii - Objectives design deficiencies were made good and arrangements for implementation were put in 8. These were pilot projects, targeted at a place. range of rainfed conditions. Their common objectives were to: identify and test methods to 13. Implementation continued to be slow raise rainfed agricultural productivity; improve because of budgetary shortages, and the projects incomes of small farmers; reduce dependence on were scaled-back following mid-term reviews imported food commodities; and provide rural (MTRs), which also reduced their complexity. employment. 14. Infrastructure was put in place before 9. As pilot projects, their designs were field activities. Some infrastructure targets were overly-complex and their budgets were large. exceeded, because costs in dollar terms fell with Each possessed four components: agricultural the devaluation of the dirham. But many development; roads and access improvement; infrastructure targets were cut back. Two large social infrastructure; and surveys, studies and village drinking water systems in Fes-Karia- training. Only Middle Atlas lacked a social Tissa were completed after long delays and cost infrastructure component. Cadastral surveys increases, but these systems now function were included in two projects, and all four effectively. included monitoring and evaluation (M&E) components. 15. Technology development occurred slowly as the modified extension system took 10. Institutional arrangements reflected the hold, but productivity gains were constrained by complexity of the projects. Each was to be lack of appropriate technologies and scarcity of coordinated by a separate directorate of the operating funds. The replacement of cereal Ministry of Agriculture and Agrarian Reform cultivation on sloping lands by perennial tree (MARA); and each had three levels of crops advanced well. Where cereal cultivation interagency coordination: central, provincial and continued, new higher-yielding soft wheats local. Eight ministries or agencies sat on most tended to replace traditional hard wheats. This of the coordinating committees. At the local change resulted mainly from new national level, the projects were implemented through programs and policies, but it was accelerated by MARA's Work Centers (CTs). the projects. 11. The main institutional objectives were 16. Many of the technologies with which the to: expand the agricultural extension force; projects set out were replaced on the basis of lower the ratio of extension agents to farmers; experience. Those which succeeded were the reduce the extension agents' administrative simpler, lower-cost technologies. Where workload; and draw the extension agents out of perennial crops replaced annuals, this was the market places (up to then, their preferred site achieved without significantly lowering incomes for meeting farmers) and into the field. Farmer during the transition. participation in rural development activities through group formation was to be strongly 17. A 1985 MTR of three projects, led to encouraged. their being scaled back and a round of loan cancellations. This was followed by a special Implementation Experience action program (SAP) in 1987, to reduce the projects further and improve implementors' 12. Despite a lengthy gestation period, access to funds. The development and approval preparation of numerous project components was of the SAP was slow. SAP provisions were not incomplete at approval, adding to slowness in effectively communicated to the implementing initial implementation. Time was lost while agencies, reducing its effectiveness. - xiii - 18. When the projects focused on single significantly in Loukkos because of improved activities, such as agricultural extension, when access and extension services. Agricultural funds were available, implementation experience credit use expanded in Fes-Karia-Tissa, the was satisfactory. Legacies of the projects are result of more liberal lending criteria. But competent teams of agricultural professionals at reliable data on the performance of the credit the provincial level, which do not appear to exist program is not available. Improvements in in provincial offices which have not received livestock productivity were generally less than external support. expected at appraisal. 19. M&E experience was disappointing 24. The most important problems of because Borrower commitment to it was weak. implementation were financial, organizational Weak M&E handicapped project managements and related to M&E, but Bank supervision was and lowered the quality of reporting. The Bank largely by technical specialists. Supervision of was aware of this deficiency from supervision social development components was negligible. reports but was unable to correct it. M&E Following the 1986 MTR, total time devoted to training was provided, but too late to make a project supervision was sharply reduced. difference to implementation. 25. Lack of reliable M&E makes the clear 20. The Bank's failure to include computers statement of results problematical. Data for M&E and statistical data handling under the weaknesses make the re-estimated ERRs projects reduced the value and timeliness of the indicative in three projects, and impossible in M&E that did take place. Numerous studies one. planned under the projects were not completed; given budget limitations, their cancellation was 26. The audit mission found evidence to justified. Much of the technical assistance indicate that the revised objectives were planned was not taken up for budgetary reasons. substantially met in Fes-Karia-Tissa and Loukkos. In Oulmes-Rommani, one project Results component (cereal production), generated enough benefits to sustain all costs of the revised 21. Infrastructure outcomes are the best project. In Middle Atlas, the ERR could not be documented. Much of the infrastructure planned re-estimated and evidence of benefits was for Fes-Karia-Tissa and Loukkos was minimal. implemented. The quality of original construction appears high, but maintenance by Sustainability Government has been hampered by operating cost shortfalls. 27. The overall outlook for sustainability is uncertain because local budgetary difficulties 22. Some new farming practices have been continue to exist. If funds for extension widely adopted to conserve soils, but the cost- activities continue to be available, gains made effectiveness of the practices is unknown. under the projects are likely to be sustained. Substantial areas of hillside have been reforested, and sylvo-pastures brought under 28. The sustainability of the perennial low-cost management. Much of the cropping systems appears likely. If not reforestation was efficiently implemented by overwhelmed by recurrent drought, the contractors, which created employment but gave collective range management in Middle Atlas rural communities little stake in the results. But, should be sustained by the shepherds' in Loukkos, a program working directly with cooperatives and the breed improvement farmers was started. associations. 23. The overall effect of the projects on 29. Sustainability of reforested areas would farm incomes cannot be well established. Crop have been enhanced if local communities had a productivity has been raised, but by less than greater stake in their maintenance, and berefitted anticipated at appraisal. Crop production grew more from their use. - Xiv - 30. With management by the National new projects should not have been approved Bureau for Potable Water (ONEP) and cost after 1980, until the projects already under recovery from users, sustainability of the implementation were adequately funded and drinking water systems appears likely. supervised. 31. Sustainability of much project-supported 37. Borrower commitment to the rainfed infrastructure seems unlikely since maintenance area development strategy promoted by the Bank is limited. As a result, the value of many and FAO, was neither complete nor sustained.11 infrastructure inputs will be transitory. But cost Bank leverage that took Morocco into the recovery arrangements for water supply schemes rainfed strategy did not buy commitment. Once in Fes-Karia-Tissa should ensure their approved, however, Borrower and Bank staff sustainability. struggled hard to implement the projects as planned, but they were coping with difficulties Findings and Issues which had been foreseen and should have been avoided. 32. Three of the projects-Fes-Karia-Tissa, Loukkos and Oulmes-Rommani-are rated 38. The rainfed strategy was well satisfactory in their revised forms. None intentioned, but it was based on limited achieved their original objectives fully. Middle experience and reflected a preocupation with Atlas is rated unsatisfactory.11 At the PCR components rather than frameworks or processes review, Middle Atlas was rated satisfactory with for change. To have succeeded, more attention a caveat that this was a generous finding. With needed to be paid to systems which underpin the exception of a successful collective range sustainableagriculture-ecological,social,public management component, the audit found little administration, and market. evidence of project benefits, and the PCR rating has been revised downward. 39. Overly-complex project designs were approved despite Bank staff misgivings. Such 33. With the exception of Loukkos, which is complexity was consistent with Bank thinking likely to be sustainable, overall sustainability of about rural development in the late-1970s. The the projects is uncertain. Institutional Borrower argued for projects with multiple development is partial or negligible. components, and the Bank agreed to them. 34. rojct erfomane ws costrine by Three projects became "Christmas trees." By 34. Project performance was constrained by 18, te dfiute soitd wt acute counterpart funding problems, weak instittnally c le asea d o t Borrower commitment to rainfed agricultural projects wee pne btea butlthe development, overly complex project designs, des n eof ues-oo m an remaine lack of detailed designs at approval, and limited g beneficiary participation. unchanged. 35.40. As pilot actions, the projects were too foreseen, but seen to be manageable. As a terve as ex pero me wasbul result, all four projects were approved without i proje ds opermment ationiof any changes in design or financial arrangements: struactites.gDta eig ofpumeous project approvals and maintenance of the lending prt cmpones wa i ete at appris program were put ahead of likely development Lac o plned proet atuppriod performance. probably contributed to the weak M&E pefomacefBtrorowraomiten.t 36. The earlier projects outperformed the permaned wea troughout men- later ones because they faced fewer local budgetary limitations. The audit concludes that tation. - xv - 41. Once the full extent of counterpart * project initiatives in areas where prior funding shortages had been formally experience is limited, especially, require acknowledged by the Bank (in 1983), the MTR clear objectives and simple project designs; and SAP should have been undertaken sooner, and approved and implemented more quickly. * integration of multiple components within a Steps to make the SAP operational should have single project is very hard to achieve, been communicated more effectively by the creating difficulties for the Bank and the Ministry of Finance to the implementing Borrower; directorates of MARA. * pilot projects require modest amounts of 42. Beneficiary participation in implemen- dependable funding. They should be small tation was generally limited, a reflection of enough to be managed intensively and given MARA's former operating style. This should sufficient time to succeed; have been anticipated and given more attention by the Bank. Nevertheless, MARA's capacity * rural and area development projects require for participatory work with farmers, and among time to be built in at the outset to prepare for directorates at the provincial level, grew with implementation; these projects: where external support was provided, a responsive style of operation was * proven agricultural technologies are instituted. frequently lacking for rainfed areas and time is required to develop or adapt them to 43. Project risks were not adequately specific settings; assessed. The risk of counterpart funding shortages was badly underestimated. Drought * prospects for sustainable improvements in risk was recognized, but the identification of natural resources management are reduced actions to take in the event of drought was where field activities have been carried out missing. Between 1978 and 1992, Morocco mainly by contractors and beneficiary experienced six years of severe drought, adding participation is low; to implementation difficulties. * sustainability of project infrastructure 44. None of the lessons of these projects for requires an on-going commitment to fund the Bank is new. The projects display maintenance and execute it; characteristics which are already well documented, including: * without full Borrower commitment, M&E suffers; * lending pressure can lead to unjustified project approvals. These projects * mid-term reviews (MTRs) can be valuable demonstrate how an approval culture can tools for course corrections, but they must be overtake concern for satisfactory development implemented purposefully and at the right results; time. When projects are recognized to be in difficulty, MTRs should be brought forward * full ownership of projects is essential from to address problems quickly; and their outset, by the government, the implementors and the intended beneficiaries, * for a special action program (SAP) to if they are to succeed; succeed, its details must be fully understood by the implementing agency, and it must * the Borrower's continuing budgetary have the full backing of the Borrower's commitment to projects is vital; financial authority. - xvi - 45. The debate continues over what variation present in rainfed and upland areas-in constitutes an effective strategy for sustainable contrast to lowland, irrigated areas. Tailoring agriculture in rainfed areas. Bank experience an approach in Morocco would have required a suggests that successful projects contain a greater capacity for experimentation, and trial number of necessary components.2 Other and error, than was built into these projects. views highlight the need for appropriate Because the value of the projects under audit frameworks and processes.