Colombia - Integrated Nutrition Improvement Project

Document of The World Bank FOR OFFICIAL USE ONLY [ E COP ' Report No. 7180 PROJECT PERFORMANCE AUDIT REPORT COLOMBIA INTEGRATED NUTRITION IMPROVEMENT PROJECT (LOAN 1487-CO) April 12, 1988 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. GLOSSARY CAR - Autonomous Regional Corporation (for the basins of the Bogota, Ubate and Suarez Rivers) DNP - National Planning Department DRI - Integrated Rural Development FONADE - National Fund for Development Projects GOC - Government of Colombia ICBF - Colombian Institute for Family Welfare INAS - National Institute of Health INCOMEX - Colombian Institute of Foreign Trade ITT - Institute of Technological Research MCH - Maternal and Child Health MOH - Ministry of Health PAHO - Pan American Health Organization PAN - National Food and Nutrition Program PANCOGER - Program in Support of Small H;me Gardens PCM - Protein-Calorie Malnutrition Currency Exchange Rates Name of Currency (abbreviation): Colombian Pesos (Col.$) US$ = Col.$ Appraisal Year average 36.00 Implementation Period Average 69.73 Completion year 142.30 Fiscal Year of Borrower January 1 - December 31 1O * *<L USE ONLY THE WORLD BANK Washington, D.C. 20433 U.SA Operations Evahuatin April 12, 1988 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Reports Colombia - Integrated Nutrition Improvement Project (Loan 1487-CO) Attached, for information, is a copy of a report entitled "Project Performance Audit Reports Colombia - Integrated Nutrition Liprovement Project (Loan 1487-CO)" prepared by the Operations Evaluation Department. Attachment This document has a vestricted distrbution and may be used! by socipients onl in the performa of their official duties. ItS contents may no othwerwise be disclosed withot WOM Ban1k autodmio. FOR OFFICIAL US ONLY PROJECT PERFORMANCE AUDIT REPORT COLOMBIA INTEGRATED NUTRITION IMPROVEMENT PROJECT (LOAN 1487-CO) TABLE OF CONTENTS Page No. Preface . . ..... ....... ..... . ............ L Basic Date Sheet ........................ ......I....... 11 Evaluation Summary .............0.......................... Vi PROJECT PERFORMANCE AUDIT MEMORANDUM I. PROJECT BACKGROUND ................................ 1 Project Formulation .............................. Project Objectives and Content ...................... 1 II. PROJECT IMPLEMENTATION ............................ 2 Design .............................................. 2 Management .o....................................... 3 Bank Supervision ............o .......... ............. 5 Cost and Finance .................................... 6 III. PROJECT OUTCOMES .................................... 6 Introduction ...... .......o..o...................... 6 PrImary Health Care Units ........................... 7 Rural Water Supply and Sanitation ................... 8 Nutrition Education o..........................0.... 8 Family Gardens (PANCOGER) ................. .0....... 9 School Gardena ...................................... 10 Food Coupons ....................................... 10 Monitoring and Evaluation ........................... 12 IV. FINDINGS ................ .... ... ................. 14 Project Content and Implementation Arrangements .e... 14 Sustainability ...................................... 14 Forward Linkages ...................... .... 15 ATTACHMENTS I-II Comments from the Borrower ................. 17-18 This document has a restricted distribution and may be use by socipiens-a on late perfoanme of their ofBicia duties. Its contents say no otherwle be disclosd withou Work Bank authoaton TABLE OF CONTENTS (cont9d) PROJECT COMPLETION REPORT Page No. I. Introduction ........................................ 21 II. Project Identification, Preparation and Appraisal ..................................... 22 III. Project Implementation and Cost ..................... 28 IV. Project Performance and Impact ...................... 32 V. Bank Performance ................................... 38 VI. Conclusions ......................................... 41 Tables: 1. Project Performance Summary - Key Indicators ........ 45 2. Price Differences due to Presence of Mobile Markets in Bogota ............................... 46 3. Actual and Appraised Project Costs by Component ..... 47 4. Actual and Appraised Disbursements by Expenditure Category ......................... 48 5. Actual and Appraised Schedule of Disbursements ...... 49 6. Project Performance Supervision Ratings ............. 50 PROJECT PERFORMANCE AUDIT REPORT gOLOMBIA INTEGRATED NUTRITION IMPROl'AENT PROJECT (LOAN 1487-CO) PREFACE This is a project performance audit of the Integrated Nutrition Improvement Project in Colombia, for which a Loan of US$25 million was approved in September 1977 and became effective in March 1978. After two extensions of the Closing Date (of one and two years, respectively) the Loan was fully disbursed and closed on June 30, 1985. The audit report consists of a Project Performance Audit Memorandum (PPAM) prepared by the Operations Evaluation Department and a Project Completion Report (PCR) dated August 18, 1986. The PCR was prepared by the Population, Health and Nutrition Department and, apart from documents available in the Bank, incorporated the findings of a Finel Report prepared by the implementing agency DRI and of a consultant's report. The PPAM is based on a review of materials in Bank files, including the Staff Appraisal Report No. 1583-CO of August 29, 1977, the President9a Report No. P-2122-CO of August 30, 1977, the Loan Agreement dated November 10, 1977, correspondence with the Borrower, the PCR and the Government's Final Report, and discussions with Borrower and Bank staff associated with the project. An OED staff visited some of the infrastruc- ture provided under the project during an audit mission in February 1987. The audit memorandum, while in general agreement with the main conclusions of the PCR, is somewhat more critical of the design of the project and the implementation experience. However, since the audit mission was facing the same data difficulties as the skthors of the PCR, steamu.ng in large measure from the absence of a proper monitoring and evaluation system, it could add but little to the findings of the PCR. Consequently, the PPM focuses on a few specific points regarding project design, Implementation and outcomes and sustainability of the various project components. As is customary in the preparation of audit reports, copies of the draft audit report were sent to the representatives of the Borrower for coments in December 1987. Comentu received were taken into account and have been reproduced as Attachments I and II. - LL - PROJECT PERPORMANCE AUDIT REPORT COLOMBIA INTEGRATED NUTRITION IMPROVEMENT PROJECT (LOAN 1487-CO) BASIC DATA SHEET KEY PROJECT DATA Item Appraisal EstImate Actual Total Project Cost (US$ million) 68.9 87.1 Overrun (2) - 26.0 Loan Amount (US$ million) 25.0 25.0 Disbursed 25.0 25.0 Cancelled - -- Repaid ) 11.52 Outstanding) as of 02/29/88 - 13.48 Date Physical Components Completed 06130181 06/30185 --in months since Loan Signature 44 92 Proportion completed by above date (2) 100 100 Proportion of time overrun (2) - 109 Institutional Performance fair CUMULATIVE ESTMATED AND ACTUAL DISBURSEMENT (in US$ million) CY 1978 1979 1980 1981 1982 1983 1984 1985 1986 Appraisal Estimate 1.3 6.9 13.6 21.8 25.0 Actual 0.3 1.8 4.6 8.2 14.2 16.4 21.7 24.8 25.0 Actual as Z of Estimate 15.0 26.0 34.0 38.0 57.0 65.6 86.8 99.2 100.0 - 11Li - STAFF INPUT (Staff weeks) S7 e 0 a 81 62 68 6 Total Pre- appraisal 80.4 - - - - - - - - - 88.4 Appraisal 96.0 18.8 - - - - - - - - 18.8 Negotialon - 7.6 - - - - - - - - 7.6 Supervison - 88.4 26.9 10.1 88.7 29.4 26.5 21.2 6.7 6.8 218.2 Other 6.9 - - -- - -- 6. - - - 1.2 Total 12.8 54.7 26.9 16.1 86.7 29.4 28.9 21.2 6.7 6.8 867.1 OTHER PROJECT DATA Original Actual or Item Plan Revision Est. Actual First Mention in Files 10102173 Negotiations 07/27/77 Board Approval 09/13177 Loan Agreement Date 11/10177 Effectiveness Date 03/09/78 Closing Date 06/30182 06/30183 06/30/85 Borrower Republic of Colombia Executing Agency Fondo DRI Ministry of Agriculture (Previously DRI-PAN and originally in the project only PAN, both under the National Planning Department (DNP)) Follow-on Project None /a z Plus 250.9 staff weeks during FY74-76. IV MISSION DATA Sent Month/ No. of No. of Staff Mission by_ ear Das Persons* Weeks** Reconnaissance Bank 03/74 11 4(DC,NS,HP,D) 8.8 Identification BaAk 01/75 10 4(NA,PO) 8.0 Preparation Bank 04175 8 2(PO,TS) 3.2 Preparation Bank 07/75 9 2(PO,LO) 3.6 Preparation Bank 09/75 4 2(PO,LO) 1.6 Progress Review Bank 11/75 9 1(LO) 1.8 Pre-Appraisal Bank 02/76 18 7(DC,PO,LO,MS 25.2 FT,NE,N) Appraisal Bank 03/76 12 10(DC,L0,EC,2HS 24.0 FT,NE,N,SW,FD) Appraisal Bank 01/77 21 10,(DC,PO,L0,EC 42.0 HS,Ff,NE,N,SW,FD) Total 118.2 Supervision 1 Bank 11177 10 2(HS) 4.0 Supervision 2 Bank 02/78 15 5(PO,HS,HG,NE,SW) 15.0 Supervision 3 Bank 05/78 17 2(HS) 6.8 Supervision 4 Bank 07/78 21 1(SW 4.2 Supervision 5 Bank 09/78 11 2(PO,HG) 4.4 Supervision 6 Bank 02/79 12 4(PO,NE,N,SW) 9.6 Supervision 7 Bank 06/79 9 1(PO) 1.8 Supervision 8 Bank 03/80 16 4(DC,PO,HS,HG) 12.8 Supervision 9 Bank 10180 6 2(PO,HS) 2.4 Supervision 10 Bank 04/81 14 5(PO,2MS,NE,N) 14.0 Supervision 11 Bank 02/82 16 4(PO,LO,HS,NE) 12.8 Supervision 12 Bank 11/82 14 6(PO,3LO,HS,NE) 16.8 Supervision 13 Bank 05/83 13 2(PO,N) 5.2 Supervision 14 Bank 04/84 12 4(PO,OA,HS,N) 9.6 Supervision 15 Bank 11/84 10 3(PO,OA,N) 6.0 Supervision 16 Bank 06/85 10 2(PO,OA) 4.0 Total 129.4 * NA: Nutrition Adviser HGs Home Gardens DC: Division Chief FT: Food Technologist POs Project Officer NE: Mass Media/Nutrition Educator LOs Loan Officer N : Nutritionist EC: Economist/Financial Analyst SW: Sanitation/Water Supply OA: Operations Analyst FD: Food Distribution MS: Food Marketing Specialist HP: Architect/Health Planner HS: Heslth Specialist D : Demographer NSa Medical Nutrition Specialist TS: Training Specialist ** Estimated number of staff-weeks attributable to this project (including travel time). VU PROJECT PERFORMANCE SUMARY - KEY INDICATORS Appraised Actual Category Target Health and Sanitation Construction/Improvement of Pri.mary Health Care Posts 281 729 Health Promotets Trained 2,028 1,368 Equipping of Health Promoters 2,400 1,297 Completed Water Works 360 326 Beneficiaries (6-8 per home) 231,250 199,227 Latrines 112,000 35,000 Nutrition Education Staff Trained 8,028 30,996 Mass Media: Radio Spots 225,000 565,660 Home Garden Plots (PANCOGER) Families with Garden Plots 20,000 45,720 Credit Recipients (families) 20,000 41,780 Food Coupons Coupons Distributed (millions) 71.8 64.8 Direct Beneficiaries 415,500 420,476 - vi - PROJECT PERFORMANCE AUDIT REPORT COLOMBIA INTEGRATED NUTRITION IMPROVEMENT PROJECT (LOAN 1487-CO) EVALUATION SUMMARY Introduction The Integrated Nutrition Improvement Project was supported by a Bank loaL of US$25 million which was approved in September 1977 and became effective in March 1978. The loan was fully disbursed and closed on June 30, 1985, three years after the planned date. Appraisal estimate of total project cost was US$68.9 million and actual cost, US$87.1 million, resulting in a reduction of the Bank's contribution from 36Z to 29%. Proiect Obiectives and Content The project was to support the Government's integrated program in the areas of nutrition, health and rural development. Components included rural health facilities and training of their staff, rural water supply and sanitation, nutrition education, input packages for family gardens, food technology and quality control, a Government-financed food coupon program and monitoringlevaluation, program management and technical assistance. Later moZifications were the addition of four Departments to the project area, the introduction of a food distribution and marketing component and the discontinuation of the food coupon and home garden schemes. Implementation Experience The loan was fully disbursed but required two extensions (of one and two years, respectively). The initial Implementation arrangements with the National Planning Department as the coordinating agency did not always give the desired results. The transfer of this program into the responsi- bility of the Ministry of Agriculture, while leading to an improved tmple- mentation performance, occurred too late in the project cycle to permit a judgement as to the effectiveness of this alternative. Project implementation was also negatively affected by a lessening of support when a new government took office in 1982, and by worsening economic conditions. This led to a project reformilation in 1983 with lesser emphasis on its distributive (welfare) aspects and a focussing on physical investments. - vil - Results to Date A major outcome of the project was of an indirect naturel it contributed to general policy formulation in the fields of health and nutrition. The development of the primary health care system as an effec- tAve vehicle for dispensing a wide array of health and nutrition services may also prove a valuable asset for similar future endeavors. The nuameri- cal achievements of the project in most cases surpassed appraisal expecta- tions, but the failure of the monitoring and evaluation system makes an impact assessment difficult. Since a majority of components were in the nature of temporary programs without accompanying physical infrastructure, it Is almost impossible to assess their degree of effectiveness years later. Moreover, the various project components were seen as an integrated package. Judging its impact and that of its individual elements would have required multivariate statistical analysis for which there was never an adequate data base (with one specific exception). Most conclusions concerning project impact are thus inevitably in the nature of general qualitative statements. Sustainability It appears that many of the health facilities provided under the project are having problems of inadequate funding for their operations (as does, to an even greater degree, the rest of the system). If continuing, this shortfall could endanger the longer-term sustainability of this component. By the same token, the user charges for the rural water supply system% are not deemed sufficient to cover major repair and replacement needs wh!ch will arise in the future. Findinzs A major handicap for a smooth implementation of this project was its great complexity. A more compact project design concentrating on a few core elements could have been expected to produce good results more expeditiously, without subtracting significantly from the total impact. The failure of the monitoring and evaluation system made a satisfactory assessment of most components all but impossible. Since this shortcoming is common to numerous projects, the Bank would be well advised to give this problem continued close attention. Finally, in the case of projects which for political or other reasons have ceased to enjoy the full support of the Borrower, it may ultimately be in the best interest of both parties if the Bank showed greater readiness to discontinue a project and, circumstances permitting, make a new start, rather than letting secondary considerations (an undisbursed loan/credit balance, the cost of project preparation, etc.) prevail and opting for project changes which may have been insufficiently prepared, hard to integrate into the existing Implementation structure, or not in harmony with the original project concept. This argument would apply a fortiori to innovative projects such as the one under review. - 1 - PROJECT PERFOR14ANCE AUDIT MEMORANDUM COLOMBIA INTEGRA4TED NUTRITION IMPROVEMENT PROJECT (LOAN 1487-CO) I. PROJECT BACKGROUND Project Formulation 1. The Colombia Integrated Nutrition Improvement Project was first mentioned in the Bank's files in late 1973. Starting with a reconnaissance mission in March 1974, project preparation extended over a long period, culminating in a two-stage appraisal (March 1976 and January 1977). Staff inputs were heavy, reflecting the complexity of project design: seven staff participated in pre-appraisal and ten each in the two appraisal missions. Both the long processing period and intensive staff work were explained by the relative novelty of nutrition projects in the Bank, the Colombia project being only the third of its kind, and accentuated by significant modifications during project preparation. The Board approved the loan in September 1977 and it became effective in March 1978. Project Objectives and Content 2. The central 'objective of the project was to support the Government'a nutrition policy as embodied in the National Food and Nutrition Plan (PAN), through a wide array of investments. In its final form, however, the project's scope went well beyond the area of nutrition and encompassed primary health care, rural water supply and sanitation. The project area included initially the Departments of Atlantico, Bolivar, Caldas, Magdalena, Risaralda, Tolima, Valle and the special district of Bogota. The principal target population weze infants and children under four years of age and pregnant and nursing women in rural and marginal (low-income) urban areas. 