Malawi - The First and Second Education Sector Credit Projects

Document of The World Bank FOR OFFICIAL USE ONLY Report No.: 20773 PERFORMANCE AUDIT REPORT MALAWI FIRST EDUCATION SECTOR CREDIT (CREDIT 1767-MAI) SECOND EDUCATION SECTOR CREDIT (CREDIT 2083-MAI) August 17,2000 Sector and Thematic Evaluation Group 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. Currency Equivalents (annual averages) Currency Unit = Malawi Kwacha (MKwacha) 1986 US$1.00 MKwacha 2.00 1993 US$1.00 MKwacha 4.45 1994 US$1.00 MKwacha 14.97 Abbreviations and Acronyms AfDB African Development Bank EU European Union FPE Free Primary Education ICR Implementation Completion Report IDA International Development Association IMF International Monetary Fund MASAF Malawi Social Action Fund MOE Ministry of Education OED Operations Evaluation Department PAR Performance Audit Report PIU Project Implementation Unit Fiscal Year Government: April 1 - March 31 Director-General, Operations Evaluation Mr. Robert Picciotto Director, Operations Evaluation Department Mr. Gregory Ingram Manager, Sector and Thematic Evaluation Mr. Alain Barbu Task Manager Ms. Linda Ankrah Dove FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operations Evaluation August 17, 2000 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on Malawi First Education Sector (Credit 1767-MAI) and Second Education Sector Credit (Credit 2083-MAI) In a difficult period of economic and political adjustment in the 1990s, the two IDA operations gave timely support to the Government of Malawi in transforming its educational policies to support economic and social development. However, only after excessive delays and costs, some of which might have been better controlled, did they achieve results from the large primary school construction components. The Ministry of Education's (MOE) investment plan called for an 85% enrollment of the primary school age-group by 1995-an ambitious target requiring thousands of additional classrooms. OED's audit focused on the outcomes of efforts to strengthen the MOE's capacity to implement this investment program and to achieve construction targets with community self-help. Annex 3 discusses alternative approaches to community participation that promise rapid and cost-effective results. It also contributes to OED's ongoing analysis of social funds projects in referring to the Malawi Social Action Fund projects (Credits 2856 and 3136). Serious external factors beyond the control of the project managers during implementation receive prominent recognition in IDA's completion reports, as was appropriate. OED emphasizes the internal factors that led to lower than expected benefits-and were obvious risks based on the experience of all the previous education projects. They included the poor status of preparation of the construction component; implementation backlogs and credit extensions due to pressures from overlapping projects under IDA and other donors; the governments failure over four years to provide a qualified project manager and staff; problems of intra-ministry communication and workload; and inadequate attention by the senior authorities to efficiency and covenant compliance. In line with its approach in two previous projects, IDA staff considered the various problems- programming, construction materials, availability of community labor-as mainly technical. But when modifications failed to produced timely results, they agreed with project staff that "the spirit of self-help had waned." After a surge of popular enthusiasm in 1994 when the new democratic government introduced "free" primary education (FPE), further delays led many disappointed communities to regard the schools as entirely the government's problem to fix. OED's review suggests, however, that the basic problem was not the community spirit but the narrowness of the IDA approach. The project's focus on input-output (community brick-making-low-cost facilities) mirrored the early engineering tradition of Bank infrastructure projects but the appraisals did not mobilize adequate engineering expertise. Equally important, the projects did not tailor the techniques used to fit organic processes of self-help. This means that socioeconomic and cultural realities were underestimated, including the unique patterns of community life and work. The fact that the concurrent social fund projects reportedly, succeed in mobilizing communities to build schools casts doubt on the notion that communities are unwilling to help themselves. 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. 2 In 1994, staff welcomed the government's FPE announcement as consistent with Bank policy and strategy and retroactively adapted the second credit to support it. IDA's decision to proceed was controversial at the time given the weak implementation capacity and, in the event, there were delays and cost overruns. The Region made a difficult decision to go with the political momentum and to reinforce IDA's own credibility in promoting primary education for all. OED Ratings OED rates outcomes as satisfactory though marginally so. Despite the problems, the projects achieved a threefold expansion in classroom capacity, well beyond original expectations. In urban areas, enrollment in primary school has become the social norm. Nevertheless, the delays discussed allowed rapid increases in the school-aged population to outpace the enrollment increases. Enrollments remain at the same level as in 1995. Primary education remains far from universal and dropout and graduation rates remain high, especially for the rural poor, while student achievement levels have declined drastically. Institutional development impact is rated as modest. Capacity is not yet adequate to handle the transition from a small school system to one covering the entire country. Compared with the 1980s, the capacities of the MOE's Planning Department have been strengthened. But the task of constructing school facilities for a mass school system requires a depth of managerial, technical and participatory skills that the projects grossly underestimated. How to build permanent implementation capacity is a critical question that still vexes the government and the Region. If a third sector credit were planned, the issues of complexity, demandingness, feasibility and economic returns would require rigorous scrutiny. OED rates sustainability as uncertain in the light of recent developments. With the recurrent budget for non-salary items currently declining, the erosion of educational quality has become a serious issue. The budgetary shortfalls cast doubt on the government's commitment to complete the unfinished business in primary education and put at risk in a pre-election period the widespread public support for the mass education policy. A related challenge to sustainability is the inadequacy of mechanisms to ensure that primary school facilities will be well maintained and repaired over the 30-year life of the investments. Though the MOE has developed a maintenance policy, the willingness and ability of local communities to contribute will be critical. The minimum requirement is that qualitative inputs must be visible to communities - to responsive school managers, well-trained teachers, well-equipped classrooms and relevant programs. Public confidence that all pupils have equitable opportunities to succeed is currently low. IDA's performance is rated as satisfactory overall, though the lack of realism is cause for concern because of the economic costs of delays and shortfalls. Although the IDA team was familiar with the recurring problems in the education projects and in the country, its confidence in government assurances that adequate funds, staff and implementation arrangements would be in place turned out to have been misguided. OED proposes that staff's preoccupation with the long-term policy reforms and their very commitment and responsiveness to the countries' educational needs may have led them to regard the implementation risks as a normal cost of doing business. In turn, this may have encouraged day-to-day troubleshooting at the expense of fundamental reassessment and more radical remedies. With the best of intentions, IDA may have held the government to less rigorous standards than it would expect from high- performing borrowers outside Africa. However, the solutions ultimately rest with the government as much as the external financiers. 3 Lessons of Broad Applicability IDA needs to consider carefully the justifications of presenting a scheduled operation for Board approval when risk analysis reveals that: (a) serious constraints within the government's control have not been removed; (b) the scale and sequencing of donor projects in the sector overload implementation capacities; and (c) the country's overall capacity has proven so weak that delays will demand credit extensions, raise costs and reduce benefits. During preparation, mutual candor and realism between IDA and the borrower are needed to ensure readiness for implementation, especially for conventional components. In the Malawi case, the innovative components were rigorously assessed in terms of benefits and risks, but the approach to implementing the more familiar school construction components was not thoroughly examined. In developing new approaches to the technical aspects of school construction, the advice and experience of local qualified personnel, such as engineers and quantity surveyors, need to be available. This was a lesson of the Bank's education sector experience in the 1970s and 1980s and is still critically important for borrowers who are still putting in place the physical infrastructure for universal primary education. In this case, the errors in design of the roofs and the selection of materials and techniques might have been avoided with more skilled local appraisal and diagnosis. Similarly, when developing components that introduce indigenous traditions, IDA needs to call upon cross-sectoral expertise. In this case, the sociocultural dynamics of community self-help required the same rigor of study as the budgetary component received. This would have allowed staff to re- examine the limitations of the instrumental approach and a search for more holistic alternatives. The performance audit emphasizes the need for IDA to balance the rapid provision of funds to fill educational deficits with systematic review of institutional capabilities and evidence from monitoring and evaluation studies about feasibility and effectiveness. For example, syntheses of local communities' experiences in self-help schemes would have been useful in developing a more effective approach. By the mid- 1990s, international expertise in participation and beneficiary assessments was readily available. Equally, within Malawi, the MASAF evaluations provide a useful basis for future policy, strategy and preparation of education sector projects. Robert Picciotto By Nils Fostvedt Attachment  Contents First Education Sector Credit (C1767)...................................................................................... iii Second Education Sector Credit (C2083) ................................................................................... iii Preface ............................................................................................................................................. v 1. The Context ..............................................................................................................................1 Country and People................................................1 The Education System....................1....... ...............1 Primary Education ........................................ .........2 Donor Assistance............................. ..................... 2 The Education Sector Credits.......................2..... .............2 2. Implementation Experience ............................................................................................... 3 Internal Factors Affecting Implementation................................4 The Status of Preparation......................................4 Overlaps, Backlogs and Bunching................................5 Organization and Communication....................... .............5 Qualified Managers and Staff....................... ...............5 Workload Pressures..........................................6 Multiple Demands of Donor Projects................. .............7 Bank Supervision...................................................7 3. Primary School Construction: Quantitative Achievements and Capacity-Building....8 Performance of the First Sector Credit....................................8 Achievements...............................................8 Delays and Project Costs........................................9 Performance of the Second Sector Credit..................................9 Achievements...............................................10 Costs of Primary School Construction ............................. 11 Linked Output under the Primary Education Project n.paid...............i.........1 Combined Impact on Classroom Access.................................11 Costs and Benefits ...................................................12 This report was prepared by Linda Ankrah Dove (Task Manager), who audited the project in February 2000. William Hurlbut edited the report. Marie Daramy provided administrative support. o 4. Community Self-Help in Primary School Construction.................................................12 Lessons of Previous Projects................................. ........12 Technical Design and Community Results................................13 Insights from the Education Project Experience...........................13 Can the Community Spirit be expanded and sustained?......... .............14 5. Findings...................................................................................................................................15 Outcomes........................................................15 Institutional Development Impact............ .................. ......15 Sustainability..................... ............................165 IDA Performance .......................................... .....176 Borrower Performance.............................................17 Lessons ....................................................... 17 Annexes A. Basic Data Sheet (C1767 MAI & C2083-MAI)................................................................19 B. The Education Sector Credits: Policy Actions; Objectives & Components..........25 C: Alternative Paradigms: Communities and School Construction...................................27 D: Comments from the Ministry of Education ..................................................................... 