Document of The World Bank Report No.: 75556 PROJECT PERFORMANCE ASSESSMENT REPORT MALAWI COMMUNITY-BASED RURAL LAND DEVELOPMENT PROJECT (GRANT NUMBERS H0810 & H5270; P075247 & P115226) June 27, 2013 IEG Public Sector Evaluation Independent Evaluation Group ii Currency Equivalents (annual averages) Currency Unit = Kwacha (MKw) 2003 US$1.00 = MKw97 2004 US$1.00 = MKw109 2005 US$1.00 = MKw118 2006 US$1.00 = MKw136 2007 US$1.00 = MKw140 2008 US$1.00 = MKw141 2009 US$1.00 = MKw141 2010 US$1.00 = MKw151 2011 US$1.00 = MKw157 2012 US$1.00 = MKw163 Abbreviations and Acronyms ICR Implementation Completion and Results Report IEG Independent Evaluation Group PPAR Project Performance Assessment Report Fiscal Year Government: July 1 – June 30 Director-General, Independent Evaluation : Ms. Caroline Heider Director, IEG Public Sector Evaluation : Mr. Emmanuel Jimenez Manager, IEG Public Sector Evaluation : Ms. Marie Gaarder Task Manager : Mr. John R. Heath iii Contents Principal Ratings ................................................................................................................. v Key Staff Responsible......................................................................................................... v Preface............................................................................................................................... vii Summary ............................................................................................................................ ix 1. Background ..................................................................................................................... 1 2. Objectives, Design, and their Relevance ........................................................................ 2 Objectives ....................................................................................................................... 2 Original ....................................................................................................................... 2 Revised........................................................................................................................ 2 Relevance of the Objectives............................................................................................ 3 Design ............................................................................................................................. 4 Components ................................................................................................................ 4 Geographic Scope ....................................................................................................... 5 Implementation Arrangements and Project Process ................................................... 6 Design of monitoring and evaluation .......................................................................... 8 Relevance of Design ....................................................................................................... 8 3. Implementation ............................................................................................................. 11 Planned vs. Actual Expenditure .................................................................................... 11 Implementation experience ........................................................................................... 11 Implementation of Monitoring and Evaluation............................................................. 14 Safeguards and Fiduciary Compliance ......................................................................... 14 4. Achievement of the Objectives ..................................................................................... 15 Outputs and Intermediate Outcomes ............................................................................. 16 Increase the incomes of about 15,000 poor rural families (original and revised objective)....................................................................................................................... 17 Increase the agricultural productivity of approximately 15,000 poor rural families (revised objective) ......................................................................................................... 20 5. Efficiency ...................................................................................................................... 23 6. Ratings .......................................................................................................................... 24 Outcome ........................................................................................................................ 24 Risk to Development Outcome ..................................................................................... 25 This report was prepared by John R. Heath, who visited Malawi in September 2012. The report was peer reviewed by Enrique Pantoja and panel reviewed by Fernando Manibog. Marie Charles provided administrative support. iv Bank Performance ......................................................................................................... 27 Quality at Entry ......................................................................................................... 27 Quality of Supervision .............................................................................................. 28 Borrower Performance .................................................................................................. 30 Government Performance ......................................................................................... 30 Implementing Agency Performance ......................................................................... 30 Monitoring and Evaluation ........................................................................................... 31 Design ....................................................................................................................... 31 Implementation ......................................................................................................... 31 Utilization ................................................................................................................. 31 7. Lessons .......................................................................................................................... 32 References ......................................................................................................................... 34 Annex A. Basic Data Sheet ............................................................................................... 37 Annex B. Other Data......................................................................................................... 41 Annex C. List of Persons Interviewed .............................................................................. 47 Boxes Box 1: Environmental and Social Risks ........................................................................... 15 Tables Table 1: Project Components .............................................................................................. 4 Table 2: The Farm Acquisition and Development Process ................................................ 7 Table 3: Planned vs. Actual Expenditure by Component ................................................. 11 Table 4: New Land Administration Outputs under Additional Financing ....................... 13 Table 5: Survey Evidence on Project Outcomes............................................................... 19 Table 6: Derivation of the Outcome Rating ...................................................................... 25 Figures Figure 1: Malawi--Poverty Headcount Ratio ...................................................................... 2 Figure 2: Population Density of Project Districts, 1998 ..................................................... 