Colombia - Community Child Care and Nutrition Project

Document of The World Bank FOR OFFICIALW USE ONLY Report No. 17168 IMPLEMENTATION COMPLETION REPORT COLOMBIA COMMUNITY CHILD CARE AND NUTRITION PROJECT (Loan 3201-CO) December 23, 1997 Human Development Sector Management Unit Colombia, Ecuador and Venezuela Country Management Unit Latin America and the Caribbean Regional Office This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY AND EQUIVALENT UNITS Currency Unit - Colombian Peso (P) December 15, 1990: US$1 - P/470 March 30, 1997: US$1 -P/1042 GOVERNMENT OF COLOMBIA - FISCAL YEAR January 1 - December 31 ABBREVIATIONS AND ACRONYMS CAIP Centro de Atenci6n Integral al Preescolar(Pre-School Child Care Centers) DANE Departamento Administrativo Nacional de Estadistica (National Department of Statistics) DHS Encuesta de Demografia y Salud (Demographic and Health Survey) DNP Departamento Nacional de Planeacion (National Planning Department) HBI Hogares de Bienestar Infantil (Child Care Homes Program) ICBF Instituto Colombiano de Bienestar Familiar (Colombian Institute for Family Welfare) ICR - Implementation Completion Report IFI International Financial Institutions INURBE Instituto Nacional de Vivienda de Interes Social y Reforma Urbana MC Madres Comunitarias (Community Mothers) NGO Non-Governmental Organization OEI Organization of Ibero-American States for Education, Science & Culture PROFAMILIA Asociaci6n Pro-Bienestar de la Familia Colombiana (Association for Colombian Family Welfare) SAR Staff Appraisal Report SENA Servicio Nacional de Aprendizaje (National Training Institute) SFP Continuous Training System SIG Sistema Integral de Informaci6n Gerencial (Integrated Management Information System) . Vice President: Shavid Javed Burki, LCR CMU Director: Andres Solimano, LCC4C Sector Leader:- Constance A. Corbett, LCC4C/HD - SMU Director: Julian Schweitzer, LCSHD Task Manager:-. Martha Laverde, LC4CO -  FOR OFFICIAL USE ONLY Table of Contents PREFACE EVALUATIONSUMMARY ...........................................................i PARTII: PROJECTIMPLEMENTATIONASSESSMENT . ....................................1 A. Project Objectives .........................................................1 B. Achievement of Project Objectives .......................................................... 3 C. Key Factors Affecting Project Implementation ....................... .................. 12 D. Project Sustainability ......................................................... 13 E. Bank Performance ......................................................... 14 F. Borrower Performance ......................................................... 14 G. Assessment of Project Outcomes ......................................................... 15 H. Key Lessons Learned ......................................................... 15 PART II: STATISTICAL TABLES ............................. .............................. 17 APPENDIXES: A. Final Supervision Mission Aide-Memoire ....................................................... 29 B. Borrower's Contribution to the ICR ......................................................... 39 C. Comments from the Government on the ICR ................................................... 47 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.   COLOMBIA COMMUNITY CHILD AND NUTRITION PROJECT (Loan 3201-CO) IMPLEMENTATION COMPLETION REPORT PREFACE This is the Implementation Completion Report (ICR) for the first Community Child Care and Nutrition Project in Colombia (estimated total project cost US$40.2 million), for which Loan 3201-CO in the amount of US$24 million equivalent was approved by the Bank on August 1, 1990. The Loan Agreement was made effective on December 5, 1990. The Project was carried out by the Colombian Family Welfare Institute [Instituto Colombiano de Bienestar Familiar (ICBF)], which also provided counterpart funds for its financing. The closing date was March 31, 1997, at which time 92-.6 percent of the total credit amount (US$22.2 million) had been disbursed. The original closing date for the Project was June 30, 1996, but was extended twice, to October 31, 1996 and March 31, 1997, respectively. These extensions were granted based on the decision to complete ICBF's First Impact Evaluation Survey of the Child Care Homes Program [Hogares de Bienestar Infantil (HBI)] and to define the internal mechanisms for analysis of the evaluation's results. The last disbursement took place on January 3, 1997. The ICR was prepared by Martha Laverde (Task Manager), Carmen Elisa Fl6rez, Patricia Avila and Gloria Maria Borrero (local consultants) of the Colombian Resident Mission, and edited by William Mayville. It was reviewed by Jacques van der Gaag, Eleanor Schreiber, Miriam Schneidman, Sandra Rosenhouse (Task Managers), and Constance Corbett, Sector Leader. Preparation of the ICR was begun during the Bank's final supervision mission, on May 8, 1997. The ICR is based on discussions with Bank staff involved in the Project, supervision reports, and related documentation in the Project Files. The Borrower and executing agency contributed to its preparation by participating in informational and analytical meetings with the consulting group in charge of the ICR, and by preparing its own evaluation (Annex B).   COLOMBIA COMMUNITY CHILD AND NUTRITION PROJECT (Loan 3201-CO) IMPLEMENTATION COMPLETION REPORT EVALUATION SUMMARY Introduction 1. The Community Child Care and Nutrition Project (Loan 3201-CO) is the first project in community child care that the Bank has supported in Colombia. Prior to this Project, the Bank helped finance seven projects in the social sectors: one each in health and nutrition for a total of US$40.3 million, and five in education for a total of US$144 million. 2. Concurrent with the Community Child Care and Nutrition Project (Loan 3201-CO), the Bank financed one project in basic health care and one in basic education. Project Objectives 3. The Community Child Care and Nutrition Project (Loan 3201-CO) was a key feature of the national plan for the reduction of poverty. It sought to strengthen the HBI (Child Care Home) Program, a national effort to improve the quality of child care for children under the age of seven, managed by the Instituto Colombiano de Bienestar Familiar (ICBF). Although under the jurisdiction of the Ministry of Health, the ICBF has administrative and financial autonomy, and supports an array of activities, of which the HBI Program forms a large part (46 percent of the total ICBF budget). The Community Child Care and Nutrition Project's overall goal was to provide support to an HBI expansion program over five years, while seeking to meet three objectives: (a) introduce and develop policies aimed at improving the cost-effectiveness of ICBF operations; (b) support the institutional development of ICBF, particularly, to strengthen its technical support, planning, and monitoring capabilities vis-a-vis the HBIs; and (c) evaluate, improve and maintain at an appropriate level the quality of HBI services during their expansion. 4. The goals of providing assistance to ICBF in the areas of management and coordination for both HBI Program expansion and the Project were incorporated into three project components: (a) policy development; (b) institutional development; and (c) service support. Implementation: Experience and Results 5. The Project was quite modest in terms of the total investment in the HBI Program, representing about 6 percent. It funded most of the policy and institutional components, but only 5 percent of the service-support component. Nonetheless, the amount allocated to the latter made up 75 percent of the Project's total costs.  - ii - 6. Under the HBI Program, coverage increased from 500,000 children in 1989 to 823,000 in 1996, but fell short of its goal of one million children over that period. It did succeed in reaching the targeted areas and social sectors: 91 percent of the total HBI beneficiary families in 1996 were in the lower-low/low strata and dispersed rural areas; and 87.3 percent of them earned an income of one minimum salary or less. The quality of the HBI services during the Project fell short of desired levels; nonetheless, Project components proved flexible enough to be adapted to changing needs. For example, the Government's decentralization initiative was supported under the institutional development component, as well as the formation of cooperatives by parents' associations under the service-support component. The Project did not establish a baseline at the outset, nor did it define quality parameters with which to measure its results. ICBFs relations with the various participating public institutions were weak with central-level institutions but effective in establishing alliances with municipalities. 