35 BOXES 1. Challenges - Economic, Fiscal & Political.....................................3 2. Preparations for the Primary School Facilities Component.......... ..................4 3. Two Decades of Staffing Issues Repeated.................... ................6 4. Relentless Pressure from IDA Lending......................................7 5. Enrollment Targets and Demographic Impact................... ..............9 6 Teacher Housing.....................................................10 Cl. How MASAF Works.....................................................27 C2. Lessons Learned in MASAF................................................28 C3. Competitorsfor Market Share..............................................29 Tables 1. Additional Classroom Provision and Pupil Capacity............................12 Cl. Sharing Knowledge and Experience for Future Strategy..........................31 iii First Education Sector Credit (C1767) Principal Ratings ICR Audit Outcome Satisfactory Marginally satisfactory Sustainability Uncertain Uncertain Institutional Development Partial Modest Borrower Performance Satisfactory Satisfactory Bank Performance Satisfactory Satisfactory Key Staff Responsible Task Manager Division Chief Country Director Appraisal N. Shields J. Scearce J. Kraske Completion N. Kulumeka J. Grawe B. Kafka Second Education Sector Credit (C2083) Principal Ratings ICR Audit Outcome Satisfactory Marginally Satisfactory Sustainability Likely Uncertain Institutional Development Partial Modest Borrower Performance Satisfactory Satisfactory Bank Performance Satisfactory Satisfactory Key Staff Responsible Task Manager Division Chief Country Director Appraisal N. Shields J.Scearce J. Kraske Completion P. Ngomba R. Kagia B. Kafka  V Preface This Operations Evaluation Department (OED) Performance Audit Report (PAR) reviews two related sector investment and maintenance operations in Malawi financed by the International Development Association (IDA). The First Education Sector Credit (Cr. 1767-MAI) was approved May 7, 1987, for SDR 21.4 million (US$27 million equivalent). The credit closed on December 15, 1993, except for the civil works and consultant categories, which were extended until December 15, 1994, with a grace period until April 15, 1995. The Second Education Sector Credit (Cr. 2083-MAI) was approved December 12, 1989, for SDR 29.5 million (US$36.9 million) and closed on June 30, 1998, after two one-year extensions. A related Primary Education Project (PEP) (Cr. 2810) was approved in 1986 and, after extensions, is due to close in 2000. It is referenced, but is not audited in this report. OED's focus is on the projects' implementation performance and the results of the construction of primary school facilities, including community self-help. The theme was selected for several reasons. School facilities constituted the largest share of project costs and were to directly benefit the very poor but delays and cost overruns turned into a major concern. OED hopes to provide a perspective relevant to ongoing discussions between the borrower, the IDA Country Office and IDA headquarters staff on this issue. The theme is relevant to an ongoing OED analysis of social fund projects that also included communities (Annex C). The projects were broad and complex in scope and IDA's role in sector-wide policy, budgetary and qualitative developments deserves more in-depth analysis than could be covered in this project audit. OED reviewed IDA's Implementation Completion Reports (ICRs) and the reviews of the Ministry of Education (MOE) and the Project Implementation Unit (PIU) (ICR nos. 14661, June 1995 and 18755, December 1998), as well as supervision reports and correspondence files. OED found that the Region's data and analysis represented a balanced and reasonably consistent account of operational experience. The audit benefited from an 8-day mission in Malawi and from the parallel mission for the Malawi Social Action Fund projects (MASAF I, Cr. 2856 and MASAF II, Cr. 3136). The findings are also informed by interviews with officials in the government and other relevant agencies, representatives of two donors, and IDA staff familiar with the projects. The mission visited six schools in one urban, three periurban and two rural communities on the way to, and in the neighborhoods of, Lilongwe, Zomba, Blantyre and Salima in the south and central regions, but did not visit the more affluent northern region. Field interviews included local officials of the MOE and local governments, an academic, two teacher trainers, and members of a school committee, 15 schoolteachers, and brief discussions with two mothers and many pupils. The interviews were not representative, though they included stakeholder and beneficiary viewpoints and interests. Following customary procedures, OED fent copies of its draft report to the relevant government officials and agencies for their review and comments. Comments received from the Ministry of Education have been incorporated into the PAR as annex D. 5 1 1. The Context Malawi is a small, landlocked, low-income and highly indebted country in Southern Africa. The First and Second Education Sector Credits were expected to be a strategic turning point for the education system. They were designed to finance development of a sector-wide policy framework; increase budgetary allocations and expenditures for education that would be equitable and sustainable; strengthen sector management; and enhance local capacities to implement large investment programs efficiently.' The broad goals of the government's education policy were to enhance the quality and relevance of the curriculum and instructional process; improve the efficiency of educational management; and expand access to schools. The Education Plan, 1985- 95, had launched the new policy direction. It aimed for 85% enrollment of the primary school age-group by 1995-an ambitious target estimated to require 8,000 new classrooms and renovation of 5,400 more, as well as investments in teachers and instructional materials. Resources for expansion, nevertheless, were severely constrained. Education expenditures constituted only 10% of total government expenditures and successive IMF and IDA adjustment operations had further constrained education budgets. Country and People Until the early 1990s, single-party, presidential rule had limited civic participation. But the country has few resources other than its people. On all socioeconomic indicators its status remains low compared to similar African countries and it has one of the worst HIV/AIDS epidemics in Africa. This has decimated schools and workplaces, affecting students, teachers and bureaucrats. The 1980s began two decades of severe economic and budgetary instability, increasing poverty and deterioration of the school system. About 90% of citizens remain very poor. They lack food security, basic health care and the education and training to participate effectively in the modern economy and society. In 1985, an estimated 3 million of the 7 million people were of school age and only about 900,000 children were enrolled in the country's 4,000 schools. In 1994, midway through implementation of the Second Education Sector Credit, citizens participated in the first multi-party, democratic elections. The new president immediately fulfilled a popular campaign pledge to abolish primary school tuition fees and school uniforms. This released pent-up demand for "free" primary education (FPE) and doubled enrollments within the year. The Education System Managed from the center by the Ministry of Education through regional and district offices, the school system features 12 years of primary, lower secondary and upper secondary education (8+2+2). In the past, poor people aspired to educational qualifications but the cost and poor access, especially in rural areas dampened demand. Communities provided their own materials and labor to build schools-both primary schools and low-cost substitutes for secondary schools. Traditionally, the community self-help schemes were directed by local leaders and the schools given teachers by the central government. 1. Annex B outlines the projects' objectives, the components, and the policy actions included in the legal agreements. 2 Primary Education Gross enrollment levels in primary schools, including under- and over-age children, rose only gradually over 20 years until the mid-1980s when they reached about 70% for boys but still only 30% for girls. About 50% of pupils typically dropped out before Standard 5 and only one in five boys completed primary school-even fewer girls. Pupils repeated classes many times, especially in Standards I and 8, taking up to 13 years to complete Standard 8. Hence, under-age children and over-age young adults occupied as much as one-third of the available classroom capacity. Overcrowding in urban schools was common, with as many as 100-150 pupils crowded into a classroom meant for 50. The average classroom-pupil ratio was over 1:60 and the average pupil- teacher ratio 70:1. Donor Assistance In 1967, three years after independence, the government became one of the first IDA borrowers for education. Since then, it has relied on donor financing for social sectors while using its own development budget elsewhere. By 1988, at the launch of the first sector credit, IDA had financed five education projects. Among many denors, the African Development Bank (AfDB) and the European Union (EU) had financed school construction. The first phase of the Education Plan, 1985-90, was to be supported mainly by IDA and AfDB. The first sector credit and the ongoing third, fourth and fifth IDA projects and were to provide about 60% of the financing, AfDB 15%, government funds 13%, and local communities about 6% in labor and construction materials. Later, in 1990, for the second phase of the plan, 1991-95, IDA was to provide 44% of the funds from the second sector credit, and nearly 22% from the first; AfDB about 24%; and the government 13%. Other donors did not contribute because of poor governance in the pre-election period. Community contributions were to be more modest. The earlier IDA projects had all been investment projects with multiple components (Annex A). They had been fairly successful in meeting their quantitative targets, but only after protracted implementation delays. In the 1980s, the Bank decided to make access to primary education its main policy priority. 2 In Malawi, the Third and Fifth Education Projects were the first to include primary school construction. Performance weaknesses, according to the project completion reports, included management, organization and local supervision of construction activity, including the communities' inputs of labor. These were highly relevant findings for the designers of the sector credits. Moreover, according to an independent review, institutional effectiveness and commitment to decentralization in the MOE were also weaker than in many other government ministries. By the mid-1980s, the IDA and the government agreed that consensus was investment projects did not have the lasting impact on sector management and effectiveness anticipated and that a sector-wide approach, also being tried out with other borrowers, was likely to have more impact. The Education Sector Credits To obtain the first sector credit, the borrower made legal commitments to achieve particular policy measures by 1988 and 1989 and to carry out the substantial investment program. The policy measures included improvements in the level and share of budgetary expenditures on the 2. World Bank, 1988, Education in Sub-Saharan Africa: Policies for Adjustment, Revitalization and Expansion, A World Bank Policy Study, Washington, D.C.; World Bank, 1980, Education Sector Policy Paper, Washington, D.C.; World Bank, 1990, Primary Education, A World Bank Policy Paper, Washington, D.C. 3 education sector and fee reductions in primary education. The main investments were the schools and other facilities, curriculum, examinations, instructional materials and teacher training. The second credit was to follow two years later in 1990. At appraisal, IDA staff justified the funding as essential because unexpected problems had put the development program at risk. * The growth rate of the primary school-age population was higher than expected and required faster expansion of the school system if enrollment targets set for 1995 were to be met. * From 1987 to 1990 only 51% of the education budget had been released and the real value of the 1988 development budget was only half that of 1982. * Policy and investment components under the first credit were behind schedule. Under the second credit, targets were modified and the legal agreement was strengthened, in particular the assurances of a higher budget share for primary education and faster progress in efficiency and decentralization. IDA's appraisals of the two sector credits were excellent in several ways: coverage of issues; thoroughness of analysis; alternative policy, budgetary and cost-reducing scenarios; cost estimates for school expansion; and reporting on the borrower's preparations. The main weakness pertained to the extreme complexity of project design relative to the known weaknesses in the country's capacities. 