5 v Principal Ratings ICR* ICR Review* PPAR Outcome Satisfactory Satisfactory Satisfactory Risk to Moderate Moderate Significant Development Outcome Bank Performance Satisfactory Satisfactory Satisfactory Borrower Satisfactory Satisfactory Moderately Satisfactory Performance * The Implementation Completion and Results (ICR) report is a self-evaluation by the responsible Bank department. The ICR Review is an intermediate IEG product that seeks to independently verify the findings of the ICR. Key Staff Responsible Division Chief/ Project Task Manager/Leader Sector Director Country Director Appraisal Rogier van den Brink Richard G. Scobey Hartwig Schafer Completion Hardwick Tchale Karen McConnell Brooks Kundhavi Kadiresan vi IEG Mission: Improving World Bank Group development results through excellence in independent evaluation. About this Report The Independent Evaluation Group assesses the programs and activities of the World Bank for two purposes: first, to ensure the integrity of the Bank’s self-evaluation process and to verify that the Bank’s work is producing the expected results, and second, to help develop improved directions, policies, and procedures through the dissemination of lessons drawn from experience. As part of this work, IEG annually assesses 20-25 percent of the Bank’s lending operations through field work. In selecting operations for assessment, preference is given to those that are innovative, large, or complex; those that are relevant to upcoming studies or country evaluations; those for which Executive Directors or Bank management have requested assessments; and those that are likely to generate important lessons. To prepare a Project Performance Assessment Report (PPAR), IEG staff examines project files and other documents, visit the borrowing country to discuss the operation with the government, and other in-country stakeholders, and interview Bank staff and other donor agency staff both at headquarters and in local offices as appropriate. Each PPAR is subject to internal IEG peer review, Panel review, and management approval. Once cleared internally, the PPAR is commented on by the responsible Bank department. The PPAR is also sent to the borrower for review. IEG incorporates both Bank and borrower comments as appropriate, and the borrowers' comments are attached to the document that is sent to the Bank's Board of Executive Directors. After an assessment report has been sent to the Board, it is disclosed to the public. About the IEG Rating System for Public Sector Evaluations IEG’s use of multiple evaluation methods offers both rigor and a necessary level o f flexibility to adapt to lending instrument, project design, or sectoral approach. IEG evaluators all apply the same basic method to arrive at their project ratings. Following is the definition and rating scale used for each evaluation criterion (additional information is available on the IEG website: http://ieg.worldbankgroup.org). Outcome: The extent to which the operation’s major relevant objectives were achieved, or are expected to be achieved, efficiently. The rating has three dimensions: relevance, efficacy, and efficiency. Relevance includes relevance of objectives and relevance of design. Relevance of objectives is the extent to which the project’s objectives are consistent with the country’s current development priorities and with current Bank country and sectoral assistance strategies and corporate goals (expressed in Poverty Reduction Strategy Papers, Country Assistance Strategies, Sector Strategy Papers, and Operational Policies). Relevance of design is the extent to which the project’s design is consistent with the stated objectives. Efficacy is the extent to which the project’s objectives were achieved, or are expected to be achieved, taking into account their relative importance. Efficiency is the extent to which the project achieved, or is expected to achieve, a return higher than the opportunity cost of capital and benefits at least cost compared to alternatives. The efficiency dimension generally is not applied to adjustment operations. Possible ratings for Outcome: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Risk to Development Outcome: The risk, at the time of evaluation, that development outcomes (or expected outcomes) will not be maintained (or realized). Possible ratings for Risk to Development Outcome: High, Significant, Moderate, Negligible to Low, Not Evaluable. Bank Performance: The extent to which services provided by the Bank ensured quality at entry of the operation and supported effective implementation through appropriate supervision (including ensuring adequate transition arrangements for regular operation of supported activities after loan/credit closing, toward the achievement of development outcomes. The rating has two dimensions: quality at entry and quality of supervision. Possible ratings for Bank Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Borrower Performance: The extent to which the borrower (including the government and implementing agency or agencies) ensured quality of preparation and implementation, and complied with covenants and agreements, toward the achievement of development outcomes. The rating has two dimensions: government performance and implementing agency(ies) performance. Possible ratings for Borrower Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. vii Preface This is the Project Performance Assessment Report (PPAR) for the Malawi Community- Based Rural Land Development Project (Grant Numbers H0810 and H5270; P075247 and P115226). The assessment aims, first, to serve an accountability purpose by verifying whether the operation achieved its intended outcome. Second, the report draws lessons that are intended to inform future operations of this nature. A grant of US$27 million was approved by the Bank’s Board on April 13, 2004 and made effective on July 13, 2004. An additional grant of US$10 million was approved by the Board on November 19, 2009, and with Board approval the project development objective was slightly amended. The operation closed on September 30, 2011, 15 months later than initially expected. Actual project costs were US$37.9 million, compared to the appraisal estimate of US$29.8 million. This report was prepared by John R. Heath, Senior Evaluation Officer, IEG, who visited Malawi in September 2012. In addition to Lilongwe, the mission traveled to the districts of Thyolo, Machinga, Mangochi and Balaka, conducting interviews in district townships and in six villages (listed in Annex B, Table B1). The findings of the report are based on a review of project documents, Bank electronic files, academic books and articles, the field visits and interviews with Bank staff, government officials, and representatives of donor agencies (listed in Annex C). IEG much appreciates the assistance of all those who participated in the assessment, including staff at the Bank’s office in Lilongwe. It would particularly like to thank the former project manager, Mr. Callistus Kharapuwa, for the support he provided throughout the mission. Following standard IEG procedures, the Government of Malawi was invited to comment on the draft report, but no comments were received. ix Summary In Malawi, a long-term policy bias in favor of large farms has helped to marginalize smallholders. Most farmers operate less than one hectare of land. At the same time, many large estates are idle. In the early 2000s, studies found that there were 2.6 million hectares of underused cultivable land that could be targeted for redistribution to smallholders. With support from the World Bank, the Government of Malawi launched a pilot operation based on voluntary land transfers between landowners willing to sell and persons with little or no land willing to buy. This was the first redistributive land reform project to be implemented in the World Bank’s Africa Region and it was the first in the world to use Bank funds for land acquisition, based on a formal exception to the decades-old policy that precluded disbursement against land purchase. The original objective of the Community-Based Rural Land Development Project was to increase the incomes of 15,000 poor rural families through the implementation of a decentralized, voluntary community-based land reform pilot