3. To these broad objectives corresponded a wide variety of invest- ments which can be grouped in several large blocks namely: (i) health services (constructionlrenovationlequipping of 281 primary health care units--health care posts or centers--and training for about 2,700 staff); (ii) rural water supply (360 local systems) and sanitation (more than 100,000 household latrines), both of these components being supported by contributions (mostly labor) of the beneficiaries; (iii) nutrition education (mass media campaigns, training in nutri- tion education for health workers and community leaders, production of relevant instructional materials); (iv) development and distribution of appropriate input packages for family gardens to benefit some 20,000 families, and establishment of some 100 demonstration school gardens; (v) investments in food technology and quality control aiming at increasing and diversifying the supply of low-cost enriched foods; and (vi) monitoring and evaluation, program coordination and adminis- tration, and technical assistance (specialist services and fellowships). 4. On the occasion of the project's reformulation in September 1983 (para. 13) an additional component was added, namely, food distribution and marketing services to benefit the poor in Colombia's major cities. 5. The project also included a Government-financed food coupon scheme to reach more than 400,000 individuals in the principal target group (para. 2). 6. The above list is a condensed summary which does not convey a complete picture of the ramifications of the project which comprised about 25 sub-components (for more details, see PCR, para. 2.20 (a-i)). 7. The project was estimated to cost US$68.9 million of which US$25 million was to be financed by the Bank loan. II. PROJECT IMPLEMENTATION Desian 8. Although the SAR does not explicitly 3ay so, this project was a field-scale experiment with a package of interventions about whose effec- tiveness and interactions little was known, but which were thought to be mutually reinforcing. This Implied a very complex project design which in turn was going to make implementation difficult. A number of components (such as the PANCOGER program and related activities) were added, in part at the Bank's urging, during the later stages of project preparation to provide a counterweight of directly productive activities to the basic distributive/ welfare orientation of this project. Later modifications in project scope and composition (PPM, para. 13) aimed in the same direction. In the process, the project acquired a somewhat awkward dual character: on the one hand a multifaceted program encompassing the areas of nutrition/ health/sanitation, on the other, several add-on activities in smallholder agriculture which were to support the nutritional objectives but later developed a dynamic of their own. 9. There can be little doubt that design complexity (which reflected the comprehensive nature of the Government's nutrition program) was one of the root causes for the difficulties encountered by this project. The difficulties of containing the project within a manageable frame were noted - 3 - by the PCR (pars. 2.07-2.10); while arguing with hind-sight is always easy, it mat surprise that this project apparently did not receive modifying contributions from sectors it partially encompassed (rural development, education), where staff had already had ample occasion to experience elsewLere the drawbacks of complex multicomponent projects. To state the case differently, a project focussing on the core components of primary health care and rural water supply/sanitation, combined with the Government-financed food coupon program, would already have been a formid- able challenge; any addition to it was bound to reduce the chances of a smooth implementation. 10. It should be stressed that it was greater than expected implemen- tation difficulties and adverse economic and political developmcnts rather than lack of technical preparation that brought on this situation: the two-volume Staff Appraisal Report was prepared with great attention to detail; however, the experience of this project confirmed once again the simple fact that careful documentation at the project design stage does not by itself assure a problem-free implementation (PCR, para. 5.07). Manaement 11. To the fragmentation of project components corresponded a large number of implementing agencies: two Ministries (Health and Education) with their departmental services, seven major agencies in the fields of health, welfare and rural development, media networks, universities and private firms, not counting institutions and individuals participating in the extensive study program. It was thought that this project would require a project coordination agency, and the prestigious National Planning Department (DNP), which had been the architect of the National Nutrition Plan (PAN) and had been the major force during the stages of project identification and preparation was the obvious choice. However, the projecz provided another example of the difficulties of coordinating the activities of independent line agencies and the institutional limita- tions of a staff agency in the combined roles of implementericoordinatorl supervisor. In retrospect, the PCR comes to the conclusion that direct links between the Bank and the individual implementing agencies would have been more productive (PCR, para. 5.12). 12. In 1980, the PAN group in DNP was merged with DNP's Integrated Rural Development Program (DRI). This move inevitably shifted the project's emphasis towards general rural development aspects, at the expense of its original specific health and nutrititn focus. In 1984, the combined PAN/DRI group was transferred from DNP to the Ministry of Agriculture, to be organized the following year as a semi-autonomous agency, the Fondo DRI, under that Ministry.41 The shift of implementation 11 It should be noted in passing that in the Bank, the location of nutrition activities posed a similar dilemmas having been established within the Population and Nutrition Project Department of the Bank, the Nutrition Group was transferred in 1977 to the Agriculture and Rural Development Department before being moved back in 1980 to form part of the Population, Health and Nutrition Department. responsibilities to the Ministry of Agriculture occurred too late in the project cycle fully to test this alternative implementation arrangement. 13. The 1982 elections brought a government to office that was more critical of large social sector spending in general and subsily programs in particular, a position reinforced by a worsening economic situation (PCR, paras. 1.08 and 3.06). This led, in 1983, to a reformulation of the project, together with the before-mentioned change in implementation arrangements. While the project underwent a partial consolidation (through the discontinuation of the subsidised food coupon and home garde s compo- nents), it experienced at the same time two additions: geographically, through the inclusion of the Department of Antioquia/ and materially, through the introduction of the food marketing and distribution component (PPAM, para. 4). 14. By this time, however, the first one-year extension of the Closing Date (from June 30, 1982 to June 30, 1983) was already coming to an end, with only 61% of loan funds disbursed. Slow disbursements had been a con- tinuous major problem. Scarcity and late availability of counterpart funds, unfamiliarity of Implementing staff with the project's administra- tive requirements and complicated Government procedures were contributing factors (PCR, paras. 3.09 and 3.13). 15. The choice facing the Bank at this juncture was whether to conti- nue with a project that, while having substantively achieved its original physical targets, had still considerable undisbursed loan funds but had largely lost its raison d'tre, or whether to close the loan and, circum- stances permitting, make a fresh start incorporating the lessons learnt. 16. The Bank opted for the former course because it felt (I) that there was still scope for investments in rural health, water supply and sanitation; (ii) that a continuation would support institutional develop- ment in the field of nutrition; and (iii) that a continued operational presence would enable the Bank to maintain some leverage in the formulation of relevant national policies (PCR, para. 3.07). 17. However, an alternative line of argumentation appears equally plausibles the project had largely run its course. As a field-scale experiment, it had not performed the essential task of providing systematic feedback on the impact of the investments, and there was no reasonable hope of activating the evaluation component. The imminent transfer of Implemen- tation responsibilities to the Ministry of Agriculture was likely to move the project further away from its original objectives. In short, it had ceased to be what its name promised: an integrated program to improve nutrition. Even with a subsidy element ranging from 401 to 60%, the food coupon program had not always been able to reach the intended participants: in 1980, for instance, while about 90% of coupons issued were redeemed, X) This followed an earlier (1981) addition of the Departments of Cauca, Huila and Norte de Santander (which had been assisted under an AID project) to the project areas. - 5 - almost 40% of entitlements were actually not requested, mostly because families did not have sufficient purchasing power to pay even the subsi- dized prices, while continuing inflation was reducing the effective subsidy. To expect the food marketing component to take the place of the coupon program was quite unrealistic, as subsequent developments proved; these activities benefitted urban middle-income families rather than the poor (PCR, paras. 4.16-4.17; Annex 2). In addition, the termination of the home garden component removed the one program that, however imperfectly it may have worked in practice, had been the only vehicle for self-help efforts to improve nutrition of the rural poor. In short, the 1983 changes all but excluded the core group of the original target population from the nutrition improvement efforts. On the other hand, the decision to continue and expand the health facilities construction/rehabilitation program (during the period 1983-85 a total of 231 units were reached, or 32% of the total number assisted) may have contributed to subsequent recurrent funding problems, an issue which had already been fully recognized prior to project appraisal (PCR, para. 2.10). Finally, it was apparent that the project (with another new component about to be added) could not be given the same degree of attention by the Bank as in the pasts PHN was experiencing increasing operational constraints as more recent projects made competing claims on staff time (PCR, para. 5.09). 18. On balance, it seems that a discontinuation of the project (with the option of resuming some of its activities under the First Health Project, Loan 2611-CO, which was then in its early stages of preparation) would have been preferable. That this did not happen appears to have been due more to extraneous considerations than to the intrinsic merits of the project in its final phase. Bank Supervision 19. There was general agreement that this proje ., because of its novelty and multi-component nature, would entail a higher than usual super- vision load (PCR, para. 2.11). This was quite apparent during the early stages of project implementation: the period November 1977 through September 1978 saw a total of five missions. However, it was also clear that the very specific types of expertise required could not be fielded for every supervision trip: whilt the two appraisal missions had ten members each, the subsequent 16 supervision missions had an average three members staying in the country about 13 days. This still sums up to a total effort well above average: 250 staff-weeks following preparation work of about 400 staff-weeks. 20. Apart from the inevitably incomplete coverage in terms of special- ties supervision missions also had to cope with the handicaps of gradually waning political support, of changes in Implementation arrangements and considerable project staff turnover (PCR, paras. 3.08 and 3.13). It also appears that supervision missions had only limited possibilities of influ- encing events as witnessed by the omission of the covenanted Bolivar pilot project to test alternative delivery systems for nutrition interventions, the difficulties and ultimate failure of monitoring and evaluation, and the ignoring by Government of the Bank's explicit misgivings over the merger of the nutrition and rural development (PAN and DRI) programs, which was seen as a serious weakening of the project's identity (PCR, paras. 3.03-3.04). -6- 21. An anticipated by-product of this project was the learning experi- ence it would afford the Bank in a new sector. The PCR holds that the Integrated Nutrition Improvement Project was an effective training ground for Bank staff and gave valuable insights that were carried over into sub- sequent related projects (PCR, para. 6.12). The audit mission sees no reason to disagree with this view but wishes to point to the considerable costa of this learning process. Cost and Finance 22. Total project cost came to US$87.1 million, more than 25% higher than the appraisal estimate of US$68.9 million. This is not, or only in part a cost overrun in the narrow sense, resulting from price increases during the extended implementation period. A larger portion is probably explained by the increase in the scope of project activities. The increase in total project cost meant a reduced share of Bank financing (29% instead of 36% as anticipated at appraisal). Community contributions to some of the physical infrastructure (latrines and water supply systems) both in cash and kind were sizeable but are not reflected in the accounts (for water supply systems, field estimates of community financing ranging from 10 to 20% of total cost were given). III. PROJECT OUTCOMES Introduction 23. It is difficult to assess (except in a general qualitative sense) the outcome of this projects most of its activities were of rather short duration and the necessary data be e to measure their impact was not estab- lished and cannot be reconstructed several years later. In the case of physical Infrastructure components, one can however observe the functioning of facilities and come at least to some general conclusions about their effectiveness. However, the combined effect of the various interventions constituting the project (much less their individual contributions) remain unknown. This is in contrast to most projects in other sectors where at least rudimentary (and often quite detailed) data on project outcomes exists increased production and incomes deriving from agricultural proj- ects, higher traffic volumes and lower vehicle operating costs in trans- port, a surge in enrollments or drop in repetition rates in education, etc. 24. This being said, it should be recognized that the project was instrumental in establishing the idea of preventive health care in Colombia--no small achievement considering the prevalence of curative health care concepts with a strong urban bias prior to the project which can still be gleaned from the early supervision reports (PCR, para. 4.02). 25. Second, it proved the feasibility of identifying, isolating and addressing directly the nutrition requirements of specific target groups through a food coupon program (PCR, para. 6.08). (Whether this interven- tion was both effective and cost-efficient is a separate question to be taken up later--PPAM, paras. 38-43). - 7 - 26. Third, the project demonstrated the viability of the primary health care structure as delivery system for a range of activities in the fields of health and nutrition, particularly in rural areas, staff short- ages and budget constraints notwithstanding (PCR, para. 4.04). 27. In one area, however, the project experience proved to be a series of disappointments: the systematic feedback of results through an effective system of monitoring and evaluation. The conclusion of the PCR that the implementation history "suggests that multi-sectoral programming can generate productive outcomes even when multi-sectoral managerial burdens prove excessive" (PCR, para. 6.06) is a far cry from the answers this project was expected to provide: what are the effects of specific policy interventions on the health and nutrition status of selected target populations, and what is the most cost-effective mix of healthinutrition activities to achieve specific targets? 