2. Implementation Experience The external factors beyond the control of the project managers during implementation receive prominent recognition in IDA's completion reports. The catalogue of challenges in Box 1 demonstrates how relentlessly they weighed down on the projects. Box 1. Challenges-Economic, Fiscal and Political * In the late-1980s, the economy remained volatile in the aftermath of crises and adjustments. * In the 1990s, the early years of implementation, the situation deteriorated. GDP dropped 25% in a year. Increased income disparity and poverty resulted from droughts and floods; declining revenues from commodity exports, exchange rate volatility and inflation; and massively increased external debt. Throughout implementation, the government struggled to reconcile requirements for fiscal discipline with the projects' requirements for increased expenditures for education, especially primary education. * In the middle years, 1992-94, the election campaign, the advent of the democratic government and heavy spending on populist programs contributed to another crisis and IMF intervention. The recruitment of 20,000 secondary school-leavers as teachers for FPE raised the MOE salary bill. * After 1994, the introduction of "cash-budgeting," (the release of funds limited to revenues in the previous month), disrupted project planning and imposed day-to-day scheduling of activities. * In the late 1990s, a Medium-term Expenditure Program, a Policy Investment Framework and actions to improve governance promised a more predictable environment for donor assistance. Given the disruptive environment, it is remarkable that the government moved forward on risky education policy reforms and managed to provide counterpart funds until the credits closed: 4 * Treasury provided the requisite counterpart funds * NEC continued to coordinate the programs and provide funds * MOE continued its efforts to add-ess complex issues related to delivering better-quality inputs such as curriculum and examination reform and teacher training. The ICRs detail the progress towards targets for all components and should be read in tandem with this audit. OED's review focuses on issues related to the most costly investments where, problems appear to defy solution. Internal Factors Affecting Implementation As the ICRs describe, implementation problems led to high administrative overheads and costly delays in expending the credit and delivering outputs. The problems arose from activities under the investment components where lessons of experience appear to have had little impact. The Status of Preparation When MOE staff were preparing the first sector credit, they were still grappling with lengthy delays in the civil works component of the ongoing fifth project. By appraisal, nevertheless, the IDA team was confident that the construction component was ready for implementation (Box 2). Box 2. Preparations for the Primary School Facilities Component The appraisal report stated that communities had donated the land for schools and the MOE had identified all the sites and provided an adequate report of its appraisal of the schools. The plans for the facilities conformed to IDA's guidelines on the sites, standard architectural designs, and norms for space utilization. At 1987 prices, unit costs for construction ($80 per square meter) were economical, while the unit cost per student place estimated at $115, was only 25% of the Bank-wide average. The confidence of IDA staff rested more on trust than realism because preparations turned out to have been less than adequate. During implementation, IDA staff reported on the unsuitability of sites (accessibility, drinking water; soil slippage, flood-prone terrain); inappropriate or inconvenient locations (site too near another school, remoteness from the community); cost overruns by contractors; and faults in materials or construction work. Added to this was the unreliability of information on the actual numbers of school-aged children enrolled, the physical status of school facilities in the MOE's inventory, the capacity of the construction industry and the involvement of communities. Given the ministry's weak technical and monitoring capacity, senior MOE staff must have lacked the means to validate reports from local offices that all sites had been visited and properly appraised by qualified staff. The lack of readiness for implementation was responsible for many of the subsequent problems and delays, yet none of them should have been unfamiliar and could easily have been anticipated. 5 Overlaps, Backlogs and Bunching As the result of the delays, costly backlogs and bunching of activities persisted throughout implementation. The two operations required simultaneous implementation during the period 1990-93, a fact not brought out in the ICRs. Under the first credit, implementation began only in 1992, four years after effectiveness and one year before the original completion date. The new project manager's priority was to complete the first project, though it meant further delaying implementation of the second project. The pace improved gradually after two years but picked up considerably only after the FPE announcement in 1994-again, only one year before the original completion date. MOE staff pointed out that the overlaps required attention to two separate schedules, administrative processes and reporting requirements. IDA agreed to extensions of the closing dates because project investments were not yet in place and funds remained unspent. Since the delays stemming from overlap were obvious shortly after effectiveness for the first credit, it appears that the risks of proceeding rapidly with the two projects together were ignored. This was the case even though they had been identified in reports of previous education project, and again were identified post hoc in the ICRs. Organization and Communication The location of project personnel was a constraint on implementation. The Project Implementation Unit (PIU) in Blantyre was responsible for project implementation, reporting to the MOE's Planning Department in Lilongwe. The pervasive problems of communication, teamwork, and oversight stemmed in part from the unreliability of telephones and the half-day's physical distance via a difficult artery road. The project manager and PIU staff controlled, de facto, all day-to-day administration and use of funds: programming, procurement, contracting, contract supervision, payments and other expenditures, monitoring, reporting and accounts. In addition, all the project-related funding needs of the ministry's regional offices, technical departments, and agencies, such as those for teacher training and textbook provision, were directed through Blantyre. The physical isolation became especially difficult after 1996 when the PIU was struggling to complete outputs under the second credit at the same time as IDA had decided to integrate PIU functions into the line ministries. This resulted in the departure of key staff and loss of status and morale for those remaining. The comments in the MOE's completion review show the depth of emotion generated. Reportedly, the decision to locate the PIU in Blantyre and allow it to remain there was based on the desperate need to attract and retain a qualified project manager. IDA's actions to resolve communication problems for future projects came too late for the sector credits and, clearly, the nine years' delay cost dearly in terms of efficiency. Qualified Managers and Staff All five previous education projects had suffered problems related to managers and staff. In 1980, at completion of the First Education Project, IDA staff had emphasized the importance of ensuring that the project management team was appropriately skilled. But subsequent projects continued to experience delays in recruiting and retaining experienced and qualified staff, as well as controlling problems of absenteeism and moonlighting. Delay in the appointment of specialist consultants and the transfer of skills to counterparts also hindered efficiency. Staffing and project implementation capacity were issues that frequently consumed time and effort. 6 Box 3. Two Decades of Staffing Issues Repeated The borrower's contribution to the ICR for the first sector credit was prepared jointly by the MOE and the PIU in 1995. It declared, "When implementation... started our [PIU] did not have the capacity to implement the project" (ICR Annex A, para.6). At credit effectiveness, is noted, the PIU had lacked a project manager, a financial analyst, the principal accountant, the principal stores officer, the procurement officer, the quantity surveyor, and some of the architects. IDA staff added the comment, "At present, PIU appears to be well staffed, but there are some lingering doubts about its effectiveness and its ability to handle larger, iitegrated projects... (ICR para. 8)." The ICR for the second credit commented in a similar vein: "At the beginning, the PIU did not have the full complement of qualified technical staff, including an architect, accountant or procurement officer... The situation greatly improved from 1992 when the Government hired with project funds [qualified] technical staff (two architects, a quantity surveyor and an accountant... as well as an experienced project manager" (ICR para. 23). The launch of the first sector credit coincided with a lengthy period of government reorganization as anticipated at appraisal. This changed the MOE's departmental arrangement and reassigned staff to new duties. The resulting confusion, even for senior officials, weakened staff accountability and morale. Moreover, the new Planning Department was small and, apart from the senior level, lacked qualified staff. IDA and the government were aware that the staffing problems stemmed not only from persistent shortages of qualified personnel but also from the reluctance of qualified staff to be transferred into the MOE. During implementation, IDA was reluctant to apply the sanctions available under the legal agreement because this would have interfered with the policy agenda. Nevertheless, frustration by both parties arose when frantic action during a supervision mission did not result in action until the next because other duties distracted MOE staff in the intervening period. In the light of all these factors, the obvious question arises why IDA presented such complex operations to the Board when it was obvious that qualified staff were stretched thin. A similar question could be asked of the borrower. Why sign the second credit agreement two full years before honoring the commitment to staff the PIU, given the well-known staff constraints under the first credit? Workload Pressures Heavy workloads for the project staff led to pressures that hindered their ability to follow through on issues and schedules agreed during supervision missions thus adding to implementation delays. IDA staff knew the problems but had little control over those internal to government. Apart from the scarcity of competent junior staff, delegation was a new practice for senior officials. This added to the pressures as central, regional and new district offices learned new ways of working. Budgetary limitations on local monitoring and supervision-viewed as low priority administrative overhead by the financial agencies-added to the pressures because local problems were big by the time they demanded attention. Government-wide measures to establish performance standards and improve staff productivity were delayed and project administrators' time and energy were engaged by the pervasive bureaucratic inefficiencies. Though all factors were not controllable from outside, IDA might, nevertheless, have reduced the relentless pressure of its lending for education. It is unclear why it chose not to consider this as an option. Instead, from 1967 to the present, the MOE has coped simultaneously with at least one, and as many as three projects, in any one year. 7 Box 4. Relentless Pressure from IDA Lending * In 1986, MOE staff were preparing the first sector credit while completing the Fourth Education Project and grappling with the fifth, which was by then a problem project. * In 1988, just as the first credit became effective, they began preparing the second sector credit. * In 1990, just when they were grappling with the budget delay and winding down the fifth project, the second credit became effective. * Finally, in 1992, the PIU manager and key staff were recruited, the completion date of the first sector credit was postponed, and staff focused on completing these activities while beginning implementation of the second credit. * In 1993, the pre-election year, staff participated in discussions with IDA on a third sector credit and prepared a revision of the credit agreement for the second. * In 1994, they accelerated school construction activity after the president's announcement of FPE. For three years more they continued to implement the second credit, as well as undertaking many new activities arising from the president's initiative. * In 1995 over a seven month period, they prepared the "fast-track" PEP. At the same time, they participated in discussions with IDA staff from the social protection sector on the first of two Malawi Social Action Fund (MASAF) projects approved in 1996 and 1998. * In 1996, a new Educational Management and Planning Unit in the MOE began to implement the PEP while the PIU staff dealt with the final years of second sector credit. * In 1997, when the credit closed, the MOE had already begun to prepare the ongoing Secondary Education Project (SEP). Multiple Demands of Donor Projects IDA knew that dealing with all the donors placed multiple demands on the MOE, including the Planning Department and the technical departments.3 The MOE staff coped as best they could the differing donor priorities, data requirements, implementation procedures, and reporting requirements because, they said, they were only too well aware that almost the entire investment program in education was dependent on donor funds. The need to welcome any willing financier was the reason they gave why the various donor-assisted projects constructed facilities to various designs and space specifications without the ministry's strong insistence on adherence to its guidelines; and why some of them were not well suited for educational activities. IDA had anticipated some modest degree of harmonization through cofinancing arrangements with DflD, but these fell through. Supervision staff said that they had little control over the demands of other donors on the MOE and that IDA projects were part of the problem. Bank Supervision The MOE reviews show appreciation for the regularity and continuity of IDA missions, and the commitment of staff to resolving problems. The six-monthly supervision reports, however, show that attention to troubleshooting unduly absorbed the scarce resources available to staff. For the first credit, the staff weeks used for supervision at were substantially lower than the Bank norm 3. In the early years, the most prominent donors were the British Council, the United States Agency for International Development (USAID), the German Technical Cooperation Agency, and the EU. 8 the time.4 Staff used about 96 weeks, an average of 14 a year for the 7 years. In contrast, in just the two years of the processing stage, staff used 123 weeks, over 20% more resources than for the entire supervision period. For the second credit, IDA devoted many more resources to supervision and slightly over the Bank norm-152 weeks or 22 a year. Significantly, the supervision resources were heavily used in the final year on procurement and expenditure issues. 