program. When additional financing for the pilot was approved in 2009, the wording of the project development objective was altered slightly, making it clear that the operation was also intended to increase agricultural productivity. The project aimed to raise the incomes of poor farmers, first, by giving groups of small farmers grants to buy land and, second, by providing the training and the inputs needed for farmers to develop viable farms. To this end, the resettlement grant that was provided financed not only the purchase of land but also essential farm inputs, as well as giving beneficiaries the cash needed to cover the cost of moving to a new location and building rudimentary shelter. Although the Ministry of Lands was the lead implementing agency, the Malawi Social Action Fund—which had been running for 8 years when the pilot was approved—was charged with disbursing project funds for farm purchase and farm development, and for financing (outside the project) the community infrastructure where beneficiaries settled. There was a practical logic to this implementation arrangement; including communal infrastructure as a project component would have made this operation too expensive for a pilot, and it made sense to work through an established program. But the arrangement presupposed a synchronization of social fund investment priorities, set by local government, and the needs of the resettled land reform beneficiaries—something that proved difficult to achieve. The project met its output target, settling 15,142 poor rural families and providing each with a two-hectare plot. The pilot exceeded both its original and its revised development objective: using a panel design, the findings of two separately-authored impact evaluations agreed that the increase in the incomes, farm output, and agricultural productivity of beneficiaries largely exceeded the comparable results achieved by matched control groups. However, the boost to agricultural productivity was largest in the year after resettlement, tapering off after beneficiaries had used up their resettlement grant. Also, diversification from maize (the main staple) into cash crops was limited, owing to shortfalls in cash and technical assistance, and the lack of access to credit. Contract farming is one avenue for diversification but farmer outcomes vary substantially according to the terms offered by the x contractors. A study sponsored by the project showed that land reform beneficiaries fared less well with tobacco contracts than non-beneficiaries, mainly it seems because most of the beneficiary sample had signed up with a different company from households in the non- beneficiary sample and faced more onerous repayment terms. Project outcome is rated satisfactory. Relevance was substantial for the original objective and high for the revised objective: because the latter emphasized increasing farm productivity— rather than simply increasing farm output—it was more closely in line with the strategies of the government and of the Bank. Achievement of objectives was rated high against the original objective because given that beneficiaries ended up with more farm land than non- beneficiaries they were able to obtain a larger increase in farm output volume. Against the revised objective, achievement is rated substantial: in terms of productivity (measured as farm yields and net incomes per hectare), beneficiaries still performed better than non- beneficiaries, but by a smaller margin than for output volume. Efficiency was rated substantial because, based on plausible assumptions, the economic rate of return exceeded expectations: assuming that only 20 percent of beneficiaries cultivated cash crops, the economic rate of return was estimated at 20 percent, above the rate of 15 percent forecast at appraisal. The flexibility of project design and the quality of the panel study relative to costs were other reasons for rating efficiency substantial. Under both the original and revised objectives, the outcome rating is satisfactory. Risk to development outcome is rated significant. This rating reflects the expectation when the project was designed that this pilot would ultimately be scaled up, an outcome that was not certain when IEG visited Malawi. On the one hand, despite the adversities faced by many beneficiaries in their new location (particularly, the remoteness from schools, health posts and markets), only about 10 percent of those resettled have abandoned these, and the majority of those interviewed in both project-sponsored surveys and by IEG are adamant that they are economically better off than they were before the relocation and in relation to non- beneficiaries in the new location. Most seem determined to make a go of their new lives. On the other hand, there is a risk that this pioneering effort at land redistribution will not be scaled up or replicated—although this aim was not stated in the project development objective, it was listed as an expected outcome in the project results framework. The government has not introduced the expected reforms that would allow for more efficient operation of land markets and, without these, scale up would be ill advised. Neither the government nor the Bank have committed to renewed efforts at redistributive land reform, as is manifest in the Country Assistance Strategy for FY13-16 that was issued three months after IEG’s mission . Bank performance is rated satisfactory. This pilot operation was the result of dedicated, persistent and patient advocacy by the project team, and the resulting project design and implementation arrangements were both thorough and eminently practical. The pilot generated a model for willing buyer, willing seller land reform that other countries may want to replicate in other circumstances where land policies are more conducive than they presently are in Malawi. Borrower performance is rated moderately satisfactory. Without the competence and the hard work of staff in the project management unit the project’s targets would not have been met. But the government contributed less than expected: progress on the land bill stalled, decisions about the rents applied to government estates were ill-advised, the xi land tax on freehold land did not progress, and there was an absence of resolve not to renew leases on idle estates. There are seven lessons from this pilot operation:  The willing buyer, willing seller approach to land redistribution can be made to work—but the institutional framework needed for success is demanding;  The finer design details of the willing buyer, willing seller model of land reform are critical to its success;  Land redistribution is but one component of a successful land reform project;  Land redistribution can raise the incomes of poor households by increasing the amount of land farmed, and it may generate a one-off boost in productivity by bringing idle land into use; further increases in agricultural productivity may be harder to realize;  The success of land redistribution projects will be enhanced by timely studies on land availability;  Giving the rural poor the option of homesteading remains a valid objective; and  The scope for contract farming to help land reform beneficiaries make the leap from subsistence cultivation to cash cropping merits closer investigation. Caroline Heider Director-General Evaluation 1 1. Background 1.1 Land allocation policies in Malawi helped to concentrate land ownership in large- scale estates. The rapid expansion of estate agriculture, coupled with land policies that favored the acquisition of vast pieces of land by a few individuals for commercial agriculture, resulted in a highly unequal distribution of land, a colonial heritage that remained unchallenged in the four decades after Malawi became independent from Britain. The policy bias in favor of large farms has been widely documented (Kydd and Christiansen 1982; Dorward 1999). This is not economically rational. Various authors have argued that, except in cases where inputs are bulky and proximity to processing and shipping is critical, there are no economies of scale in agriculture, such that there may be efficiency as well as equity grounds for supporting the subdivision of large estates (Binswanger, Feder and Deininger 1995; Eastwood, Lipton and Newell 2010). 