-1 Certainly, was lowered by budgetary problems, lessons interventions need to be closely tailored to learned were specific, rather than contributions local circumstances because there is so much to the overall debate. 1/ The Bank's MNA regional office disagrees with these conclusions. 2 OED, Area Development Projects. Lessons and Practices, No. 3, September 1993. 21 J.A.N. Wallis, Some Intensified Systems of Farming in the Tropics and Sub-tropics. World Bank Technical Paper, December 1993. PERFORMANCE AUDIT REPORT KINGDOM OF MOROCCO FES-KARIA-TISSA AGRICULTURE PROJECT (LN. 1602-MOR) LOUKKOS RURAL DEVELOPMENT PROJECT (LN.1848-MOR) MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT (LN.2082-MOR) OULMES-ROMMANI AGRICULTURAL DEVELOPMENT PROJECT (LN. 2217-MOR) I. BACKGROUND 1.1 This audit of four rainfed area development projects in Morocco - Fes-Karia-Tissa (Ln. 1602), Loukkos (Ln. 1848), Middle Atlas (Ln. 2082), and Oulmes-Rommani (Ln. 2217) - evaluates them as representatives of a significant class of Bank projectY It also puts them in their historical context in Morocco, where the development of irrigated agriculture had, and continues to have, the highest priority. 1.2 Morocco is geographically distinguished by its extensive mountains, coastal arable lowlands, and broad exposure to the Atlantic Ocean. The Rif and Atlas ranges capture orographic precipitation from moisture-laden storms originating over the Atlantic, and westward-draining rivers carry the water to the coastal lowlands, where Mediterranean climatic conditions predominate. East of the mountains exists a significant rainshadow. Rainfall is concentrated in the cool winter months of October to April; summers are hot and dry. Located on the southern margin of northern mid-latitude frontal storms, there are sharp irregularities in the amount and timing of precipitation. Droughts are common, and rainfed agriculture is precarious in most regions of the country. Rivers and streams draining the Rif and Atlas ranges are also marked by irregular flows that limit their use. Cereal production in Morocco is marked by low and fluctuating yields, directly attributable to irregular precipitation.2/ 1.3 In 1977, prior to a roval of the first of the projects under audit, Morocco's population was estimated to be 17.5 million. Two-thirds of this number lived in rural areas. Agriculture was dominated by cereal and livestock production, and was responsible for more than 25 percent of GDP and 60 percent of employment of the economically active population. Agro-industries and agri-businesses accounted for a further 10 percent of GDP and about 30 percent of the labor force in the industrial and service sectors. Virtually all of Morocco's 1.2 million private landowners were rainfed farmers. Nearly 80 percent of the farmers owned less than five hectares, accounting for less than 30 percent of the cultivated area. It was representatives of this large and important segment of poor farm households, in upland and mountainous areas, at whom the audited projects were aimed. 1.4 The audited projects were conceived as large-scale pilot actions to identify and test approaches to improving rainfed agriculture. Until the mid-1970s, modern agricultural development took place almost exclusively in Morocco's broad plains and plateaus which border its coasts. Most of 11 Nearly 900 agriculture projects have been evaluated by OED since 1970, of which 216 were area development projects. 2/ Morocco's average combined yield for its two principal cereals, wheat and barley, for the period 1974-84 was only 920 kg/ha, compared to a global average of 1,930 kg/ha for the same period. In 19 years of low rainfall between 1938 and 1984, Morocco's average cereal yield fell below 700 kg/ha. FAO Monthly Bulletin of Statistics, 1980-85. 1' See Morocco: Survey of the Agricultural Sector and Prospects for the Development of Rainfed Areas, The World Bank, March 16, 1977. -2- Morocco's good soils, level lands and water resources are located in this zone, as are most of the urban centers. Agricultural development, both during the protectorate period (1912-56) and since Independence, has been focused on the management of Morocco's water resources, to create a thriving irrigated agriculture in the Atlantic plains. In the two five-year development plans, 1968-72 and 1973-77, preceding the implementation of the projects under audit, 69 percent of the Government's financial commitment to agriculture was to irrigationA' despite the fact that more than 76 percent of the gross value of crop production and almost all livestock production came from the rainfed areas.11 The budget commitment to rainfed agriculture and forestry was 18 percent, plus 4 percent to the livestock sector. Furthermore, between 1968/72 and 1978/80, the proportion of planned government investment in forestry and soil conservation was halved, while the investment in irrigation increased slightlyfi/ 1.5 Morocco's emphasis on large-scale water resource development started nearly 70 years ago when "la politique des grands barrages"-the big dams policy-was established. By the end of the 1930s, the aim of the Protectorate administration was to harness all of Morocco's surface waters. The irrigation plan formulated then became known as the "century plan" because its horizon was the year 2000, when one million hectares were to have been transformed by perennial irrigation. *2 Since Independence, the priority of irrigation has been undiminished, with two significant exceptions: a) "Operation Labor", launched by the newly independent government in 1957; and b) two externally funded projects: the Project for the Economic and Rural Development of the Eastern Rif (DERRO) supported by UNDP and FAO, and the Meknes Agricultural Development Project (Cr. 0555) supported by IDA. 1.6 The aim of "Operation Labor" was to raise rainfed agricultural productivity quickly, through the large-scale introduction of tractors to enable timely land preparation, the more widespread use of fertilizer and improved seeds, and farmer organization in cooperatives. But "Operation Labor" failed and became unpopular with Morocco's farmers, making the Government cautious about further investment in rainfed agricultural developmentY DERRO and the Meknes project appear to have been more carefully adapted to conditions in the areas at which they were aimed. 1.7 DERRO was genuinely experimental and, while not always successful, provided directly relevant technical experience for the four projects under audit, in an area adjacent to the Loukkos project. This legacy is acknowledged in Bank SARs. 1.8 The Meknes project was considered by the Bank to have been satisfactory but had numerous shortcomings stemming mainly from an overly-optimistic expectation that lar e-scale transfer of state-owned and state-managed farms to land reform beneficiaries would take place. Te sensitivity of land reform issues in Morocco was evident in Meknes, but the lessons for rainfed agriculture which See page 31 of Morocco.: Survey of the Agricultural Sector and Prospects for the Development of Rainfed Areas, The World Bank, 1977. Ibid, page 21. See page 20 of The Agricultural Development of Algeria, Morocco and Tunisia by Kevin Cleaver, The World Bank, 1982. By 1986, the total irrigable area in Morocco's lowland plains under the command of large dams had reached 591,750 ha. Source: Morocco Travaux Public, unpublished documents. See Hermann Van Wersch, Rural Development in Morocco.: Operation Labor in Economic Development and Cultural Change, Volume 17, (1968). 2 OED Report No. 6167, Project Performance Audit Report - Morocco Meknes Agricultural Development Project (Cr. 0555-MOR). -3- were learned there are not referred to in the background to the projects under audit, even though they represent its logical descendants. 1.9 The performance of both DERRO and Meknes was mixed, and Government's commitment to rainfed agriculture remained tentative throughout the 1970s. After "Operation Labor", the initiative to place rainfed development higher on Morocco's agricultural development agenda came from donors, led by the Bank and FAO. The Bank's concern stemmed from the need to justify further investment in irrigation development, the poverty of large numbers of farmers in the rainfed zone, and the emerging recognition of the environmental link between upper watershed management and the sustainability of downstream impoundments. 11. THE PROJECTS 2.1 The four loans were approved between 1978 and 1982. The projects were conceived as pilot projects, designed to test and demonstrate methods to raise the productivity of rainfed agriculture across a range of agroecological conditions. All the projects contained agricultural development components, road and access improvement components, and components for surveys, studies, and training. In addition, three of the four projects contained large social infrastructure components. Costs of major project components at appraisal are shown in Table 2.1. Table 2.1: Cost Structure of the Projects, at Appraisal (US$ millions)A1 Fes-Karia-Tissa Loukkos Middle Atlas Oulmes-Rommani US$ m % US$ m % US$ m % US$ m % Agricultural Development 43.1 26.7 22.7 33.5 53.8 58.4 48.6 46.0 Roads 33.0 20.4 9.5 14.0 8.8 9.5 5.7 5.4 Social Infrastructure 20.4 12.6 7.3 11.2 --- - 11.4 10.8 Surveys, Studies and Training 18.0 4.2 24.2 13.8 18.3 6.5 28.4 9.2 Physical Contingencies 9.1 5.6 3.2 4.7 6.0 6.5 7.8 7.4 Price Contingencies 53.8 33.3 18.4 27.2 19.8 21.5 23.6 22.4 177.4 102.8 85.3 104.4 106.7 102.4 125.5 101.2 Physical and price contingencies have been omitted. 2.2 The projects' designs reflected the resource endowments of the project areas and the priorities of the coordinating directorates. Fes-Karia-Tissa aimed to provide a new model for a cereal growing area that needed to conserve soils and moisture in order to raise yields profitably. With densely populated and eroded mountains, Loukkos aimed to develop sustainable farming systems in the watersheds of strategically important dams, and to reforest high elevation hillsides. Middle Atlas focused on livestock grazing, and forest and forest range improvement in areas of low potential, and Oulmes- Rommani on crop and livestock systems, forests and sylvo-pasture in areas of higher potential. Each project started with a menu of potential technological alternatives which needed to be tested and extended. Each project served clients who were mostly resource-poor farmers, likely to be highly risk-averse. -4- Fes-Karia-Tissa Agriculture Project (Ln. 1602-MOR) 2.3 Loan 1602-MOR was approved for US$65.0 million in June 1978. The main objective of project was to improve the living standard of 34,000 farm families who cultivated 256,000 ha, 96 percent of the project area. This was to be achieved by increasing agricultural production and farm incomes. Agricultural initiatives were to be supported by investment in road construction and social infrastructure, including school classrooms, health facilities, and drinking water systems. Soil and water conservation measures were to be implemented on a trial basis. The project was to be implemented over 6 years at a total cost of US$161.5 million, with implementation coordinated by the MARA Directorate of Agricultural Development. 2.4 Project implementation was initially delayed by an incomplete design. Further delays were caused by a shortage of counterpart funds, which became apparent within two years of approval. Faced with budgetary difficulties, a mid-term review (MTR) was initiated in 1985, and a Special Action Program (SAP) was concluded in 1987 to expedite the flow of funds. The revised project was completed with a four-year delay in 1990. Actual costs were US$86.75 million, and 44 percent of the Bank loan was canceled. Because of US dollar appreciation against the dirham (DH) during implementation, many project components were implemented within the reduced budget. 2.5 The project established an improved agricultural extension system, based on the extension system. The ratio of extension agents to farmers was improved from 1:1,200 to about 1:350. Agricultural credit use was expanded from DH 2.2 million to about 2,400 farmers in 1979, to DH 22.0 million to 11,100 farmers in 1982. This rate of increase, however, overtook the capacity of the local banks to manage the loan portfolio adequately. Expanded credit use coincided with a period of recurrent drought, lowering farmers' ability to repay loans. 2.6 Hard wheat yields in the project area were increased from 1.0 ton/ha to 1.5 ton/ha, and soft wheat yields from 1.7 ton/ha to 2.3 ton/ha, less than anticipated at appraisal. The project successfully established 14,000 ha of tree crops, mostly olives, 6,000 ha more than foreseen in the SAR. 2.7 The project constructed 252 km of tertiary roads, 68 primary school classrooms, 22 dining halls, and 140 houses for teachers. The health component was only partially completed. Construction of the two drinking water systems was completed but both systems became operational only after project completion. 2.8 The PCR, issued in 1992, concluded that the modified extension system and the revised credit criteria introduced by the project had been beneficial. The integration of agriculture and social infrastructure components required a level of coordination among ministries and agencies which was not achieved. Design deficiencies slowed down implementation because the design of complex infrastructure was incomplete at approval. Six years was too short a period for a project of this type and level of complexity. A well designed and easy to use M&E system is an essential management tool, and especially important for a pilot project. The near absence of monitoring and evaluation (M&E) made it difficult to evaluate project performance. The ERR could not be reliably re-estimated, but an "indicative ERR" was re-estimated for the agricultural component to be 9 percent (21 percent at appraisal). Loukkos Rural Development Project (Ln. 1848-MOR) 2.9 Loan 1848-MOR was approved for US$34.0 million in May 1980. The overall objective of the project was to transform the traditional upland agriculture of the Loukkos watershed through soil conservation measures, and thereby raise agricultural production and incomes of poor farmers. The -5- project area included 31,000 ha of agricultural upland of moderate potential, 15,000 ha of steeper slopes, and 6,900 ha of forest. Project actions included forest plantations, erosion control works, crop and livestock development, roads and social infrastructure, and a cadastral survey component. The project was to be implemented over six years at a total cost of US$67.6 million. Implementation was to be coordinated by the MARA Directorate of Water and Forests. 