7. In retrospect, there was little institutionalization of the original Project aims or values. On the other hand, at the local level many HBIs are still functioning. The model of HBIs was based on the successful experience of UNICEF using this approach. After the Project was scaled down, its priority seemed to lessen for ICBF, which exacerbated the difficulty of achieving expected outcomes, especially given the central link of Community Mothers to the Project's overall success in reaching the targeted beneficiaries. This may reflect to an underlying difference between the central role played by ICBF, which is a semi-autonomous government agency, and the Project objective of building linkages into local communities at the family level. Policy Development Component 8. Subcomponent: Establishingfinancial arrangements for Pre-school Child Care Centers [Centro de Atencion Integral al Preescolar (CAIP)i. ICBF met the goal of reducing its contribution to the CAIP Program by 30 percent, as planned. Alternate financing sources were identified, and a percentage of the cost was transferred to parents, municipalities, and private companies. 9. Subcomponent: Improving the efficiency ofproduction and distribution of Bienestarina. Based on studies made, the production and distribution of Bienestarina (a nutritional supplement) was directly retained by ICBF, with efficiency improvement through plant modernization, purchase of raw materials through the farminig exchange (which resulted in a significant price reduction, by substituting alternate solutions of raw materials for product diversification) and by dividing the country into regions for the distribution. The three production plants were kept in operation and the micronutrient composition of Bienestarina was also improved. Management Development Component 10. Subcomponent: A study on ICBF structure and staff The Project supported two studies on ICBF restructuring in 1993 and in 1995, which tailored the ICBF organizational structure to new Constitutional requirements. Concurrently, control of ICBF Programs by  - 111 - civil society occurred through citizen surveillance groups, along with co-management with the municipalities of some of its programs. During the life of the Project, activities were initiated to decentralize selected functions as a strategy to strengthen regional centers. 11. Subcomponent: Establishing an Integrated Management Information System [Sistema Computarizado Integral de Informaci6n Gerencial (SIG)] for the planning, programming, budgeting and monitoring of ICBF Programs. Despite significant delays, the SIG was developed as a management tool for planning, programming and budgeting of ICBF Programs. For the HBI Program, the SIG will facilitate the monitoring of coverage and internal efficiency of the Program. 12. Subcomponent: Introduction of Monitoring and Impact Evaluation Systems Nationwide. After a four-year delay, the First Impact Evaluation Survey was designed and conducted in the first quarter of 1996, to determine the direct effects on those children covered by the Program. The survey suffered some design limitations: since there was no control group, the survey comprises baseline information and is not a true impact evaluation, as intended. (The ICR refers to the impact evaluation as the First Impact Evaluation Survey.) Nevertheless, this survey provides a useful conceptual model and sampling base for the implementation of a continuous impact evaluation system of the HBI Program nationwide. It should facilitate future monitoring and interventions in the Program. 13. With the First Evaluation Survey, the lack of a systematic monitoring design for the life of the Project became evident. The responsibilities for service quality follow-up at national, regional and local levels needed to be clearly defined, as the First Evaluation Survey shows. However, with the consolidation of the SIG and the evaluation survey, the ICBF will have the required tools to perform systematic, qualitative supervision, such as training and technical assistance, as needed. Service Support Component 14. Subcomponent: Provision of regular trainingfor community mothers and other community HBIparticipants in the management offood, nutrition, health registers, child growth and household management. The Project funded the execution of decentralized training plans, including procurement of teaching materials. The training and materials helped mothers to manage households and carry out activities to develop children's psycho- social learning process. However, the results of the First Evaluation Survey showed that the training was inadequate. Though positive results were reported for home management, the survey revealed that the training had a very limited effect on improving children's welfare in terms of meeting nutritional, health, and psycho-social development needs. 15. Concurrent with this training process, ICBF developed a campaign to eradicate illiteracy among Community Mothers (MCs) and to provide them with basic education, with the intention being that a high percentage of them would finish primary or lower secondary school. However, less than half of the MCs in urban areas complete the 9th grade, which is ICBF's minimum schooling requirement; in rural areas, the percentage of  - iv - those meeting the minimum requirement (third grade of primary school) is somewhat higher. It appears that the HBO Program's quality standards established for the provision of services may have been unrealistic, and that other service support models might have been more useful. 16. Leadership training was also provided for the establishment of HBI-related microbusinesses in order to improve family income. This training, along with household improvement loans, the increase of allowance, and the organization of MCs into unions and community groups, boosted MCs into positions as community leaders. This may be considered as one of the most significant by-products of the Program. 17. Training of Other Actors: Training processes did not emphasize the role of Community Mothers over that of the parents themselves; the CMs would rather act on behalf of the parents by providing child care and nutritional needs. The Project did support a study to turn the Parents Associations into cooperatives, as part of the training activities in the administration of HBIs. 18. Training of ICBF Staff The Project funded training of ICBF personnel at all levels in a number of areas: ICBF restructuring, decentralization and community participation strategies to attain co-management with municipalities, human rights awareness, utilization of impact evaluation, and internal control tools in computer operations and the SIG. Nevertheless, the Program did not develop a permanent supervision system to allow for trainee and training follow-up and assistance, or the means to detect problems and take corrective actions in a timely manner. 19. Subcomponent: Loans to Community Mothers for Household Improvement. Of the 50,000 loans planned, 60,711 MCs have benefited from home-upgrading loans to date. However, 28 percent of the loans covered other needs related to housing outside the Project guidelines, such as payment of arrears or land purchases. This suggests the need for a more strict control in the use of home improvement loans. Currently, the household improvement program for MCs has funds of nearly US$10 million. 20. Even though the MCs who received loans now enjoy better housing conditions, the HBIs' physical conditions generally remain poor. The amounts lent may not have been sufficient to achieve the home improvement necessary to provide adequate conditions for child care. The fact that 51 percent of the HBIs have less than one square meter per child may indicate that the Program's regulations should tailor the guideline of 15 children per MC to reflect the actual capacity of each HBI. 21. Technical Assistance to ICBFfor Program and Project Management and Coordination. To avoid creating a parallel structure, the Project was supported by a Management Unit responsible for supporting ICBF in administrative matters. The technical aspects of the Project were entrusted to the corresponding sections of ICBF. Essential technical assistance in this area was provided, generally toward the end of the project, in such areas as definition and development of quality standards for the Program, analysis of  v educational strategies and the planning of training, an integrated management information system (SIG), and impact evaluation. 