1.2 In the early 2000s, most farmers cultivated less than one hectare and, at the same time, it was widely reported that estate land was substantially underused (Chirwa and Chinsinga 2008). The public discussions about land reform following the transition to a multi-party democracy in 1994 and, specifically, the setting up of a Presidential Commission on Land Inquiry in 1996, raised expectations about a government response to land hunger, and may have encouraged invasions of freehold land, particularly in the southern districts of Thyolo and Mulanje. In 1999, the government estimated that illegal land occupations of privately owned farms covered from 5 percent of farms less than 20 hectares to 52 percent of farms with 500 hectares or more (Kanyongolo 2005:129; Chinsinga 2008). 1.3 The designers of the Community-Based Rural Land Development Project hypothesized that recent policy changes would increase the supply of farmland on the market. The opening up of tobacco cultivation to small farmers in the 1990s created competition for tobacco estates; coupled with the fall in tobacco prices, this seemed to make it more likely that some large tobacco producers would sell up. In 2002, a new National Land Policy was introduced, partly to address inequalities in land distribution. Based on separate land use studies supported by the European Union, United Kingdom, United States and the World Bank, the government concluded that there were 2.6 million hectares of underused cultivable land that could be targeted for redistribution. The confluence of an expected increase in land sales and a progressive trend in government policy created an opportunity for introducing a new approach to land redistribution based on voluntary land transfers between landowners willing to sell and persons with little or no land willing to buy (Simtowe, Mendola, and Mangisoni 2011). 1.4 There was a sound case for experimenting with new approaches to reducing rural poverty. While the proportion of the urban population that was poor declined sharply after 1998, the more modest downward trend in the prevalence of rural poverty ceased after the 2004, with the poor’s share of the rural population sticking at 56 percent (Figure 1). 2 Figure 1: Malawi--Poverty Headcount Ratio 70 66.5 Percent of Rural/Urban Population 60 54.9 55.9 56.6 50 40 Rural 30 25.4 Urban 20 17.3 10 0 1998 2004 2010 Source: World Development Indicators Database, 2013 2. Objectives, Design, and their Relevance Objectives ORIGINAL 2.1 The project appraisal document and the development grant agreement use similar language to describe the project development objective. According to the grant agreement, ―the objective of the Project is to increase the incomes of about fifteen thousand (15,000) poor rural families through the implementation of a decentralized, voluntary community- based land reform pilot program on Eligible Land in the Project Districts‖ (World Bank 2004b: 19).1 REVISED 2.2 When additional financing was approved by the Bank’s Board in October 2009, the project development objective was expanded to add a reference to agricultural productivity. The revised wording of the objective was: ―To increase the agricultural productivity and incomes of approximately 15,000 poor rural families through the implementation of a decentralized, voluntary community-based land reform pilot program on eligible land in the Project districts‖ (World Bank 2010: 17). 1 The project appraisal document says: ―The project development objective is to increase the incomes of about 15,000 poor rural families by implementing a decentralized, community-based and voluntary approach to land reform in 4 pilot districts in southern Malawi‖ (World Bank 2004a: 2). 3 Relevance of the Objectives 2.3 The project’s objectives responded to the need to redress a long-run tendency toward the marginalization of smallholders in Malawi (Kydd and Christiansen 1982). Following the establishment of Malawi as a British protectorate in 1891, the colonial administration promoted the transfer of customary land to European estate owners, and then forced Africans to seek wage employment on the estates by imposing a hut tax. Following independence from Britain in 1964, government policy retained its large-farm bias, by banning peasant cultivators from growing the main cash crop, tobacco, and by using marketing board surpluses to finance the development of commercial agriculture based on the estates. In spite of the support received from government policy, many of the estates were subject to weak financial management and lacked the internally-generated savings needed to invest in raising productivity. 2.4 Given the unequal distribution of land, most of the farmers in Malawi were unable to support their families with the income generated from their small plots: in 1998, 60 percent of the 1.8 million smallholders cultivated less than one hectare of land. When this project was prepared it was estimated that one-third of arable land in Malawi was underused (World Bank 2004a). Small, overworked parcels lay next to idle land that was locked up in private or government-owned estates. Land distribution was particularly uneven in the south, where the original project's four districts were located. In this region, rural population density was higher than in most other parts of Africa. Evidence from household surveys pointed to the severity of rural poverty and also showed that poverty was driven to a large extent by the lack of access to land. The decision to use land redistribution as the means to reduce poverty was partly influenced by lessons learned from the Land Reform and Poverty Alleviation Pilot Project in Brazil (P006475), which had created viable family farms out of larger properties that had been willingly sold by their owners (World Bank 2003b; 2009). It also took account of previous unsatisfactory attempts by government to redistribute land in Malawi, attempts that had been marred by political interference and corruption, aggravating social tensions (World Bank 2012). 2.5 Increasing poor people's access to land was at the heart of the government's approach to reducing poverty. The 2002 Poverty Reduction Strategy Paper argued that most Malawians depend on farmland for their living, noting that restricted land access is one of main causes of poverty. The same point was made in the FY07-10 Country Assistance Strategy, which was the most recent statement of Bank strategy for Malawi when the project closed. This document noted that small landholdings and degraded land reduced agricultural productivity, citing the Community-Based Rural Land Development Project as one of the instruments that the Bank was using to raise smallholder agricultural productivity (World Bank 2007a: 36). 2.6 Project objectives are also consistent with long-run corporate strategy. In its 1975 Land Policy Paper, the Bank argued that redistributive land reform is a valid intervention, ―consistent with the development objective of increasing output, improving income distribution and expanding employment‖ (World Bank 1975; see also, Deininger and Binswanger 1999). The Bank’s most recent land policy paper (World Bank 2003) also makes the case that land redistribution can promote growth and reduce poverty. 