28. It follows from the foregoing that the quantitative achievements of this project and its individual components have to be discussed mostly in terms of inputs realized. In this respect the project has fared rather well: the project recorded large increments over appraisal targets in the areas of nutrition education, home gardens, and primary health care units built or improved. The rural water supply and food coupon components experienced only minor shortfalls (9% and 10% respectively, and in terms of beneficiaries, the coupon program was even 1% above target). By contrast, the training of health promoters and rural sanitation (latrines construc- tion) showed considerable implementation deficits (from 331 to 692). The following comments on individual components are mostly derived from the project files and supplemented by very limited field observations and global national figures3l neither of which can claim to mirror faithfully the conditions of the various project areas and target populations. Primary Health Care Units 29. According to data supplied by the Ministry of Health, only four health centers and two health posts assisted under the project (out of a total of 729) were not functioning in April 1987, but some of the units visited were understaffed and showed signs of supply shortages. The statistics for the entire system point in the same direction and suggest even more serious difficulties for non-project institutions: in 1984 there was understaffing in 441 of the health posts, 37Z of the health centers and 30% of the local/regional hospitals which form the apex of the primary health care system. More than half of the health posts had incomplete or defective equipment, 83% of the health centers and 59% of the local hos- pitals. Less than 20% of the health promoters had the necessary health kit (down from 80% in 1979), and only 452 of those kits were in good condition. Only 12% of health promoters were satisfied with their working conditions, compared to 39% in 1979. Among the most frequent complaints were lack of equipment (51%), transport (48%) and medicines (352). These observations 3l The principal source was: Ministerio de Salud, Atenci&n Primaria, Evaluaci6n de extensi;n de cobertura en Servicio de Salud 1979-1984, Bogota, December 1985. - 8 - are supported by the relevant budget figures: while the physical size of the primary health care system (facilities and staff) almost doubled between 1979 and 1984, expenditures for supplies decreased by 28Z and for general non-personal items, by 8%. 30. These shortcomings are the result of the extended period of economic difficulties and of enforced budgetary stringency Colombia has experienced since the late 1970s. It is to the credit of the health staff that the various services were being expanded during this difficult time. This shows in the coverage of the priority target groups and the extent of vaccination campaigns, to name just two examples. Similarly, the efforts to teach oral rehydration and improvements in water supply/sanitation no doubt contributed to the drastic reduction of infant mortality due to diarrhea (up to 1979 the most important cause of infant mortality). There is good reason to assume that the improvement of these indicators in the project areas was ahead of national averages, given the considerable investments in facilities, equipment and staff training and a lower start- ing point. Rural Water Supply and Sanitation 31. The national statistics mentioned above (footnote 3) also contain information on targets and achievements for specific investments during the period 1979-84 (which is largely congruent with the project Implementation period). A comparison of these figures with the project achievements shows that the project counted for about half of the water supply systems built during that period and for more than 20% of home latrines. Unfortunately, the latrine building program, for a variety of reasons (supply bottlenecks, cost escalation, insufficient information efforts), fell far short of expectations, so that the combined impact of water supply and sanitation facilities was achieved only rarely. In addition, regional implementation personnel, in their desire to spread project benefits as widely as possi- ble, often purposely separated the two subcomponents and gave them to different villages, thus weakening further the hoped for impact. 32. Nevertheless, for those villages which received either type of infrastructure, the positive effects on general child development indi- cators (weight/height) seem to have been considerable. The one available multivariate analysis of the impact of the various project elements4/ suggest, that improved water supply and sanitation had a greater effect on child development than the nutritional enhancements provided under the food coupon program. Nutrition Education 33. Staff training was given to 31,000 persons (instead of the targeted 8,000) and 565,000 radio spots (instead of 225,000) were broad- cast, a commendable achievement. However, a consultant participating in one of the earlier supervision missions observed that a good deal of the 4/ P. Pinstrup-Andersen, El Impacto Nutricional del PAN en el Cauca, Informe de Asesoria para el DRI/PAN, Washington, January 1984. - 9 - radio messages appeared to explain and popularize the National Nutrition Program rather than attempt to bring about lasting behavioral changes known to have a positive influence on the Lealth and nutrition status of specific populations (infants, pregnant women, etc.) (see also PCR, para. 4.10). While a systematic impact assessment of specific campaigns has not been made, it would still be a worthwhile and valid exercise (and not necessari- ly a very costly or time-consuming one) to undertake a survey of women in a few project areas to determine their knowledge and continued use of speci- fic practices, such as breast-feeding, oral re-hydration, etc., in order to determine the durability of basic nutrition and health care messages. By including in the survey younger women not yet of child-bearing age during project implementation, one could learn whether these behavioral changes are firmly established i.e., whether the broadcasting program and follow-up work by rural health staff have had a catalytic effect beyond the initial project population. With careful survey design, one may in addition be able to reveal specific interactions (with age or education of mothers, family income, size of location, extent of complementary investments under the project, etc.). In short, in this particular area one could still hope to learn important lessons as to longer-term effects of specific informa- tion campaigns in health and nutrition. Family Gardens (PANCOGER) 34. This component was included at a later stage of project prepara- tion to add a production-oriented dimension to the project. Since it did not form part of the original integrated nutrition program, its linkages with the rest of the project were tenuous and it soon began to follow an independent course. In its original form it was designed to provide 20,000 very poor families with a basic package (valued at US$171familylyear equi- valent) of improved seeds (initially mainly for vegetables), fertilizers and p-sticides to improve nutrition and to generate a modest cash income to permit repayment of the subsidized credit.51 The Colombian Institute of Family Welfare (ICBF) was to implement this program. 35. Before long, this component underwent several changes: it was decided to diversify the input packages and to increase the credit ceiling to Col.$10,000 (then about US$275 equivalent). This soon raised a number of issues and illustrated the difficulties of helping the absolutely poor: first, the additions (or alternatives) to the initial package--grains and pulses, fruit trees, small animals (chicken, pigs), fish ponds and ulti- mately even improved housing--had not been tested with a very poor clien- tele. Second, the new items were in large part slow-maturing, relatively high-cost and high-risk investments and not always advisable for marginal subsistence farmers; third, by orienting the credit recipients more strong- ly towards market production, there was a distinct risk that this would have an adverse effect on their nutritional status, by inducing them to make more non-food purchases and/or to switch to less nutritious "modern" 51 During the first two years inputs were to be provided free of charge, with cost coverage by the beneficiaries rising to 50Z in the third year and 1001 thereafter. Interest charged would be one half of the market rate. - 10 - food items. These potential risks were compounded by an unwieldy adminis- trative super-structure resulting in high overhead costs.61 It also seems that neither ICBF not Its companion agency for the credit activities, Caja Agraria, had the staff resources to provide the necessary technical assist- ance to support the more sophisticated packages among a large number of small marginal farmers in remote areas. There were thus a number of factors at work that would have tended to favor farmers that did not belong to the most disadvantaged group. 36. Before this component was abandoned in 1983 (because of its high cost and the subsidy element it contained), it had reached a total of almost 42,000 families (compared to an appraisal target of 20,000), which would indicate a strong social demand for such a program. However, since data on the income profile of credit recipients are unavailable, it is not possible to say to what extent the component realized i. initial objec- tives of helping the absolutely poor to help themselves.71 School Gardens 37. This activity was initially a demonstration effort complementary to the PANCOGER program but was later transferred to the nutrition educa- tion component. Although it comprised many more than the 100 gardens initially envisaged (a 1980 figure of 1,064 school gardens was given in a supervision reports), it nevertheless seems to have had little lasting impact. There are a number of possible explanations for thist the inter- ruption of necessary gardening work through school holidays, lack of interest or know-how of teachers from urban backgrounds or insufficient acceptance by the pupils and their families. This disappointing experience has been made before by other countries attempting to "ruralize" primary education. Food Coupons 38. This Government-financed component was perhaps the core element of the project--and at the same time, because of its subsidy character, a source of controversy from the beginning. 39. Among the beneficiaries the judgement was overwhelmingly favorable: A national survey of coupon recipients showed that 90% considered this an Important economic assistance, 67Z valued the nutrition 6/ In 1981, an average figure of Col.$1,700/credit was reported. This &epresented about one-third the amount of a typical credit of Co.$5,000. /Z A 1984 survey of families with home family gardens -evealed that more than 92Z of them thought that the program had permitted them to increase their food supplies. While this figure attests to the wide acceptance of the PANCOGER program, it does not say anything about its physical Impact (Ministerio de Salud, op.cit., see footnote 3). - 11 - information thet went with it, 59% thought it increased their food supply, and only 1% deemed it useless.81 40. This component was the subject of an in-depth study in the Department of Cauca9J which gave the following results: on average the value of food coupons amounted to 1% of monthly family incomes, and augmented calorie consumption by 160 caloriesicoupon, or 1920 caloriesl family/month; this translated to 10.6 caloriesiday, or 0.5% of the average daily calorie requirements of each family member (assuming that the food was entirely be given to the family member most in need, the figure would have risen to about 64 caloriesiday). On balance, the food coupons did not seem to have had a significant impact on the weight and size of children; this may reflect a diversion of part of the additional food supplies to other family members. Since the survey measured family rather than indivi- dual consumption, the results are difficult to interpret. 41. In retrospect, it seems that the circle of beneficiaries was drawn too wide and that the (processed) foods included in the food coupon program were too expensive (about US$0.8011,000 calories in 1981 prices) to permit provision of a significant food supplement for every beneficiary at accept- able cost. It was also noted during one supervision in 1980 that infants under two constituted only 15% of the beneficiary population whereas pregnantinursing women accounted for 252 and children between two and four years for 60%. This compared to appraisal estimates of 43%, 29%, and 28%, respectively, for the three groups. This pointed to the difficulties of a differentiated treatment of beneficiaries,0Q although there were a number of special food items for infants.111 There were also severe quality problems which surfaced during the quality controls organized and financed under the project which however do not seem to have led to sanctions against repeat offenders. Since these controls were carried out at the point of production rather than distribution, ultimately the problem of sub-standard or contaminated food was probably more serious at the level of the consumers. Finally, there were references in supervision reports that families deliberately (possibly because of unfamiliar texture or flavor) removed added nutrients such as soybean oil from food thus reducing its nutritive value. 1/ Ministerio de Salud, op.cit. (footnote 3). 21 P. Pinstrup-Andersen, op.cit. (footnote 4). l0/ The initial issuing of separate coupons for infants was discontinued in 1980 because of cost considerations. Ill It was related that one of these foods was the object of alleged food coupon fraud in two departments (Supervision Report dated March 8, 1982). - 12 - 42. The diversity of these problems show that even a seemingly straight forward activity (whose main perceived difficulty prior to the start of the project had been to find an efficient mode of distribution)p12! could give rise to a host of complications which were neither easily foreseeable nor subject to prompt remedy. 43. It bears repeating that in this project measures to ensure safe water supply and adequate sanitation facilities appeared to have had a much greater positive impact on children's health status than direct nutrition supplements (PPAM, para. 32). This suggests that the measures accompanying the coupon program (water supply, sanitation, nutrition education, primary health care) were either per se more conducive to better child development than food coupons, or that the coupon program was Insufficient or did not reach its intended recipients in equal measure. Monitorina and Evaluation 44. Monitoring and evaluation was the one component where the gap between expectations and achievements was extreme, despite continuous efforts by supervision missions. Some of the reasons for this failure are evident, others less so. First, there can be little doubt that the project's complexity engendered complexity in its M&E componentt at one point it was reported that the regional PAN directors had to provide infor- mation on some 800 indicators every month;U2/ on another occasion, DNP field staff were said to have to make and report about 2,500 observations on a monthly basia,14/ clearly excessive workloads for staff whose main responsibilities lay elsewhere. Bank supervision missions succ*eded in cutting down on the volume of M&E data, but even then the size and complexity of the data compilation and analysis tasks must have acted as a deterrent for a thorough evaluation. 45. Second, project Implementation was characterized by frequent modificationst project components were added, deleted or restructured, implementation arrangements changed and physical input mixes reshuffled. While a measure of flexibility is a desirable feature in a "regular" project, lack of constancy is basically inimical to the concept of a controlled experiment. 12 This concern with delivery systems was so prominent that a pilot piogram to test alternative modes of coupon distribution was included in the project and backed by a covenant. It was however not implemented (PCR, paras. 4.20). 131 Supervision Report dated May 8, 1980. j41 Supervision Report dated July 2, 1979. - 13 - 46. Third, in a project where delays in physical Implementationg insufficient flows of funds and difficulties with interagency coordination were recurrent problems, project managers could not be expected to give more than intermittent attention to an exercise whose usefulness may not have been apparent to them. Here again, project complexity took its toll, albeit indirectly. 47. Fourth, from 1982 onwards, political support for the project was waning. It goes without saying that this development must also have weakened commitment to monitoring and evaluation. 