2.10 Project implementation started slowly and experienced continuous financial difficulties. Slowness was due to the absence of necessary infrastructure for project staff in the project area, the need for time to secure the participation of the population in the project area, and the almost continuous shortages of local funds. The project was subjected to an MTR at the same time as Fes-Karia-Tissa (and Middle Atlas) in 1985, and benefitted from the SAP of 1987. "Non-essential components" (roads, education infrastructure and the cadastral survey) were cut back, and the project was reduced in scope to a core agriculture and forestry program. The project was completed in 1989, with an 18 month delay. Actual total project cost is unknown, but only US$14.4 of the Bank loan was disbursed and 58 percent was canceled. 2.11 Agricultural yields improved over the life of the project although less than anticipated at appraisal, except for fruit tree yields which equalled appraisal estimates. Significant changes in land use occurred. Most notably, the area under hard wheats declined, more than replaced by a six-fold expansion of soft wheats. Soft wheat production expanded greatly during the project period throughout Morocco, however, but the project accelerated the pace of change in the Loukkos area. The area under fruit trees (olives and figs) exceeded the appraisal estimate of 3,370 ha by 35 percent. Livestock development was below appraisal estimates. A sugar beet component was abandoned as ill-suited to the project area and uneconomic. 2.12 Appraisal targets for reforestation and soil conservation measures were achieved with a few exceptions. Some targets (fruit tree planting, forest plantation establishment, construction of check dams and stone walls, and forest roads) were exceeded. Many of the project's physical infrastructure targets were met. Achievement of appraisal estimates was 64 percent for road construction; 80 percent for irrigation rehabilitation; 100 percent for building and office construction; 74 percent for vehicle purchases; 36 percent for school classroom construction; 100 percent for construction of rural dispensaries; 133 percent for construction of village markets; and 92 percent for construction of village water points. Most of these measures were project inputs, however, implemented by contractors, not by farmer beneficiaries, as intended at appraisal. The cadastral survey component was canceled. 2.13 The PCR, issued in 1990, recognized that the primary objective of the project kept shifting between rainfed agriculture development, poverty alleviation, livestock development and erosion control. This contributed to a lack of focus that hampered implementation. The Directorate of Water and Forests was unable to provide the necessary methodological support for M&E, forestry research and erosion control. The Bank was criticized as slow to recognize the implications of the continuous fiscal crisis. The Ministry of Finance did not comply fully with the agreements reached under the SAP for minimal financing of the project. Financial sustainability was considered to be likely because some activities will be continued by a Bank-supported Research and Extension Project (Ln. 3036-MOR). The PCR stated that the project relied on a high level of government services which were maintained through a period of acute budgetary restraint by Bank intervention. The cost-effectiveness of soil conservation measures could not be determined because environmental monitoring had not taken place. Weak M&E performance made much PCR data unreliable, and the PCR admitted that it was impossible to re-estimate an ERR. An "indicative" re-estimated ERR of 15 percent was calculated, equal to the ERR at appraisal. -6- Middle Atlas Agriculture Development Project (Ln. 2082-MOR) 2.14 Middle Atlas was designed to increase production of forest, rangeland, and arable land in a relatively poor and mountainous area in Central Morocco. Coordination of implementation was to be by the Directorate of Livestock Development. Management plans would be prepared for 36,000 ha of forest land, and implemented on 124,000 ha. Management plans would also be applied to 24,000 ha of forest range, and 51,000 ha of collective rangeland. Cropping development would be addressed to an unspecified area of arable land, 9,600 ha of small irrigation perimeters would be rehabilitated, and 2,000 ha of irrigable land would be de-stoned. Roads and access would be improved, but no social infrastructure components would be included. The project would be implemented over five years at a total cost of US$92.2 million. A Bank loan was approved in January 1982 for US$29.0 million. 2.15 Project implementation was handicapped by budgetary constraints and frequent staff turnover. The project was subject to an MTR, and a SAP which scaled back the project, but these had little effect on implementation performance. Implementation took two and a half years longer than anticipated and Bank disbursements (US$9.0) lagged throughout the entire period. Sixty-nine percent of the Bank loan (US$20.0) was canceled because of counterpart funding shortages, the appreciation of the dollar against the dirham, and low use of technical assistance. The project experienced very high staff turnover. 2.16 The PCR, issued in June 1992, found that much of the project's impact on forestry development could not be substantiated because of weak M&E. Major forest production components (fuelwood and sawlogs) appeared to achieve between 39 percent and 63 percent of appraisal estimates. Range improvement components achieved between 23 percent and 64 percent of appraisal estimates. Collective range management suffered substantial delays because of lack of information concerning land rights, lack of appropriate technologies, and weak initial commitment of the Government and shepherds. Once implemented, however, this became the single most successful project activity. At Timahdite, a range management plan was prepared for 25,000 ha of collective range, and 7,500 ha were placed under a deferred management grazing plan. Under the plan, plant biomass production per ha was almost doubled. The animal health component reportedly cut annual sheep death rates in half. 2.17 The PCR indicated that twenty-eight pilot extension centers were established under the project, but significant improvements in extension service were not possible until late in the project because of frequent changes of staff and the lack of financing for vehicles and operating expenses. Credit use by farmers increased by 76 percent, but no information was available on the number of beneficiaries, the nature of the investments made, or the repayment performance. Neither of the two M&E units proposed under the project were established. Budgetary difficulties reduced technical assistance to 12 percent of the appraisal estimate. 2.18 Lack of M&E data made re-estimation of an ERR impossible. (The ERR of the Timahdite component, however, was estimated to be 37 percent.) The PCR admits that the project had, at best, a marginal impact on agricultural and forestry production. Some components of forest and collective range management were likely to be sustainable because the technology had proved effective, and beneficiary participation was high. Bank performance was criticized in the PCR because the SAP was implemented only in 1987, not in 1985 at the time of the MTR. Oulmes-Rommani Agricultural Development Project (Ln. 2217-MOR) 2.19 The objective of Oulmes-Rommani was to increase crop production and productivity in a rainfed area of good agricultural potential. This was to be achieved through the provision of improved -7- agricultural and social services, better integration of livestock and forest production, and introduction of soil conservation measures. The project was to be implemented by the MARA Directorate of Crop Production, over five years at a total cost of US$105.0 million. A loan of US$30.0 million was approved in December 1992. 2.20 Main project components included: crop development on 80,000 ha of private rainfed land in conjunction with the development of 12,000 ha of forest range; improved nutrition and genetic improvement of livestock; development of 570 ha of small irrigation perimeters and development of groundwater; establishment of 8,000 ha of forest production plantations and 4,000 ha of protection plantation; provision of 4,000 ha of soil erosion control measures; and management of 12,000 ha of sylvo-pasture. This would be achieved through expansion of the extension service, and provision of research, inputs, credit and technical assistance. In addition, major components were included for social infrastructure (schools, health centers, water supplies, and rural electrification), market place development, and forest and rural roads. A 320,000 ha tranche of cadastral survey would be undertaken. 2.21 As the fourth rainfed project in this group, Oulmes-Rommani ran into budgetary constraints almost immediately after approval. Initial delays in implementation persisted, and the project really began in earnest in 1987, five years after approval, having been restructured and reduced in scope. Restructuring entailed the elimination or reduction of most of the non-agricultural components, and the cancellation of US$8.0 million of the loan within two years of effectiveness. The project account was closed in December 1992, three years behind schedule, with US$0.8 million still undisbursed after total cancellations of US$16.5 million. 2.22 Despite financial difficulties, agricultural credit use and the uptake of seed and fertilizer was doubled during implementation. Wheat yields have been increased from about 1.6 ton/ha to 2.4 ton/ha, in drought-free years. Some of this increase, however, would have occurred without the project, but the rate of adoption of the new technologies which made this possible would have been slower. Incremental staff were assigned to training and extension beginning only in 1987, when a modified extension approach was applied. A three-year research program was initiated in 1986, and some monitoring of extension activities was initiated in 1987. An effort to evaluate performance began in 1991 with a farm survey, but the data have not been analyzed. Expansion of small irrigation perimeters was abandoned when it was found that water was insufficient. 2.23 Project activities had no appreciable impact on livestock development. Sylvo-pastoral and range improvement activities did not take place on a significant scale, and the distribution of improved animals was too small to make a difference. The forestry plantation component met its modest physical targets, but evaluation would require an inventory which has not taken place. Erosion control by planting of olive trees exceeded the appraisal target by 100 percent because it appeared to be profitable to farmers. Sylvo-pastoral development was limited to 5 percent of the appraisal estimate. Aerial photography for the cadastral survey was completed, but the analysis was canceled. Most of the technical assistance proposed was canceled (83 percent), as were half the studies. 2.24 Because of the virtual absence of data on project impact, it is not possible to re-estimate the ERR. The draft PCR does, however estimate an indicative ERR for the agricultural component of 13.9 percent, almost equal to the appraisal estimate of 14.0 percent. The indicative ERR, however, is based on generous assumptions about wheat yields, the spread of new technology to 100,000 ha of rainfed land, and the pace of yield increases attributable to the project. Since comparable yield increases have been achieved in other rainfed areas without the project, the indicative RERR must be interpreted with caution. No contribution to the indicative RERR was attributed to the forestry component because of weak performance and data deficiencies. -8- The Main Conclusions of the PCRs 2.25 The PCRs present a reliable picture of implementation and a number of common conclusions. First, the projects were all subject to continuing local funding shortages. Second, efforts to rectify budgetary problems were late and slow to be applied, waiting until an MTR in 1985 and a SAP in 1987. Third, project designs were overly complex, requiring a level of coordination that was beyond the capacity of the implementing agencies. Fourth, most project designs were incomplete at approval, causing early delays in implementation. Fifth, pilot projects require a higher than normal level of M&E, and M&E in these projects was deficient. III. THE ORIGINS OF BANK SUPPORT FOR RAINFED AGRICULTURE IN MOROCCO 3. 1 The performance of the projects under audit was severely constrained by financial and design problems. Most of the financial problems which arose were foreseen, but not recognized as serious. The problem of overly complex designs was built into the projects, and partially removed when they were restructured and scaled back. To understand the origins of these problems, and how they were responded to, the example of the Loukkos project is analyzed in detail. The Evolution of the Loukkos Project 3.2 The first reference in Bank files to the Loukkos project dates back to July 1972. While Fes-Karia-Tissa was the first to be approved, Loukkos was the first to be identified and begin preparation. It was through the Bank's dialogue with the Government on Loukkos that the overall rainfed strategy took shape. 