4 2.7 Relevance of the original objective to raise income is rated substantial and for the revised objectives of both raising income and productivity it is rated high: by raising the bar to emphasize increased agricultural productivity, not just more income from agricultural production, the revised objectives were more in line with the emphasis in the FY07-10 Country Assistance Strategy on increasing smallholder productivity. Design COMPONENTS 2.8 The four project components are captured in Table 1. The funding of the project was increased through additional financing approved in 2009 but the components were not revised. Component 1 financed land purchase and farm development. It did not fund the off- farm infrastructure investments (schools, health posts, etc.) that were deemed essential complements to the beneficiary settlements; these investments were to be funded outside the project, through an established institution, the Malawi Social Action Fund, the intention being that approval would be synchronized with approval of the farm development plans. Components 2 and 3 financed capacity building, including beneficiary training and support to the operation of the Ministry of Lands, Physical Planning and Surveys. Table 1: Project Components COMPONENTS 1. Land Acquisition and Farm Development Formation of Beneficiary Groups, identification of land available for redistribution, negotiation of price with willing sellers, financing of land purchase, and preparation of farm development plans. 2. Land Administration Development of the capacity of the Ministry of Lands, Physical Planning and Surveys at both the national and the local government levels to assist the Beneficiary Groups in the purchase and transfer of land acquired under the Project, involving the financing of training, acquisition of equipment and vehicles, and operating costs. 3. Capacity Building Financing of public information campaigns, community mobilization, participatory rural appraisal, training and technical support to communities, district and national participating institutions and stakeholders. This will include financing environmental and social impact assessments and action plans. 4. Project Management, Monitoring and Evaluation Administration, coordination and supervision of activities to be implemented under the Project. Design and implementation of a monitoring and evaluation system, to assess (i) the number and cost of farms acquired under the Project by Beneficiary Groups; (ii) the number of poor, rural families that benefit from the Project; and (iii) the speed with which Beneficiary Groups commence agricultural production on purchased land. Source: World Bank 2004a; World Bank 2004b. Note: The steps in the project process are set out in Table 2 below. 5 GEOGRAPHIC SCOPE 2.9 When the project was appraised the intention was to launch the land reform pilot in the districts of Mulanje, Thyolo, Machinga, and Mangochi, where there were ―simmering tensions‖ over land rights: ―The history of the carving out of estates from customary lands, combined with the current juxtaposition of idle, unused estates with overcrowded lands has resulted in localized tensions‖ (World Bank 2004a: 100). The districts of Mulanje and Thyolo had population densities above the nationwide average (Figure 2). While there were other districts in the southern region with higher population densities than Mulanje and Thyolo, encroachments and invasions (sometimes associated with extreme violence) were most frequent in these two districts, and were widely covered by the media, creating pressure for their inclusion in the project, irrespective of how much land was likely to be offered for sale. At appraisal, the Bank maintained that there were unused estates with lands available for redistribution in Mulanje and Thyolo. According to a project team member, one of the advantages of including Mulanje and Thyolo was that estate owners in these districts policed their boundaries, making land invasions less likely; if estates were invaded, project rules prevented them from being considered for redistribution. Most of the estates in these districts were tobacco and tea plantations, often operated with foreign capital. In central Malawi, idle estate land was more abundant but the owners were more likely to be absent from the property, increasing the propensity for invasion. Figure 2: Population Density of Project Districts, 1998 Mangochi 96 Machinga 97 MALAWI 104 Ntcheu 108 Persons per km2 Balaka 115 Mulanje 208 Thyolo 267 0 100 200 300 Source: World Bank 2004a: 100. 2.10 Initially, the project also included other districts in the south, Machinga and Mangochi, where population density was lower, allowing for the possibility of resettling persons from the two more crowded districts. The PAD noted that ―there was already some natural migration from Thyolo and Mulanje to Machinga, which has an ample supply of idle estates. Mangochi is somewhat farther away, but has a substantial number of unused estates. Even though relatively less densely-populated, both Machinga and Mangochi have significant numbers of land-poor households, implying a substantial potential for land redistribution within these districts‖ (World Bank 2004a: 102). 6 IMPLEMENTATION ARRANGEMENTS AND PROJECT PROCESS 2.11 Project implementation was handled by the Ministry of Lands, Physical Planning and Surveys (henceforth, the ―Ministry of Lands‖) and the Malawi Social Action Fund. 2 The Ministry of Lands was responsible for strengthening land administration institutions at the central and district levels (Components 2, 3, and 4), while the Malawi Social Action Fund was responsible for implementing the land acquisition and farm development subprojects (Component 1), which accounted for 55 percent of estimated project costs (Table 2). At the Ministry of Lands, the Project Manager reported to the Principal Secretary and a project steering committee. At the Malawi Social Action Fund, project oversight was assured by an inter-ministerial National Technical Advisory Committee which was responsible for final approval of the farm development subprojects developed in the communities and vetted by district governments. 2.12 Under the project there were twelve steps needed to buy a farm and develop a farming plan (Table 2). Beneficiary eligibility was checked on three occasions (Steps 1, 6 and 8), and the farm development plan was appraised twice (Steps 6 and 12). Smooth implementation of the project presupposed good coordination between the Ministry of Lands and the Malawi Social Action Fund because, while the former was the lead implementing agency, the latter was responsible for disbursing project funds for farm purchase and farm development, for financing (outside the project) the community infrastructure in areas where the farms were established, and for informing the public about what the land reform pilot offered and the conditions for participation. Project effectiveness also depended on the availability—both in the districts and at the center—of technical expertise in land administration, planning and agriculture. 2.13 In each locality, Community Oversight Committees and Project Management Committees were to be set up to promote the formation of beneficiary groups, and help with the acquisition of land and the resettlement process. Community Oversight Committees, made up of villagers openly chosen by a majority of the members of originating and receiving communities, worked with the Group Village Headmen at origin and destination. In the sending communities, these committees were intended to promote the formation of beneficiary groups, and check that would-be beneficiaries met the eligibility criteria (Step 1). (Would-be beneficiaries had to be Malawian citizens, landless or near landless, and with low levels of income.)3 In the receiving communities, the committees were expected to help smooth the integration of the newly-arrived families (Step 12). The beneficiary groups were designated as trusts. Each trust had a name and was required to develop its own constitution and to elect a Project Management Committee, comprising a subset of beneficiaries, which would lead the search for land and negotiate a price with potential sellers (Step 4). Project 2 During implementation, the Ministry of Lands, Physical Planning, and Surveys was renamed the Ministry of Lands, Housing, and Urban Development. 3 The eligibility of the beneficiaries was partly location-specific. The Project Implementation Manual indicated that beneficiaries needed to be both ―poor‖ (according to a predefined list of criteria and the Field Appraisal’s own assessment of local conditions) and ―land poor‖ (meaning either that they had no land, or less land than needed for own subsistence, given the characteristics of the agro-climatic zone in which they were located (World Bank 2004a: 38). 