3. Primary School Construction: Quantitative Achievements And Capacity-Building The MOE's goal for the plan period was to accommodate 200,000 pupils in 4,000 new and replacement classrooms in primary schools at a pupil-classroom ratio of 50:1. Together, the two sector credits were expected to contribute substantially towards these goals. In 1994, on the declaration of FPE, the government estimated that 38,000 additional classrooms would be required within 10 years to maintain a pupil-classroom ratio even at 60:1. A revised net enrollment target of 90% was set for 2004. Performance of the First Sector Credit Under the first credit, the physical investments in primary school buildings, furniture and textbooks were expected to amount to 47% of the total project costs and roughly 70% of IDA funding for all physical investments. The project was to provide 1,500 classrooms at 300 dispersed school sites in disadvantaged areas, plus furniture and equipment and houses for 750 teachers and their families (1 house for every 2 classrooms in a school block). At appraisal, the component was expected to benefit 75,000 pupils (only about 4% of total enrollments by 1993). The main thrust was to enhance the learning environment for children already in school by reducing classroom overcrowding, providing shelter for classes taught outside and improving physical facilities in poor shape, as well as the complementary qualitative improvements. A related thrust was to achieve more efficient use of the infrastructure. The minimum age for entry to Standard I was to be set at five years and repetition was to be phased out beginning in Standard 8. By accelerating pupil flows through the schools, the space released would provide accommodation for more out-of-school children aged 5-13 years. Class sizes would be reduced from an average of over 60 to 55 pupils by project completion. Achievements The ICR reports that when the PIU found the original targets too taxing, the project staff reduced them. The project finally constructed 1,074 classrooms (70% of the target). These facilities accommodated 64,440 pupils (86% of the target) who would otherwise have continued to study in a poor learning environment. MOE officials said that, after years of decline, the dilapidation was so extensive that reconstruction was necessary. The new school facilities were the most visible MOE achievements for the communities, many of which expressed appreciation at the time for the brighter prospects they offered their children. Also, the 520 houses completed (69% of the target) improved working and living conditions for teachers assigned to them, including household members. 4. The low level possibly reflects the Bank-wide slippage in supervision quality before 1992 when President Lewis Preston initiated improvements. World Bank, 1993, Getting Results: The World Bank's Agenda for Improving Development Effectiveness. Washington D.C. 9 Delays and Project Costs The delays almost certainly raised the direct cost of the investments. The project delivered only 75 classrooms on average in the first two years, 400 a year in 1992 and 1993 and the remaining 175 in 1994 after the credit extension. This meant that about 900 schools were actually delivered between 1991 and 1994, four to five years later that originally estimated. But because the project could not provide a breakdown of the cost among all types of civil works, the size of the cost overrun on the primary schools is obscured. Overall, however, the share of civil works components in total project costs rose considerably, from nearly 51% at appraisal to nearly 65% at completion (Annex A). In addition, the appraisal cost of architectural, engineering and auditing consultant costs rose from an estimated 2.8% of the civil works costs to 13.6%. The ICR did not explore the educational costs of the delays but they too must have been large. Four hundred and twenty-six classrooms were never built. Therefore, if the improvements in classroom environments had all been targeted at Standard I, about 10,000 pupils in their first year would have foregone the benefits for at least five more years of schooling-if they had not already dropped out. Similarly, if the classrooms had been targeted at Standard 4, nearly 7,000 pupils would have suffered. The lack of classroom space, combined with non-delivery of furniture to the project schools and the late arrival of textbooks in 3,000 schools, would have made it almost impossible for children to study. The costs would have been private and social in terms of self-esteem, learning achievements, civic competence, productivity and income potential. Moreover, the gains expected in efficiency and effectiveness would have been reduced for the teachers assigned to the 230 school blocks when they remained without adequate housing and taught in cramped classrooms. Overall, the re-estimated economic rate of return to primary education would certainly have been lower than the 20% estimated in 1990/91, and possibly the 14% estimated in 1988. No definite price for the delays can be given because the studies planned were not done. Box 5. Enrollment Targets and Demographic Impact Even before FPE, demographic pressures on classroom space were overwhelming because the numbers of children attaining school age were rising rapidly. Estimates of the natural growth rate were revised upward from 4.3% to 4.8% a year in the early 1990s. Enrollments doubled to 1.9 million between 1985 and 1993, but at least one million children were still out of school. Repetition rates remained high; one child in five repeated class at least once. Classrooms with 100 or more pupils remained common and the average pupil-classroom ratio remained around 60:1. The average pupil- teacher ratio was estimated at about 70:1. The primary school pupils who benefited from increased access to classrooms under the first sector credit would have constituted only about 3.4% of the total number of pupils enrolled countrywide. Throughout the decade, estimates of the population growth rate continued to rise from an estimated 3.10/ to 3.7% a year. As the uncontrollable demographic pressures overtook the pace of implementation, the project's contribution to meeting the targets for access to primary schools under the plan became less significant than was originally intended. Performance of the Second Sector Credit In 1990, for the second sector credit, IDA decided to consolidate on the basis of an ongoing MOE census of unfinished schools. The project's contribution would be to finish the incomplete classrooms rapidly and economically. The credit was to provide locally produced roofing sheets for 228 of the unfinished school buildings and 406 teacher houses in rural areas. Communities and public works programs would provide the bricks. Small contractors were to provide the other materials and do the construction. The roofing work was scheduled over four years. The cost was estimated at about $8,700 for rural classrooms and $21,000 for urban. Locally produced desks and chairs were also to be provided for the 3,000 classrooms that lacked furniture. IDA staff said 10 that the government had little option but to rely more heavily on community contributions and construction but that a major problem with previous self-help construction was the inability of communities to buy roofing sheets. By providing these expensive items through the credit, the government intended to motivate communities to contribute. Achievements The goal of providing usable classrooms was attained, but neither rapidly nor economically. The school structures were weak and unable to bear the heavy steel selected as roofing material-a weakness that sound technical appraisal should have identified. In late- 1993, therefore, the project manager proposed that the project should include extensive rehabilitation of the facilities. IDA agreed to include the installation of concrete reinforcement beams, the roofing, and surfaces for dirt floors left unfinished by communities, reallocate funds from a cancelled component, and defer the credit closing date by a year. Thus, four years into implementation, just as the President announced FPE, the schools remained unfinished and unusable. IDA then expanded the component to include rehabilitation of 2,000 classrooms and ancillary amenities, and the roofs for more teacher houses. Time was too short for the MOE to identify and appraise the school sites, but dates were set for this to be done. The borrower's review, however, states that the MOE's efforts to compensate for lost time did not result in any positive results by 1996 and that this was the reason for a further one-year extension to 1997. As the ICR says, the eventual achievements were considerable. The precise output as reported, however, is unclear. * The ICR says that sheets for roofing 841 blocks (2,218 classrooms) were provided and that 75% of the classrooms were fully rehabilitated. The borrower's review, however, says that re- roofing of 641 classroom blocks comprising the same number of classrooms was achieved. Though this discrepancy is probably typographical, the MOE's figure adds up to an average of 3.5 classrooms per block, while IDA's estimate adds up to 2.5 classrooms. * The ICR says that 75% of the classrooms were fully rehabilitated, in addition to acquiring new roofs. If measured against the original target of 2,000, the output at 75% would have been 1,500 classrooms. If measured against the 2,218 classrooms rehabilitated, it would have been 1,663, almost the same number as targeted under PEP! * The ICR reports that the roofs for the teachers' houses were eventually "not required." (Table 5). But this belies the strong case made for including the housing in the projects. Were the 520 houses provided under the previous credit sufficient, despite a shortfall of 230? Did another donor provide them? Did communities prove unwilling to provide them? If they were not provided, how were the funds spent? Box 6. Teacher Housing The mission visited a handsome school facility. Four of the seven teachers had young families and two of the young men were in bad health. They explained that the headmaster and the first-comers were lucky enough to be assigned to the houses. The houses had been built about 10 years but the roofs leaked and the water supply was not connected. The other teachers lived in rented accommodation. But houses were scarce and community members required high rents for the thatched, one-room homes. The teachers explained that most of them were not from the local community and did not feel welcome. They said that the community did not respect them in part because the school had not recently produced any graduates with success in gaining entry to the secondary school. 11 Costs of Primary School Construction The ICR does not report actual expenditures on the primary school construction component, since they are not disaggregated from the total for civil works. The share of total civil works in final cost increased from 21% in 1990 to nearly 45% by project completion in 1997, and the share of the credit devoted to civil works doubled, from 21% to 42% (ICR Table 8A and PAR Annex A). The sector credits only partly provided the qualitative inputs needed to ensure that teaching and learning would be enhanced. To accommodate the additional pressures under FPE, the amount of training for new teachers was halved and furniture and equipment for the 3,000 schools was delayed. These credit funds were reallocated to school rehabilitation. Educators interviewed complain that the shorter period of entry-level teacher training is not adequate and that instructional materials are still in short supply. A visit to an IDA-financed school on a good access road demonstrated that basics such as legible chalkboards and chalk are scarce. Management of resources remains weak. Old books, paper and garden tools littered classroom closets. Teachers arrived at school later than many of the children. The majority of pupils, nevertheless, had arms bulging with textbooks. The qualitative efforts are still underway but, after the credits closed, budgetary resources for non-salary expenditures have declined. This is serious because everyone interviewed agreed that the funds are urgently needed to complement the heavy investments in physical facilities and to maintain universal support for the democratization of primary education under the government's policy. Linked Output under the Primary Education Project IDA extended the second sector credit for the final year and was thus able to make it the umbrella for PEP. While qualitative improvements were a major thrust, this project directed about 50 % of the funding to the construction of 1,600 classrooms. The aim was "to provide as many well-built classrooms as quickly and cost-effectively as possible, and to provide teacher houses and administrative blocks in proportion" (SAR para. 3.4). At appraisal, IDA staff said that, with the addition of over one million pupils on the rolls, urban classrooms already had to cope with as many as 200 pupils (an enormous increase over the 60 planned in 1994). Nearly 100 of the classrooms would be constructed during 1995 under a preparation facility. About 600 would be classrooms in 75 new schools in remote rural sites. In justifying PEP, IDA estimated the social rate of return as a result of the increased participation expected from FPE at about 20%, comparable with the estimate for the early 1990s. About 950 classrooms (60%) are expected to be ready by the completion date extended until December 2000. Combined Impact on Classroom Access The sector credits and PEP eventually will achieve about 92% of their revised targets in terms of classrooms constructed and additional pupil capacity. If achieved by 1995, as originally planned, this accomplishment would have substantially reduced overcrowding. Because of the delays, however, further demographic pressures and FPE offset the positive impact on classrooms. Of the original I million children in school in 1988, the sector credits planned to benefit nearly 32 percent with improved classrooms. Of the estimated 3 million children of all ages in school in 2000, the credits would have actually benefited about 23 percent, if all classrooms were filled to 12 capacity. Table 1 shows the combined achievements in increasing classroom provision and pupil capacity, including PEP.5 Table 1. Additional Classroom Provision and Pupil Capacity (60:1 Ratio) Targets Actuals % Actual (latest) Classrooms Pupil Capacity Classrooms Pupil Capacity Classrooms & Pupil