3.3 The Bank's initial interest in Loukkos began with the realization that the development of the downstream irrigation potential of the Loukkos River had to be seen in a watershed context for economic reasons: irrigation development was proving so costly that further investment in the plains hinged on the profitability of agriculture in the rainfed area upstreamA1 As a result, the potential for projects in the upper watershed was investigated. This started formally with a reconnaissance mission in January 1974. The Bank had expressed an initial interest in a livestock or range management project, but after reconnaissance these ideas were replaced by a focus on small rainfed farmsL11 and improved watershed management. The Government, however, was at first reluctant to follow the Bank's lead. A spokesman for the Ministry of Cooperation was reported as "extremely negative (about rainfed agricultural development) ... practically any form of technical assistance was unwelcome in Morocco, with the exception of particular studies for large irrigation schemes.".1W 3.4 Throughout 1975, with the support of FAO, the Bank proposed alternatives for rainfed development which the Government resisted, saying that "... in view of the high rate of interest of Bank LO2 "In conclusion, it appears that the Loukkos (integrated irrigation/rainfed agriculture) project hinges on the profitability of its rainfed component. In view of the low return on investment in the irrigation component and the questions raised about its technical and organizational feasibility, it is recommended that consideration should be given to the implementation of (1) a smaller scale integrated irrigation/rainfed project based on groundwater exploitation; or (2) an integrated rainfed agriculture project." WB Office Memorandum, January 30, 1973. .11/ In 1975, 72% of farm households in the Loukkos watershed owned less than 29% of the cultivated area, in raided holdings of less than 2 ha. 12/ WB Office memorandum, October 31, 1974. -9- money, such projects did not warrant Bank lending but only cheap money or bilateral grants".31 Government enthusiasm for the project was muted at best. 3.5 During project identification, "The institutional vacuum that characterized the first week (of identification).... was followed by a limited participation of the regional and local technicians of the Ministry of Agriculture in the field work."1" By January 1976, FAO advised postponement of the project because little was known of the local reaction to such a project, technical information on erosion control and pasture development was insufficient, and "the Ministry of Agriculture (was) apparently reluctant to carry out rainfed rural development with World Bank financing".l But the project was not postponed. The Bank believed that its support would spur the Government to raise its budget for rainfed development, and began systematically to address the social and technical weaknesses FAO had identified. 3.6 Government resistance to a Bank-supported rainfed agriculture project in Loukkos waxed and waned, but preparation of Loukkos and Fes-Karia-Tissa was initiated following Government's acknowledgement of the need for watershed management "primarily ... to preserve and extend the useful life of dams that are being constructed..".' L61 Following this breakthrough in 1976, Bank efforts were focused on specific project design issues, in particular on the growing complexity of the projects. The Government had finally agreed to Bank support for rainfed projects, but it urged that their scope be greatly expanded by the addition of road, health and village water supply components.12 This was well beyond the Bank's original concept, transforming the projects into multi-sectoral area development projects. (Of the four projects approved, only Middle Atlas retained a single-sector focus.) 3.7 The Bank acknowledged the issue of over-complexity, but did not challenge it. In October 1977, it was clear to the Bank that "the number of ministries and departments concerned with the (Loukkos) project makes coordination difficult. A strong and continuous support from the Prime Minister's Office would be needed to overcome this..."A.8 The cost and financial justification of the project were also becoming problematical, with recognition that cost recovery should not be expected from farmers in the upper watershed because of their level of overty.191 The cost per family was "estimated at US$12,000, for marginal agricultural increases".20P Bank operational staff recommended application of an ERR lower than the commonly accepted threshold ERR,211 and management agreed that individual sub-components might have ERRs equal to or lower than the threshold, as long as the ERR of the entire project was 10 percent or greater.221 13 / WB Office memorandum, June 18, 1975. 14/ WB Office memorandum, November 19. 1975. 5 / WB Office memorandum, January 22, 1976. 16/ Back-to-Office Report, August 3, 1976. 7/ WB Office memorandum, May 13, 1977. L8/ Project Brief, November 24, 1977. 9/ WB Office memorandum, March 31, 1978. 20/ WB Office memorandum, May 26,1978. U1/ WB Office memorandum, May 25, 1978. 22" WB Office memorandum, June 7, 1978. - 10 - 3.8 In July 1978, Bank staff realized that the proposed road component, estimated to cost US$9.5 million, had not been subjected to engineering design or economic appraisal. But it was accepted by Bank management as an "indisputable part" of the project; detailed engineering studies could start after appraisal. Knowing Morocco was about to embark on a three-year fiscal austerity plan (1978-80), the Bank also acknowledged that with "no economic analysis or traffic forecast", upgrading of tertiary roads "in the present financial climate in Morocco... would probably have low priority for available funds and is unlikely to be undertaken by the RD (Roads Department)".31 "The main issue now is the Government's three-year plan which aims at reducing drastically new investment expenditures. (Current) erosion control works in the Upper Loukkos have already been halted because of the present financial situation. Although the project is now better justified economically, it is still doubtful whether the Government will include it in the three-year plan".41 In November 1978, the Bank learned "from discussions with both the Prime Minister's Office and the Ministry of Agriculture it appears that the project has been partly included in the three-year austerity plan, now under review by Parliament".l Five months later, the Prime Minister's Office in a memo to the Ministry of Agriculture indicated that "la rentabilite economique du Loukkos ne semble pas justifier le projet".26 During appraisal of Loukkos (and just prior to approval of Fes-Karia-Tissa), Government support for rainfed projects was less than complete. 3.9 In the final stages of appraisal, key questions about the Loukkos project's economic rationale remained unanswered. In particular, could a 60 percent level of subsidy for pilot activities in rainfed agriculture and an outlay of about US$1,000 per hectare be justified? Was expenditure at this level likely to be replicable?. The Bank concluded that the answer to both of these questions was "yes". Meanwhile, the Borrower was more concerned about the terms of the loan. Could the loan period be extended? The Bank agreed that it could, from 17 to 22 years. Could US$9.3 million of the proposed Bank loan be used to cover local currency costs? If it could not, the "austerity program and combination of a scarcity of foreign exchange and local financial resources would increase the pressures on the Government to postpone execution of this project until the situation eases". 2-7 The Bank agreed that US$3.6 million of the loan could be used to cover local currency costs, justified on the grounds of social impacts and the existence of the austerity program. The Bank now saw this project as having a strong poverty focus. 3.10 The Loukkos project loan was finally approved in May 1980, almost eight years after the first correspondence appeared in the Bank's project file, and more than six years after the reconnaissance mission. The summary of the discussion of the project by the Bank's Board indicates overwhelming support for rainfed rural development in Morocco, but reveals two important concerns': a) could Morocco provide the necessary counterpart funds for Loukkos? Implementation of several Bank projects had already been delayed for lack of counterpart funding; and b) despite lengthy project preparation, had 3/ Office memorandum, July 12, 1978. By the following year, road cost estimates were still vague but the consultants employed to undertake the economic analysis assumed that rural roads to provide basic transportation would be "an absolute necessity": Office Memorandum, July 13, 1979. 24 / WB Office memorandum, July 17, 1978. 25 / Project Brief No.5, November 13, 1978. L6/ "The economic rate of return of Loukkos does not seem to justify the project". Undated memorandum received by the Bank on April 2,1979. L7' Office memorandum, February 28, 1980. 2 Summary of Executive Director's Meeting, June 27, 1980. - 11 - Bank staff identified the optimal model for Morocco's rainfed rural areas? If the project was to be replicable, it had to be deliverable at minimum cost, in terms of time and money. 3.11 Staff assured the Board that the Government had pledged that the necessary counterpart funds would be forthcoming. Staff also said that the long gestation of this project had been helpful because the characteristics of the project area were now well understood by the Bank. While project activities were expensive, it was likely that the unit costs of replication would be lower. The project would provide the opportunity to discover lower cost approaches, and the M&E unit to be set up in MARA would find ways to reduce costs in future projects. Experience showed, however, that operating funds for the Provincial Directorates of Agriculture (DPAs) in the Loukkos project area were already being received six months late in 1980, six months before the project was approved; cost estimates, which had been criticized for appearing high, were acknowledged by the Borrower to have been seriously underestimated;Lg/ and Bank supervision reports show that the Ministry of Finance began to restrict the release of project funds in January 1982, only six months after loan effectiveness. The Austerity Program, 1978-80 3.12 The origin of the Government austerity program can be traced to financial disequilibria experienced prior to 1977.1/ Between 1968 and 1977, Morocco went through a period of significant social and economic development, but the growth in savings and exports did not keep pace with government expenditures, investments and imports. Fluctuations in the prices of Morocco's principal imports and exports, the collapse of the early-1970s phosphate price boom, the cost of modernization of the armed forces, and a major defence build-up added to the financial disequilibria. The balance of trade registered a US$546 million deficit in 1975. Morocco drew on IMF facilities for SDR 115 million in 1976 and began a determined effort to re-establish a new equilibrium, with cutbacks in investment, imports and employment, incorporated in the Three-Year Austerity Plan of 1978-80. Agriculture was not exempt from the Plan. "The Three-Year Plan reflected this austerity for agriculture, with many projects either delayed or cut back in size. This was particularly true for new projects in non-irrigated areas. Therefore, any increase in Government investment in agriculture compared to that budgeted in the Plan was difficult to implement at that time, despite the evident desirability of such an increase for that sector. The Plan emphasized the completion of ongoing projects for which considerable investment had already been made. Since most ongoing public investment in agriculture had been directed to equipping the irrigation zones, the Plan emphasized the completion of such investments (57 percent of planned investment in agriculture was foreseen for irrigation areas).... Only DH 50 million of the investment foreseen in the Three-Year Plan was allocated... to rainfed project(s) (1. 7 percent of Government investment in agriculture). The Government of Morocco was, therefore, implementing a budget austerity plan which known by the Bank not to favor of new initiatives in rainfed areas before the four projects under audit were approved. 9 I Bank Supervision Report of July 14, 1981. 30/ See Morocco: Economic and Social Development Report, Christian Merat, Coordinating Author, The World Bank, 1981, p 59, and A Multisector Framework for Analysis of Stabilization and Structural Adjustment Policies: The Case of Morocco by Abel Mateus and others, World Bank Discussion Paper No. 29, 1988. Ll/ See Morocco: Economic and Social Development Report, 1981, pp 161-162: a WB report published shortly after approval of the Loukkos project and before the approval of Middle Atlas and Oulmes-Rommani. - 12 - Only DH 50 million was allocated under the austerity budget to all rainfed agriculture for 1978-80, when Fes-Karia-Tissa alone required DH 163 million for 1979 and 1980. The second year of austerity was underway when the Fes-Karia-Tissa loan became effective, and the Government had been imposing budget austerity for more than five years when Oulmes-Rommani was approved. 