22. Bank Performance. Bank performance revealed some deficiencies in monitoring and evaluation during the initial years of the Project life, primarily delays in implementation of the information and evaluation components. Dated loan covenants for SIG and evaluation in design were not enforced strictly and broad sanctions provided in the Loan Agreement for non-compliance with objectives were not applied, except for an interruption of disbursements in 1992. In the initial years of the Project, security issues complicated supervision efforts. Also, the somewhat volatile political environment, including vacillating government commitment, and scaling-down of the project, also contributed to delays and problems with monitoring and evaluation. 23. Borrower Performance. Frequent turnover among ICBF staff and in the Project Management Unit and the low priority given by ICBF to the Project, especially to monitoring and evaluation efforts, contributed to delays. Successive ICBF directors generally agreed, for example, to mount an impact evaluation, but never followed through to complete its design. Toward the end of the Project, ICBF interest in and commitment to the Project began to increase, and planned activities were completed. 24. Evaluation of Project Results. The First Impact Evaluation Survey showed disappointing results for the Program's impact on children's welfare in terms of their nutritional and health status and psycho-social development. Results are explained by the inadequate physical conditions and overcrowding in HBIs and by inadequate nourishment provided to the children. Lack of appropriate training on a regular basis and supervision within the Program did not permit these problems to be resolved. These results suggest that the HBI Program model - - which at the time, was a tested prototype - - should be reevaluated in terms of the criteria for selecting MCs, number of children per HBI, appropriate housing loan amounts, and the suitability of lower-income homes as day-care sites. Moreover, the model may be better suited to deployment by NGOs in local communities (as originally conceived) than to a large-scale program run by a central agency of the Government, with weak institutional linkages both centrally and in local jurisdiction. Lessons Learned * Understanding both the political and institutional climate has important implications for the immediate impact of the project and sustainable outcomes. Projects such as this one that involve different participating institutions to support the achievement of Project objectives should establish agreements with authorities or institutions, and provision should be made for periodic review/revision /ratification of institutional arrangements or agreements throughout implementation. This is especially germane in the case of first- time projects in sectors with weak or nonexistent institutional linkages at project inception.  -vi- * The scale and location ofprojects has a direct bearing on whether government agencies or NGOs are better positioned to achieve optimal project results. Centralized procedures, characteristic of the older operations supported by the Bank and the Colombian Government, were poorly suited to a highly decentralized and participatory service delivery model such as the HBIs. In this case, the involvement of NGOs might have led to a greater project impact. * Performance indicators, evenfor already-piloted models, such as the UNICEF-HBI model, need to be developed independent ofprevious experience and adapted to institutional arrangements under specific country and project conditions. Projects need adequate baseline data on which to compare progress. Design should reflect and already incorporate these data including plans for evaluation follow-up. In addition, project design should include provision for periodic evaluation of standards and testing of assumptions, and for revision of guidelines in response to impact and experience. More importantly, when using a "model" approach, sustainability on the scale, scope, and other key factors need to be evaluated to guide a tailoring to the country, institutional, and project-specific environment. * New information systems may prove difficult to design and implement, which needs to be reflected in implementation timing milestones. The delays in implementing ICBF's information systems prevented it from reaching the scale intended. The system was difficult to design, and, above all, to implement, as in the case of SIG and the impact evaluation, where insufficient attention was given to practicability and unrealistic compliance times were specified.  COLOMBIA COMMUNITY CHILD AND NUTRITION PROJECT (Loan 3201-CO) IMPLEMENTATION COMPLETION REPORT PART I: PROJECT IMPLEMENTATION ASSESSMENT A. Project Objectives Background 1. The Community Child Care and Nutrition Project (Loan 3201-CO) was a key part of the national plan for poverty reduction, seeking primarily to strengthen child care provided in Child Care Homes [Hogares de Bienestar Infantil (HBIs)]. The HBI Program is conducted at the national level by the Instituto Colombiano de Bienestar Familiar (ICBF). ICBF, an institution attached to the Health Ministry, has both financial and administrative autonomy and is funded mainly by a three-percent tax payroll levied on both public and private institutions, a third of which is assigned to finance the HBIs. (The HBI Program is a community-managed, home-based child-care program aimed at promoting development of poor children under seven years of age. The model was designed to provide up to 15 children per home with food, social stimulation and health services based on need. Each HBI is run in the home of a Community Mother (CM) selected by parents. 2. The HBI Program is the most important ICBF Program and was initiated in 1987. In the first year of the Program, nearly 123,700 children were served in 8,247 homes. By the end of 1992, 858,000 children were served in 57,000 child care homes. Expansion of coverage stabilized in 1993, and by year-end 1995, 890,295 children were being served in 59,353 homes. Currently, the Program is being executed in about 1,025 Colombian municipalities, with the participation of 77,875 MCs. 3. The Pre-School Child Care Centers [Centro de Atencion Integral al Preescolar (CAIP)] Program is currently the second most important ICBF Program. The CAIP Program is center-based and operated by professionals. It provides food, health services and pre- school education to children two months to seven years of age who are at risk of abandonment, malnutrition or abuse, or from lower-income families with working parents. In contrast to the HBIs, coverage under the CAIP Program declined from 236,500 children under age seven in 1989 to 177,000 in 1993. 4. The Colombian Government investment plans for the HBIs for the 1989-94 period sought to improve the efficiency and equity in the allocation and use of ICBF funds through better targeting and strengthening the quality and efficiency of services (SAR, Annex 7). To improve equity the 1990-95 plans for the HBI Program sought to broaden the coverage of  -2- lower-income children. This expansion was to be financed in part by a reduction in the ICBF subsidies to the CAIPs (SAR, par. 1.17). CAIPs, though high in quality, were considered comparatively expensive and inaccessible for a majority of lower-income families, and were funded almost entirely by ICBF. 5. To improve efficiency, the production of the national nutritional supplement Bienestarina was to be made more cost-effective by seeking alternatives for its production and distribution by ICBF or others. Under ICBF's overall management, the supply of Bienestarina was unreliable and costly, due to delays in input materials and the lack of adequate funding, as well as ICBF's inexperience. By increasing the access of HBI users to effective health care and consolidating children's nutritional improvement, the HBI Program sought to improve both equity and efficiency. 6. Of the three types of service provided through the HBIs, the nutrition component was the most developed. Health monitoring and early childhood development received less emphasis and resources and so lagged behind (SAR, pars. 1.31 and 1.32). The support provided through ICBF and the local organizations for the establishment and operation of the HBIs was effective; however, reinforcement was needed to achieve a regular training system for ICBF staff and the community participants, especially the MCs. Project Objectives 7. The main objectives of the Project were to support ICBF in the implementation of the HBI Program, and in particular to: (a) introduce and develop policies to improve cost- effectiveness of ICBF operations; (b) support the development of ICBF's institutional capacity to manage the HBI Program by strengthening its technical support, planning and monitoring capabilities; and (c) evaluate, improve and maintain, at an appropriate level, the quality of HBI services (SAR, par. 2.2). Initially the Government asked the Bank to help finance a significant share of HBI service delivery. Following the Government's decision to allocate the one percent share of payroll to the HBI Program, the smaller, more limited Project evolved (SAR, par. 1.0). 