Capacity. Sector Credits 3,600 216,000 3,818 229,080 106 Including PEP 1,600 96,000 950 57,000 59 Totals 5,200 312,000 4,768 286,080 92 Costs and Benefits By how much the unit costs of the classrooms exceeded estimates is not known. The 2000 Public Expenditure Review estimates that the cost of producing a primary graduate has doubled just since 1994. In 1999, the pass rate in the Standard 8 examinations was only 10%. These data alone suggest that the 1988 estimate of $115 for the economic cost of primary education should almost certainly be revised upward. The indications are that primary qualifications alone are currently of much less benefit to students than more advanced qualifications in terms of subsequent wages. All this means that the economic rate of return to primary education, with capital costs and educational performance included, has very likely declined. Nevertheless, if the numbers of children who attain literacy and numeracy on a permanent basis in primary school, they may well try to ensure that access to education increases their own children's prospects for a better life. 4. Community Self-Help In Primary School Construction Community self-help was a critical component for primary school construction in both sector credits and in PEP. Through this mechanism the projects achieved much in quantitative terms, though with lower than expected speed and efficiency and with high transaction costs. The outcomes for communities have been mixed because the delays and inefficiencies have damaged trust, enthusiasm and delayed local opportunities to develop skills and capacities. Following is an analysis of factors that produced these results. Lessons of Previous Projects When the first sector credit was being prepared, self-help was an element in two ongoing education projects and many lessons were already available. The third project, approved in 1979, constructed 1,000 classrooms with materials and labor contributed by communities. The fifth project was similar. The objective was to build many schools rapidly and cheaply. But problems arose repeatedly. Conditions at school sites were difficult. Local partnerships did not work smoothly. Contractors proved unreliable; communities were unable or unwilling to provide their time or synchronize their contributions with construction schedules; and MOE supervisors lacked skills to provide technical guidance and oversight. In 1986, the ICR commented that 5. Using the higher, revised targets and assuming the higher ICR output figures are correct. 13 management, organization and supervision of community participation is a difficult challenge and that a strong institutional framework, especially at local level, is essential for success. Technical Design and Community Results The design of the first sector credit took account of the emerging experience but with modest results. The project required the communities to provide only the traditional kiln-fired bricks and not their time for the construction work. The appraisal team expected increased efficiency because the central PIU would purchase other materials and small contractors would compete for the business. An innovation was the public communication campaign planned. Incentives were built in so that the provision of furniture and texts would be conditional on the completion of classrooms. When problems still arose, IDA agreed to adapt the approach mid-course. Studies were made of the community brick-making process and the construction work was contracted in large packages to attract experienced contractors.6 Despite some improvement, problems persisted. The MOE's review says that, at many project sites, the communities were unable to provide sufficient numbers of bricks of acceptable size and quality and that, in future, local communities should be asked to contribute only sand and aggregates. "It was evident," declare both parties, "that the spirit of self-help had waned" (ICR para. 21(b) and Appendix A, para. 8). How many schools remained unfinished due to community issues is unclear. Nevertheless, community contributions in defraying the costs is certain to have been considerable, especially the opportunity costs of their time. In the same pattern, the second sector credit achieved a large output after delays but it is not clear how far community self-help actually contributed. Initially, communities were to provide bricks but not roofing materials. In 1994, when FPE led to an expansion of the school facilities component, IDA and project managers reluctantly agreed that community input would be essential but were not optimistic about output. Meanwhile, in the effort to learn from hard experience, PEP introduced an adapted technique. Contractors were to construct the school frames and roofs (the "shell") and school committees organize communities to make hydraform blocks rather than bricks. The hydraform is easily molded from local materials with machines loaned to each community. At appraisal, IDA commented that the technique had been successfully tried elsewhere and would be more environmentally friendly than bricks.' Results achieved so far have proven disappointing mainly because of the familiar scheduling problems, and the approach has not been expanded. Insights from the Education Project Experience Among a number of observations in a recent OED review of PEP, three are also applicable to design of the sector credits.' * Logistical and scheduling problems make efficient programming of labor and technology inputs difficult to achieve. 6. Weak capacity among building contractors is a familiar problem in education because experienced, large firms with sufficient capital are not attracted to small school projects. In reviewing a long list of the contractors used, an MOE official identified only eight still in business and none were small firms. 7. In fact, delays in finishing the schools elsewhere turned out to be a chronic problem. The environmental advantages and disadvantages of hydraform and bricks are debated among construction technology experts. 8. Wendy S. Ayres. "Malawi: Primary Education Project." Background paper to OED "Participation Process Review," Draft, June 6, 2000. 14 * Self-help is predominantly supply-driven. Communities are left out of the identification and design. They are invited only to share in implementation. * The design lacks appreciation of what community participation really means and what is needed to successfully involve communities in projects. The focus in all the education projects on techniques and technology is in line with the early engineering traditions of Bank operations in infrastructure. The difference is that the education projects adopted techniques such as brick making but did not factor in the technical complexities as would be routine, for example, in designing a bridge. Equally important, the projects did not tailor the techniques to fit into organic processes of self-help. The input-output paradigm (community brick-making-low-cost facilities; community labor-community ownership) is deficient by itself because it does not ensure compatibility with socioeconomic and cultural realities: unique patterns of community 'ife and work; decision-making and business practices; and the balance of power between local administrations, commercial interests and others. The appraisal documents allude to traditions of self-help in Malawi as justification for introducing community-based components. However, they do not show the same high quality of technical and socioeconomic preparation as the economic and budgetary analysis. This deficiency led to poorly designed components that ignored political and social realities. In the past, the bureaucratic management practices of central government were supportive of self-help traditions. The government's paternalism and control mechanisms supported traditional leaders in mobilizing local communities and organizing and disciplining them to maintain participation until the job was done. When IDA decided to take the local self-help schemes to scale, observers predicted that large, one-size-fits-all projects would erode the delicate balance of paternalism and participation. The implementation delays in the 1980s appear to have eroded communities' willingness to contribute. The introduction of FPE initially restored enthusiasm, but further delays and increasing problems of poor quality teaching and learning outcomes led to further frustration. The sudden introduction of "free" primary education without sufficient preparation and communication led people to "disown" the schools and regard them as entirely the government's problem to fix. Can the Community Spirit be expanded and sustained? The MOE hosts a small, community-based, USAID-financed project regarded as successful by educators and evaluators. The Girls Attainment in Basic Literacy and Education project has improved teaching and learning levels in its schools through involving parents and communities, though not primarily in constructing schools. Significantly, GABLE found that, "People are receptive to dialogue and involvement. They are not receptive to being told."9 The successor project built on the experience. It is unclear, however, whether the approach is sustainable or replicable on a national scale. Outside the ministry, also, know-how in strengthening community capacities has developed rapidly, and the use of local school construction subprojects is one of the most popular vehicles for doing so. The Social Action Program is one example. Evaluations show that the communities develop a participatory spirit when they achieve positive results for themselves. The purpose of 9. Quoted in, Lessons Learned. Social Mobilization Campaign for Educational Quality (SMC-EQ). A Government of Malawi Project, funded by USAID through Abel 2. Creative Associates International, Inc. no date. 15 introducing the MASAF experience (Annex C) is to explore perceptions of the commonalities and differences between the MOE's self-help schemes and MASAF's approach. The main question posed is whether or not the establishment of stronger mechanisms for sharing experience, knowledge and mutual learning across institutional boundaries would be useful and practical in order to achieve new primary school enrollment targets for 2004. 10 5. Findings The projects were highly relevant to sector needs and were moderately efficacious in establishing the foundations for needed sector-wide adjustment and long-term development. Given the inauspicious context before, during, and after implementation, this has been a major achievement. Continuing efforts are needed to expand and improve education at primary and secondary levels, to strengthen specialist institutions and to achieve an appropriate balance in financing and budgetary allocations for all levels. In assessing the feasibility of a third sector-wide credit, the issues of complexity, demandingness, feasibility and economic returns would require scrutiny. Outcomes The credits achieved a three-fold increase in primary school classroom capacity, well beyond original expectations. Outcomes were satisfactory in this regard. In urban areas, enrollment in primary school has become the accepted social norm. Nevertheless, the delays discussed allowed rapid increases in the school-aged population to outpace enrollment increases. Enrollments remain at the same level as in 1995. Primary education remains far from universal and dropout and graduation rates remain high, especially for the poor, while student achievement levels have declined drastically. Institutional Development Impact Implementation capacities have not yet proven adequate to cope with the transition from a small school system to a mass school system and remains modest. Compared with the 1980s, the capacities of the government and the MOE's Planning Department have been strengthened as a result of the project experience. But the task of constructing school facilities requires a variety of managerial, technical and participatory skills that the projects vastly underestimated. How to building permanent implementation capacity, centrally and locally, is still a vexing question for the government and IDA. Remedies to solve staffing problems, retain institutional memory, enhance morale and commitment and provide sufficient monitoring and supervisory resources appear to be the minimal requirements that require joint efforts by the MOE and senior government agencies. Sustainability Sustainability remains uncertain in the light of recent developments. With the recurrent budget for non-salary items currently declining, the erosion of educational quality has become a serious issues. The budgetary shortfalls cast doubt on the government's commitment to complete the unfinished business in primary educa".ion and put at risk in a pre-election period the widespread 10. This PAR is not intended as an evaluation of GABLE or MASAF. An OED case study on the program is being developed and is intended to contribute to the forthcoming OED study of social fund projects task-managed by Soniya Carvalho. 