3.13 Some stabilization policies were put in place in the early 1980s,2-2 but social unrest followed efforts to increase the prices of subsidized commodities in 1981. Bank proposals for a SAL operation in 1981 were not accepted by the Government, although three IMF Standbys were later used to support the Government's stabilization program between 1983 and 1988, prepared in close cooperation with the Bank. 3.14 Four conclusions can be drawn from this extended background: (i) the long-established and principal priority for agricultural development in Morocco has been lowland irrigation; (ii) with support from FAO, the Bank persuaded the Government to adopt a rainfed agricultural strategy, but Government commitment was less than complete; (iii) the four rainfed projects under audit were approved after Morocco's entry into an extended period of fiscal austerity, which was known to the Bank; and, (iv) both factors combined to curtail activity in each project through a lack of counterpart funds, from very shortly after their approval. Budgetary problems were foreseen by Bank staff who reviewed the project in the final stages of preparation, and during Board discussions, but they were seen by project staff to be surmountable. In the early 1980s, the second oil price shock, a debt crisis and four years of drought, compounded Morocco's financial difficulties, deepening the need for austerity in the rainfed projects. IV. IMPLEMENTATION EXPERIENCE The Effect of Budgetary Shortfalls 4.1 The most striking aspects of implementation of these projects was the shortage of counterpart funds. This led to both a slow pace of implementation and a need to scale-back the scope of the projects. Loans totalling US$158.0 million were approved, US$72.4 million was disbursed (45.8 percent), and US$85.6 canceled (54.2 percent). After two or three extensions for each project, the loan cancellations ranged from 44.3 percent in Fes-Karia-Tissa to 69.0 percent in Middle Atlas. (See Table 4.1.) Table 4.1: Loans, Disbursements and Cancellations (US$ millions) Fes-Karia-Tissa Loukkos Middle Atlas Oulmes- Total Rommani Approved 65.0 34.0 29.0 30.0 158.0 Disbursed 36.3 14.4 9.0 12.7 72.4 Canceled 28.7 19.6 20.0 17.3 85.6 Cancellations as % of approvals 44.3 57.7 69.0 57.7 54.2 2/ Jean-Jacques Dethier and Lynn Salinger, Agricultural Sector Adjustment in Morocco, 1989. A paper presented at The World Bank Seminar on Policy-based Lending in Agriculture, Baltimore, May 17 - 19, 1989. - 13 - 4.2 Counterpart funds became scarce for each project in turn, shortly after loan effectiveness. Fes-Karia-Tissa encountered counterpart funding problems 18 months after loan effectiveness, Loukkos after only four months, and Middle Atlas after eight months. Oulmes-Rommani recorded financial difficulties one month after loan approval. A formal admission by Bank Operations that financial difficulties would affect Bank-supported projects came only after the 1983 debt crisis.3' This coincided with the approval of the fourth project loan.The scarcity of counterpart funds dwarfed all other aspects of project implementation, and much of the implementation experience became a process of strategic and tactical adjustment to the problems this caused. Figure 4.1 summarizes the timing of major project activities, the external factors which influenced progress, and the main steps taken to adjust to them. 4.3 Lack of operating funds, and lack of timely funding, lowered the effectiveness of field personnel and the pace at which fieldwork was implemented. Funding problems worsened in step with Morocco's overall budgetary crisis. The phrase "(i)n spite of the budget approval of the Ministry of Finance, it is likel that only a part of it will be available to the project", was repeated in Bank supervision reports. Funds required in January became available only in June, and the 1983 budget for Oulmes-Rommani was released only in October of that same year. The Government imposed systematic delays in processing contracts.-5 Coupons needed by MARA staff for gasoline were obtainable only late in the year. In 1987, 65 percent of the annual allocation of gas tickets for extension agents for Middle Atlas became available only in November. 4.4 Each directorate of MARA tried to cope with financial difficulties as best it could, including seeking the support of the Bank in the search for releases from the Ministry of Finance. The Bank successfully lobbied the Ministry of Finance to release funds on behalf of Fes-Karia-Tissa as early as July 1980. In the absence of a public expenditure review, this lobbying should be questioned since it singled out one Bank-supported project for special treatment. 4.5 In the face of progressive financial difficulties, the level of achievement of physical targets was surprisingly high, assisted by the appreciation of the US dollar against the dirham. In their reduced forms, the projects achieved or approached many of their targets for infrastructure and civil works, and the acquisition of vehicles and equipment. Targets for field-based activities which directly supported farming systems development or natural resource conservation were, in contrast, less well achieved. There were three related causes for this. First, infrastructure and vehicles were needed to create a MARA presence in several project areas where it previously did not exist, for example in Loukkos and Middle Atlas, where the DPAs were newly established, or where the presence was initially weak, as in Fes. This was consistent with Government policy at the time. Second, in the late-1970s and early-1980s, MARA operated along fairly traditional lines: the direction of agricultural development came from the top and was implemented through an hierarchical structure that required a distinct physical presence before it could deliver its services. MARA was an agency accustomed to initiating action and carrying it through. Throughout the life of the projects, and in large part because of them, this approach was challenged and modified. Third, the fact that these were pilot projects meant that implementors did not have a broad array of established technical packages waiting "on the shelf" to be applied. Time was needed to test the existing menu of technical packages and expand it, but vehicles could be purchased quickly. L WB Office Memorandum, May 1985. I / Supervision report, Oulmes-Rommani, November 21, 1984. L5 / Ibid. Figure 4.1: Chronology of Significant Events in Morocco and in the Projects under Audit MOROCCO FES-KARIA-TISSA IOUKKOS MIDDLE ATLAS OULMES-ROMMANI 1972 - First oil shock phosphate boom - Origin of project in Bank files 1974 - Drought - Reconnaissance 1975 - Phosphate price collapse: BOP - Preparation I, 10/75 deficit US$546 million 1976 - IMF SDR utilized - Preparation, 11/76 1977 Appraisal, 1 I/M 1978 - Three-year austerity budget Approval, 6/78 - El Makhazine Dam completed introduced - Preparation Hf, 7/78 1979 Effective, 2/79 - Appraisal, 1/79 - Preparation, 9/79 1980 Counterpart funding - Approval, 5/80 - Preparation, 1/80 problem evident, 7/80 1981 - Drought - Effective 9/81 - Appraisal, 10/81 1982 - Drought - Counterpart funding problem - Approval, 1/82 - Appraisal, 3/82 evident, 1/82 - Effective, 9/82 1983 - Drought - Major financial difficulties, - Major financial - Approved, 5/83 - Debt and forex crisis 6/83 difficulties, 5/83 - Major financial difficulties, 6/83 1984 - Government staff - Loan cancellation 11/84 - Recognized need to - Effective, 3/84 - Recruitment frozen reduce objectives, 4/84 1985 - Drought - Loan cancellation, 4/85 - MTR - MTR - ASAL 1, 6/85 - MTR - Loan cancellation, 5/85 1986 Planned closing, 6/86 - 'A shadow budget - SAP exists?", 12/86 - Loan cancellation, 10/86 - SAP 1987 - SAP - SAP - Loan cancellation, 11/87 - Loan cancellation, 11/87 - ASAL II, 11/87 - Loan cancellation 4/87 - Planned closing, 6/87 1988 - Loan cancellation, 11/88 - Planned closing, 6/88 - Actual closing, 12/88 1989 - Loan cancellation, 9/89 - Planned closing, 12/89 1990 - Loan cancellation, 6/90 - Loan cancellation, 12/90 - Actual closing, 6/90 - Actual closing, 12/90 1991 - Drought - laan cancellation, 12/92 - Actual closing, 12/92 1992 - Drought - 15 - 4.6 As budgetary problems deepened, the likelihood of the projects achieving their physical SAR targets declined. The projects which came on-stream later suffered the most. In Fes-Karia-Tissa, achievement in field-based activities of direct benefit to farm households or soil and water conservation (livestock improvement, hillside farming and wasteland improvement) exceeded 50 percent of the SAR estimates in 9 of 14 categories of activities; in Loukkos, achievements (pasture improvement, reforestation, gully control, farm mechanization, and improved seed and fertilizer utilization) exceeded 50 percent of SAR estimates in 14 of 19 categories; and, in Middle Atlas, achievements (forest production improvement, insect control, forest range improvement and livestock development) exceeded 50 percent of SAR estimates in 7 of 15 categories. For Oulmes-Rommani, however, achievements (sylvo-pastures, forestry, on-farm developments, and flood control) exceeded 50 percent of SAR estimates for only 3 out of 19 categories of activity. 4.7 Because of the depreciation of the Moroccan dirham (DH) against the US dollar,& dollar costs of most activities were lower than expected at appraisal. An exception to this was the higher cost of civil works for village drinking water systems in Fes-Karia-Tissa, attributable to incomplete design before approval and protracted implementation. The Effect of Inadequate Technology 4.8 Adequate and well-understood technology was generally limited throughout the early and middle stages of the projects. In several instances, technologies promoted at the outset of implementation proved to be expensive and inappropriate. Examples included the development of sugar beets in the Middle Loukkos, and the replacement of moderately degraded oak forest in Middle Atlas with contour planting of cactus or acacia. The sugar beet component appears to have been no more than a theoretical possibility, and was soon abandoned for both technical and economic reasons. Forest and sylvo-pastoral improvement only began to take-off when high-cost measures, such as replanting with species (such as acacia) which were not maintenance-free, were replaced by low-cost alternatives, which were. Technologies proposed during project preparation were gradually abandoned or modified in favor of cost- effective alternatives. The extended periods of implementation, however, gave time for several successful adaptations to be learned and applied. 4.9 Olive and fruit tree planting eventually replaced hillside cereal cultivation in large areas of Fes-Karia-Tissa, Loukkos, and Oulmes-Rommani. The evidence of recent expansion was apparent to the audit mission. Fodder and pulse production expanded in Oulmes-Rommani. The rate of expansion of soft-wheat production was accelerated in Fes-Karia-Tissa, Loukkos and Oulmes-Rommani. In all cases, the technologies which took off were relatively simple, well-adapted to the upland environment, generated revenues relatively quickly, and allowed for subsistence cereal production to continue until less- erosive perennial crops started bearing fruit. It is difficult, however, to isolate fully the effect of the projects on these developments because the M&E components of all projects was weak. Expansion of olive tree planting and soft wheat production also took place outside the project area, with the support of successful national programs. 4.10 Environmental benefits stemming from the projects are also impossible to evaluate, although it is clear from field observation that in many cases they are positive. This is disappointing because the origins of these projects can be traced in part to concern for irrigation catchment management, and because proposals to monitor environmental impacts were, somewhat progressively for the time, to be built into projects from the outset. But the monitoring proposals were not followed through. 36/ The dirham depreciated from 3.8 to the US dollar in 1978, to 8.3 to the US dollar in 1990. - 16 - 4.11 Most of the livestock technologies introduced under the projects were of questionable value. With the very significant exception of collective range management activities in Middle Atlas, the Government remained committed to free drug delivery and vaccination services to livestock owners,7' lacked appropriate animal nutrition recommendations, and employed methods of genetic improvement which had little broadly-demonstrable effect. Narrow concerns for animal health and genetic improvement overshadowed a broader concern for improved animal husbandry. Public insemination centers, shown earlier to have been failures in Algeria and Tunisia, were no more successful in Morocco.38 The Effects of Incomplete Preparation 4.12 At the design stage, Bank staff input into each of these projects was substantial: between 121 and 325 staff weeks for each project, through loan approval. Detailed design of some infrastructure components, however, was still incomplete at the time of project approval. This contributed to early delays in implementation while design work was completed, particularly for roads and drinking water systems. Once detailed designs were completed, however, infrastructure was put in place relatively quickly whenever funds were available. The exceptions to this were the drinking water systems in Fes- Karia-Tissa. These were subject to lengthy delays in construction and commissioning. Only since the PCR has been completed have the projects become operational, 14 years after project approval. 4.13 Incomplete preparation adversely affected project costs. Two months before the SAR of Fes-Karia-Tissa was finalized, road cost estimates jumped from US$50,000 per kilometer to US$140,000 without supporting justification, and five years after appraisal, water and road system costs were twice the appraisal estimates.39/ The Effect of Complex Institutional Arrangements 4.14 Institutional arrangements for implementation were complex. Each projects had three layers of coordination: central, provincial and local; and eight separate ministries or agencies typically sat on each central or provincial committee. At the local level, the projects were implemented through MARA's Work Centers (CTs), other existing local bodies, or through local ad hoc commissions or cooperatives formed for the purpose. The identities of the principal players are set out by project in Figure 4.2. 4.15 Complex institutional arrangements might have handicapped project implementation more than they did, but the relationship at the provincial level between the implementing DPAs and their respective Provincial Governors was effective. Senior level participation was limited in some cases. Project components were assigned by provincial coordinating committees to each responsible ministry or agency, which implemented the tasks. Components were implemented individually and there was little if any synergy between components. But funding issues overrode institutional issues: components were implemented as and when funds were available. Infrastructure components tended to be completed first. L2/ Oulmes-Rommani Supervision Report, April 1989. 