7 Management Committees were intended to continue operating even after the beneficiary group moved to the new community, but would defer to the traditional leaders of that community. The Malawi Social Action Fund was to approve and finance grants to the beneficiary groups, synchronizing this decision with approval and funding of off-farm infrastructure in the receiving community. Table 2: The Farm Acquisition and Development Process Steps Agents Working with Beneficiary Groups 1. Would-be buyers of land (self-selected) form a Beneficiary Group and the Community Oversight Committee Community Oversight Committee checks their socioeconomic status to ensure (―sending community‖). they are eligible to participate in the project. 2. The Beneficiary Group submits its application to the district government. Project District Lands Officer. 3. The government advises the Beneficiary Group what properties are for sale District Lands Officer; Ministry of Lands and how to proceed. Each Beneficiary Group elects its own Project (which maintains Land Data Base); Social Management Committee. Fund (which issues monthly newsletter). 4. The Project Management Committee selects and visits a farm that is for sale, Project management Unit; Sellers; Estate provisionally negotiates a price with the owner or leaseholder within a range agents; NGOs. set by the Project Management Unit, and uses a standard template to prepare an outline farm plan. 5. The seller gives a letter of agreement to the Beneficiary Group, which passes Seller; District Lands Officer. it to district government; public notice is served. 6. The selected farm is appraised from a legal, financial, environmental and District Land Officer leads Field Appraisal social perspective, and the eligibility of the would-be buyers is confirmed. Team. 7. If appraisal is positive, the government in the district where the farm is District Executive Committee (the Social located (the ―receiving district‖) approves the farm development subproject Fund implementing agency), augmented to after evaluating the fairness of the asking price and confirming that the include Traditional Authorities, in the subproject is viable in legal, financial, environmental, and social terms. The receiving district; independent valuers. scope for obtaining (outside the project) Social Fund financing for communal infrastructure is one of the approval criteria. 8. The district where the Beneficiary Group originated (the ―sending district‖) District Executive Committee in the verifies the would-be buyers’ eligibility. sending district. 9. The farm acquisition proposal is forwarded for higher-level review. Project Management Unit, Ministry of Lands, in consultation with the National Technical Advisory Committee, which includes representatives from other line ministries. 10. After high-level approval, the Beneficiary Group signs a grant agreement Social Fund Management Unit. with the Social Fund; the agreement includes detailed farm development plan and disbursement schedule. 11. Farm seller is paid by the Social Fund; the property is transferred, Social Fund; Ministry of Lands; District surveyed, and a group land title is issued by the Ministry of Lands. Lands Officer 12. The farm development plan is re-appraised and, if approved, the Social District Lands Officer; Field Appraisal Fund disburses farm development funds to Beneficiary Group and the Lands Team; Social Fund Management Unit; Officer monitors implementation of the subproject, with oversight of the Community Oversight Committee receiving community; the community becomes eligible for Social Fund (―receiving community‖) infrastructure development (not funded by the project). Source: World Bank 2004a and 2004b “Ministry of Lands” refers to the Ministry of Lands, Physical Planning and Surveys. “Social Fund” refers to the Malawi Social Action Fund. 2.14 Each beneficiary household was to be provided a resettlement package (Step 12). This included a grant in cash of US$1,050 per household, with 30 percent earmarked for buying land, 10 percent to cover the costs of transport and shelter, and 60 percent for farm development. This was not paid out as a lump sum, but in three tranches (40%-30%-30%), with tranche release keyed to pre-defined steps in the agreed farm development plan. The 8 resettlement package also included a starter pack of farm inputs in kind: each household received six bags of fertilizer, hand implements, and maize and tobacco seeds. 2.15 Beneficiary groups were designated as trusts. (The number of households per trust would vary according to the size of the estate acquired; at project closing, the mean was 23 households per trust.) Households in the group were allocated two hectares of land each. The project provided a group land title to each trust (Step 11), which was deemed to provide sufficient land tenure security in the short-term. Families had the option to subsequently obtain individual land titles, providing that they were able to cover the costs themselves. Also, beneficiary groups were not allowed to dispose of their land in the five years after it was allocated to them, and individual households were not permitted to subdivide their holding into parcels smaller than two hectares. DESIGN OF MONITORING AND EVALUATION 2.16 Given the pilot nature of this operation and the need to build a case for eventual scaling up, it was essential to prepare for an impact evaluation at the outset. The appraisal document made clear the commitment to designing a panel study that would create a credible counterfactual and allow for rigorous measurement of the difference made by the project. It was envisaged that evaluation would be carried out annually, starting with a baseline survey, which would be implemented as soon as the grant was made effective. Data would be collected from representative sample surveys of beneficiaries and non-beneficiaries. A management information system would be established to keep track of project inputs and outputs. Qualitative data would be supplied by civil society organizations, using beneficiary assessments and participatory rural appraisals. Allowance was also made for a land market database that would be used to track trends in land prices and the number of hectares offered for sale. 2.17 There were two design shortfalls. First, the designers of the project opted to place M&E in the hands of a stand-alone project management unit whose staff lacked the relevant expertise. Second, by the time the project grant became effective, the design of the management information system and the broader monitoring and evaluation effort had still not been fully specified. This contributed to subsequent implementation delays. Relevance of Design 2.18 The project logic was as follows. To a large extent, rural people were poor because they lacked adequate access to land. Agricultural productivity was low because large estates were underused and small parcels were overworked and therefore subject to erosion and low soil fertility. The project would raise the incomes of poor farmers, first, by giving groups of small farmers grants to buy land and, second, by providing the training and the inputs needed for farmers to develop commercially viable subprojects. The design of the project acknowledged that to increase incomes and agricultural production it was necessary to provide a basic level of land tenure security plus a package of inputs, in addition to land: the resettlement grant that was provided to each beneficiary financed essential farm inputs (particularly fertilizer) as well as giving beneficiaries the cash needed to cover the cost of moving to a new location and building rudimentary shelter. 