8. The Project was organized into three components: * Policy development: Studies and technical assistance designed to enhance effectiveness of ICBF resources use by: 0 assessing and establishing revised alternatives for the funding of CAIPs and reducing the related ICBF budget contribution; and 0 improving the efficiency of production and marketing of Bienestarina. * Institutional development: Strengthening ICBF management capacity by: 0 conducting a study of ICBF organizational and staffing needs;  -3 - O establishing a computerized, integrated management information system (SIG) to support ICBF planning, programming, budgeting and monitoring of programs and activities; and 0 introducing and implementing a monitoring and impact-evaluation system for the HBI Program on a national scale. Service support: Improvement of the nutritional, health and pre-school development aspects of the HBI Program, through: 0 in-service training for MCs and other HBI community participants for guidance in food management, nutrition, health registers, child development and home management; and 0 providing MCs with access to home improvement credits. 9. The Project provided a relatively modest contribution to the total ICBF investment in the HBI Program. According to the SAR, the total estimated national investment of the Program for the 1990-95 period was US$674.6 million, of which US$40.2 million (six percent) represented the Project (Table 8A). However, the Project's share of total ICBF Program cost varied greatly by component and subcomponent. The Policy Development component was financed in its entirety (100 percent) by the Project, as was nearly 76 percent of the Institutional Development component. 10. On the other hand, the Service Support component represented 75 percent of the total Project costs, but accounted for only 4.5 percent of service support in the overall Program (Table 8A). Tables 8B and 8C show the planned and actual distribution of Project expenditures. During implementation, realignment of priorities led to reallocation of resources from the training category to technical assistance. Actual disbursement for training represented 83.3 percent of the original allocation, while disbursements for technical assistance were 120 percent of the original plan (Table 8B). The loan amount of US$24 million represented 60 percent of total Project costs and 48 percent of local costs (Table 8C). 11. The Project provided support for improvement of ICBF's capacity in the areas of: (a) long-term planning and programming to respond to feedback from groups as well as from quality control and financial control systems; and (b) regular monitoring and a evaluation of the HBI Program (SAR, par. 1.29). B. Achievement of Project Objectives 12. The Project partially met its objective of supporting coverage expansion within the HBI Program. Of an enrollment goal of one million children by the end of 1994, to be maintained until Project completion (Table 5), enrollment reached 799,000 children in 1994 (HBI Census, 1994), and 827,000 in 1996 (First Impact Evaluation Survey, 1996). (Accurate figures for actual coverage of the HBI Program are difficult to obtain. Estimates are based on the number of HBIs reported by each regional office and an assumption of a constant number  -4 - of 15 children per home.) Coverage increased most in the early years of Project execution, stabilizing after 1993. Thereafter, efforts and resources were concentrated on increasing food rations (the daily ration was improved, and night, weekend and vacation rations were created), and provision of durable teaching materials (ICBF, 1996). 13. The Project should be evaluated not only in terms of its impact on the beneficiary groups but also on the target population, the lower-income families. The First Evaluation Survey shows positive preliminary results. According to the survey, of the total HBI beneficiary families in 1996, 91 percent belong to the lower-income strata or live in rural areas; 87.3 percent of the families' earned income was at the minimum salary or less. These results match those found by the "Study of Incidence in Social Public Expenditures," conducted by the Departamento Nacional de Planeaci6n (DNP) in 1993, which indicated that the HBI Program had high incidence on the target population: over half this population live in conditions of extreme poverty and over 80 percent are poor. However, successful targeting did not equate with full coverage of the priority group. The best estimates available suggest that overall coverage reached about half the target population (First Evaluation Survey, ICBF, 1996). 14. It is not possible to distinguish the achievement of project objectives from those of the HBI Program as a whole. The Project did not support any specific geographical areas or particular activities. Moreover, no indicators were defined for the Project other than those of the HBI Program, nor was any causal relation established between the results of the Project and those of this Program. 15. For the purposes of the ICR, the achievement of the three specific Project objectives is assessed based on the results of the First Impact Evaluation Survey and the evaluation processes developed during Project implementation. The survey reveals Program results that will clearly become valuable instruments in the strengthening of the HBI Program, especially with respect to quality of services and impact on child nutrition. 16. Project objectives were broad, and the specific strategies used in some cases may have been too limited to obtain any global results for the Program. During Project identification, for example, risks were detected concerning ICBF management capacity, but were inadequately specified. Once implementation began, it was clear that the SIG was far less advanced than anticipated. The lack of an evaluation culture in ICBF also became clear. 17. The components proved to be flexible enough to permit required adjustments as the Project developed. At the end of 1995 and early 1996, the Government undertook an- initiative that, among other aspects, sought to aid children who were in hiding from the guerrilla movement. All departments and municipalities were to be involved in this effort. In this regard, ICBF took part by funding training activities under the Project to advocate MCs and local jurisdictions in dealing with such children under the human rights initiative of the Government. Although progress was not significant in overall institutional strengthening, the flexibility of the components did enable the introduction of new issues such as human rights and decentralization, and the formation of cooperatives among parents' associations.  -5- Policy Development Component (Original Baseline Cost: US$0.3 million) 18. Subcomponent: Establishing Financial Arrangements for the CAIPs. The Project design included a study to evaluate alternatives for the long-term financing of CAIPs, including the possible privatization of the Centers or their transfer to municipalities. Based on the study, recommendations to obtain alternative funding sources for CAIPs are being pursued. More important than the study results was the compulsory reduction of the Program budget and the pressure to maintain its quality standards, forcing the Institute to find alternative funding sources such as municipalities and private institutions, and charging higher monthly tuition fees. This realignment of funding sources for CAIP was helpful to the HBI Program, the Project, and the ICBF as a whole, as it reduced the subsidy to better-off families, thus permitting expansion of the better-targeted HBI Program. 19. Concurrently with conducting the study, ICBF met the goal of reducing its contribution to the CAIP Program by 30 percent. For 1992, this reduction had already reached 15 percent, exceeding by two points the percentage established by CONPES (Council on Economic and Social Policy) and agreed with the Bank. To date, CAIP accounts for 12 percent of the ICBF investment expenditure budget. Even though other alternative funding sources exist for this Program, a deficit of approximately US$2 million remains for benefits payable to personnel assigned to CAIPs. 