3&/ Ibid. L/ Supervision report, March 7, 1983. Figure 4.2: Organizational Arrangements for Implementation: Membership of National, Provincial and Local Coordinating Mechanisms Fes-Karia-Tissa Loukkos Middle Atlas Oulmes-Rommani CENTRAL PROJECT CoMMrrrEE CENTRAL COORDINATING COMMFrEE NATIONAL COORDINATING NATIONAL PROJECT COORDINATING COMMITrEE COMMITTEE MARA - OMV (Chair) MARA - DEF (Chair) Min. of Public Works Min. of Public Works MARA - DE (Chair) MARA - DPV (Chair) Min. of Public Health Min. of Public Health CNCA Min. of Public Works Min. of Education Min. of Education INRA Min. of Education CNCA CNCA Min. of Interior SONACOS Min. of Finance FERTIMA CNCA Nat'l Cadastral Service SONACOS FERTIMA H. PROVINCIAL PROJECT INTERPROVINCIAL PROJECT INTERPROVINCIAL PROJECT PROVINCIAL PROJECT COORDINATING COORDINATING COMMrITEE COORDINATING COMMITTEE COORDINATING COMMITTEE COMMITTEE Provincial Governor (Chair) Three Provincial Governors (Chair) Two Provincial Governors (Chair) Provincial Governor (Chair) DPA (Project Director) DPA (Project Director) Project Management Units (2) DPA Min. of Public Works Min. of Public Works DPA Min. of Public Works Min. of Public Health Min. of Public Health CNCA Min. of Education Min. of Education Min. of Education SONACOS CNCA CNCA CNCA FERTIMA SONACOS SONACOS FERTIMA FERTIMA III. LOCAL IMPLEMENTATION LOCAL IMPLEMENTATION LOCAL IMPLEMENTATION LOCAL IMPLEMENTATION CTs (8) CTs (2) Local management units (5) CTs Forest Stations (3) Range improvement commissions Administrative Councils Technical Committees - 18 - 4.16 When it was admitted in 1983 that Morocco's financial crisis was "real"A1 each project's coordinating committee decided which components had priority and which were not essential. Principles or criteria employed to guide these choices are unclear, apart from defining the "core" of the projects in terms of their agricultural priorities, which corresponded to the priorities of the respective directorates of MARA which were coordinating implementation. For example, by June 1984, less than one year after the Oulmes-Rommani loan became effective, the irrigation, research, rural roads, social infrastructure, education and health components were defined as "non-essential" and slated for cancellation. Budgetary pressure forced the simplification of project designs and arrangements for implementation. 4.17 Within MARA, however, the staff engaged in the field learned to work together at the provincial level, in most cases for the first time. Through the effort to implement an extension system, based on a mix of traditional methods and a modified extension system, a new capacity was established for collaborative work among front-line field staff and back-stopping specialists. This capacity continues to function in the project areas after project completion. Some "old-style", mass extension approaches were retained because they afforded the DPAs a measure of control over the extension agenda, which they were reluctant to give up. But many of the barriers which had existed between directorates were removed as DPA staff were required to collaborate in order to respond to farmers' needs and to support and train field staff. 4.18 Important premises of all four projects were that extension staff had to be relieved of their administrative burdens, to work directly with their clients, and become accountable for development impacts. Under the projects, extension agents learned to work directly and responsively with farmers on production and marketing issues, throughout the farming year. Extension agents admit now that they did not know what to do when extension started. Prior experience had not prepared them to work responsively in farmers' fields, and they began with limited relevant knowledge or confidence. Knowledge and confidence have clearly grown, and a cadre of agents able to work effectively with farmers has been put in place under the projects. The Effects of Weak Monitoring and Evaluation 4.19 Monitoring and evaluation (M&E) was designed to take place both at the central level in Rabat, and in each project in the DPA. Despite detailed plans at appraisal, the Borrower's alleged commitment to M&E, and the technical competence of MARA staff, the M&E experience was disappointing in all projects. The PCRs suggest that M&E was virtually non-existent. The audit mission found that all DPAs carried out some M&E activities, but focused mainly on counting inputs, rather than outputs or effects. The data collection that did take place, however, was processed by hand and not computerized, making it less accessible to management and Bank supervision missions in timely ways. Failure to provide computers under the projects was a design flaw which exacerbated poor record keeping, and limited management's use of data which was collected. 4.20 In Fes-Karia-Tissa, sample surveys of project households were undertaken in 1981 and 1983 as proposed in the SAR, but the data collected still remained unprocessed in 1988, as the project approached closure. The survey originally proposed for 1985 was not carried out. In 1984, the Bank expressed frustration at the limited progress on M&E in all projects, but criticism had little impact. Early supervision reports dwell on M&E deficiencies and, following the mid-term review, overseas training in M&E was provided for staff of all four projects, followed by practical training in Morocco. This was, however, too late to make a difference, and the weak commitment to. M&E was still apparent at audit. E WB Office memorandum, May 5, 1983. - 19 - 4.21 Environmental monitoring was specifically included in the design of the Loukkos project, but not implemented. Much of the original justification for Loukkos stemmed from concern for soil conservation above important dams, such as the El Makhazine Dam. The failure to implement environmental monitoring meant that the effectiveness of the project to conserve soil and prolong the life of impoundments remains undetermined. Particular technologies, especially reforestation and the replacement of cereal cultivation on sloping lands with olive tree plantations, undoubtedly reduced soil erosion, but it is unclear by how much or at what cost. Supervision 4.22 Bank staff time devoted to supervision ranged between 70 and 140 weeks per project, declining in the order the projects were approved: highest for Fes-Karia-Tissa and lowest for Oulmes- Rommani. Staff time devoted to supervision was significantly higher for the first half of the implementation period of each project than for the second. After 1986, time devoted to supervision declined sharply. Seventy-three percent of all project supervision was completed by 1986. (See Figure 4.3.) Between 1982 and 1986, an average of 57 staff weeks was devoted to supervision annually. For the second half of the implementation period, this average fell to 22 weeks. Following the MTR, the Bank took a formal decision to reduce normal supervision for these projects in order to focus remaining staff resources on particular impediments to implementation. It is unclear how effective this strategy was. But in the final years, supervision in Fes-Karia-Tissa was devoted almost exclusively to completing the drinking water projects, contributing to their completion. Figure 4.3: Project Supervision Time (Staff Weeks) a0 70 - 60 - 50 - 40 30 20 10 0 1992 1993 1994 '1995 1988 1997 1999 '1989 1990 1991 '1992 1993 4.23 The overall frequency of supervision missions was adequate, but a single supervision mission was mounted in 19 of 53 separate project years. With the exception of the MTR, supervision missions were typically composed of one or two members, who were usually technical specialists in agriculture, livestock and forestry. The dominant issues identified in supervision reports, however, were - 20 - financial, institutional, or related to M&E. A financial analyst was a member of only one mission (to Loukkos in 1980); no institutional specialists were included; an M&E specialist was included in only one mission (to Fes-Karia-Tissa in 1982); and a highway engineer was a member of only one mission (Loukkos in 1981). No specialists in health or education were used. The Mid-Term Review and the Special Action Program 4.24 It was not until 1985 that the Bank decided to address directly the implementation problems caused by counterpart funding shortages, with a mid-term review (MTR).4-1 Despite its share of difficulties, Oulmes-Rommani was excluded form the MTR because it was considered to be too new. The MTR reported good progress with infrastructure development and more limited progress with agricultural developments and M&E. It focused on chronic financial difficulties and the poor outlook for financial sustainability. The MTR forced the Bank to identify and apply criteria to extend the life of the projects, identify inevitable cancellations, and reach agreement on guidelines to avoid future commitments to low priority project components. 4.25 The MTR concluded that the financial situation prevailing in Morocco at the time would preclude additional rainfed projects in the foreseeable future. By the time of the MTR, some funds for Fes-Karia-Tissa had already been canceled and plans to reduce the scope of Loukkos were only four months away. 4.26 For Middle Atlas, the MTR concluded that the project suffered from a notable lack of support by the Borrower. Middle Atlas had failed to have an approved budget for most of 1984, permission to spend project funds had been systematically blocked, and consulting contracts were not being approved. The same problems had affected all four projects at different times, but Middle Atlas was the most seriously affected at the time. 4.27 Discussion of a possible Special Action Program (SAP) to help cope systematically with Morocco's budgetary crisis and expedite the flow of funds to the projects, first took place in November 1984, before the MTR. Nineteen-eighty four was Morocco's fourth drought year in a row, and the year following formal recognition by Bank Operations staff of major financial difficulties in the projects. The SAP was, however, not agreed until April 1987, 29 months after the initial discussions took place. No explanation has been provided for this delay. 4.28 Fifteen Bank projects in Morocco ultimately benefitted from the SAP, including the four under audit. Portions of loans were canceled, special project funds were established or enlarged, and percentages of expenditures to be financed from loans were substantially increased, retroactively. Under the SAP, large cancellations were made in the loans for Fes-Karia-Tissa, Loukkos, and Middle Atlas. Special accounts of US$2.0 million were established for each project in local currency to be drawn from directly, and disbursement percentages were raised for several categories of project expenditure from 60- 65 percent to 80-85 percent.421 4.29 The effects of the SAP were limited. While it is not possible to say how much less would have been disbursed in the absence of the SAP, payment delays persisted in all the projects. Projects continued to be deprived of operating funds, and the rate at which disbursed funds were claimed from Oulmes-Rommani was deemed too recent to be included in the MTR, although it was also intended to benefit from the Special Action Program (SAP) which followed. 42/ These percentages were subsequently increased further for Oulmes-Rommani, to 85-100%, under a Gulf Crisis Assistance Program in July 1991. -21 - the Bank continued to be slow. In both Middle Atlas and Oulmes-Rommani, the SAP had little effect. In Middle Atlas, "all the measures taken in 1987 to augment and accelerate project finances had no effect",!' and in Oulmes-Rommani "payments did not increase significantly (+6 percent) in spite of measures taken to improve the financial situation".A" The DPA in Oulmes-Rommani did not understand the operation of the new systems put in place to accelerate funding. The result was that only US$220,500 of the US$2.0 million special account was drawn upon in the 14 months following its establishment. The Ministry of Finance had not informed MARA or the DPAs that special accounts were available for their use, nor how to access them. V. PROJECT OUTCOME Farm Impacts: Cereals and other Crops 5.1 The overall effect of these projects on the productivity and profitability of rainfed farms was limited. Drought conditions, which prevailed in 6 out of 14 implementation years, reduced the impact of the projects, but drought has to be expected in the Mediterranean zone. The projects resulted, however, in several positive changes to rainfed farming systems. Most notably in Loukkos and Fes- Karia-Tissa, cereal production on sloping ground has been progressively replaced by tree crops and forages which conserve soils. 5.2 Cereal production shows improvements in yields, but these are generally lower than anticipated at appraisal. Yield increases are mainly attributable to use of improved seed, increased use of fertilizer, better land preparation, and weed control: all promoted by the projects. Shifts within cereal production have also taken place, from traditional hard wheat to new soft wheat varieties, with significant increases in yield in those years when rainfall was adequate and timely.Ail While the projects accelerated the pace of the shift to soft wheat production in the project areas, similar developments took place elsewhere in Morocco under the influence of national programs and new pricing policies. 5.3 Overall, hard wheat yields increased from about 1.0 ton/ha to about 1.4 tons/ha, and yields of soft wheats increased from about 1.5 tons/ha to over 2.0 tons/ha, in good rainfall years. Within each project area, yields also varied considerably by location. On steep hillsides, where cereal growing was being replaced by perennial tree crops, soft wheat still commonly yielded only 0.4-0.5 tons/ha. Farm Impacts: Livestock Production 5.4 Production of forage crops increased somewhat. Trials of annual Medicago and Australian 'ey farming systems were not successful. The techniques which succeeded best were those which required the lowest investments and the least change in farming systems. 