9 2.17 The project results framework was predicated on the following assumptions:  Good coordination between the Ministry of Lands and the Social Action Fund;  Availability of technical expertise, including agricultural extension;  Sufficient incentive for landowners to be ―willing sellers‖ of their land; and  Feasibility of government actions (for example, on ground rents and land taxes) intended to encourage landowners to sell. 2.19 With respect to the original project development objective, incomes would rise as a result of increased agricultural production: given that beneficiaries would have more land to work than in their previous location, and that the land they were receiving was previously idle or underused, it was realistic to expect that the project would lead to incremental farm output. When the project development objective was revised, the bar was raised: there was a new expectation that agricultural productivity would increase. Given that the project made some (limited) provision for agricultural extension—which included advice on introducing cash crops—and, given that the grant was sufficient to finance improved seeds (hybrid maize) and fertilizer, it was reasonable to expect that an increase in output and incomes per hectare. 2.20 Buyers and sellers were both willing: the beneficiaries who were assisted to buy land were free to choose where they settled; the estate owners were not coerced into selling their land, nor were they subject to outright expropriation—and, moreover, project rules excluded purchase of estates that had been invaded (removing the incentive for preemptive invasion by those seeking land). These considerations, respectively, helped to ensure that private investment in agriculture would not be discouraged, and that land invasions would not be rewarded, further helping to protect the investment climate, which was conducive to attainment of the agricultural productivity objective. A parallel, Bank-supported development policy operation (Fiscal Management and Accelerating Growth, introduced in 2004) complemented the Community-Based Rural Land Development Project, because it contained an agreement that government would introduce ground rents and land taxes, with a view to reducing land speculation and increasing the incentive for estate owners with idle land to sell up (World Bank 2007b). 2.21 The provision of group title to the land provided the minimum tenure security needed to encourage beneficiaries to invest in their holdings. It was appropriate that beneficiaries were given the option to subsequently acquire individual title at their own expense: financing individual titling under the project would have pushed up costs and strained the implementing capacity of the land administration agencies—moreover it was not clear that beneficiaries would feel a need for individual title in the first years after resettlement. 2.22 The project provided farmers with a grant to buy and develop farms, rather than a loan. (In other countries where the Bank financed willing buyer, willing seller land reform— such as Brazil and Guatemala—land purchase was financed through loans that were not part of Bank funding for the project.) In Malawi, there were no financial institutions either interested or able to administer loans for small farmer land purchase and farm development. At the outset at least, it was expected that farm production would be mainly subsistence- oriented, meaning that cash incomes from crops would probably not be sufficient to amortize 10 loans—and the need to repay a loan might compromise attainment of the objective of raising the beneficiary’s income. Moreover, there was a social justice argument for providing a grant: the beneficiaries were descended from Africans who had had their customary land expropriated from them without compensation during the colonial period. 2.23 The project components did not cover the provision of amenities external to the farm plots—roads, schools, health posts and communal drinking water system. Including these elements in the project would have substantially increased the cost and was not appropriate for a pilot of this nature whose primary purpose was to test the viability of a process for acquiring and developing farm land. The project objectives did not promise poverty reduction in the fullest sense—which would have meant improving the education and health status of the beneficiaries—but they committed to boosting incomes. Using the existing social fund to provide coverage outside the project for extra-farm investments was a logical design choice, although it necessarily created risks that coordination between the project implementing agency and the Malawi Social Action Fund might be less seamless than expected. A key consideration is that the project communication strategy did not make false promises to would-be beneficiaries about the amenities that they would have in places where they were resettled, and the project provided funding for beneficiary groups to inspect potential resettlement sites before committing. When it transpired that the adequacy of drinking water supply could not be taken for granted, the project design proved flexible enough to allow for a swift adjustment that permitted the resettlement grant to be partly used to cover installation of shallow wells or boreholes. 2.24 Finally, appropriate to a pilot, the project design made thorough provision for a rigorous evaluation of results, allowing for a panel-based impact evaluation that would compare beneficiaries with a matched control group. Financing was provided for a baseline and follow-up surveys. 2.25 The project design was pioneering. This was the first redistributive land reform project to be implemented in the Africa Region and it was the first in the world to use Bank funds for land acquisition, based on a formal exception to the decades-old policy precluding disbursement against land purchase.4 Although, in the first instance, success would depend on whether the pilot met its objectives, over the longer term what counted was if the project developed a model that could be scaled up in Malawi, and adapted to conditions in other countries where land of arable potential was underemployed. 2.26 The relevance of the design to both the original and the revised objectives was high: the package of components and activities provided by the project was necessary and sufficient to both to increase agricultural production (original objective) and to raise agricultural productivity (revised objective). 4 Paragraph 2(b) of OP 12.00 on Disbursement. The exception was granted by the Managing Director in March 2002 and a Land Committee was established to supervise application (World Bank 2012a: 6-7). 11 3. Implementation 3.1 The project was approved by the Bank’s Board in April 2004 and the grant was made effective in July of that year. At appraisal, it was estimated that implementation would span five years, but ultimately it stretched to over seven, with closing dated September 30, 2011.5 Additional financing, approved in 2009, was used to cover the higher than expected cost of the components and activities in the original design; there was no expansion of project scope. The actual counterpart paid by the Borrower was US$1.1 million, or 73 percent of the planned amount. US$1.8 million was canceled (Annex A). Planned vs. Actual Expenditure 3.2 At project closing, total costs exceeded appraisal expectations by 27 percent (Table 3). The cost increment was funded through an additional grant that was approved by the Bank in November 2009. The extra cost derived from facilitation of farm development (including beneficiary selection and training, plus surveying and boundary demarcation), rather than from land purchase. Table 3: Planned vs. Actual Expenditure by Component COMPONENTS COST (US$ millions) (1) (2) Expecteda Actual (2)/(1) (%) 1. Land Acquisition and Farm 16.5 16.4 99 Development 2. Land Administration 3.0 6.5 217 3. Capacity Building 3.4 5.2 153 4. Project Management, Monitoring 6.9 9.8 142 and Evaluation TOTAL 29.8 37.9 127 Source: World Bank 2004a; World Bank 2004b; World Bank 2012. a. Original Costs, as estimated at appraisal. In November 2009, the Bank approved Additional Financing of US$10.00 million, distributed as follows: US$1.23 million to Component 1, US$4.0 million to Component 2, US$1.93 million to Component 3, and US$2.84 million to Component 4. Implementation experience 3.3 Progress was slow at first, owing to delays in recruiting staff and setting up the management information system, and the limited coverage of public information campaigns that were needed to prime the buying and selling of land under the project. Initially, potential sellers of idle or underused estates were skeptical that government had the means to finance 5 In May 2009, the closing date was extended by six months to December 31, 2009, giving the time needed to prepare additional financing. When the additional financing was approved, the closing date was extended for a second time to September 30, 2011. 