48. However, these project-specific factors overlie a more basic problem with ME in Bank-assisted projects which has been observed t" other countries and sectors: presented in a nutshell, borrowers do not always seem to share the Bank's strong conviction that a systematic feedback of project experience is an essential element in the development process. "'%. Here again, the underlying reasons are many sided: urgent day-to- dry implementation problems besetting the project staff, an overambitious L Ltial M&E design which may have led field staff to believe that this was a largely academic exercise of little practical usefulness and rather for tie bene&fit of the Bank or, at the level of the policy-makers, a hidden concern that negative M results might diminish the chances for a follow- up project. 50. There are a number of steps the Bank could take to overcome or bypass these problems. First, in shaping MB components, it should make simplicity and practicality of design its guiding principle. This applies to the number and sophistication of indicators as well as to the coveraget in most cases, a thorough evaluation in one or a few typical areas would be preferable to a comprehensive evaluation based on data of uneven or uncer- tain quality. (For routine Implementation monitoring and reporting purposes, coverage would of course continue to be complete.) 51. Second, preparatory work for ME should be moved further upstream. The conceptual and administrative framework should be fully developed, and adequate baseline data, including a survey of intended beneficiaries, be available before the final stage of project processing. If necessary, this could be made a condition of negotiations. The financing of such upstream work should not present a major problems in most countries the Bank has now a variety of relevant financing mechanisms in place or can activate them at relatively short notice. Examples would be financing under an existing project or a technical assistance loanlcreditg a project prepara- tion facility; or retroactive financing. 52. Third, for innovative projects in which the Bank has a much higher stake, a larger extent of loanlcredit financing of ME work would be justified--fur this project the relevant disbursement rate was 552. 53. Fourth, individual Bank staff should make every effort to explain to their borrower counterparts the importance and usefulness of ME for the borrower country. By the same token, the Bank should reassure the - 14 - Government that negative evaluation results would not per se lead to a cessation of Bank lending to the sector but rather to a strengthening of future projects. 54. Finally, in cases where a timely feedback of evaluation results is crucial (e.g., for a mid-term review of an innovative project or component), the Bank might consider the use of a general or partial (component-specific) disbursement condition. IV. FINDINGS Proiect Content and Implementation Arrangements 55. These aspects have been discussed in detail in paragraphs 8 through 18. To recapitulate, the project design showed great complexity and the implementing agency was not in an ideal position to coordinate the activities of various line agencies. The transfer of the PAN/DRI program to the Ministry of Agriculture probably Improved the situation to some extent but occurred too late in the implementation process to provide a full test. Sustainability 56. Budget data for the health services suggest that the system, like other government services, is at present underfunded which, unless corrected, could affect the longer term sustainability of this component. A similar comment applies for the water supply systems the user charges (typically around Col.$300/family/month) are not sufficient to cover anything beyond routine maintenance and minor repair works. This problem is more vexing for small systems facing diseconomies of scale and ior pumping (rather than gravity) systems which are inherently more costly to maintain. As the infrastructure ages, repair and replacement needs will rise, and the question of adequate operation and maintenance funds wifl become more acute. On various occasions Bank supervision missions recom- mended the setting up of roving maintenance and repair teams at the depart- mental level but this proposal does not seem to have been acted upon. The upkeep of the systems appears to be left to local arrangements. 57. As to the various ad hoc programs, the issue is replicability rather than sustainability as these activities have already been discon- tinued. This question is of particular interest for the food coupon and home garden components. Successive supervision missions came to the conclusion (which is also reflected in PHN's 1984 Nutrition Review) that the coupon program was too costly to be expanded nationwide. Two possible responses to this problem would have been (i) a more restrictive selection of beneficiaries (e.g., setting the cut-off point for children's participa- tion at, say, age three instead of age four) and/or (ii) a greater use of cheaper unprocessed staples (instead of the more costly processed foods) in the food coupon basket. The early demise of the food coupon program precluded an exploration of these possibilities. - 15 - 58. The home Sasdez component was also effected by an early disconti- nuation, after having involved more than twice the intended number of bene- ficiaries. It Is not clear, though, to what extent this component reached the intended target population of the poorest rural familesa the range of activities and the sise of credits suggest an expansion of the component towards the somewhat better-off members of rural communities. 59. Yet the question of replicability is moot unless the effectiveness and cost-efficiency of these two interventions are established. Since the few available MEB results do not permit this, the project experience in these two areas must be considered inconclusive. Forward Linkaes 60. As related in para. 21, this project was eonw Oered by the PCR as a valuable traintng ground for later projects elsewhere- It also provided relevant experience to be applied to a subsequtut Health Project (Loan 2611-CO). Finally, the recently observed increasing attantion to nutrition Issues in Colombia opens the perspective of a future operation in the sector. The Bank is working closely with the Government on a comprehensive analysis of health and nutrition Issues. The findings of this review could form the basis for a possible follow-up project. - 17 - (Translation of Incoming Telex dated January 28, 1988) MR. GRAHAM DAALDSON CHIEF, AGRICULTURE, INFRASTRUCTURE AND HUMAN RESOURCES DIVISION OPERATIONS EVALUATION DEPARTMENT WORLD BANK, WASHINGTON D.C. Us PROJECT PERFORMANCE AUDIT REPORT ON THE INTEGRATED NUTRITION DEVELOPMENT PROJECT (LOAN 1487-CO) WE APPRECIATE YOUR SENDING THE ABOVEMENTIONED REPORT. WE CONSIDER YOUR COMMENTS VERY MUCH TO THE POINT AND IN SOME ASPECTS THEIR RELEVANCE EXTENDS TO PROJECTS WHICH ARE STILL BEING IMPLEMENTED, FOR EXAMPLE: (1) THE INSTITUTIONAL LIMITATIONS OF AN AGENCY IN THE ROLES OF COORDINATOR/ SUPERVISORI (2) THE PROBLEMS RELATED TO AUDIT REPORTS; AND (3) THE BASIC PROBLEM OF MONITORING AND EVALUATION OF BANK-ASSISTED PROJECTS. WE ACKNOWLEDGE THE FIRM CONVICTION OF THE BANK THAT A SYSTEMATIC FEEDBACK OF PROJECT EXPERIENCE IS AN ESSENTIAL ELEMENT OF THE DEVELOPMENT PROCESS. AS REGARDS THE ASSESSMENT THAT PROJECT IMPLEMENTATION WAS AFFECTED NEGATIVELY BY A LESS EFFECTIVE SUPPORT FROM THE NEW GOVERNMENT IN 1982 WE WOULD LIES TO STATE THAT IT IS A UNIVERSALLY ACCEPTED PRINCIPLE THAT EVER! ADMINISTRATION DURING ITS TERM IN OFFICE TAKES ON RESPONSIBILITY FOR THE PROGRAMS AND PROJECTS AND ANSWERS FOR THEIR IMPLEMENTATION FROM THE MOMENT OF ITS INAUGURATION. IN SPITE OF THE LARGE VOLUME OF INDICATORS AND OF THE DIFFICULTY OF EVALUATING MULTIPLE VARIABLES, WE ARE CERTAIN THAT LOAN 1487 HAD A SIGNIFICANT REDISTRIBUTIVE IMPACT ON THE POPULATION IT TRIED TO BENEFIT AND THAT THE EXPERIENCE ATTAINED THROUGH THIS PROJECT WILL BE VERY VALUABLE FOR FUTURE PROJECTS OF THIS NATURE. CORDIALLY, JORGE BUSTAMANTE ROLDAN GENERAL MANAGER, FONDO DRI BOGOTA, JANUARY 28, 1988 - 18 - COMMENTS FRO THE BORROHER ATTACHMENT II (Translation of Incoming Letter dated February 8, 1988) Republic of Colombia Ministry of Health Bogota, D.E. General Secretariat No. 04164 February 8, 1988 Mr. Graham Donaldson Chief, Agriculture, Infrastructure and Human Resources Operations Evaluation Department World Bank, Washington Dear Mr. Donaldson: With reference to the Project Performance Audit Report for the Colombia Integrated Nutrition Improvement Project (Loan 1487-CO) I would like to make the following comments: One can not regard the introduction of food distributionlmarketing as a negative aspect of the project. It would be important to analyse it in detail and evaluate its actual effects. In the process of project appraisal the aspect of nutritional impact was omitted because the necessary quantitative indicators were not included. Although they were obvious the failure to take them into account from the beginning suggests that future projects should be ready to deal with this subject. Project complexity, because of its constraining character, ought to be an element to be mindful of in future projects in the sense that all activities and efforts should concentrate on fundamental aspects, aiming to make complementary elements form part of the natural functioning of the entities. There seems to be some conceptual contradiction with regard to the project's characteristics: while the Bank looks at them as inputs into an analytical exercise, the Borrower considers them as contributions to the solution of real problems, resulting in different approaches and criteria for an assessment. Thank you for your attention. Cordially, /sgd/ German Perdomo Cordoba Secretary General - 19 - COLOBIA INTEGRATED NUTRITION IumROvEINT PROJECT (LOAN 1487-CO) PROJECT COMPLETION REPORT August 18, 1986 Population, Health and Nutrition Department (PVND3) - 21 - COLCMBIA INTOGRATED NUTRITION IMPROVEEHNT PROJECT - LOAN 1487-O PROTECT COMPLETION REPORT I. 1ITROUCTION 1.01 The Colombia Integrated Nutrition Improvement Project approved in September 1977. was the third Bank-assisted project in the nutrition sector. The Bank decision to lend in this sector was part of the basic needs strategy adopted in the early '70s and was taken on "the ground that a well- directed nutrition investment can lower mortality and morbidity rates, contribute to increased productivity, assist in achieving family planning and education objectives, serve as a direct means of income redistribution and, in general, improve the social and economic conditions of disadvantaged portions of the population". 1.02 A nutrition project seemed particularly opportune in Colombia because of the persistence and extent of malnutrition (especially in the rural areas), the Colombian leadership's strong commitment to addressing the problem, the adoption of an innovative National Food and Nutrition Plan (PAN) for combatting malnutrition in a truly integrated fashion, considerable Colombian human and institutional resources in the nutrition field, atd the imminent phase-out of the nation's external food aid program. 1.03 Preparation of the project began in February 1975. and the Board approved the loan in September 1977. The borrower had desired to process the loan more rapidly, but substantial delays occurred during project preparation. Following several preparation missions and two appraisal missions, negotiations were conducted in July 1977. 1.04 The Loan Agreement was signed on November 10. 1977. Total project cost was estimated at US$68.9 million over four years. Financial provisions were a US$25 million Bank Loan (Loan 1487-CO) to cover both foreign and local costs. 1.05 The loan became effective March 9. 1978. The project was completed in June 30, 1985. after two extensions of one year and two years respectively, and a reprogramming of activities in 1983. The loan was fully disbursed against actual total project cost of US$87.1 million equivalent estimated by DRI-PAN (Integrated Rural Development-National Food and Nutrition Program). 1.06 The main project goals were: (a) improving the nutritional status of targetted population groups; (b) systematic monitoring and evaluation to ensure sustained project effectiveness; and (.) strengthening Colombia's managerial base in the nutrition sector. -.22 - 1.07 The project was prepared during the period when the Bank was starting to lend in the nutrition sector. Consequently, project design was influenced by staff's recognition that "nutrition iLa a new field for the Bank and many possible policy questions have not yet been addressed." Equally important influences on project design were the strong country commitment and desire to address the diverse causes of malnutrition in an integrated manner. The project was complementary to an existing Bank Loan for integrated rural development, also being implemented by DRI-PAN. 1.08 The project was initiated at a time when Colombia's commitment to nutrition was strong, and economic conditions were favorable. The nutrition program was a top priority within the 1974-78 Development Plan and was an important one in the subsequent national development strategy for 1978-82. In 1982, the election of a new President with a more conservative approach to social sector spending, and the continuation of an economic decline that had begun in 1980. brought ebout a policy reversal for nutrition and for subsidized activities in particular. 1.09 By July 1983 most project targets had been met (and in several cases substantially surpassed, see Table 1). and government expenditures for nutrition and related work over the previous four years were well in excess of original projections. In late 1983, due primarily to the US$8 million undisbursed loan balance, accumulated through DRI-PAN's failure to demand documentation of expenditures by implementing agencies and request reimbursement, the project was extended for another two years. At this time the loan was reformulated at Government's request, reflecting the new Government's belief that the nation's malnutrition problems could be more effectively addressed through rural development and increased production of low-cost food than through specific nutrition programs. The Bank agreed to this reformulation and extension, despite misgivings over diluting the loan's original nutrition focus, both because an undisbursed $8 million loan balance remained, and because the new activities proposed were consonant with the project original goals. 1.10 Much was learned about the advantages and pitfalls of implementing multi-sectoral nutrition projects under contrasting political and economic circumstances, the management of complex projects, the use of multiple and secondary Implementing agencies, and the Bank's own potential and limits. II. PKMC IDE ICATION. PREARAIO AlD APPRAISAT A. Project Origin 2.01 In the early 1970s. gratification over rising GNP in many developing countries was tempered by growing recognition that economic development was by-passing large numbers of people. Still caught in the cycle of poverty, unemployment, malnutrition and disease, the "poorest of the poor" were neither benefitting from nor contributing to economic growth. - 23 - With the aim of strengthening both production and equity goals, the Bank decided to complement its traditional lending for infrastructure investment with loans designed to meet basic needs of the very poor. 2.02 Within this context and recognizing that "malnutrition in low- income countries is a problem of substantial proportions, with significant and wide-ranging developmental consequences, "the Policy Guidelines for Back Nutrition Activities were adopted. 2.03 At the time, however, data availability was limited. Most countries lacked information about food consumption. spending patterns, nutritional status of households and degree of malnutrition. The impacts of public feeding programs were largely unknown. Faced with this information gap, the Board encouraged the Bank to finance cert.n high priority experimental actions and their evaluation. The hope was that through "learning by doing" in a few selected countries, the Bank would help to meet immediate needs, while simultaneously laying the foundation for larger-scale activities in the future. 2.04 Several factors created promising opportunities in Colombia. The administration that assumed power in August 1974, had developed a broad. highly innovative and carefully planned strategy designed to become the world's first full fledged multi-sectoral effort for combatting malnutrition. It included a nutrition program (PAN) and an integrated rural development program (DRI). In addition, the country had relatively strong human and institutional resources in nutrition, as large-scale nutrition programs had operated in Colombia since 1954. Despite these previous efforts, however, and the fact that Colombia's aggregate availability of calories per capita appeared adequate, malnutrition remained widespread among the poorest 20% of the population. Concern about the persistence and extent of malnutrition was heightened because Colombia's external food aid - then the third-largest in the world (after India and Egypt) - was due to be phased out in 1978. 