20. Subcomponent: Improving the Efficiency of Production and Distribution of Bienestarina. The study to evaluate the potential for private production and distribution of Bienestarina and the resulting action plan were completed in August 1992, almost one year behind schedule. Subsequently, other studies proposing improvements in the supply and distribution of Bienestarina were funded between 1993 and 1995. ICBF reviewed and improved the composition of the micronutrients in this product, to fulfill the nutritional needs by age groups and food habits by region. 21. These studies provided ICBF with the essential decision-making information and criteria for the production and distribution of Bienestarina. Although the initial study recommended selling the manufacturing plants to the private sector, other analyses of the private market allowed the opening of bids for food production and pointed to the existence of an oligopolistic market in the private sector, as well as price manipulation practices for raw materials and production. As this would have jeopardized the goal of efficiency improvement, it was decided that ICBF should retain control of production and distribution of Bienestarina, while modernizing its plants, purchasing raw materials through the farming exchange, lowering prices significantly, using raw material substitute alternatives for product diversification, and regionalizing distribution. Institutional Development Component (Original Baseline Cost: US$2.2 million) 22. Subcomponent: Study of ICBF Organizational Structure and Staffing Changes. The Project supported two phases of study and restructuring of ICBF, in 1993 and 1995. The  -6 - first, which focused on the organization and personnel assignments, failed to achieve the expected results, in part owing to lack of support for change at the central level of the ICBF. The second helped ICBF to adapt its organizational structure to the new Constitutional requirements of Colombia, including administrative decentralization and community participation. No specific evaluation has been performed to assess the benefits of the 1995 restructuring in improving ICBF performance. However, it was noted that the restructuring strengthened local centers, established global staff groups as recommended in the SAR and improved the technical level of staff. 23. Decisions concerning the decentralization of services were postponed due to the lack of: (a) appropriate channels to guarantee an equitable distribution of funds collected; (b) political, institutional and legal arrangements; and (c) establishment of municipal administrative structures and capacity for provision of primary services, such as education and health. Nevertheless, decentralization is an ongoing, if gradual, process, and ICBF maintains its strategy for expansion of coverage through partnerships with municipalities. 24. As part of the ICBF restructuring objective, the Project supported an internal reorganization of processes. The main results were the adoption of procedures for auditing controls, adjustments to management control indicators and SIG results, as well as training personnel in the field in self-management. Likewise, support was provided for the introduction of controls to Programs under ICBF responsibility by the civilian society through citizen surveillance. 25. Despite these efforts, it appears that the institutional modernization process of ICBF remains unfinished. The ICBF institutional mission should be simplified, and its scope narrowed. This key issue for the organization and operation of ICBF was not resolved but actually aggravated in 1995, when its restructuring was meant to reflect a national law based on the restructuring of the Family Welfare System in general; however, this further strained the already weak institutional capacity of ICBF to meet its objectives. 26. Building on the organizational restructuring objective, a national strategy and awareness plan encouraged ICBF to promote a human rights "culture" within the institution, as well as within the institutions contracted by ICBF to execute its programs. An information system on children's rights violations was also created. This system will serve at the same time to follow up on the enforcement of the Convention for Children's Rights. Its initial application to the HBI Program has not yet been evaluated. 27. Subcomponent: Establishing an Integral Computerized Management Information System (SIG) for the Planning, Programming, Budgeting and Monitoring of ICBF Programs. The SIG was designed as a tool for the management of administrative and monitoring processes of ICBF Programs (planning, programming and budgeting), including the HBI Program. For the HBI Program, the SIG supports monitoring of coverage and internal efficiency by providing: (a) annual updates on the evaluation of needs; (b) basic information gathered from MCs on children's nutritional and health conditions; (c) selected information on parents and other participants; (d) information on the allocation and use of resources; and  - 7 - (e) information required for financial control and monitoring. The SIG has proved to be a useful monitoring tool for the HBI Program. 28. Although the SIG was to be introduced by June 1991, its development began in 1992. It has yet to be fully completed due to: (i) problems in recruiting sufficient qualified staff; (ii) delays in procurement of equipment at the central level; (iii) differences of opinion on technical grounds; (iv) delays in defining the variables feeding the system and the lack of prototypes (as identified in 9/92 and 12/94 missions); and (v) lack of proper physical and technical infrastructure at the decentralized level for implementation. Nevertheless, hardware and software were procured in 1992-93, and the system is currently installed in 15 percent of the total 245 sites where it will be used. The services subsystem for the HBI Program is in the process of being designed and tested. Related training for its application has been scheduled. 29. Based on ICBF's experience so far, the SIG is one of the most complete information systems available among Colombia's public institutions. It has provided ICBF with many positive results, including significant reduction in the processing time for child adoption and tax collection processes. However, emphasis should be given to developing a planning culture that permeates the entire organization and ensures the proper use and updating of the SIG, as well as to extending planning processes beyond simple programming and monitoring of budgetary and physical goals. 30. Subcomponent: Introduction of Nationwide Monitoring and Impact Evaluation Systems. Originally, the Project sought to monitor ICBF operations, particularly those of the HBI Program, essentially depending on information gathered from the SIG (SAR, par. 2.12). The impact evaluation, on the other hand, was to be based on a national survey, conducted annually in several cities, with both HBI and non-HBI families participating. This information was to be supplemented by a broader annual survey undertaken by the Departamento Administrativo Nacional de Estadistica (DANE) in 13 cities. Both direct and indirect effects of the Program were to be measured, including the impact on participants and other family members. In particular, the income and price effects on household consumption would be examined to estimate total impact on family nutrition and health, as well as the income effect of additional free time for the mother (dedicated to work, studies, leisure) that could be attributed to the Program. However, very few of these goals were met. 31. The design and implementation of a comprehensive impact evaluation started in 1995. Earlier, only limited efforts at program evaluation had been made. In 1991, DNP conducted an ex-post evaluation of the HBIs (Sarmiento, 1991), which focused on administrative processes more than impact. In 1991-92, UNICEF (United Nations Children's Fund) and ICBF conducted an "Evaluation of HBIs" (Ortiz, Nelson et al., 1992), which lacked quantitative rigor and, in particular, a control group. In 1994, ICBF performed a National HBI Census to generate a data base to be used for the design and implementation of regional qualification plans. However, although the census was a useful tool, it did not include data regarding direct care provided to children, the target group of the Program. In 1995, standards were established to monitor quality and outcomes of the HBI Program, as part of an  - 8 - ICBF Management and Results Evaluation System design. Finally, ICBF initiated in 1995 the design, development and implementation of a Permanent Quality and Impact Evaluation System of the HBI Program. The First Impact Evaluation Survey was conducted nationwide during the first quarter of 1996. 