5.5 Impacts on livestock system performance were well below appraisal estimates. Collective range management in Middle Atlas, however, proved highly successful at Timahdite and partly successful L3 Supervision report aide memoire, May 2 - 5, 1989. !H/ Supervision report, June 6, 1989. L5 The most progressive farmers in Oulmes-Rommani now achieve soft wheat yields in excess of 5 tons/ha, but this is not typical of rainfed Morocco, where overall performance is erratic. This is clearly demonstrated in Figures 4 and 5 of the Oulmes-Rommani PCR. - 22 - at Ain Leuh, after a slow start. At Timahdite, a cooperative of shepherds and flock owners was successfully formed around the management of water systems which serve both the shepherds' communities and their pastures. The cooperative charges fees per head of livestock for access to the improved pastures on an annual basis. Rules to avoid overgrazing are established and enforced by the cooperative's members, who also undertake routine maintenance of the water system (although the fountains visited at roadside by the audit mission were in need of repair). Grazing systems at Timahdite utilize both collective pastures and forest grazing lands, which extend the grazing resource and contribute to its durability. An effort was made under the project to improve the Timahdite sheep breed, through the national breed improvement association. Some of the improvements sought are in conformation and appearance, however, unrelated to economic performance. Project activity at Timahdite started slowly, caused by the lack of appropriate technical packages for range improvement and the difficulty of organizing shepherd tribes and clans, which had been underestimated. 5.6 The productivity of the collective pasture at Timahdite is affected by the erratic rainfall. At the time of the audit mission, the pastures were suffering their second successive year of drought, but reportedly were capable of recovering if winter precipitation in 1993/94 was adequate. More prolonged drought could, however, significantly affect pasture productivity, putting stress on cooperative members to break the grazing rules which have heretofore safeguarded the resilience of the improved pastures. It appears that the cohesiveness of the cooperative depends most upon the maintenance of the water system which serves shepherds' homes and families, and this is likely to be assured, since it is so highly valued by all the parties involved. 5.7 Since Timahdite's shepherds also manage large numbers of sheep and goats for outsiders, much of the benefit accruing from pasture improvement is acquired by people who are absentee herd and flock owners. Many of these were said to be urban residents from Casablanca, resulting in the transfer of some project benefits to absentee flock owners. 5.8 Overall, however, the Timahdite range improvement component stands out as a significant success in an otherwise bleak performance. Livestock system improvements in other project areas show little effect. Project efforts to improve the genetic potential of farm animals, such as in Loukkos, have had little effect. The existing herds and flocks are large and of poor potential, and the injection of modest amounts of new genetic potential will have little benefit. Progressive and better-off farmers have for generations cross-bred their animals to improve their quality, but the overall impact of these actions in the project areas is not widespread. These actions continue, however, under Government livestock programs managed mainly by veterinarians, rather than animal husbandry specialists. Emphasis continues to be placed on genetic improvement though cross-breeding and health care provision. Infrastructural Outcomes 5.9 All projects included significant expenditures on construction of offices, stores and staff houses. The initial quality of construction appears to have been high, but maintenance has been limited, the result of continuing budgetary difficulties. Some residences built by the projects, such as at Moulay Bouazza in Middle Atlas, still lack water connections several years after construction because of inappropriate siting. Other buildings visited by the audit mission appear to be less than fully utilized. The justification for much of this construction stemmed both from the need to establish a government presence in the under-serviced project areas, and MARA's desire for a substantial physical presence, as a prerequisite to the provision of services to farmers. But this approach has heavy capital and maintenance cost implications, which have not been met during a period of budgetary restraint. - 23 - 5.10 Construction and improvement of roads and tracks were significant activities in all four projects. In all projects, appraisal targets for road and track construction were reduced, but the revised construction programs were still substantial. These appear to have had a significant effect on the accessibility of rural people to markets and services, and on MARA's access to its clients, in Fes-Karia- Tissa and Loukkos. Many planned additions to rural education and health infrastructure were also cut back: the earlier these components were started, the more likely they were to be built. Virtually all of the proposed expenditure on education and health infrastructure in Oulmes-Rommani was canceled, and the roads program cut back to less than one-quarter of the SAR estimate. 5.11 Social development components, which had been argued to be essential during appraisal, were curtailed or canceled when counterpart funds became scarce. The exceptions to this were the drinking water systems in Fes-Karia-Tissa, which were persisted with by MARA even though it was not responsible for water systems. They were eventually completed at significantly higher-than-estimated costs, and transferred to ONEP, the national agency responsible for drinking water system development. The project, now serves at least 56,000 rural residents, who pay for the metered water they receive. The completion of both systems (Ain G'dah and M'kansa) is a tribute to the persistence of the implementors and Bank staff, but their value as models is questionable. Economic Rates or Return 5.12 Undoubtedly, the combination of six years of drought during the implementation period and the slow pace of implementation affected project ERRs. The paucity of reliable data in all four projects prevents a full re-estimation of their ERRs. This is a serious deficiency. For Fes-Karia-Tissa, the appraisal estimate of the ERR was 21 percent and the "indicative" re-estimate of the PCR is 9 percent, which is just below the generally accepted threshold of 10 percent. For Loukkos and Oulmes-Rommani, appraisal estimates were 15 percent and 14 percent respectively, and "indicative" re-estimates of similar levels are provided in the PCRs. The Middle Atlas PCR admits that it is not possible to re-estimate the ERR because data are inadequate, and this was still the case at audit. Inadequate M&E in all four projects made it impossible to track their performance closely during implementation and to re-estimate their ERRs on completion. 5.13 The indicative re-estimated ERRs are based on provincial-and district-level data, spot observations by the audit mission, and limited project monitoring data. For soft wheats, yield increases attributable to the projects made the greatest contribution to the re-estimated ERRs. Overall, soft wheat yields increased by about 56 percent with the projects (from 1.6 to 2.5 ton/ha), compared to 25 percent (from 1.6 to 2.0 ton/ha) without the projects. The assumptions behind the re-estimated ERRs are realistic. In Oulmes-Rommani, the net benefits from improvements in soft wheat production were sufficient to sustain all costs of the revised project. Environmental Efects 5.14 A major objective of the projects was to conserve upland soil and water resources. The absence of environmental monitoring data makes it impossible to quantify the projects' environmental effects - erosion control, in particular - although they are undoubtedly positive. The cost-effectiveness of the conservation measures employed under the projects is unknown. 5.15 The most obviously beneficial environmental effects of the projects stem from the soil and water conservation effects of contour farming techniques; the conversion from cereal production down-the-slope on hilly lands to permanent tree crops on terraces on the contour; the reforestation of steeply sloping lands in the Izarene forest of the Loukkos; the low-cost, low-maintenance management - 24 - of some Middle Atlas oak forests as sylvo-pastures; and the collective management of common pastures at Timahdite. 5.16 The lack of environmental data is disappointing. A requirement for environmental monitoring was included in the Loukkos SAR, which grew out of concern for catchment management: "the life of the (El Makhazine) reservoir would be greatly influenced by the rate of erosion upstream in the Loukkos catchment, so erosion control studies were initiated as part of the FAO/UNDP project". / The Loukkos project aimed to continue the monitoring program of the Directorate of Water and Forests in the El Makhazine watershed, as a source of data for environmental benefits' estimation. This was not done, and project's contribution to increasing longevity of the reservoir is unknown. Institutional Development 5.17 These projects were institutionally complex. Multiple agencies were involved in implementation of each project; project coordination was required at three levels - national, provincial and local; and the lead agency for each project was an agricultural directorate with a narrow range of experience - crop production, livestock development, water and forest resources management, or agricultural development. At the national level, MARA's directorates continue to work largely independently, but a substantial degree of cooperation and collaboration is evident in the project areas. Similar capacity appears not to have evolved in areas of Morocco which have not received external support. 5.18 As the projects evolved, their formal institutional structure remained unchanged. But so- called "unnecessary components" were reduced or eliminated by the funding shortfalls and the MTR, and the institutional arrangements needed for implementation were correspondingly reduced. Projects designed to be complex were simplified during implementation. Experience as Pilot Projects 5.19 At the time of their appraisal, experience in rainfed agricultural development in Morocco was limited. The projects under audit were pilot projects, of potential importance to larger areas of rainfed Morocco. With limited exceptions, however, the projects did not provide satisfactory vehicles for developing and testing either new technical packages or new organizational arrangements. The relatively large size of the projects and the weak M&E performance reduced their experimental value sharply. The frequent turnover of staff and limited use of technical assistance in Middle Atlas exacerbated these weaknesses. 5.20 Extension of the results of these projects to other areas of Morocco has been restricted by the continuing shortage of funds. Where external funding has been available, the experience gained in these projects has been built upon, as in the Bank-supported National Agricultural Research and Extension (Ln. 3036) and Second Forestry Development (Ln. 3156) projects. But the value of the innovation that did occur under these projects has not been fully realized because funding limitations continue to restrict rainfed agricultural development. !/ Staff Appraisal Report, Loukkos Rural Development Project, April 23 1980, page 13. The FAO/UNDP project referred to is DERRO, the national development program for the Western Rif. - 25 - VI. FINDINGS AND ISSUES Overall Findings 6.1 The overall finding of the audit is that Fes-Karia-Tissa, Loukkos and Oulmes-Rommani were satisfactory in their revised forms, and Middle Atlas was unsatisfactory. The PCR review had earlier rated Middle Atlas as satisfactory: this judgement has been reversed by the audit. None of the audited projects achieved the quantitative objectives established in the SARs: all were scaled down as a result of counterpart funding shortages. With the exception of Middle Atlas, all the projects had overly complex designs, and the resulting reductions in scope were undoubtedly beneficial. Despite funding problems, Fes-Karia-Tissa and Loukkos achieved many of their physical targets for infrastructure and agricultural development. Oulmes-Rommani was less successful overall, but achieved a satisfactory performance largely on the basis of its impact on expanding cereal production on better land. Middle Atlas was the least successful. 6.2 The lack of reliable M&E data makes the re-estimation of ERRs difficult or impossible for all the projects. The indicative re-estimated ERRs for Fes-Karia-Tissa (15 percent), Loukkos (9 percent), and Oulmes-Rommani (14 percent) are plausible. For Middle Atlas, the ERR cannot be re- estimated. Overall sustainability is considered to be likely in Loukkos, where institutional development is also judged to have been substantial. Sustainability is considered to be uncertain in the three other projects, where institutional development was found to have been partial or negligible. Table 6.1: Project Performance Summary Fes-Karia-Tissa Loukkos Middle Atlas Oulmes-Rommani SAR ERR 15 21 21 14 Revised ERR 9h/ 1597 n.a.d' 14-' Overall Assessmenta/ Satisfactory Satisfactory Unsatisfactory Satisfactory Sustainability Uncertain Likely Uncertain Uncertain Institutional Partial Substantial Partial Partial Development Overall assessment is of the revised projects. Indicative only. Indicative only. Not able to be re-estimated. Indicative only. Commitment and Political Will 6.3 Prior to the these projects, Government's interest in rainfed area development was limited, and the Bank and FAO made a deliberate effort in the mid-1970s to raise its importance. The primacy of irrigation was not seriously challenged, but it was successfully argued by the Bank that the relative neglect of rainfed areas was no longer efficient. With these projects, rainfed agricultural development was placed back on the national agenda. But its timing could not have been worse, since their implementation coincided with the beginning of a long period of budget austerity. - 26 - 6.4 The Bank knew that budget austerity was to be initiated before the first of these projects was approved, and the weakness of Government commitment showed up quickly in budgetary reductions, the slow release of funds, and delayed approvals. These projects and the Forestry Project (Ln. 2110) were the hardest hit of all Bank-supported projects in Morocco during the period 1978-92. The level of commitment to the rainfed projects shows in the record of Bank loan and credit disbursements to Morocco. In descending order, the proportion of original principal disbursed, to be as follows:7' Completed Bank-supported Projects, Implemented between 1978 and 1992: % Original Loan Disbursed Agricultural credit 100 Hotel and real estate development 95 Economic development banking 93 Irrigation and water resources 91 Phosphate, oil shales and petroleum development 90 Education 86 Housing and urban development 86 Small-scale industry and mining 50 Rainfed agriculture 46 Forestry 35 6.5 Between 1978 and 1992, the proportion of original Bank loan principal disbursed to agricultural credit and irrigation remained high, but area development and forestry were reduced below 50 percent, and by more than all other sectors. 