16 public support for the mass education policy. A related challenge to sustainability is the inadequacy of mechanisms to ensure that primary school facilities will be well maintained and repaired over the 30-year life of the investments. Though older MOE schools have lasted over 40 years, many new schools will be of much less durability unless they are regularly upgraded. Though the ministry has developed a maintenance policy, the willingness and ability of local communities to contribute will be critical. Their contribution will in turn depend on whether they derive value from primary education. For this to happen, the minimum requirement is that qualitative inputs must be visible- effective school managers, well-trained teachers, well- equipped classrooms, and relevant programs. Public confidence that all pupils have equitable opportunities to succeed is currently low. Political stakeholders may exploit these concerns. In addition, the needed investments in secondary schools currently underway must not supersede the priorities still remaining in primary education. IDA Performance IDA's performance was satisfactory though the lack of realism is cause for concern because of the economic costs of delays and shortfalls. Although IDA was familiar with the recurring problems in education projects and the country, its confidence in government assurances that adequate funds, personnel and implementation arrangements would be in place turned out to be misguided. Preoccupation with the policy and budgetary reforms that were critical to long term appears to have been at the expense of rigorous scrutiny of the feasibility of the investment components and of routine aspects of project administration. Staffs very commitment and responsiveness to the country's educational needs may have led them to overlook the implementation risks as a normal cost of doing business. In turn, this may have encouraged day- to-day troubleshooting and damage control at the expense of fundamental reassessment and more radical remedies. With the best of intentions, IDA may have held the government to less rigorous standards than it would expect from high-performing borrowers outside Africa. As one observer put it, presented with a request for a loan for a large and complex investment, a reputable commercial Bank would insist on evidence of a sound track record and business plan, including operating systems, financial controls, and value-added. At the strategy level, it is unclear what factors encouraged IDA to allow so many overlapping education projects to proceed in the face of clear evidence of implementation overload. At the very least, this situation should have led to candid scrutiny of the internal decision-making process on the number and phasing of education sector projects, as well as budgetary transfer issues and donor coordination. IDA was far-sighted in appreciating the necessity to expand basic education for the long-term health of the economy and society. In the early years, IDA's performance was highly satisfactory in preparing the government to expand access to primary education in a financially sustainable manner and to proceed as rapidly as political and economic constraints allowed. However, staff overestimated the MOE's capacity to deliver a large and complex construction program within schedule and cost, in part because of its reliance on the ministry's weak appraisal of the construction industry capacity and community readiness. Views are divided on how far IDA's support was critical to FPE after 1994. Critics say that the new government might have decided to proceed with more preparation in making the transition to mass education if IDA had signaled reluctance to support the FPE announcement. On the other hand, one government official said that other donors would have stepped into the gap even without IDA's lead. No one suggested that the government might have found the means by reallocating its own resources from other areas. In retrospect, it is reasonable to question whether staff should have urged more attention to implementation options well in advance, especially since FPE was widely anticipated from the previous year. At the time, however, staff considered that the political momentum was too strong 17 to be resisted and that the opportunity to advance primary education should not be foregone despite the inevitable implementation and expenditure-control risks. From the staff perspective, IDA had little option but to support the new government. To do otherwise would have been flatly inconsistent with its education policies and social sector priorities; would have made IDA unwelcome in the country; and would have damaged its credibility with other borrowers and donors. In sum, IDA staff made a difficult decision that is easier to criticize in hindsight. In OED's view , the fundamental question is whether children who gain access to primary education as a result of FPE would have this opportunity if IDA had declined to finance a big push forward. Borrower Performance Borrower performance was mixed. In committing to the expansion of primary schooling and following through to construct many classrooms, performance was highly satisfactory. But the government lacked a realistic plan and sufficient commitment to efficiency and development outcomes during implementation. At the very least, the government appears not to have fully appreciated its legal requirements under the credit agreements that require adequate project management, cost controls and accountability for efficient use of project funds. By signing the financial and management covenants, senior officials vouched that the implementing agency would have the capacity to perform efficiently. But they did this even at the same time as they were concerned that previous projects performing poorly. The government exacerbated the problem by delaying counterpart funds and key staff appointments until 1992. Moreover, the MOE only partially complied with those covenants that require project managers to follow IDA guidelines for procurement, supervision, monitoring, and timely, audited accounts. These issues are recurring and remain a high concern among committed individuals in the MOE and government. However, the persisting nature of the problems leads observers to concluded that the MOE is unwilling to exact high standards of performance from its staff. Some even imply that the government is not committed to ensuring the delivery of project benefits intended for the children of the poor. They charge that officials piesume that IDA will step in to resolve problems as they arise; that IDA funds will continue to flow despite poor performance; and that credit closing dates can always be extended. However much these views distort the real situation, the fact that they are held at all implies the need for a candid assessment by the government of its business relations with IDA in the primary education subsector. Lessons of Broad Applicability The insights offered as a result of this audit review relate only to the issues of project implementation and the primary school construction. They are offered for the consideration of IDA and the Borrower as an independent perspective on a challenging set of issues that still face the education system. IDA needs to consider carefully its justifications for presenting a scheduled operation for Board approval when risk analysis reveals that: (a) serious constraints within the government's control have not been removed; (b) the scale and sequencing of donor projects in the sector overload implementation capacities; and (c) the country's overall capacity has proven so weak that delays will demand credit extensions, raise costs and reduce benefits. During preparation, mutual candor and realism between IDA and the borrower are needed to ensure readiness of a project for implementation, especially for conventional components. In the Malawi case, the innovative components were rigorously assessed in terms of benefits and risks. 18 But the approach to implementing the much more familiar school construction components was not thoroughly examined. In developing new approaches to the technical aspects of school construction, the advice and experience of qualified local personnel, such as engineers and quantity surveyors, need to be available. This was a lesson of the Bank's education sector experience in the 1970s and 1980s and is still critically important for borrowers who are still putting in place the physical infrastructure for universal primary education. In this case, errors in design and selection of materials and techniques might have been avoided with more skilled appraisal and diagnosis. Similarly, when developing components that introduce indigenous traditions, IDA needs to call upon cross-sectoral expertise. In this case, the sociocultural dynamics of community self-help required the same rigor of study as the budgetary component received. This would have allowed staff to re-examine the limitations of the instrumental approach previously and a search for more holistic alternatives. The project experiences emphasize the need for IDA to balance the ready provision of funds to remove educational deficits with the need for systematic review of institutional capabilities, and evidence from monitoring and evaluation studies about feasibility and effectiveness. In this case, a synthesis of local communities' experience involved in self-help schemes would have been useful in developing more effective approaches. By the mid- 1990s, international expertise in participation and beneficiary evaluation was readily available. And in the country, the GABLE and MASAF evaluations are two of a number of sources of ideas for future policy and strategy development (Annex C). Education sector projects need to incorporate similar levels of systematic data gathering, monitoring and analysis for policymakers and financiers. 19 Annex A Basic Data Sheet MALAWI: FIRST EDUCATION SECTOR CREDIT (CREDIT - 1767) Key Project Data (US$ millions) Source of funds Planned at Revised Actual Appraisal IDA 27.0 27,0 28.6 ODA 2.6 - - Government 4.4 4.4 5.7 TOTAL 34.0 31.4 34.3 Cumulative Estimated and Actual Disbursements Bank FY 1988 1989 1990 1991 1992 1993 1994 1995 Appraisal estimate 1.6 7.3 14.9 21.3 25.7 27.0 -- -- Actual 0.5 3.5 6.8 13.6 25.6 26.9 28.6 28.6 Actual % of Credit 1.7 12.2 24.2 47.3 87.9 92.3 98.0 98.0 Date of Final Disbursement: April 1995 Project Costs Category SAR Estimate Revised Estimate GoM IDA Total GoM IDA Total GoM IDA Total Civil Works 0.7 7.4 8.0 0.7 7.2 7.9 2.4 16.3 18.7 Equip., furn., veh., books 0.4 11.9 12.3 0.4 13.9 14.3 0.1 9.1 9.2 and materials Arch/ Eng. services 0.1 0.6 0.6 0.1 0.3 0.4 0.2 1.2 1.4 Technical assistance 5.1 5.1 3.9 3.9 0 3.2 3.2 Training inclucl. Fellowships 0.2 1.9 2.1 0.2 2.1 2.3 0.3 1.1 1.4 Project administration and 2.6 9.5 12.1 2.6 4.1 6.7 0.4 10.4 10.8 recurrent costs Unallocated 0.1 0.6 0.7 0.1 5.3 5.4 0 0 0 TOTAL 4.1 36.9 41.1 4.1 36.9 41.0 3.3 41.3 44.7 Project Dates Steps in project cycle Actual Date (Month/Year) Reconnaissance Mission 9-10/84 Identification Mission 5/85 Preparation Mission 12/85 Appraisal Mission 7/86 Negotiations 2/87 Board Approval 3/87 Credit Effectiveness 6/87 Project Completion 3/94 Closing Date for categories 2, 4, 5, 6, and 7 12/93 Closing Date for categories 1 and 3 12/94 Staff Inputs (staff weeks) Stage of project cycle PLANNED ACTUAL HQ FIELD HQ FIELD Through Appraisal -- - 55.1 16.3 Appraisal Board 18.2 29.6 Board Effectiveness - -- 1.4 2.1 Supervision 61.4 50.2 51.8 43.8 20 Mission Data Annex A Stage of Projects Mont/Yr. Specialized staff skills Performance ratings by activities cycle represented (1) General Procurement Managemen Financia t Pre-Identification Oct-84 E, A, Identification May-85 E, GE Preparation Dec-85 E, GE. A.A Preparation Mar-86 E Appraisal Jun-86 3E, A. 2GE Post-Appraisal Dec-86 2E.A Supervision Sep-87 E, d 1 1 1 1 Mar-88 2E, GE, 2 - 1 3 Oct-88 E, GE, A 2 2 1 3 Mar-89 E, A 2 2 1 3 Oct-89 GE 2 2 1 3 Feb-90 E 2 2 2 2 Jun-90 GE, A 2 2 2 1 Oct-90 2E, A 2 1 2 1 Mar-91 E, GE, A 3 1 3 2 Jun-91 E. A. 3 1 3 2 Sep-91 E, GE, A 3 1 2 2 Apr-92 GE. 2 - 1 2 Sep-92 GE, E, A 2 1 1 2 Mar-93 GE. 2 1 1 2 Oct-93 GE/ 2 2 1 2 Feb-94 GE 2 2 1 2 Oct-94 GE, A 2 2 1 2 Completion Mar-94 E, A 2 E Economist GE = General educator A = Architect 21 Annex A Other Project Data Borrower/Executing Agency: Related Bank Loans/Credits Credit Achievement/Purpose Yr. of Amount Status approval First Credit Secondary school expansion, primary teacher training, 1967 None Closed 12/93 Cr. 102-MAl introduction of technical subjects in secondary schools. available Second Credit Primary teacher training, expansion of secondary 1975 None Closed 65/82 Cr. 590-MA! education and establishment 'model' primary schools. available Third Credit Expansion and improvement of primary education; 1979 14.5 Closed 5/86 Cr. 910-MAI expansion of primary teacher training; expansion of correspondence/distance learning facilities; establishment of Institute of Education (for teacher upgrading) and strengthening educational management and planning. Fourth Education Enrollment expansion in secondary schools, and increased 1981 41.0 Closed 12/86 Project output of secondary school teachers from the university Cr. 1123-MAI was achieved. Heavy reliance on expatriate staff for project implementation was noted. 1330-MA Fifth Expansion of primary teacher training, secondary school 1983 30.9 Closed 9/90 Educatgion enrollment including MCDEs and enrollment of Project accountancy students was successfully achieved. Heavy reliance on expatriate staff for project implementation was noted. Second Primary school expansion and quality improvement; 1987 27.0 Closed 12/93 Education Sector expansion of secondary schools; and institutional Credit development. Cr. 2083 Primary Improve the quality of primary education by providing 1996 22.5 12/00 Education teaching and learning materials, and pedagogical support Project to in-service training for untrained teachers. Cr. 2810-MAI Malawi Social Poverty alleviation through the construction of demand- 1996 56.0 12/01 Action Fund driven social infrastructure (including schools). Cr.2856-MAI Secondary Expand access through the construction of secondary 1998 48.2 12/03 Education schools; improve educational quality secondary schools Project through provision of teaching and leaming materials; Cr. 1670-MAI improve efficiency and effectiveness of the secondary education system by strengthening school-based management through training strengthen project implementation capacity through technical assistance; support Government's policy to encourage private secondary schools, and provision of training and materials to stem the spread of HIV/AIDS among teachers and students. 22 23 Annex A Basic Data Sheet MALAWI: EDUCATION SECTOR CREDIT II (CREDIT 2083-MAI) Key Project Data (US$ million) Source of Funds Appraisal Estimate Actual IDA 36.9 41.2 Domestic Contribution 4.1 3.3 TOTAL 41.0 44.5 Note: Fluctuations between the SDR and US$ explains the difference in the amount financed by IDA. Cumulative Estimated and Actual Disbursements BankFY 1990 1991 1992 1993 1994 1995 1996 1997 1998 Appraisal estimate 0.0 4.9 12.9 24.9 30.9 35.9 36.9 Actual 0.0 4.1 10.3 13.7 17.1 19.5 27.4 33.8 41 Actual as % of estimate 0 84% 80% 555 55% 54% 74% 92% 111% Date of Final Disbursement: November 10, 1998 Project Costs Category Appraisal Estimate Actual Local Foreign Total Local Foreign Total 1. Civil Works 8.6 8.6 17.2 2.9 19.3 22.2 2. Equip., furn.., vehicles, books 1.1 4.2 5.4 0.0 5.6 5.6 3. Archit., eng.& auditing cons. 0.2 0.3 0.5 2.5 1.0 3.5 Serv.