32. One of the objectives of the survey was to determine the effects of ICBF interventions on the welfare of HBI children. While more rigorous than the earlier efforts, the HBI study also had no control group. In a joint analysis with the Bank, two complementary information sources, although insufficient, were used. To overcome this limitation the 1995 Demography and Health Survey (DHS-95) conducted by PROFAMILIA (Association for Colombian Family Welfare), where HBI users and non-users could be identified along with a "continuous duration in the Program" variable. This variable helped to compare groups with different periods of exposure to the Program. The DHS-95 Survey sought to identify health and nutrition conditions in the lower socio-economic strata and HBI non-user population, and compare them with the Program beneficiaries, taken in the First Impact Evaluation Survey conducted by ICBF. 33. The ICBF survey enabled the direct effects on children covered by the Program to be analyzed. The indirect effects on, and benefits for, the child's family were not identified, although this was one of the original objectives of the Impact Evaluation System proposed under the Project. Overall, the ICBF survey developed a good conceptual instrument and sampling tool for a permanent impact evaluation system of the HBI Program on a national scale, which will allow future follow-up. Even though this survey is cross-sectional, it will permit a longitudinal focus through the establishment of an HBI sample size that guarantees a sample turnover for a cycle of four surveys. It will be important to include in the survey a control group of non-beneficiaries in the target population of the Program. 34. Despite the progress made in developing the Evaluation System and the SIG Service subsystem, which represent the basic tools for regular monitoring and evaluation of the HBI, the two systems have not yet been integrated. Both systems have partially used the indicators defined in quality standards for the HBIs, but no structured information support system is in place for using the hardware procured or the software that has been developed. Service Support Component (Original Baseline Cost: US$30.1 million) 35. Subcomponent: Provision of Regular Trainingfor Community Mothers and Other Community HBI Participants in the Management of Food, Nutrition, Health Registers, Child Development and Home Management. The training of MCs and others was identified as a priority need to achieve and maintain adequate quality of the HBI Program during expansion. Training for MCs focused on nutrition, food management (cooking and hygiene), child growth and health monitoring, and activities related to pre-school development (see SAR 2.14). The continuous training system (SFP), which combines individual study and group discussion, used by ICBF at the time of Project design, was judged appropriate, though deficiencies were acknowledged in the results of the training provided to MCs (SAR 1.3). The Project initiated the SFP, which was centralized until 1994 and later decentralized to  -9- enable the regions, municipalities and NGOs to contribute. Educational centers were identified to support the training, creating the need for a coordinating agent who would be responsible for monitoring training and updating the Program as educational needs were detected. 36. The formulation of local training plans coincided with increased flexibility in the management of loan resources based on an agreement with UNDP. The HBI Program quality standards became a fundamental tool in guiding the MC training processes and profiles. Successful experiences were gathered as a means of providing feedback to the process. Noteworthy among these experiences was the case of the Municipality of Manizales, with its strong local and private sector linkages. This enabled a significant improvement both in training provided under the Program as well as in the services provided by the MCs, based on the institutionalization of supervision and monitoring in HBI homes by local (municipal) centers. 37. In 1996, economic problems forced the central level of the ICBF to resume training management. In late 1995, an agreement was reached between the Bank and the Borrower to cease the promotion of events that were detracting from the essential purpose of the Project, and it was decided to redirect those resources to reinforce the technical assistance required to drive the impact evaluation process. 38. The Project financed the execution of decentralized training plans and the procurement of teaching materials to support the MCs training and the design of materials and activities related to psycho-social development. ICBF also developed a literacy and basic education program for MCs. The First Impact Evaluation Survey reveals that 27 percent of the MCs still have not finished primary school and 1.3 percent of them are illiterate. 39. Leadership and microbusiness management training was also provided to MCs and parent participants seeking to improve their family income. Partly as a result of this training, many MCs emerged as leaders in their communities. The training strategy was reoriented in 1995 to focus on improving outcomes for HBI. 40. The ICBF survey indicates that child malnutrition and morbidity rates and risk profiles for psycho-social development are similar to non-beneficiary groups. Significantly, while the Survey showed positive results with respect to MCs' performance in the administrative aspects of home management, it revealed deficiencies in the quality of care provided. For example, only 35 percent of MCs could prepare home-made rehydration salts, only 29 percent provided proper nutrition to children according to the standard menu, 18 percent kept no record of children's weight and height, 57 percent did not take the standard measurements of children upon their enrollment, 35 percent could not interpret the child growth chart, and less than 10 percent could interpret the individual and collective growth curves. The main reasons given were that the charts were unavailable or that they did not receive needed training.  - 10- 41. The survey indicated high turnover of children in the Program: 51 percent participated less than one year; 26.5 percent spent 12-23 months in the Program, and 22 percent remained in the Program for 24 months or longer. This high turnover of children in the HBIs remains unexplained and requires firther analysis. 42. From the survey results it may be concluded that the Program had no clear effect on the nutritional condition of children. Duration in the HBI Program is associated with some reduction in overall malnutrition (weight/age) and of acute malnutrition (weight/height), but there is no apparent impact on chronic malnutrition (height/age). The low impact of the Program on child nutrition may be based, in part, to deficient diet provided by the HBIs. Only 10 percent of the HBI homes complied with the standard ICBF requirements of providing an appropriate diet, 65 percent provided a deficient diet, and 25 percent provided a severely deficient diet. 43. The results showed that 70 percent of the children surveyed were ill two weeks before the evaluation, and the same percentage of HBI beneficiary families were not registered for health care. However, major progress has been made recently in the Social Security Health System coverage and the expansion of the Basic Health Care Plan (Plan de Atenci6n Bdsica en Salud), with the support of the Social Solidarity Network (Red de Solidaridad Social). Program support for the health of MCs and their families improved, and 85 percent of the HBI beneficiary families reported receiving health care. Nevertheless, the Survey evaluated 12 psychological processes, but the results were inconclusive. It should be noted that the Program's design did not include clear, systematic interventions for stimulation of children's psycho-social development. 44. Multivariate analysis indicates that the following factors have major effects on nutritional status, psycho-social development, and morbidity of the children participating in the Program: (a) physical condition of the HBI; (b) MC's educational level; (c) children's growth evaluation; (d) food preparation; and (e) monitoring activities. The training received by the MC had no significant effects on the children's health and nutrition. 45. In view of the above results, the relative impact and cost-effectiveness of HBIs as compared with CAIPs should be reevaluated. CAIPs have been found to be of better quality, and the cost advantage of HBIs has decreased over time, as consideration has been given to indirect costs and the need for additional inputs. At inception of the HBI Program in 1987, unit costs for CAIP were 2.6 times those of HBI, but in 1995, this ratio was estimated at 1.2 (ICBF, 1996). 