6.6 Bank leverage drew Morocco into rainfed area development but it did not buy commitment. Each project was subjected to stringent restrictions in counterpart funding, cancellations and postponements. The Government's desire for multi-sectoral projects also suffered when social sector components were curtailed. Two questions can be asked: first, should the Bank and the Borrower have proceeded over a period of five years to approve a set of four rainfed projects which were so strongly Bank-influenced? The answer is probably "no." In particular, the wisdom of approving two new projects after 1980, when Morocco was clearly having difficulty with counterpart funding for the first two projects can be questioned. Secondly, would the projects have fared better if the level of commitment from the Borrower had been higher? The answer is clearly "yes". The Bank persuaded the Government that a rainfed strategy should be a priority but, even after this was accepted, the Government expressed its misgivings about the use of Bank loan funds for this purpose. When the budget austerity was enforced, funds for these low priority purposes were reduced. When Bank-supported projects of higher priority to Government were being funded more or less fully, support for rainfed agriculture was cut back sharply. Implementing Pilot Projects 6.7 To succeed as pilot projects, the projects under audit required a high degree of flexibility and good M&E. For most of the non-agricultural project components, implementation was straightforward, provided the financial resources were made available. Agricultural components, which aimed to build new capacities for responsive work with farmers, and to serve farmers required more time, L/ See Statement of Loans and Statement of Development Credits, The World Bank, July 31, 1993. - 27 - flexibility and good management information. Time became available to project implementors as the result of funding shortages and delays, but this was not planned for. The weak M&E in all projects handicapped the Borrower's ability to manage the projects as experiments. Time to Start Implementation Up 6.8 Only in Oulmes-Rommani were project start-up activities planned for the first year of implementation, even though new administrative and technical capacities had to be initiated in all projects. In Fes-Karia-Tissa, which was well funded at the start, the absence of an adequate start-up plan forced implementors to go more slowly than envisaged, requiring almost two valuable years. Financial difficulties forced the projects to go slowly, and time which had not been budgeted became available to prepare for implementation. 6.9 Provision of lead time for implementation might have improved the likelihood of M&E components being effectively introduced, but a solid commitment to M&E did not exist. The absence of computers diminished managers' accessibility to the M&E work that was undertaken. Coping with Shortages of Counterpart Funding 6.10 The projects provide a vivid case study of coping with counterpart funding shortages. Various attempts were made by the Bank to keep the projects funded - lobbying the Ministry of Finance, an MTR, and a SAP. Bank staff showed persistence in their efforts to keep implementation moving ahead. But these problems could have been avoided or reduced. Clear indications existed from 1978 that Morocco could have budgetary problems funding rainfed agricultural projects fully, but the loans were approved despite them. Four projects entered the lending program, and four projects were approved. But by 1982, the year Middle Atlas and Oulmes-Rommani were approved, Morocco was in its fifth year of budget austerity. Middle Atlas and Oulmes-Rommani should have been candidates for postponement or cancellation until Morocco's budgetary problems were overcome. This would have been preferable to approving them and funding them poorly. 6.11 When it was formally acknowledged by Bank Operations in 1983 that counterpart funding problems would not be short-lived, the MTR and SAP should have been introduced sooner and implemented faster. They should have been implemented sooner because all the audited projects were in financial difficulty before mid-1985, when the MTR was initiated. The preparation of the SAP should also have been accelerated, and finalized in less than the 29 months that elapsed between initial discussions and completion. Once approved, however, the SAP proved ineffective for the rainfed projects because the implementors in the DPAs did not know how to access its provisions. Complementarity of Components 6.12 The complex design of the projects was built upon the then acceptable assertion that rural development required joint implementation of complementary economic and social development components, and that there was potential synergy in their concurrent implementation. Services provided by agricultural extension agents, and access to markets, were improved by the roads provided under the projects. Overall complementarity between agricultural and social development components was, however, minimal. The inclusion of so many components made project coordination difficult. The addition of social infrastructure helped to raise the quality of life in rural areas, and lower the rate of outmigration. But they also raised Government's operating and maintenance costs during a period of budget austerity. Improvements in household incomes were not used to offset the cost of provision of social services. - 28 - Beneficiary Participation 6.13 Beneficiary participation in project planning and implementation was minimal. The active participation of women in project activities was not sought. But through project implementation, MARA learned to work more effectively with its many of its clients. This was a highly positive change for an agency which previously expected farmers to come to it for advice and services. In a few cases, notably in collective pasture management at Timahdite, MARA worked in a genuinely participatory way with farmer organizations, to the point where responsibility for most aspects of collective range management was turned over to cooperatives during the project period. This example should be more widely known both inside and outside the Bank. 6.14 In forestry, effective participation with forest land users was more limited. Implementation of forest works by contractors was typical of most project components, and the opportunity to use project funds to build better working relations with forest and sylvo-pasture users was not taken up. Substantial amounts of project funds were used to develop local infrastructure and to terrace and reforest hillsides, executed by contractors. If these funds had been channeled through farmer organizations - which would have had to have been created - a stronger relationship between MARA and project beneficiaries might have developed. Local organizations would have a stronger sense of ownership of project works, their operation and maintenance. This would have represented a more staff- intensive strategy for MARA than the approach taken, but it could have led to the establishment of a joint capacity for local management of natural resources. Necessary Conditions for Rural Development Success 6.15 Rural development experience inside and outside the Bank has shown that successful projects tend to be built around a three-part model1, comprising an economic or incentives model to provide returns sufficient to overcome the risk-averse behavior of rural households; a technical or production model to raise productivity dependably; and an institutional or social model to permit household and village organizations to establish effective relationships to markets and rural service providers. In the projects under audit, the conceptual underpinning was not clear. With the exception of a few components, such as collective pasture management at Timahdite, these projects were not strong in terms of these necessary constituents. 49 6.16 As part of the incentives model, improvements in roads and access have created benefits for rural communities which were previously isolated. This is noticeable in Fes-Karia-Tissa and Loukkos, but the reduction of this component caused by shortage of funds lowered its overall impact. As part of the production model, rainfed technical packages for farms, forests and pastures were not well developed at the outset. Technical progress has been modest, hampered by limited and late agricultural research, and missing monitoring. 6.17 With the exception of team building within the project DPAs, institutional development under the project has been limited. Farms are understood as farming systems, and extension agents do now leave their offices and market places for farmers' fields, to work responsively with farmers. Apart from the cooperatives managing collective ranges, new formal institutional capacities among farm households for resource management have not been developed. See Chapter 5 of OED's Annual Review of Evaluation Results, 1990 issued in 1991. Lq In Timahdite, the water points provided the new economic incentive available to all shepherds; pasture improvements raised overall productivity; and, the shepherds' cooperative and breed improvement associations provided the mechanism to operate and maintain the system sustainably. - 29 - 6.18 The question remains, however, if the projects had been conceptually stronger, would their impact have been greater, given the chronic funding shortages? The answer to this is probably "yes". Fewer, smaller, and better integrated projects would have provided a better test of Morocco's rainfed development strategy, and better survived the long period of budget austerity. A stronger demonstration effect in these projects would have raised the relative priority of rainfed agriculture. Bank Performance 6.19 Comments on Bank performance are addressed to issues of project design, supervision and the PCRs. Project design was overly complex. It was clear that many Bank Operations staff believed this to be the case prior to project approval. Numerous components were, however, included because this was the expressed wish of the Borrower and was believed to be consistent with rural development thinking at the time. Three of the projects became "Christmas trees". By the early 1980s, however, poor Bank experience with projects of this kind was more widely understood, making the design of Oulmes-Rommani harder to accept. 6.20 The detailed design of numerous infrastructure components was incomplete at approval. The time required to earn beneficiary participation in project implementation was also generally underestimated. Some DPAs took this responsibility seriously and built strong and continuing relationships to their farmer clients, which continue after project completion. 6.21 Frequency of project supervision was satisfactory through April 1986, but declined sharply thereafter in line with a Bank decision to limit general supervision and focus on problems of implementation. The benefits of this strategy were few because the overall counterpart funding problem was not eliminated. 6.22 The PCRs were generally satisfactory, but their completeness was severely limited by the lack of M&E data. Too much information in the PCRs is interpretive, rather than factual. The field portion of the audit of Oulmes-Rommani took place with the others in July 1993, before its PCR had been received by OED, in response to the Region's proposal that OED audit all four projects as a group.  。，&:．」．.．슨  IBRD 14879R APRIL 1980 KINGDOM OF MOROCCO LOUKKOS RURAL DEVELOPMENT PROJECT OOrouro ý PROJECT ZONES UPPER LOIKKOS CHAOUEN MIDOLE LOUKKOS (ZA IZARENE-NEW PLANTATIONS ZARENE-EXISTING REFORESTATION FOREST STAT.ONS o VILLAGES Eouieduane W NORK CENTERS SUB CENTERS Souk - WATER POINTS Et Kollo HEALTH CENTERS ® PROVINCIAL HEADOU.RTERSRS01FIN, COUNTY HEADODUARTERS R o TOWNSHIP NEADOUARTERS CATCHUENT BOUNOARY KSAR El KBIR 8 N PROVINCIAL BOUNDARY -- - -----SUBDIVISION BOUNDARY 1 -- TERTIARY ROADS N TE PROJECT ~MORISSET FARM ACCESS ROADS IN TRE MIDDLE LOUKKOS edo ---FARM ACCESS ROADS IN UPPER LOUKKOS -£PRIM)char A,ab 0 Arboou Dr SECONDARY ROADS ------ TERTIARY ROADS A cer c 0 M zefroun - 0 C0 £T QZoumni UEZZANE 5< 0". *001 0035 2 ofoV1 R  井  N s PM I<dyn- AA .1-11 - r AIN LEUH ITE ElA 0 -ý 0' -t 4 , N I 1CA WAER SUPPtY 9.1 ____ 2 ~.nooI 1oZod KINGDOM OF MOROCC -\ , ROYAJME Du MAROC AZROU MIDELT MIDDLE ATLAS AGRICULTURE DEVELOPMENT PROJECT PROJET DE DEVE1OPPEMENT AGR/COLE DU MOYEN ATIAS INFRASTRUCTURE AND PROJECT WORKS INFRAST~XTUIE ET TRAVAUX LU PROJET A .,Acewe,,a tuu.e TZ EiRl-k C) n c..... . 0 0 o o. - ... 5. e.ii.y ... alto sti.. .. .5... .5.  s^Jr lx k .N t r i5  二汰  IBRD 16474 6.y',e30- 6-15 6.00" PonTuG SPAIN 14 KINGDOM OF MOROCCO OULMES- ROMMANI AGRICULTURE DEVELOPMENT PROJECT M°R°C°- LAND USE -- ALGERIA 0 5 10 KitOMETERS ME K NES 3-3'30 33-30- Es1 ul wve AeaK H E N l F R A 5.45' R PROVINCE OF R PROJECT AREA KHOURIBGA