2.05 There was great interest in a Colombian nutrition project at the highest levels of the Bank and the Government. Accordingly, in January 1975, an identification mission visited Colombia to explore the Government's request that the Bank finance components of the Nutrition Program, along with the Integrated Rural Development Program, both under PAN. Initial consideration was also given to include a major family planning component* but the GOC deemed this approach too politically sensitive. B. Project Preparation, Appraisal. Negotiations and Appproval 2.06 This project was an integral part of Colombia's overall National Food and Nutrition Plan (PAN) and the Government of Colombia assumed primary responsibility for project preparation. The process was coordinated by the Social Development Unit of the National Planning Department (DNP). in close collaboration with Bank staff and consultants. - 24 - 2.07 Preparation began in February 1975, and during that year was assisted by four Bank missions. By the end of the first year of preparation, eight project components for Bank financing had been identified under PAN's integrated, multi-sectoral attack on malnutrition: (1) subsidized food distribution to lactating mothers and to children under five: (2) production of processed foods with high nutritive value; (3) quality control of food processing; (4) construction of primary care health facilities; (5) training of para-professional workers to provide basic nutrition and health services; (6) safe drinking water to communities of 2.500 persons or less; (7) nutrition education (mass-media, inter-personal and professional training sad (8) the creation of an organizational structure for project manasement and administration. 2.08 Nevertheless, due to the size, complexity and innovativeness of the project, inadequate inter-agency coordination, and changes in the Bank's own approach to nutrition while the project was being prepared, it took longer to move from the conceptual to the operational stage than the borrower or the Bank had expected. Suchdelays should have been anticipated, given the Bank's experience with other projects of similar scope and complexity. However, the GOC feared that further delay would endanger implementation of the Nutrition Program under the Administration in power at that time. 2.09 In this context, Colombia urged the Bank - and the Bank agreed - to send a pre-appraisal mission in February 1976. with the hope of completing all the necessary preparation work. An appraisal mission followed in March 1976, but a number of unresolved issues prevented the completion of appraisal until the visit of a second appraisal mission in January 1977. 2.10 During the preparation and appraisal process, the Bank repeatedly raised a number of sulstantive and procedural issues. Substantive issues included the heavy burden of recurrent costs (especially in the food and health components); the importance of developing quantified objectives and indicators of effectiveness; the need to refine the selection of beneficiaries so as to assure the highest possible coverage of the neediest 30% of the population; and the necessity of close monitoring and evaluation. including an appropriate evaluation methodology and staff to permit timely project adjustment in light of operating experience. Procedural concerns focused on operation and management, due to the greater than usual need for strong coordination among implementing agencies; among national, regional and local officials; among different sectors and project components to achieve the mutually reinforcing effects expected from this multi-sectoral program. 2.10 Many of these issues were resolved during appraisal, enabling the second appraisal mission to conclude that "the long Bank dialogue with Colombia has resulted in serious and sustained Government efforts to improve t7re organization, management content and coordination of its nutrition prcgram." The mission further concluded that "constructive extension of that dialogue can best take place by shifting to project implementation - 25 - since preparation is as complete as possible, given the innovative, complex and comprehensive nature of the Colombia nutrition effort." 2.11 At the same time, there was explicit recognition that the untested features of the project posed higher than usual risk, and that consequently this loan would involve the Bank in unusually intensive supervision. 2.12 Government and Bank consultation with external aid agencies in Colombia ensured that the project would complement investment commitments and assistance to the nutrition program already made or scheduled by the U.S. Agency for International Development (USAID). the United Nations Development Program (UNDP), the United Nations Childrens' aund (UNICEF). the Pan American Health Organization (PAHO) and other donors. 2.13 The loan was approved by the Board in September 1977, and it became effective on March 9, 1978. Total project costs were estimated at US$68.9 million over four years. Financial provisions were a US$25 million Bank Loan (Loan 1487-CO) to cover foreign exchange costs and US$12.1 million in local costs. Project beneficiaries would provide direct financing totalling approximately US$4.8 million including contingencies. The Government of Colombia would finance the remaining US$39.1 million. C. Pr iect Targets and Goals 2.14 The Appraisal Report set forth three basic goals for the proposed nutrition project: "(a) to improve nutrition. living standards and the productive capacity of 1.8 million Colombians now living in absolute poverty in seven selected regions and the Federal district of Bogota; (b) through systematic monitoring and evaluation, to enable the Government to refine PAN and ensure its sustained effectiveness and financial feasibility as coverage expands during the project and thereafter, and (c) to strengthen the managerial base for planning and executing the nutrition program." 2.15 By the end of the project period, the Bank anticipated around a 25% reduction in mortality rates for children under four and around a 502 reduction in the incidence of first, second and third degree malnutrition in that age group among project beneficiary families. 2.16 The Report noted that "these goals are ambitious but probably could be met substantially where the proposed project operates as scheduled." 2.17 The Appraisal Report enumerated a range of benefits to be derived from the project. Immediate returns would be the increased productivity of economically active persons through reduced numbers of sick days. - 26 - translating into a potential annual increase of US$2.4-$4.8 million in family incomes for project beneficiaries. Another short-term benefit would be the increased efficiency of health spending and reduced beneficiary expenses for curative drugs and health care due to better preventive medicine. It was estimated that a 5-10% reduction in family health care expenses would save an additional US$0.5-$1.0 mi.11on for those covered by primary health care services. 2.18 The intermediate term benefits included: improved availability and utilization of existing food through better processing, storage and transport; the likelihood that improvements in infant and child mortality would lead to greater acceptance of family planning efforts; and reductions in the per capita unit costs of schooling due to fewer dropouts. less grade repetition and more efficient use of existing facilities by a more responsive student population. 2.19 The major long-term benefit was expected to be "the improved quality, vigor, and productivity of the future work force. Reduction in the rate of manintrition and serious disease should result in better mental and physical growth, and consequently a better educated and more able adult population." Admittedly such gains would be difficult to measure; however, provisional calculations indicated a minimum annual return of US$10-13 million exclusive of lower mortality benefits. D. Project Description 2.20 The project at appraisal consisted of a combination of action programs and supporting activities affecting food demand, supply and utilization. The action programs were directed primarily at populations in the target communities, while supporting activities we,uld take place nationally. The Appraisal Report described the following components: a) Health Servicess The project would finance the strengthening of the primary health care system as a mechanism of integrated delivery of nutrition/health services including nutrition education, coupons for subsidized food supplements, safe drinking water supplies, and latrines. The 1.8 million beneficiaries (at 90% coverage) would include around 573.000 pregnant and nursing women, and preschool children at particularly high nutritional risk. Project funds would be provided to construct 56 and renovate 225 community health posts in rural and urban areas. The project would also finance training for around 2.000 health workers and 677 auxiliary nurses and promotoras to staff project health posts. The new health system represented a major departure from previous Government health programs which bypassed most of the rural poor and emphasized curative services provided by physicians ratler than preventive care by paramedical personnel. b) - ater Supply ad nvirenmnntal Semitatief: Project area population with access to potable water would be doubled through the construction of 360 rural water supply systems serving 231,000 persons in 372 small communities also served by health posts. Installation of the - 27 - rural water supply system would be carried out by the National Institute of Health (INAS), a decentralized semi-autonomous agency funded through the Ministry of Health (MOH). Benefitting communities would contribute at least 10% of the initial investment cost, and would operate and maintain them principally with local resources. The component would also include installation, by the MOR, of 112.000 household latrines serving 627,000 persons. Benefitting families would provide at least half the estimated total cost of each installed latrine (US$33) in labor, cash or materials. c) Nutrition Education: This component was designed to improve nutritional habits, particularly breastfeeding, better weaning habits, and nutritious, low-cost food choices. Mass media financing would support the development, production, pretesting and dissemination of radio messages, nutrition films, and a television program on nutrition. In addition, the component would finance training in nutrition education for health personnel, community leaders, and others involved in nutrition education campaigns, as well as production of educational materials, equipment. monitoring and administrative costs. d) Pamily Gardens: This component would support a pilot effort to increase home production and consumption of nutritious vegetables and legumes by 20.000 families. It would include demonstration centers at 100 rural schools; the development of technical packages of inputs for home garden plots; training and materials for extension agents and teachers; and salaries for the extension agents. e) Applied Food Technology and Quality Control: This component would improve the use of available food in low-income areas, and stimulate production of new varieties of low-cost enriched processed foods. The Institute of Technological Research (ITT) would undertake research and development studies, pilot projects and feasibility studies in the manufacturing, processing. packaging, storage, marketing and fortification or nutritious foods. The Food Quality Control Sub-component would finance Government installation and operation of a system to ensure that enriched foods produced by commercial manufacturers conformed to prescribed quality standards. f) ionitoring and Evaluation: This component was designed to strengthen the government's capacity to monitor, evaluate and refine the national nutrition program through financing of studies, surveys, and the establishment of an evaluation data system. Monthly and quarterly monitoring reports from the regions would form the basis of DNP's semi- annual reports to the Bank, and would provide project management with early warning signals of problems in individual components. Evaluation would include semi-annual baseline surveys of the nutritional status of target groups. Special studies would evaluate the community health system, the management effectiveness of the nutrition delivery system, and project impact on worker productivity, employment and income. The Monitoring and Evaluation unit would also conduct an intensive four-year evaluation of the nutrition delivery system in one low-income region. Bolivar, with particular emphasis on the food coupon program. The Bank would not disburse against - 28 - the subsidy element itself, but would finance an evaluation designed to determine the system's nutritional impact, its operational problems, and the relative cost effectiveness of different variants. Because all elements of the nutrition program would operate in Bolivar, the "Bolivar Test" was expected to yield particularly fruitful data for evaluation and course correction before PAN expanded to national coverage. g) ProrM Coordinmtion and Administration: In order tc strengthen the managerial base for the national nutrition program, this component would finance the salary and operating costs of the PAN group within DNP, which was set up in 1976 to coordinate, evaluate and refine the national nutrition program and prepare plans for its further implementation. b) Technical Assistance: The project would finance a total of about 45 person-years of consultant services to assist in implementing the various components, as well as about 10 person-years of fellowhips in nutrition education and applied food technology. L) fo Distribtion and Maketing: This component was added when the loan was reformulated in September 1983, largely as a replacement for the discontinued food coupon and home gardens components. Designed to lower the cost of the family food basket, particularly for urban consumers, this "commercialization" component included: modernization of the retail sector by improving the packing and supply of staple foods to the small retailer serving the urban poor; training and technical assistance for at least 80 small food retailers (tenderos); a pilot project in Cali of urban warehouse distribution to small food retailers; six feasibility studies for urban food marketing projects; and technical assistance to small and medium- sized industrial food producers in five Colombian cities. The new component also included plans for associations of small retailers (proveedurias) in five major cities. An average of six mobile markets in which wholesalers assemble weekly at various sites were planned in each of five major cities. The component included construction of central marketing facilities in these cities. Initial baseline studies, assessing marketing conditions, food prices and consumer needs, were to precede initiation of the proposed activities. The project planners asserted that the reductions in food prices to be achieved through the combined effects of the commercialization activities would equal or exceed the benefits to beneficiaries previous.ly obtained through food coupons and PANCOGER. I. PRDJBCT IMR.mRRMA1ram All OSf A. Effectiveness and Start-up 3.01 There was some delay in start-up because of presidential elections in August 1978. that created uncertainty and personnel turnover, though key project ztaff remained unchanged. Economic conditions also became less favorable. The revenues realized by the Government's tax reforms amounted - 29 - to only a fraction of what had been expected. The coffee bonanza came to an end; although it had significantly improved the country's balance of payments situation, it also accelerated inflation. 3.02 Despite these problems, the new administration 4it maintain previous nutrition policy and program commitments, and toward the end of the first project year the pace of execution picked-up. The major problem was slovaess in the organization of some components, due to the lag in disbursements that would continue to plague the project. Most components were proceeding satisfactorily, the evaluation component was singled out for especially high praise, and the leadership of the nutrition program was deemed outstanding. 3.03 In 1980 PAN and DRI were merged. In response to strong expressions of concern, the GOC assured the Bank that the merger would not weaken or divert the project's central emphasis on meeting the nutritional needs of targetted beneficiaries. Soon thereafter, the newly integrated program was extended nationwide. 3.04 These dievelopments had two important consequences. First, despite prior assurances, the merger of PAN and DRI rather quickly shifted the project's primary emphasis from consumption to production, in response to DRI's more politically powerful constituency. Second, the extension to additional departments was premature; in retrospect, it would have been preferable to implement the program's on-going activities fully in municipalit-lis already included but where coverage was still incomplete. 