46. Training of Other Actors: Training aimed at strengthening the responsibility of parents in child rearing and care was weak. The Project supported a study to form cooperatives of Parent Associations and to integrate these cooperatives with the microbusinesses that might emerge and with individual parents. This was intended to improve supervision of the resources received by HBIs from ICBF as well as to provide opportunity for higher family income. Workshops were conducted in all regional offices, and 102 cooperatives were created. Currently, a shortage of resources has slowed the creation of  new cooperatives. This strategy has increased social benefits and improved income distribution through the formation of multi-action cooperatives which generate and redistribute surpluses among their members. Surpluses may cover health services for HBI parents, procurement of teaching materials for children, and higher investments in household improvements. 47. Training of ICBF Staff The Project funded training for ICBF personnel at all levels. Areas covered included restructuring, decentralization, citizens' surveillance groups and. citizen participation, co-management strategies with municipalities, human rights awareness, the use of impact evaluation and internal control tools, the training MC trainers, and computer and SIG management. 48. Subcomponent: Loans to Community Mothers for Household Improvements. Support for the loan Program for MC home improvement was considered essential to guarantee better child care and attention. Project resources were to permit at least 75 percent of the MCs to improve their homes (SAR par. 1.32). To expedite the process, resources were to be managed through a trust fund. The community participated in selecting the beneficiary mothers, as well as in the supervision of works, initially through Parent Associations (SAR 2.23). This role was later assumed by universities and NGOs, as a means to improve technical assistance and supervision. Today, 75 percent of the ICBF regional offices are receiving that technical assistance. The improvement of 50,000 homes were set as the original goal. To date, some 60,711 MCs have benefited from the loans, including some recipients who recieved more than one loan. 49. The system chosen to develop the HBI Program was subsidized credit, paid directly through discounts on grants given by ICBF to the MCs, a system which has proven to produce healthy collections and Program capitalization. Currently, the Program for MC home improvements has resources of about US$10 million, as a result of collections on MCs' debts and interest generated through trust fund management. The HBI Program gained from the participation of NGOs, universities and others, all of which provided technical assistance and works supervision, which strengthened the administrative, technical and financial decentralization by ICBF. As of 1997, the Program information systematization was initiated to improve collections management and provide better knowledge of the specific use of the loans. 50. Despite meeting physical goals of the housing Program, the First Evaluation Survey indicates that, of the 48 percent of the MCs who received a housing loan, 72 percent used the funds as intended, i.e., to improve bathroom facilities, kitchens, and children's areas. However, 28 percent used the funds to satisfy other household-related needs not originally considered in the Project's guidelines. These included payment of arrears and land purchasing, indicating a need to improve control over the use of housing loans. 51. The survey indicates overcrowding of children at many HBIs: 51 percent have less than one square meter per child. This may indicate that the Program needs more flexible regulations regarding the maximum number of children per home: while the Program sets a  -12- ceiling of 15 children per HBI, this may exceed the capacity of some HBIs. Moreover, 15 children under seven years of age may be too much of a workload for the MCs guarantee qualified attention. This aspect of the HBI model needs to be reevaluated. 52. The physical conditions shown by survey results were precarious in many HBIs: 36 percent of them lacked a sewage system; 24 percent had no water supply system; and 37 percent had feces inside and outside of the home, and floors in bad condition. The food preparation and storage sites were equally deficient: 16 percent of homes had the kitchen on the same site as that of child care, and 14 percent lacked the proper area for food storage. Although the conditions of HBIs that made use of the loans were better than those of non- beneficiaries, the prevalence of poor physical conditions suggests the need for improvement. The maximum loan amount should be studied to gauge whether it is enough to satisfy the required household improvements to provide proper child care. Furthermore, the MC and housing eligibility standards need to be reviewed. Given that the Program operates in a low socioeconomic context, greater efforts are required to ensure adequate child care. 53. Subcomponent: Technical Assistance to ICBFfor Program and Project Management and Coordination. The Project provided for a small Management Unit to provide support to the ICBF in project administration, while assigning responsibility for technical aspects of the Project to ICBF departmnents, thereby avoiding parallel structures. Until 1994, ICBF was in charge of resource management, and goods procurement was contracted with UNIDO (United Nations Industrial Development Organization). In September 1994, an agreement was signed with UNDP (United Nations Development Programme) for the management of Project resources, and in November 1995 with OEI (Organization of Ibero-American States for Education, Science & Culture), for the administration of training resources. C. Key Factors Affecting Project Implementation 1. Organization and Management 54. By virtue of being a part of the HBI Program, and not having its own structure, the Project was subjected to changes at the ICBF top management structure and the various guidelines given by five different directors appointed during its execution period. Performance was also affected by high turnover in the Management Unit (four coordinators in six years). During much of the Project, the Unit lacked sufficiently qualified staff. Delays in establishing the information system and the lack of an evaluation culture in ICBF influenced Project development. 2. Project Design and Scope 55. Many external factors affected Project development, including: (i) constitutional changes, which defined a new role for the central level and decentralization and public participation at all levels; (ii) application of revised economic policies and decentralization; (iii) the outburst of social instability and increased violence, narcotics trafficking and armed conflict; (iv) political crisis; and (v) an economic crisis and recession, with subsequent  - 13 - application of fiscal measures that often caused delays in loan disbursements. The risk of these factors was not considered at the Project design stage. 56. The Project was designed to support the expansion of a Program that had been underway since 1987, which in turn had its own design and strategies. No indicators were designed specifically for the Project other than those of the Program, which ultimately made supervision, management, and monitoring activities difficult to perform during project implementation. There were also internal security issues that complicated project supervision. 57. The objectives established for the Project were broad support for, and expansion of, the HBI Program. The Program's resources for the past five years were only 48 percent of the investment expenditure funds, and 40 percent of the total ICBF funds. 58. As a result of significant delays in the first years of the Project life, only 92 percent of the loan funds were disbursed (Table 4). The main reason for these delays was ICBF's institutional weakness (see par. 54). 59. Lack of clear quality standards was a weakness of Project design. 