& training 4. TA. Cons.serv. and training 0.9 3.4" 4.3 0.0 0.4 0.4 5. Project admin. (exclu. salaries) 0.9 0.9 1.9 0.3 1.0 1.13 6. Univ. students loan scheme 0.0 0.0 0.0 0.0 1.1 1.1 7. Unallocated 0.1 0.6 0.7 0.1 5.3 5.4 TOTAL 13.9 19.9 34.0 5.7 28.6 34.3 Figures do not add up due to rounding. this includes ODA co-financing of one TA staff, which was later changed to parallel financing. Project Dates Project Phase Planned Actual Identification (Executive Project Summary) March, 1988 March, 1988 Preparation Mission October, 1988 October 1-21, 1988 Appraisal Mission February/March, 1988 March, 1989 Negotiations September, 1989 September 6-10, 1989 Board Approval December, 1989 December 21, 1989 Signing March, 1990 April 6, 1990 Credit Effectiveness April 19, 1990 April 19,1990 Completion Date December 31, 1995 December 31, 1997 Closing Date June 30, 1996 June 30, 1998 Staff Inputs Stage of Project Cycle Actual Staff Weeks US$'000 Pre Appraisal 105.0 95.2 Appraisal-Negotiations 37.1 72.8 Supervision 151.6 443.6 Completion 12.3 11.9 Total 306.0 623.5 24 Mission Data Stage of project cycle Month/Yr. Specialized Staff Skills mpementation Status Represented Preparation 10/1988 2 E, 5 ES, 2 A, FS Appraisal 3/1989 2 E, 4 ES, 2 A, FS Board Effectiveness 1/1989-4/1990 E, ES, A Supervision 10/1990 E, ES, A Supervision 2/1991 E, 2 ES, A 2 Supervision 6/1991 E, ES, A Supervision 9/1992 2 ES, A, OA 2 Supervision 3/1993 ES, EQS 2 Supervision 10/1993 ES, TS 2 Supervision 2/1994 ES, E Supervision 10/94 ES S Supervision 2/1996 ES S Supervision 5/1996 PS S Supervision 12/1996 TS, E Supervision 2/1997 EE Supervision 6/1997 EE, TS, A, E, TNS S Supervision 10/1997 EE, PS, ES, 2 E, IS, CPS S Supervision 2/1998 EE, TS, PS, A, E S Supervision 4/1998 E S Completion 6/1998 Ed, E combined with other missions CPS = Community participation specialist FS = Financial specialist Ed = Education OA= Operations analyst EE = Education economist PS = Procurement specialist ES = Education specialist TNS = Technical/vocational specialist EQS = educational quality specialist TS = Textbook specialist IS = Implementation specialist Related Bank Credits (See Annex A, page 19) 25 Annex B The Education Sector Credits: Policy Actions, Objectives And Components First Education Sector Credit (FY88) Objectives * Increased access to primary and secondary education through the expansion of facilities, increased allocation of budgetary resources and implementation of efficiency measures to improve the flow of students entering these levels of education. * Improved quality of primary and secondary education through the upgrading of teachers and teaching materials and the provision of adequate textbooks. * Improved sector management and measures to improve educational quality and efficiency through TA and training. Components * At the primary level, the construction and equipping of 1,500 classrooms and associated staff housing, provision of two million textbooks, and in-service training for teachers. * At the secondary level, the construction and equipping of two boarding schools and one day school, conversion of three study centers to day secondary schools and improvement of the umbrella institution's organization and operations. * Education management improvement through the merger of two national teacher examinations systems and studies to monitor and evaluate the efficiency of implementing project investments, the costs effectiveness of educational inputs and instructional methods, and a tracer study on the income effects of post primary schooling. * At the University level, establishment of a student loan scheme. Covenanted Policy Actions * A minimum age for entry to primary school of 5 years. * One repetition only allowed in primary standard 8. * Reform of the primary school leaving examination. * Introduction of double shifts in 13 urban secondary schools. * Decisions on secondary school facilities based on a school mapping study. * Provision in the education recurrent budget for 2 kwacha per student for primary textbooks. * Introduction of an in-service teacher training program for teacher upgrading. 26 Annex B * Increased user charges for boarding to MK135 and thereafter to full cost-recovery. * No new boarding facilities to be constructed after the 10 in process unless IDA agreed * An increase in university fees net of living allowances to MK200, the real value to the maintained. The Second Education Sector Credit (FY90) Covenanted Policy Actions * The government was to adopt development and recurrent budgets for education satisfactory to IDA * The MOE was to: - implement the decentralization plan; - introduce new rules on the age of admission to primary school and on grade-repetition; - reduce the scope of the secondary curriculum; - increase tuition fees in secondary schools and the University; - increase boarding fees in secondary schools; and - discontinue the provision of free exercise books and writing materials for secondary school students. Objectives * Improved quality of education at all levels. * Expanded access to primary and secondary education. * Strengthened sector management, budgeting and planning. * Improved resource mobilization and allocation policies by reversing the decline in the sector's budget share in accordance with stated government policy. Components * Quality Improvements. Development of the curriculum, textbooks, science materials, libraries and examinations; a pilot educational radio program; increased numbers of qualified primary teachers and teacher training; and vehicles for school inspectors. * Access Improvements: upgrading of school facilities by providing concrete roofing sheets, furniture and equipment for school buildings constructed through community self-help schemes; introducing changes in grade-promotion criteria and more intensive use of facilities (double shifts at secondary schools and the admission of day students at the University); and expanding the University's capacity to produce accountants-a skill in extremely short supply. * Management Improvements: capacity-building for the Malawi Institute of Education and the MOE, and a ministry decentralization program. * Improved Resource Mobilization and Allocation: fee structure reforms at all levels. 27 Annex B Alternative Paradigms: Communities and School Construction MASAF is supported under the two Malawi Social Action Fund Projects. The Ministry of Economic Planning and Development coordinated preparation in 1995 with the help of IDA's social protection sector staff. The program includes community involvement in the construction of basic social infrastructure such as health, water supply, and school facilities. It has gained a reputation for rapid and inexpensive results due to the close involvement of the communities who stand to benefit (Box Cl). Box C1. How MASAF works MASAF channels resources directly to small-scale, community subprojects. These must be proposed, designed, and implemented by the beneficiary communities. They must meet eligibility criteria established for all MASAF investments, including a sound proposal similar to a business plan, which MASAF appraises. About 60% of the community subprojects so far have followed proposals for schools, teachers' houses, latrines and bore holes for water. Within four years, communities have managed the construction of 1,230 schools and 1,460 teachers' houses. The finished facilities vary in sophistication because the design and quality of the construction work relies on communities' capacities. The program is still learning how to mobilize and support communities in expressing and following through on their needs. As well as coaching in technical, management and business aspects, the program builds-in supervision, incentives to encourage accountability, and access to collaborative networks on local governments, business interests and other stakeholders. At the central level, the program has begun to develop public recognition and support at home and abroad through disseminating the findings of participatory surveys and evaluations. Distinctive Missions and Purposes Both programs share the aim of providing communities with school infrastructure as rapidly and efficiently as possible. But, beyond this, their emphasis diverges in terms of missions and purposes. * As a line agency, the MOE's mission is to make primary schools available and accessible on a mass scale. Its responsibility is to plan, manage and supervise the school construction process, administer the budgeted funds, and encourage communities to make their expected contributions in partnership with district education offices and the contractors that they hire. From the ministry's perspective, the guiding purpose of the community "self-help" schemes is to meet physical targets under the education plan within schedule and budget. If involvement encourages communities to remain committed to the schools in the long term, their sense of ownership holds promise that enrollments will rise and the local people will maintain the facilities as additional benefits. * Under its charter, MASAF's mission is to encourage participation as a core competency essential to community development and effective citizenship for the poor. The guiding principle is that when communities leam how to assert their own developmental priorities, these should be the driver for investments in community infrastructure; and when people identify their own needs and manage the process of fulfilling them, the process is efficient and the results are relevant. The agency's responsibility is to provide the funds and technical assistance in response to communities proposals for subprojects. The main purposes of the MASAF program are to help local people build confidence and skills to manage community development. At present, this has the related benefit of filling gaps in social infrastructure. 28 Annex B Different Mandates and Business Cultures Both programs are funded within the overall government budget and develop strategic plans in conformity with government policies But this similarity belies differences in mandate and business culture. Overall, MASAF aims to be participatory, market-driven and service-oriented. MASAF's senior managers report administratively to the Cabinet and have direct access to the President. Established under its charter to a semi-autonomous, nonprofit agency, it is not accountable to any sector ministries and makes efforts to engender non-partisan support. The Executive Director enjoys the flexibility to develop policies, business strategies, operating rules and procedures and human resource policy, including recruitment (Box C2). MASAF has developed institutional processes for program monitoring and beneficiary feedback, financial accounting and auditing, staff performance evaluation and accountability and management by results based on teamwork. It aspires to international standards of sound business practice. Box C2. Lessons learned in MASAF An early priority for MASAF was to develop a strong framework of policies, performance standards and business procedures at the central agency level in order to guide business practice as it evolved at local level. The central office is endowed with leadership skills and technical competencies including technical specialists in subproject design, appraisal and development, participation, training, budgeting, accounting and auditing. The common purpose is to achieve results through teamwork, high quality support services, monitoring and evaluation and attention to value. An important lesson learned is that the program's effectiveness also depends on strong local teams alongside the central institutions. At local level, field staff need skills not only to mobilize communities but also to persevere in coordination, facilitation and provision of technical assistance. MASAF invited the mission to visit a primary school. It has only two classrooms-for Standard 7 and 8-and six classes are held outside. Children sit on upturned bricks. Most teachers and pupils were absent that day because of the heavy and prolonged rain. In a meeting with the school committee, it turned out that the MASAF officer had not kept the members informed about the progress of their proposal to build classrooms and they expressed considerable concern about this. On hearing of this experience, central office staff said that the officer's competencies were known to be below standards expected and that remedies were being considered. Overall, the MOE's culture tends to be bureaucratic, supply-driven and process-oriented. As head of a line ministry within a sovereign government, and as a member of the Cabinet under the governing party, the Minister's mandate is to manage and develop the entire education system. Unlike MASAF, he must compete with other line agencies for budget, staff and priority status. He must juggle complex political and technical priorities and special interests. Primary school infrastructure is but one responsibility. As employees of an institution bound by administrative rules and procedures, managers have limited control over personnel policies and operational rules and procedures. The government has only just begun to develop performance standards to enhance management effectiveness and efficiency-checks and balances to keep proper track of costs and expenditures and mechanisms to enhance accountability and responsiveness to the public. Constraints on Knowledge-Sharing Overall, neither agency has so far invested heavily in knowledge-sharing and the mechanisms to encourage this remain weak. MASAF's view appears to be that the transaction costs are higher than the modest benefits. Efforts to engage the Ministry in sharing basic information and technical guidelines have not generally produced useful results, with the exception of those districts where staff are mutually committed to cooperation. Similarly, opportunities for generating mutual insights among MASAF and primary education experts are difficult to organize, in part because of the busy schedule and pressing day-to-day matters for government 29 Annex B staff. The sector ministries used to take part in MASAF's management steering committee but do so no longer. They do retain technical responsibilities on the advisory board but MASAF finds that the quality of inputs is variable because turnover of sector representatives is rapid and their level of familiarity with the topics to be discussed is variable. MASAF has learned to build its knowledge base. For basic sector information, it develops working relationships with district offices. For development of its own praxis, the agency relies on monitoring data and analysis, action research and evaluations. For global theory and practice, managers turn to the international experts, including the social fund literature and IDA staff. To maintain sector contacts on operational matters at central level, MASAF has invited sector ministries to appoint desk officers and is currently developing communication mechanisms at local levels, especially with the new local governments. The MOE consulted