3.05 Still, a year-end supervision mission in 1982 was able to conclude that "the project had been relatively successful over its first four years." Several important achievements were noted, particularly the project's catalytic effect on the provision of health care. Performance in evaluation and reporting, however, had fallen far short of its initial quality. preventing a clear view of project impact in either quantitative or qualitative terms. The mission recomiended to extend the loan for an additional year and the Bank accepted the mission's recommendation. B. Revisions 3.06 In August 1982, a conservative President took office and the DRI/PAN program was significantly reoriented under a new director. Strongly opposed to food subsidies, the new administration rapidly phased out the food stamp program (not financed under the loan) and the PANCOGER home garden component. They were replaced with a food distribution and marketing component designed to reduce the cost of the family food basket, primarily for urban consumers (see paras. 4.15-4.17). In 1984 overall responsibility for PAN/DRI was transferred from DNP to the Ministry of Agriculture (MINAG). This organization change reflected the government's increasing priority on rural development and food production, and its corresponding shift away from the nutrition objectives in the original project. In 1985 PAN/DRI became fondo DRI, a semi-autonomous agency located within the Ministry of Agriculture. - 30 - 3.07 In 1983 the GOC requested a two-year extension of the loan. The case for extension was two-fold. First, although physical targets had been largely met and Government counterpart funds had exceeded their original target, accounting problems and related disbursement shortfalls left a loan balance of $8 million. Second, the prevailing view was that the government's proposal constituted an important agenda of nutrition and health activities consonant with the loan's original objectives. Despite remaining concerns that the recent substantive and organizational changes would likely dilute the project's original nutrition focus, an extension was approved in September 1983 - in the hope of implementing the project more fully, maximizing the lessons learned regarding nutrition interventions. advancing the policy dialogue in nutrition, and consolidating the foundation for a subsequent health project. 3.08 Most of the revised project components were completed on time. Despite PAN's absorption into DRI, the erosion of the Government's nutrition priority, the shift of emphasis from consumption to production, and the dilution of focus on the most nutritionally vulnerable population. considerable nutrition-related activity was conducted during the project's final two years. C. Implementation and Disbursements Schedules 3.09 Slow disbursements from the GOC to the executing agencies was a significant managerial problem throughout the life of this project. These disbursement lagi resulted in slowing the level of resource transfer to Colombia. hampering project implementation by reducing the liquidity of the executing agencies, and undermining project management's authority in relation to these same "line" ageozies. The root cause was that PAN was being run by the planning agency rather than by a line mi:nistry. DNP's unsuitability for managing and monitoring project activities implemented by numerous entities in eight different jurisdictions, as well as high turnover of personnel at project headquarters, were all compounded by Colombia's excessively bureaucratic and cumbersome procedures for transferring funds in the public sector. The Bank's two concurrent Colombian projects encountered equally serious disbursement problems. 3.10 This problem was identified as a potential bottleneck at the time of appraisal, and the Bank organized training seminars in an effort to prevent its occurrence. To expedite matters further the Bank agreed to approve disbursements for under US$5000 as statements of expenses. Even so, project management never fully succeeded in matching reimbursement requests to expenditures. As a result, after all possible withdrawal applications had been processed by the end of the fifth year, a loan balance of $8 million remained. D. Raporting 3.11 Reporting was also a chronic problem, due to the absence of a functional management information systim for the project as a whole, and the - 31 - high turnover of staff responsible for project management. Headquarters staff had difficulty aggregating. analyzing and acting on the voluminous information received from the field. Project management attempted to correct the situation, but poor technical assistance, delays, high staff turnover, and changes in the political environment limited the results. 3.12 Audit Reports. The project had ten implementing agencies that were required to present annual audit reports through Contraloria General. There were generally long dela)a in presenting audits throughout the project's life. The Ministry of Health. FONADE (Fondo Nacional de Proyectos de Desarrollo), CAR (Corporaci6n Aut6noma Regional de las Cuencas de los Rios BogotA. Ubate y Suarez). and DRI-PANIDNP had difficulty following satisfactory accounting practices which caused the general auditors to abstain from an opinion on various occasions. Because of poor accounting. many expenditures remained undocumented and hence ineligible for reimbursement, and 55% of the loan balance remained undisbursed by the loan's original closing date. The general auditors made specific recommendations to improve accounting and provided training to improve professional standards. By project completion, substantial improvement had been made in accounting systems, especially in the Ministry of Health at the Regional level. E. Procurement 3.13 Project procurement was always prolonged due to the cumbersome bureaucratic procedures required by the Colombian Government. Approval to import equipment was delayed by INCOMEX in certain instances due to import bans during economic austerity periods. Most procurement was done locally. Minor problems were encountered with Bank Procurement Guidelines regarding the maximum cost for which local procurement was permitted in the case of pumps and pipes conflicting with local procurement regulations, mainly because of the large number of suppliers in Colombia. F. Project Costs 3.14 Total project cost estimated in constant prices is US$87.1 million equivalent, or 126% of the appraised cost. Tables 3 and 4 summarize actual and appraised project costs by component and expenditure categories. These are not absolute cost overruns, as the final project implemented additional activities and quantities not envisaged at appraisal. G. Financial Sources 3.15 The Bank contributed 29% of total project cost, and the Government the remaining 71% of costs. The community participation financial contribution envisaged at appraisal for health, water and sanitation works was not quantified but was substantial. Obtaining the approved budgetary funds was difficult for DRI/PAN. especially during strict economic austerity measures. Budgetary allocations were a problem after the second year of implementation, and became critical during the final two project years. - 32 - H. Borrower's Performance 3.16 In terms of the objectives and commitments in the Appraisal Report, the borrower succeeded rather well with regard to physical implementation of component activities. Many targets were met or exceeded (see section IV. Project Impact). and the GOC allocated substantially more financial resources to the program than planned. 3.17 Borrower's performance fell short in the areas of coordination. monitoring, evaluation, and reporting and auditing of accounts throughout the project's duration. The underlying reasons, discussed in greater detail in Section IV, include the failure to develop a functional reporting system. personnel changes and organizational shifts contrary to covenants in the loan agreement. rivalry between DNP and certain of the implementing ("line") agencies and, as economic conditions worsened and political priorities changed, disenchantment with smaller and partial budget approval, and administrati- e difficulties of implementing a complex multi-sectoral nutrition project for the poor. 3.18 PAN represented a bold step on the Borrower's part. At the time of effectiveness the program enjoyed outstanding leadership and was supported by an able, dedicated and enthusiastic staff. Frequent staff turnover weakened program management over time. Nevertheless, many aspectZ of the project represent substantial accomplishments and the borrower deserves credit for breaking new ground in the nutrition field. IV. PRTZCT PRONCB AND IMPACT 4.01 The Colombia nutrition project fell short of the goals described in the Appraisal Report, and seldom met its prescribed timetable. However. assessments outside these parameters suggest that implementation, though slower than originally planned, created useful health and nutrition infrastructure and achieved important sectoral outcomes. A. Nutrition Interventions 4.02 Primary HeLth Care. The project had a catalytic effect on the development of primary health care in Colombia. Because the nutrition loan placed high emphasis on extension of the nascent primary health care system, emphasized inclusion of nutritional services in it, and allocated substantial funds to it, the health component generated impressive impacts. In fact, the prestige of DNP and PAN helped to maintain both the funding levels and the preventive, primary-care orientation of the new health system in the face of professional pressures to favor the curative, hospital-based approach. - 33 - 4.03 The project resulted in the construction and improvement of 729 primary health care posts - 259% of the original appraised target of 281. It was estimated that 6.490 persons benefitted by each post, or 101% of the appraised target. 4.04 Not surprising in light of implementation experience worldwide, however, the Government found it easier to build and remodel health posts than to staff and manage them effectively. Moreover, financial stringency in Colombia reduced project impact by causing difficulty in the funding of recurrent costs at new primary health care facilities. Nevertheless, the nutrition project's emphasis on linking nutrition and primary health care resulted in considerable improvement over earlier MOR efforts. Food coupons were distributed by the promotoras which increased their acceptance by the beneficiaries. The project trained 1.368 new promotoras. or 67% of the number estimated at appraisal. As a result of the program, all promotoras are now better trained in nutrition-related services, including growth monitoring, and MCH services include greater stress on breast feeding and oral rehydration. 4.05 The nutrition project also helped the MOH to establish an adequate monitoring and eveluation system. Though far too complicated and, hence. slow to come into operation, the system did enable MOH to document construction, staffing and equipment of facilities. It also generates regular data on use of services and is now moving toward regular monitoring of nutrition status and incidence of disease. 4.06 Through the nutrition project, the Ministry of Health was also able to develop plans that served as the basis for Colombia's first Bank health project, which was approved in July 1985. 4.07 Nutrition activities envisioned under primary health care could not be fully implemented due to the continued rivalry between the Ministry of Health and ICBF, the agency with primary responsibility for nutrition. This rivalry, hampered implementation of the mass media component and impeded the Ministry's ability to make full use of ICBF's capacity to provide technical assistance and training related to nutrition services. It also impeded further nutrition interventions under the new health project. 4.08 Water Supply and Sanitation. In financial terms, this was the largest component of the project (27% of project cost). Implementation lagged in the early stages, but by completion all funds allocated to this component were spent and disbursements claimed. A total of 326 water supply systems were built, or 88% of those programmed; 35.000 latrines had been installed, representing only 31% of the original target of 112,000. The communities upheld their commitments to finance at least 102 of installation costs, although community maintenance of water systems still presented problems and INAS was obliged to supplement community funds to perform major unexpected repairs. - 34 - 4.09 Nutrition Education. The mass media part of this component, administered by the DNP, was off to a strong start in mid-1978 with a diarrhea information campaign. A second campaign in 1979-80 to promote breast feeding was particularly well planned and implemented, and a survey conducted nine months after the campaign began showed a significant change in attitude toward breast feeding among target families. In addition, these campaigns helped to change prevailing dogmas within the health community itself concerning breastfeeding. oral rehydration and appropriate food choices. It was disappointing that the impact evaluation for this component was designed but was never carried-out. 4.10 Production of mass media materials was well above targets, and the techniques of message pre-testing were introduced to Colombia for the first time. Radio announcements made under the project came to 251% of appraised targets. Unfortunately, management failed to focus sufficiently on promoting behavioral change within the target population. Nevertheless, the 302 messages prepared by DRI-PAN and broadcasted daily, plus posters. TV documentaries, booklets, and other audio-visual efforts of the component constituted a useful package of nutritional and other consumer guidance. Following reformulation, however, the Mass Media Division's role as the public relations arm of DRI-PAN reduced substantially the education activities addressed to nutrition improvement contemplated initially in the loan. 4.11 Under the second part of this component, administered by the Colombian Family Welfare Institute (ICBF). training was provided for 30.996 agricultural extension workers, health workers, nurse auxiliaries, home economists, teachers, shopkeepers and others. This figure is 386% of the appraised target of 8,028 trainees. Except for the breast-feeding campaign. the coordination of the training and the mass media effort, and the relations between the agencies responsible for them, were poor. B. Development and Production of Low-Cost Nutritious Food 4.12 Family Gardens. Incorporated into the project at the urging of the Bank's Board of Directors to strengthen the project's productivity orientation, this component became much larger than foreseen in the Appraisal Report. What was planned as a modest program for 20,000 families expanded into a small-scale rural development effort. It proved very popular and was gradually broadened to include livestock, fish farming, and model sanitary homes. The program, known as PANCOGER. eventually reached over 45,700 families, or almost two and one half times its appraised objective. Although clients included many small farmers traditionally outside the reach of previous credit schemes, a 1982 evaluation reported that 89% of all recipients kept up on their payments. The impact on nutritional status has not been evaluated. Because this component had gone so far beyond the original objectives, because its nutritional focus was diluted in the process, and because the subsidized interest rate of 8Z conflicted with the government's policy, the program was suspended at the end of 1982. -35 - 4.13 Applied Food Technology and Quality Control. Under this component a network of eleven regional quality control laboratories was developed, to serve as quality control support for the food coupon program. Following that program's termination, the laboratories became responsible for analyzing all new food products offered for legal registration. All procurement was completed and the network was deemed to be coordinated effectively. 4.14 The Institute for Technological Research (IIT). also originally focused on the coupon-subsidized products, conducted nutritionally important investigations, financed by the loan, while simultaneously responding to expanded DRI requirements over time. The Institute completed 43 studies, as planned, involving a range of food processing and packaging improvements for the output of DRI producers. The loan's contributions to the strengthening of IIT and the quality control laboratories, as well as its outreach to the food industry, illustrate institution-building impact important for nutrition and broader concerns. 