3. Inter-Sectoral and Intra-Sectoral Relations 60. Many of the Project's activities described in the SAR depended on outside agencies such as the Ministry of Health (for the health care for the children) and the Ministry of Education (pre-school education for children participating in the HBI Program), INURBE and IFI (housing component and formation of cooperatives by parent's associations under the HBI Program), United Nations Industrial Development Organization (UNIDO), United Nations Development Programme (UNDP) and OEI (ICBF's contractors for the management of Project resources), SENA (MC training), DNP and Ministry of Finance (disbursement effectiveness). Territorial Entities (municipalities) and civil society (NGOs, universities, parent's associations) were also involved. However, the relations with the above institutions were not always effective or sustained. No formal agreements were established to clarify responsibilities of other agencies and confirm their commitment to carry them out. 61. On the positive side, ICBF initiated a co-financing process for welfare services with the municipalities, through which strategic alliances with the public and private sectors, NGOs, and universities increased the quality of the services provided and extended their coverage. D. Project Sustainability 62. ICBF's financing structure (mainly from tax revenues) ensures the continued financing of the HBI Program. The loan's contribution to the financing of the Program was quite modest (six percent).  - 14- 63. The Impact Evaluation System, introduced with the First Impact Evaluation Survey, still in process, is important for ICBF. It is expected that, given the strength and impact of its results, the system can provide the basis for sustainability, not only for the HBI Program, but also for other ICBF Programs. 64. Although encountering delays, the monitoring system is also likely to be maintained through the SIG in the future. It has become institutionalized and its application has demonstrated benefits. 65. The credit subcomponent for the improvement of MCs' homes yielded sufficient returns (US$10 million from the payment of loans to the MCs and the interest generated through the trust fund) to allow for its continuation. The necessary arrangements have been made for this program to supplement the Government housing programs for lower-income populations. 66. Though sustainable, training activities should be reviewed using the results found in the survey to reformulate the process, methodology, contents and supervision of the training for MCs. Permanent supervision should also be introduced to maintain quality standards and ensure continuous training. This action has been undertaken by allowing the participation of NGOs and universities in the training and monitoring process, to improve the quality of training. 67. The co-financing of projects with municipalities will continue to be of importance to sustain the HBI Program, as additional resources are required to broaden coverage and qualify services. E. Bank Performance 68. Bank support and guidance was key in the definition and development of quality standards for the HBI Program, technical assistance for the impact evaluation, suggested educational strategies, training plans and process systematization (SIG). However, the Bank could have employed more rigorous control, follow-up and analysis during the first years of Project life with respect to some components (e.g., timely development of the SIG, the Impact Evaluation System, and the training process for MCs). Given the technical support (rather than investment) focus of the Project, typical loan sanctions provided (i.e., dated covenants) were not particularly useful in ensuring effective Project implementation. In retrospect, the Bank should have provided more support for design of the impact evaluation prior to the Project, given its importance for both the Project and the Program. Stronger emphasis on the design and implementation of the impact evaluation was applied toward the end of the Project life, although a baseline was not clearly established. F. Borrower Performance 69. The relatively small financial contribution of the Project to the HBI Program was sometimes reflected in low priority assigned to it by some of the ICBF administrations.  - 15 - ICBF's application of the Bank's recommendations on some components (especially MC training and impact evaluation processes) was slow. This disregard by successive ICBF and Project managers of activities that were of key importance to the Program was reflected in both delays (impact evaluation) and low quality (training). ICBF's inter-institutional coordination was generally weak with central-level institutions, but effective in establishing operational alliances with municipalities. G. Assessment of Project Outcome 70. Overall, the results of the Project are satisfactory. The Project achieved its objective of influencing coverage expansion in part, with high incidence on the target population. The HBI Program had only a limited effect on the welfare of children, expressed in terms of their health and nutrition and their psycho-social development. However, the Project ultimately succeeded in introducing an evaluation and information culture in the ICBF. These two elements should help ICBF improve quality in the future in terms of home improvements and knowledge of home management that should benefit local communities and MCs in particular. Actions taken to support ICBF institutional strengthening, along with the SIG and the impact evaluation processes, are also positive outcomes. H. Key Lessons Learned U Understanding both the political and institutional climate has important implications for the immediate impact of the project and sustainable outcomes. Projects such as this one that involve different participating institutions to support the achievement of Project objectives should establish agreements with authorities or institutions, and provision should be made for periodic review/revision /ratification of institutional arrangements or agreements throughout implementation. This is especially germane in the case of first- time projects in sectors with weak or nonexistent institutional linkages at project inception. * The scale and location ofprojects has a direct bearing on whether government agencies or NGOs are better positioned to achieve optimal project results. Centralized procedures, characteristic of the older operations supported by the Bank and the Colombian Government, were poorly suited to a highly decentralized and participatory service delivery model such as the HBIs. In this case, the involvement of NGOs might have led to a greater project impact. * Performance indicators, even for already-piloted models, such as the UNICEF-HBI model, need to be developed independent ofprevious experience and adapted to institutional arrangements under specific country andproject conditions. Projects need adequate baseline data on which to compare progress. Design should reflect and already incorporate these data including plans for evaluation follow-up. In addition, project design should include provision for periodic evaluation of standards and testing of assumptions, and for revision of guidelines in response to impact and experience. More  -16- importantly, when using a "model" approach, sustainability on the scale, scope, and other key factors need to be evaluated to guide a tailoring to the country, institutional, and project-specific environment. * New information systems may prove difficult to design and implement, which needs to be reflected in implementation timing milestones. The delays in implementing ICBF's information systems prevented it from reaching the scale intended. The system was difficult to design, and, above all, to implement, as in the case of SIG and the impact evaluation, where insufficient attention was given to practicability and unrealistic compliance times were specified.  - 17- COLOMBIA COMMUNITY CHILD AND NUTRITION PROJECT (Loan 3201-CO) IMPLEMENTATION COMPLETION REPORT PART II: STATISTICAL INFORMATION Table 1.1: Summary of Assessments Table 1.2: Assessments over Time Table 2: Related Bank Loans Table 3: Project Timetable Table 4: Loan Disbursements: Cumulative Estimated and Actual Table 5: Key Indicators for Project Implementation Table 6: Key Indicators for Project Operation Table 7: Studies Included in Project Table 8.A. 1: Project Costs and HBI/ICBF Program Costs by Component: 1990-95 (US$ million) Table 8.A.2: Estimated Project Costs by Item: 1990-96 (US$ million) Table 8.A.3: Estimated Project Costs and Credit Executed by Item: 1990- 96 (US$ million) Table 8.B. 1: Project Estimated Funding (US$ million) Table 8.B.2: Actual Project Costs (US$ million) Table 9: Economic Costs and Benefits Table 10: Status of Legal Covenants Table 11: Compliance with Operational Manual Statements Table 12: Bank Resources: Staff Input Table 13: Bank Resources: Missions  -18- Table 1.1: Summary of Assessments A. Achievement of Objectives Substantial Partial Negligible Not-Applicable Macro Policies E E El Sector Policies El E1 El Financial Objectives rl EJ C1 Institutional Development Q1