MASAF on the design of the PEP communication campaign and MASAF consulted the ministry on the preparation of the IDA projects. Since individual officials have worked within both agencies, they are in contact informally but not systematically. On a formal level, such as official meetings, taskforces and workshops, the quality of contact depends on commitment and mutual understanding. For the MOE staff, contact with MASAF is only one among many demands on them. For more than a year, the MOE desk officer and MASAF had no contact. The MOE affirms that it provides teachers to all schools without distinction. From the MASAF perspective, however, the MOE's criteria for rationing supply when faced with shortages are neither clear nor consistent. Without accurate information, communities perceive delays in receiving teachers as deliberate discrimination where none may be intended. The perception that ministry guidelines and procedures are unclear or out-of-date creates problems for MASAF and MOE field staff when communities want to try out new approaches in meeting their needs. The ministry concedes that MASAF has a superior reputation for mobilizing communities and achieving speedy results but considers that these advantages are offset by low quality construction. Does Competition make sense? Overall, MOE self-help is not the same thing as community participation under MASAF. The contemporary paradigm of participation as a grassroots, demand-driven process adopted by MASAF differs from the instrumental approach of all the education projects. Moreover, the programs differ in their main objectives, their styles of management, the scale of their operation and their time-frames for achieving results. For the MOE, the construction of schools is a high priority; for MASAF it is a means of meeting the priorities expressed by communities. The MOE's behavior suggests that it is ambivalent about whether MASAF is a strategic partner or a competitor. MASAF's behavior suggests that it welcomes guidance on education policy and processes but views the administrative style of the ministry as incompatible with closer relations and possibly hurtful to its own need to be agile and responsive. When observers suggest that the two agencies should work together or make judgments about comparative performance and achievements, these differences are not always appreciated. The metaphor in Box C3 is a deliberate exaggeration to encourage scrutiny of the underlying assumptions perpetuated by both parties. Box C3. Competitors for Market Share? Metaphorically, MASAF may be likened to a small, entrepreneurial company with a driving vision, an evolving business strategy and growing brand-recognition. The start-up company aims to sell distinctive products in small and difficult markets that are undeveloped by a large monopoly firm. Similarly, the ministry may be likened to a large, established company with market dominance and a well-known commodity. The large company sees an assertive newcomer breaking down the barriers to entry to invade its traditional business, threaten its dominance with fast delivery at low cost and take over as yet unexploited markets. 30 Annex B The apparent reluctance of both agencies to invest in mechanisms to exchange of knowledge and experience appears to come from a shared perception that the differences are too wide and do not need to be bridged. There are, however, some areas of commonality. Both share the objective of involving communities in the development of social infrastructure that will benefit them directly. Their central agencies are responsible for shaping policy and maintaining quality assurance mechanisms. Their local staff need tc work in partnership with contractors, stakeholders and local governments. Does Partnership make sense? Since the MOE has experienced difficulties with school construction for a prolonged period, and especially with the shift from a small to large school system, the question is sometimes raised whether it would make sense for the long-term for the MOE to move out of primary school construction completely. This is usually premised on the assumption that the required supervisory and technical capacity may be more easily fostered through local management and community networks. MASAF is slated as the most obvious candidate to lead such a program (just as GABLE is sometimes perceived as the model for quality improvement). Alongside this view is juxtaposed the proposition that, in the context of government decentralization, the MOE should put top emphasis on developing sector policies, standards of performance for education managers, teachers and students, and the supportive specialist institutions, such as teacher training, curriculum and testing and supervision which are necessary to improve teaching and learning. The immediate question, nevertheless, is whether closer collaboration in sharing knowledge and experience between the MOE and agencies such as MASAF would be productive. Asked for their views, people interviewed raised more questions than answers. The issues raised are grouped below, not as recommendations or priorities-that would be beyond the scope of the audit-but as policy and strategy dilemmas as perceived by stakeholders and beneficiaries. They may be helpful as feedback for decision-makers but they would require a deep sharing of knowledge and experience by both partners in terms of desirability, feasibility and sustainability. To address issues such as these satisfactorily, systematic exchange of experience and ideas between the two agencies would be required. But this is unlikely to happen automatically. The critical first step would be for both agencies to agree that their shared aims are more critical than their differences and that an agenda for knowledge-sharing would be mutually beneficial. If the benefits are clear and a firm commitment is made, the next step would be to develop mechanisms to create a business partnership, operational plan and time-frame for a pilot program. The aim would be to develop mutual appreciation of each other's strengths and replicable features, find solutions for areas where implementation problems have occurred in the past and establish collaborative evaluations for future planning. 31 Annex C Table Cl. Sharing Knowledge and Experience for Future Strategy Ministry of Education MASAF To ensure that enrollment targets for 2004 would be met, Would it make sense, or not, for MASAF to offer to would it make sense, or not, for the MOE to adopt the expand the community school construction program on MASAF approach in its countrywide program of primary behalf of the MOE? school construction and rehabilitation? Would it make sense to broaden the program's reach to What lessons would it need to gather about how to the poorest communities in very remote rural areas in mobilize communities, especially those in hard-to-reach order to help meet national targets? areas? What lessons of experience would it need to gather about What legal and technical capacities would it require that it replication on a large scale? currently lacks? What changes in its charter and governance would be needed? If the ministry devolved responsibility for the If MASAF became one of several agencies collaborating construction of primary schools to other agencies and to attain national targets, what policy, procedural and communities, at what stages would it need to retain technical guidelines, if any, would it find helpful? responsibility? If MASAF chose to form a strategic alliance with the Would the ministry be involved in setting standards for sector ministry, what legal, financial and other safeguards school site-selection and building design, procuring would it require to ensure its autonomy, flexibility and construction materials, or managing building contractors? distinctive mission? Would the ministry need to have a direct supervisory presence at local level, in order to ensure that school facilities developed through programs such as MASAF complied with national standards? Since the increasing number of primary schools require Since MASAF schools may not be physically durable, regular maintenance and repair, can the ministry expand how far should the agency become involved in developing its current maintenance program to cover the entire capacities in preventive maintenance and repair so that system? communities may derive long-term value from their investments in education? Since communities need encouragement to support How far has MASAF succeeded in building community investments in houses for teachers, could other agencies' support for investments in houses for teachers? Would its experience be useful in adapting current approaches? experience be useful for the nation-wide program?  33 Annex C Table Cl. Sharing Knowledge and Experience for Future Strategy Ministry of Education MASAF To ensure that enrollment targets for 2004 would be met, Would it make sense, or not, for MASAF to offer to would it make sense, or not, for the MOE to adopt the expand the community school construction program on MASAF approach in its countrywide program of primary behalf of the MOE? school construction and rehabilitation? Would it make sense to broaden the program's reach to What lessons would it need to gather about how to the poorest communities in very remote rural areas in mobilize communities, especially those in hard-to-reach order to help meet national targets? areas? What lessons of experience would it need to gather about What legal and technical capacities would it require that it replication on a large scale? currently lacks? What changes in its charter and governance would be needed? If the ministry devolved responsibility for the If MASAF became one of several agencies collaborating construction of primary schools to other agencies and to attain national targets, what policy, procedural and communities, at what stages would it need to retain technical guidelines, if any, would it find helpful? responsibility? If MASAF chose to form a strategic alliance with the Would the ministry be involved in setting standards for sector ministry, what legal, financial and other safeguards school site-selection and building design, procuring would it require to ensure its autonomy, flexibility and construction materials, or managing building contractors? distinctive mission? Would the ministry need to have a direct supervisory presence at local level, in order to ensure that school facilities developed through programs such as MASAF complied with national standards? Since the increasing number of primary schools require Since MASAF schools may not be physically durable, regular maintenance and repair, can the ministry expand how far should the agency become involved in developing its current maintenance program to cover the entire capacities in preventive maintenance and repair so that system? communities may derive long-term value from their investments in education? Since communities need encouragement to support How far has MASAF succeeded in building community investments in houses for teachers, could other agencies' support for investments in houses for teachers? Would its experience be useful in adapting current approaches? experience be useful for the nation-wide program?  35 Annex D COMMENTS FROM THE MINISTRY OF EDUCATION MINISTRY OF EDUATION SCEINCE AND TECHNOLOGY COMMENTS ON DRAFT PERFORMANCE AUDIT REPQ_RY ON MALAWI : FIRST AND SECOND EDUCATION SECTOR (9-EDITS 1767 - MAI & 2083 - MAl INTRODUCTION In general Ministry agrees with the report and its findings. There have truly been problems in implementing the projects under the two credits. The problems generally affect conceptualization, consultation, capacity determination in terms of expertise, and numbers of qualified people and the availability of part two financial contributions. However, it is equally fair to point out that the Bank staff had contributed a share of these problems. SPECIFIC PROBLEM AREAS i. PROJECT CONCEPTUALIZATION There is every indication that the projects were agreed upon and signed before thorough conceptualization. let alone wide consultation. This process requires the full involvement of all relevant stakeholders, local communities included. As a result of the ill-conceptualization and consultation we have ended-up realizing later that we made hasty decisions and commitments when Government was .not ready. These are expensive project omissions. 36 Annex D While accepting the weaknesses cited above it was very often observed that the Bank had influenced the transfer of whatever practices which worked elsewhere to assume they would work in Malawi. This put pressure on the borrower to hastly accept the project and unduly commit Government. ii. IMPLEMENTATION CAPACITY It is true that under these two credits the borrower had no or inadequate capacity to deliver a large and complex construction programme. Both technical and administrative staff to implement and support it were recruited late and in piecemeal. The situation was worsened by the weak and unrealiable local and small construction industry capacity and local community readiness to deliver. These proved to be ill-prepared in terms of adequate resources and lack of expertise. iii. INABILITY TO TURN DOWN PROJECT SUPPORT WHEN NOT READY The observation that the borrower accepted an overload when it was clear that it was already struggling to implement the first credit is correct. However, this does not seem to be deliberate. It is obviously because of mounting pressure from both the community and political needs and expectations. This became even more 37 Annex D pronounced after the 1994 democratic change and with the introduction of Free Primary Education as a centre-piece of poverty alleviation. However, IDA should have been able to prepare the borrower even under such pressures. IDA knew the borrower's limitations but decided not to advise probably for fear of not embarrassing Government. iv. BORROWER'S PERFORMANCE Admittedly the Borrower was caught in between two worlds. There was mounting community and political pressures to provide access to and quality education. This situation forced the borrower to hastly put together poorly thought-out and unrealistic plans and accepted and signed legal covenants with the Bank which later it could hardly fulfill. The borrower was unable to recruit the project management and administration staff it committed on time nor was it able to provide the part two funding it committed itself to. All these contributed to its poor performance and straining relationship with the Bank.