4.15 Food Market (Camercialisation). This component implemented two new strategie? added to the project with its 1983 reformulation. They were: (a) mobile markets in four cities, aimed at lowering prices by eliminating intermediaries; and (b) small shopkeepers' associations to improve their efficiency, product quality, and in the future, pricing. 4.16 Mobile markets were very popular among middle-class consumers, some small retailers, and others who could afford to buy in bulk at these weekly wholesale markets. However, mobile markets failed in their primary objective of increasing food consumption among the urban poor, largely because the poor could not mobilize enough purchasing power to make the markets sufficiently profitable for the wholesalers. The markets lowered prices for vegetables, fruits, and dairy products substantially while other products that make up a larger part of the food basket were slightly higher (see Table 2). The markets were well run by the wholesalers, with support from DRI-PAN. Consumer and nutrition information was displayed, including all prices announced for that week. Prices were fixed weekly and published in major newspapers. However, the fact remained that the wealthier the area, the more profitable and more attended the market. In Bogota for example, 62% of the markets are located in middle- and high-middle-income neighborhoods, and only 3% in low-income neighborhoods. In Cali, the markets have been much more successful in benefitting the poor, with 61% of the markets located in low-income neigborhoods. In the past two years, however, nine markets in poor areas of Cali closed because wholesalers found it easier and more profitable to locate in higher income neighborhoods. Total share of these markets is still very low in terms of total Colombian foods sales (2.5%). Anecdotal evidence suggests that the poor, in cities other than Bogota, are benefitting somewhat from downward pressure on retail food prices, thanks to mobile markets. 4.17 The small shopkeeper's program tried to reach the principal food suppliers in Colombia - small shopkeepers move about 70% of food sales. Small shopkeepers are also the suppliers of the poorest, and provide credit - 36 - to their clients. DRI-PAN coordinated training, credit, and technical assistance in organizing associations and establishing warehouses per association. Shopkeepers have welcomed the program and have benefitted from training in marketing efficiency, accounting, and general organization. Five hundred and fifty-seven shopkeepers received training and 17 chains were formed in 17 cities where they continue to receive technical assistance from Fondo DRI. In the long-run. Fondo DR1 plans to have stricter price controls and to develop a separate program based on a food basket containing "inferior goods" (that is, products with negative income elasticities of demand) for sale to the very poor. C. Monitoring and Evaluation 4.18 From the earliest stages of preparation, the Bank placed unusual emphasis on monitoring and evaluation because several project components were innovative and untested. Successive missions stressed the importance of deiining koy indicators of physical and financial progress, service coverage and quality, and incorporating them into a format which would allow project management to make speedy course corrections as required. The original project leadership shared the Bank's evaluation goals and recognized their relevance as a management tool. 4.19 Unfortunately, this component never lived up to its promising start. At the outset data were satisfactorily collected and transmitted. and the system could have been further improved through technical assistance under the loan. Unfortunately, the GOC's commitment to systematic monitoring and evaluation was progressively diluted with PAN's absorption into DRI, staff and administrative changes, and the erosion of nutrition's preferred position. 4.20 The promised baseline studies and other data-collection efforts proved very inadequate, in part because of their failure to utilize recent developments in household consumption surveys and analytical methodologies. Another limiting factor was the lack of specific attention to the development of nutritional status surveillance systems (unlike the Indian and Indonesian nutrition projects). The Government did not carry out the covenanted Bolivar Test of Nutrition Delivery (or the proposed substitute), which was to provide extensive testing and evaluation of the targetted food coupon system in one department. Consequently, results from completed evaluation activities do not provide valid and reliable data regarding the impact of PAN on the nutritional status of the target population. 4.21 The Bank's expectations for evaluating this project were highly ambitious, given the number and dispersion of project entities. The Colombian experience argues for restraint in setting evaluation goals. particularly for never projects, and concentration on process and impact indicators. - 37 - D. Program Coordination and Administration 4.22 The Colombia project required the collaboration and coordination of nearly 20 agencies in several sectors, and of eight geographic departments. Consequently, one of the project's central goals was to develop an institutional capacity in DNP to organize nutrition projects, and to create a permanent locus for nutrition concerns that could transcend the responsibilities of individual line ministries. This seemed a feasible goal when national commitment to nutrition was strong. The PAN group often succeeded in using its political clout and prestige to assure that line agencies adhered to PAN's nutritional goals in the face of competing pressures, to mediate among many different entities, and to stimulate substantial progress in regional planning and administration. Moreover, PAN's effectiveness at coordinating many participating agencies was enhanced by the staff's absence of disciplinary or ministry allegiances, and their ability to leave aside individual disciplinary backgrounds and follow the integrated approach implicit in the project concept. However, policy and organizational changes in 1982 ended any hope for the establishment of a permanent PAN group within DNP. 4.23 The Colombia experience confirms the difficulties of converting planners into implementers, and of making a staff agency rather than a line agency responsible for actual program execution. These difficulties were compounded by lack of adequate managerial expertise within PAN, and by the continued autonomy of the implementing agencies. 4.24 PAN did a better job of coordinating program activities than in assisting participating agencies to conform with requirements for reporting physical and financial progress. Some PAN regional office staff succeeded in winning the confidence of participating agencies and helped them work more effectively. However, even the best PAN regional offices failed to convey to local staff of participating agencies the importance of, and methods for, separating expenditures reimbursable from the loan and forwarding appropriate documentation promptly. Financial documentation weaknesses and budgetary constraints seriously slowed disbursements, contributed to project ineffi:iency and disenchantment of local offices with PAN. The rapid program expansion aggravated coordination problems by causing PAN to neglect much of the preparation, training and systems development that was essential for effective performance. 4.25 Technical Assistance. PAN staff expressed considerable satisfaction with outside consultants suggested by the Bank and with technical assistance provided by Bank staff. Project preparation benefitted substantially because a number of consultants had been involved with nutrition work in Colombia under USAID sponsorship, prior to the Loan. Bank staff also provided useful help. Technical assistance from Colombian consultants was more uneven, as was the case for poorly-prepared local evaluation specialists proposing inappropriate, overly-complex evaluation models that compounded PAN difficulties. - 38 - V. BAW I A. Project Justification and Objectives: Bank's and Borrower's 5.01 Malnutrition had long been a serious problem in Colombia, due to various causes. Despite steady increases over time. low incomes restricted effective food demand. Limited access to health services, poor environmental sanitation and unsafe water contributed to diseases which reulted in significant nutrient losses, particularly among pre-school children, compounded by inappropriate food habits such as premature weaning. 5.02 At the time of the loan application, the most acute nutrition problem was protein-calorie malnutrition (PCM), thought to affect approximately four million or an estimated two-thirds of all children under seven and at least an estimated 1.5 million women of child-bearing age - despite data indicating that Colombia's aggregate availability of calories per capita appeared to be adequate. 5.03 Considering malnutrition among its poorest populations a major obstacle to improving human productivity, the GOC incorporated a National Food and Nutrition Plan (PAN) into its general development strategy. PAN drew on previous Colombian nutrition experience to formulate a program carefullly targetted to the most nutritionally vulnerable groups, and integrating combined measures in the agricultural, industrial, health, education and social welfare sectors. The GOC requested the Bank's assistance in developing and implementing the nutrition program in its initial stages. 5.04 The Bank, for its part, was interested in exploring ways of combatting malnutrition in order to increase the human productivity required to maximize the development impact of its resource investment. 3. Project Content and Scheduling 5.05 Upon the Bank's recommendation during the appraisal process. Colombia agreed that the loan would be phased in over four rather than three years. The Government originally intended PAN to achieve full national coverage by 1978. In consultation with the Bank during appraisal, the Government did agree to a slower implementation pace but even that proved too ambitious. More investment was carried-out than would have been if the project had expanded more slowly. -nt more systematic expansion could have led to better services and perhaps nore demonstrable effect. 5.06 This project has been critized for its complex content, involving several sectors and nearly 20 separate agencies. A closer look indicates that the multi-sectoral aspect of the project was manageable when participating agencies shared PAN's nutrition goals and were required merely to expand rather than to modify existing activities. For instance the water construction agency, INAS, and the Ministry of Health responsible for primary health care, were assisted by the project to institutionalize and - 39 - improve construction capability. Including existing activities within this multisectoral project targeted them to nutritionally vulnerable groups and infused them with concern for nutritional consequences. Moreover. despite considerable administrative difficulty and interagency friction. PAN frequently succeeded in coordinating activities of diverse entities, thereby producing synergistic effects that increased impact. For example, the health services component and the food coupon program were mutually reinforcing: distribution of the coupons by health post personnel proved an efficient way of reaching target groups, while the availability of the coupons motivated more regular attendance at health posts. To have financed these activities separately, without the multi-sectoral nutrition project umbrella. would have sacrificed much of this outcome, On the other hand. the involvement of several agencies did increase financial, administration, and reporting difficulties, since PAN was obliged to train several different agencies and to assist each with development of a project information system. C. Project Implementation and Operating Outcomes 5.07 Preparation Regairements. Because the initial appraisal mission jud4eZ project preparation to be inadequate. PAN was delayed one year and was appraised twice despite the Colombian government's protestations to senior Bank management. The result, however, was the appointment of a most capable new director for PAN, more professional attention within DNP, and substantial improvements in service delivery. Thanks to this extensive preparation, the project as a whole was designed with unusual care and thoroughness. Nevertheless, the Appraisal Mission warned that "it must be accepted that even with the Government's extensive preparation." some of the components "have not been prepared to normal Bank standards. Given the complexity and novelty of this project that must be expected. The 'innovative activities', home and school gardens components and part of the food technology project fall into that category. In these circumstances, the Bank should be constructive rather than dogmatic." Ultimately, careful project preparation proved a necessary but not sufficient condition for the project's successful implementation, when the erosion of political support led to PAN's demise. 5.08 Borrower's Implemtation Capacity. Many Bank and Colombian staff argue that the project's complexity generally strained the borrower's implementation capacity. An alternative interpretation, advanced by other Bank, Colombian and consulting staff, is that it was not complexity, but rather the speed of implementation and expansion, that taxed the implementing agencies. The Colombian staff was overextended, with the project going into too many areas too quickly. The rapid expansion to national coverage left PAN inadequate time to establish regional offices and do the necessary preparatory work with participating agencies. For example. INAS and the Ministry of Health were under excessive pressure to increase construction capability beyond their existing adequate, though limited. capacity. The two construction agencies eventually responded well, so that by project termination both had more and much improved building teams, as well as reasonably adequate information systems and flows. - 40 - 5.09 Bank Supervision. The Project preparation documents frequently noted that multisectoral projects in general, and this project in particular would require greater than usual supervision. Accordingly, a total of 240 staff-weeks were spent on supervision over 8 years. or an average of 30 staff-weeks per year (see Basic Data Sheet). The Bank's extensive and above-average supervision did not necessarily correlate with improved performance, however. Reduced supervision over the years did not itself result in negative performance; it should also be noted that supervision was difficult in the later years. when changing program priorities led to a total turnover of projt:t staff. 5.10 Working Relationship. During preparation and the project's first five years. the Bank-Borrower relationship was excellent, by all accounts. The Appraisal Mission noted, for example. "that the Colombian government was exceedingly receptive to suggestions of the mission including those of an extremely sensitive nature." PAN and other Colombian officials, for their part, expressed high praise for both resident and visiting Bank staff. Unfortunately, when a combative new Director for PAN was appointed in 1983. the Bank-Borrower relationship became very difficult. When he was replaced one year later, the Bank and Borrower resumed a productive and cordial working relationship, but considerable damage to the program had been done. D. Implications of Project Outcomes 5.11 The Colombia project is instructive regarding the importance of distinguishing between research tasks for which the borrower feels a real need and is committed to. versus those of primary interest to the Bank. The Borrower was never really committed to the Bolivar study that was of such keen interest to the Bank and other donors. When Bank research priorities are not fully accepted by the borrower or, though accepted, exceed the capabilities of the national agency, the Bank needs to consider financing this essential research on its own. 5.12 The Colombia case also illustrates the risks of vesting coordinating responsibility in a staff agency. The Bank was over-optimistic about DNP's innate capacity to get the line bureaucracy moving, and it would have been more effective to deal directly with implementing agencies rather than a supra planning directorate. DNP's effectiveness declined dramatically when a new administration rejected its predecessor's commitment to nutrition. PAN might have been able to preserve more of its programmatic nutritional focus if it had built a stronger political base for itself (encompassing participating line agencies, Colombian community organizations, and PAN beneficiaries.) Unlike DRI, PAN failed to cultivate support or to create an identity at the regional or local levels. Communities receiving PAN-financed water systems, for example, associated the benefits with INAS and not with PAN. PAN's "invisibility" left the program relatively defenseless given the new administration's reorientation of priorities. - 41 - 5.13 The Bank must be alert to the drawbacks as well as the advantages of funding projects to which there is a strong country commitment. Without the GOC's deep commitment to overcoming malnutrition, there would have been no project in the first place. But this high national commitment also generated some negative effects on the program. For example, the Government pressured for premature appraisal by the Bank; PAN and DRI programs were put in different departments for maximum political visibility rather than being synergistically linked in fewer geographic areas; the program was rapidly expanded in lieu of consolidation and careful course correction based on evaluation. Related implications for the Bank are that political commitment at the highest levels of the Government does not guarantee similar commitment on the part of implementing agencies. In PAN's casea this commitment was particularly hard to come by when those agencies had a hard time getting the promised funds for their role in implementing the project. Nor does high-level commitment even guarantee effective back-up support by the intended beneficiaries, who often lack a power base or political clout. 5.14 When funding politically sensitive nutrition projects. the Bank can exercise important leverage at various stages. In Colombia the Bank was instrumental in assuring more thorough project preparation, despite pressures to accelerate the program prematurely. When the policy environment shifted away from nutrition, the Bank played an important role in encouraging the Borrower's adherence to original project goals, in the face of competing pressures. The Bank's involvement lent considerable authority to PAN in the Colombian context. With four-year presidential terms, the Bank contract assured more stability for the provisions of PAN and for the program's internal financing through counterpart funds, than DNP or other government agencies could have achieved without Bank support. VI. CONCLU 6.01 The project as described at appraisal set forth three major objectives: (a) improving the nutritional status of targetted population groups; (b) systematic monitoring and evaluation to ensure sustained project effectiveness; and (c) strengthening Colombia's amanagerial base in the nutrition sector. While there is clear evidence that the project met or surpassed many of its specific appraised targets, there is little reliable data concerning the effect of these activities on the food consumption or nutritional status of the intended beneficiaries. Several attempts at evaluating the project were made by the PAN evaluation unit, but design flaws and poor quality data seriously curtailed the usefulness of the information collected. A comprehensive impact evaluation of PAN was never done. this failure is particularly disappointing, in view of the program's experimental nature. - 42 - 6.02 Small surveys and close observations indicate that individual project components did have significant impacts. though not necessarily in nutritional terms. The positive health and nutritional consequences of these outcomes were not quantified. Because the results of component activities were recorded so poorly, it is even more difficult to assess their combined nutritional impact. It must be noted that the steady economic deterioration during implementation, with a demonstrable negative impact on nutritional status, masked positive nutritional consequences of the project. 6.03 From an institutional point of view, the project contributed to the strengthening of Colombia's infrastructure in nutrition. There now exists: a) a network of primary health care delivery units, and sound plans for continued expansion of coverage (under Loan 2611-CO), a cadre of promotoras and auxiliary nurses with training in nutrition, and a masters- level program in nutrition at Javeriana University; b) a substantial cadre of government officials, economists, health professionals, and others, sensitized and committed to nutritional goals, and with increased technical capacity; c) a national network of food quality control laboratories, a more adequate food technology and research facility, and a food industry more sensitive to nutritional concerns; d) a more extensive, better-equipped and capable agency for constructing water systems: e) a parallel marketing structure targetted specifically to small growers and merchants, that has contributed to a permanent increase in marketing efficiency. f) The project's institutional development also includes major progress toward the development of a health and nutrition evaluation system in the Ministry of Health. It is important that the new Colombia health project build on this important work by continuing to encourage attention to nutrition monitoring and surveillance. 6.04 The Borrower's views are recorded in a Final Report presented to the Bank in February 1986. The Report states that the project's financial and physical implementation was satisfactory despite three problems affecting implementation, namely late availability of funds due to inadequate budget allocations from the GOC; slow disbursement of the loan; and Colombia's financial difficulties from 1981 on. The Final Report presents a summary review of activities conducted under the loan, concedes that there has been no overall impact evaluation of the project, and eschews any analysis or assessment of its own regarding project outcomes and present nutrition policy. - 43 - 6.05 The Borrower's disappointing failure to draw any conclusions about lessons learned or consequences derived from PAN (e.g with regard to nutritional status, health care delivery, national food policy or future program directions) reflects the current Colombian administration's rejection of its predecessor's nutrition/consumption/subsidy orientation, in contrast to its own emphasis on rural development, food marketing and food policy. This shift in official priorities is probably the single most important factor accounting for the absence of a follow-up loan to Colombia. although the country's economic crisis since the early 1980s is also fundamental. Economic conditions became far less favorable during the duration of the project than they were when PAN was larnched. It is not surprising that subsidized nutrition programs for the lowest income groups, designed within the context of economic expansion. would be an early casualty of fiscal austerity. However, should political and economic conditions become more propitious for nutrition, such activities will benefit substantially from the strengthened human resource base and institutional infrastructure that evolved under the aegis of PAN. 6.06 The Bank has devoted considerable attention to the content of nutrition projects. One school of thought maintains that the potential of multi-sectoral projects to achieve maximum synergistic results justifies their additional risks and resource requirements; the other asserts that their success demands optimum conditions seldom realized in practice, that they invariably exceed the borrower's managerial capabilities, and that less complex projects are more likely to achieve their stated (albeit more modest) objectives. The Colombia project suggests that multi-sectoral programming can generate productive outcomes even when multi-sectoral managerial burdens prove excessive. 6.07 A key lesson of PAN's experience is that nutrition interventions can be targetted successfully and economically. Colombia's development of a national poverty profile. coupled with the use of the health system for even more precise selection of beneficiaries, dramatically improved the effectiveness and efficiency of the food stamp component: avoiding the leakages common elsewhere, without imposing the high administrative costs typically associated with targetting efforts. 6.08 Regarding the ideal institutional locus for nutrition projects, the Colombia experience suggests that the effectiveness of a supra- ministerial coordinating agency depends heavily on a favorable political climate for nutrition, along with adequate management capabilities for assuring the prompt flow of funds to implementing agencies. PAN's fate also suggests that the coordinating entity would benefit from closer ties with a line agency, while retaining some degree of autonomy - an institutional arrangement similar to that of Fondo DRI, created in 1985 within the Ministry of Agriculture. 6.09 A four year implementation period for a multisectoral project is not realistic. The Bank had been successful in changing the GOC proposed project time table from three to four years. The project took seven years to implement requiring two extensions. - 44 - 6.10 The Colombia project affords a lesson regarding the importance of distinguishing between research tasks for which the borrower feels a real need and is committed to. versus those of primary interest to the Bank. When Bank research priorities are not fully accepted by the borrower, anA exceed the capabilities of the national agency, the Bank needs to provide and to finance its own research. 6.11 When funding politically sensitive nutrition projects, the Bank can exercise important leverage. In Colombia the Bank was instrumental in assuring adequate project preparation, despite pressures to accelerate the program prematurely (para. 5.13). Later, when the national climate became less favorable to nutrition. Bank involvement brought about greater adherence to original project goals than project management would otherwise have achieved. 6.12 Many of these lessons were built into subsequent nutrition projects in India and Indonesia, and are reflected in the high performance of these subsequent Bank funded programs. 6.13 The Bank's decision to finance this project was well justified, in the light of the Bank's own goals, contemporary thinking within the international nutrition community, and the opportunities within Colombia at the time of project identification and preparation. - 45 - cOMmazA ,RAa MTRnM raniscr (La. 1487--0) PRnECT COmrTmO as1O1r Table 1: Project Performance Summary - Key Indicators Activity Appraised Actual Actual as Target % of Appraised ALTB & SAKITATION - Construction/Improvement of Primary Health Care Posts 281 729 259% - Health Promotors Trained 2.028 1.368 672 - Equipping of Health Promotors 2,400 1.297 54Z - Completed water works 372 326 882 Beneficiaries (6-8 per home) 231,250 199,227 86% - Latrines 112.000 35,000 31% UTRITION EUCATION - Staff Trained 8.028 30,996 3862 - Mass Media: Radio Spots 225,000 565,660 2512 W GaE ROTS (PANCOG) - Families with Garden Plots 20.000 45.720 2282 - Credit Recipients (families) 20,000 41.780 209% T0M COU-OS - Coupons Distributed (millions) 71.800 64.800 902 - Direct Beneficiaries 415.500 420,476 101% Source: Final Project Evaluation Report, Fondo DRI, January 1986. - 46 - CO1NBIA INWRATED NUTRITION - POJCT - L. 1487-0 PMJCT CPIATION 88PORT Table 2: Price Differences Due to Presence of Mobile Markets in Bogota Products Supermarket "Cajas de Market Shopkeepers Compensacion" Squares (1) (2) MBILB MARBT PRICS INDBI = 100 (3) Grains & Processed Food 101.1 94.4 99.8 100.0 101.0 Vegetables 131.9 121.0 108.5 102.5 110.0 Fruits 116.6 109.4 98.2 103.2 107.6 Meat 105.2 103.0 99.8 101.2 103.5 Eggs & Dairy Products 124.4 125.0 99.7 102.3 107.3 Average Price 112.0 105.7 101.8 101.2 104.7 (1) Shopkeepers surveyed on days when mobile market was operating. (2) Shopkeepers surveyed in locations where the mobile market was not operating. (3) The index considers prices of about 44 products of the food basket, taking into consideration their importance in family food spending, using the methodology for price indexing of DAN. Note. Data has not been adjusted for difference in quality and service. Source: Calculations based on price survey of retailers in Bogota, "Mvaluaciom de los Programes de Mercedos Moviles", Bogota, Sept. 1985. Final Project Evaluation Deport, Fondo DRI, January 1986. g RV mEITIS WWR WCT - ,08 1487-0M paw3~ Ontma- Table 3: Actual and Appraised Project Costs by Component (Constant 1977 Pricest US$ Millions Bquivalent) Component 1977 1978 1979 1980 1981 1982 1983 1984 1985 Total Appraised Actual as 2 No. % No. 2 of Appraised Primazy Health Care 1.4 1.1 3.4 4.2 6.8 1.7 1.3 1.3 0.3 21.5 25 15.1 22 142 Basic Rural sanitation 1.5 1.4 2.0 3.5 3.5 3.8 2.1 0.8 - 18.6 21 18.4 27 101 Nutrition Education 0.8 0.6 0.6 0.6 0.5 0.1 0.5 0.3 0.02 4.0 5 6.1 9 65 Rome Gardens/PANCOGVR 0.3 0.2 1.2 0.9 2.9 3.7 3.0 0.2 - 12.5 14 1.4 2 893 Tood Technology 0.2 0.1 0.1 0.1 0.2 0.1 0.1 0.3 0.02 1.2 1 2.4 3 50 Tood Quality Control 0.1 0.1 0.2 0.1 0.1 0.1 0.2 0.04 0.01 0.9 1 1.4 2 64 Nut.System Test - Bolivar - - - - - - - - - - 3.4 5 0 School Cafeterias 0.8 0.3 0.6 1.0 0.6 0.03 - - - 3.3 4 - food Distribution 3.1 4.4 2.5 5.7 0.7 - - 0.6 0.03 17.1 20 - Monitrta & Evaluation 0.8 0.3 0.7 0.3 0.3 0.1 0.3 0.1 - 3.0 3 3.1 4 97 Administration/Coordination 0.2 0.5 0.3 0.4 0.5 0.9 1.0 1.1 0.2 5.1 6 4.0 6 128 Contingencies 13.8 20 W2EA 9.2 9.0 11.6 16.8 16.1 10.6 ga 4.7 06 87.1 100 60 100 126 Coupons * 1.3 0.5 1.0 1.6 3.0 2.5 1.5 - - 11.4 9.3 123 Sourcet Final Project Rvaluation, Fondo DRI, January 1986. Ir Note. Data was taken from PAN payment data, and does not include activities financed by USAID. * The appraisal report estimated the cost of the coupon distribution activities would cost US9.3 million equivalent over four years, and would be financed by the Government. - 48 - COLOMBIA 1WRA T RITICH IPRON PROTECT - LOAN 1487-00 PROJECT COKMEIION REPORT Table 4: Actual and Appraised Disbursements by Expenditure Category (US$ THOUSANDS) Expenditure Category Appraised Actual No. % No. I Civil Works for Health and Sanitation 9.000 36 7.056 28 Equipment. Materials, Supplies for all of the project except Monitoring & Evaluation 6.600 26 5,479 22 Water pipes and pumps 1,700 7 4,858 19 Technical Assistance and Fellowships for Nutrition Education & Food Technology and Quality Control 700 3 2.255 9 Training for Health, Nutrition Education and Family Gardens 1.200 5 2.653 11 PAN Group Salaries for Coordination and Administration & operating expenses including office equipment 2.150 8 1.464 6 Project Management Fund for Innovative Activities 150 1 172 1 Mass Media Production and Transmission 1.700 7 1,063 4 Unallocated 1,800 7 TOTAL 25,000 100 25,000 100 - 49 - COLMIA INTHGRATEM MOTRITION IM RPatCT - LOAN 1487-00 PRWCT COMPLBIGi RKPO Table 5: Actual and URraised Schedule of Disbursements (US$ THOUDANDS) APPRAISED AMMENDED ACTUAL % Difference: YEAR ---- ---- - Actual - Appraised QUARTER Amount % Amount % Amount I (- Ammended thereafter) 1978 1. 150 1 0 0 - 1 2. 520 2 0 0 - 2 3. 930 4 0 0 - 4 4. 1920 8 263 1 - 7 1979 1. 3060 12 760 3 - 9 2. 4310 17 1047 4 -13 3. 5430 22 1407 6 -16 4. 6850 27 1809 7 -20 1980 1. 8490 34 2182 9 -25 2. 10300 41 3304 13 -28 3. 11760 47 4220 17 -30 4. 13600 54 5600 22 4605 18 -36 1981 1. 15725 63 6850 27 5076 20 -43 2. 18060 72 10090 40 5735 23 -49 3. 19710 79 10865 43 6798 27 -52 4. 21810 87 12840 51 8157 33 -54 1982 1. 23710 95 14410 58 10763 43 -52 2. 25000 100 17760 71 11354 45 -55 3. 18535 74 13098 52 -22 4. 20410 82 14216 57 -25 1983 1. 21860 87 14493 58 -29 2. 25000 100 15217 61 -39 Loan Ntension: 06/83 3. 16839 67 16202 65 - 2 4. 17619 70 16364 66 - 4 1984 1. 18849 75 16929 68 - 7 2. 20079 80 17792 71 - 9 3. 21309 85 20084 80 - 5 4. 22539 90 21690 87 - 3 1985 1. 23769 95 22740 91 - 4 2. 25000 100 23223 93 - 7 3. 24491 97 4. 24841 99 986 1. 25000 100 - 50 - COLOMIA U1w 9-RAT W MihRITZG PRATCT - L. 1487-00 PRTECT COMPL0TIN RIPO Table 6 : Prolect Performance Supervision Ratings Year: 1978 1979 1980 1981 1982 1983 1984 1985 Item Month: 9 2 6 3 10 4 2 11 5 4 11 6 Status 1/ 2 2 2 2 2 2 2 2 222 Trend 2/ 2 1 2 2 2 2 2 3 1 1 1 Main Types of Problems 3/ M NM MM N MM N P P P P F F 7 7 7 P P -- New MIS Ratings: 4/ Available Funds 2 Project Management 2 Development Impact 3 Implementation Status: I/ Disbursements 3 3 3 3 3 3 1 2 2 1 Estimated Cost 1 1 1 2 2 1 1 1 1 2 Anticipated Completion 1 2 2 2 2 2 2 2 2 1 Compliance with Loan Conditions 1 1 1 1 1 2 2 1 1 1 Project Finances 2 2 2 2 2 2 2 2 1 2 Management Performance 2 2 2 2 2 2 2 2 2 2 Procurement Progress 2 2 2 2 1 1 1 1 1 1 Performance of Consulants 1 1 1 1 1 1 1 1 1 2 Reporting 1 2 2 2 2 2 2 2 2 1 Development Impact: I/ Expected Benefits 2 2 2 2 2 2 2 2 2 2 Institution Building 1 1 1 1 1 1 1 1 2 2 11 1 = Problem-free or minor problems 21 1 = Improving 3/ WManagerial 2 = Moderate Problems 2 = Stationary F=Financial 3 = Major Problems 3 = Deteriorating P=Political / New MIS Ratings: Available Funds: 2 = Although there are some difficulties and delays, overall funding is likely to face only moderate problems. Project Management: 2 a Despite some shortcomings, project management is capable of doing its job with only moderate problems. Development Impact: 3 = Major objectives will probably be achieved only in part, but sufficiently so to justify the project.