12 land purchase, and speculated that the project was a cover for involuntary land acquisition. Implementation picked up after November 2005, when the first four beneficiary groups settled on new farms purchased under the project. The project management unit was moved from Lilongwe to Blantyre in May 2006, bringing it closer to the district governments and beneficiary groups involved in the project. 3.4 By February 28, 2007, a month before the midterm review mission, only 2,948 beneficiary households had been relocated, and it was projected that only 10,000 families would be resettled by closing. Along with the Board’s approval of additional financing in October 2009, changes were made to the project development objective, new indicators and targets were added, and more funds were allocated to the land administration component. The additional resources were also used to fund resettlement of 1,000 families—to ensure achievement of the appraisal target of benefiting 15,000 poor families. Additional financing of US$10.0 million increased by about one-third the funding provision at appraisal, and was declared effective on April 29, 2010. 3.5 By the midterm review in March 2007 it had become clear that, contrary to expectations, no land was being offered for sale in Mulanje and Thyolo. (A land availability study was not completed before the project was approved; such a study was only carried out in 2007, in preparation for the midterm review.) The grant agreement was subsequently amended to include two additional districts, Balaka and Ntcheu where, according to the land availability study, there were good prospects of estates being offered for sale. 3.6 Even with the two added districts the overall availability of land for sale did not increase as much as expected because the government chose not to lift the moratorium on the collection of ground rent and land tax, and decided not to cancel expired leases.6 This reduced the incentive for the owners of underused land to sell up. Also, the government was not able to commit to the project the 4,000 ha of land that it had originally pledged. (The government had bought idle estate land in Mangochi and Machinga districts, with the intention of parceling it out. Local chiefs responded by encouraging people to invade. Project rules excluded invaded land from inclusion in the redistribution program.) The government offset this loss by increasing counterpart funding for land acquisition and farm development by US$0.6 million. 3.7 Throughout the project, the biggest obstacle to implementation was the weak capacity of land administration: it was difficult to conduct timely searches in the land registries owing to the disarray in the land records held by district and central government; surveying capacity was limited; and land titles were issued late. The additional financing, approved in October 2009, was mainly used to strengthen the framework of land administration, through legal reform, capacity building of central and district land agencies and an assessment of land tax and ground rent collection (Table 4). This change was prompted by the slow progress in granting title to beneficiary groups. The additional financing also provided the funds needed 6 The government complied with the land policy tranche release conditions attached to the Fiscal Management and Accelerating Growth Program (2004-2006), conditions that were consistent with the objectives of the project assessed here (World Bank 2007b). However, after FMAGP closed, the government backtracked. 13 to resettle 1,000 households, thus making it possible to achieve the original target of benefiting 15,000 families. Table 4: New Land Administration Outputs under Additional Financing ORIGINAL PROJECT ADDITIONAL FINANCING “Land Administration” (Component 2) “Land Administration” (Component 2) ―Strengthening the Recipient’s land ―Strengthening the Recipient’s land administration system through the development administration system, including the of the capacity of the Ministry of Land Physical applicable legal framework, through the: (i) Planning and Surveys at both the national and development of the capacity of the Ministry of the local government levels to assist the Land Housing and Urban Development at Beneficiary Groups in the purchase and transfer both the national and the local government of land acquired under the Project, through the levels to assist the Beneficiary Groups in the provision of training, acquisition of equipment purchase and transfer of land acquired under and vehicles, and Operating Costs‖ the Project; (ii) enhancement of district, regional, and national land registries through (World Bank 2004b: 19). capacity building of the Department of Lands, the Department of Physical Planning, and the Surveyor General’s Department in surveying, registration, and titling services, and land information technology; and (iii) assessment of issues relating to ground rent collection, freehold land taxation, and finalization of the Recipient’s new land law, and identification of possible solutions, all through the acquisition of information and land technology and office equipment and vehicles, provision of technical advisory services and training, and the financing of Operating Costs‖ (World Bank 2010: 18). Italics added, to emphasize difference between original project and additional financing. 3.8 Project implementation was hampered by the difficulty of synchronizing land purchases with the delivery (outside the project) of community infrastructure funded through the Malawi Social Action Fund. Also, the project design initially failed to provide for an adequate supply of drinking water. However, once this problem was identified, some of the first tranche of the grant (for Component 1) was swiftly made available for financing boreholes and protected shallow wells and, by the mid-term review, 52 beneficiary groups had obtained secure access to potable water in this manner.7 7 IEG was unable to verify what percentage of beneficiary groups had problems obtaining drinking water, and what proportion of these were the subject of remedial action under the project. 14 Implementation of Monitoring and Evaluation 3.9 There were delays in setting up the management information system. On the other hand, although there were delays and some gaps in executing the surveys and building the panel, a solid basis was laid for impact evaluation. The baseline survey was administered in 2006, after the (later than expected) relocation of the first beneficiary groups. Follow-up surveys were conducted in 2007, 2008 and 2009. A 2009 impact evaluation study, conducted under IEG auspices, compared survey results from the 2006 baseline with those from 2008 (Datar, del Carpio and Hoffman 2009). A final impact evaluation compared results over two time periods, 2006/2007 and 2006/2009 (Simtowe, Mariapia, and Mangisoni 2011). 3.10 Part of the information obtained from the surveys—16 percent of items in 2006, 29 percent in 2007, and 17 percent in 2008—was based on respondent recall, which may have reduced measurement accuracy. Thus, for the baseline survey, all information on agricultural production and livestock holdings referred to the previous agricultural cycle. Data on the number of meals taken in the ―lean season‖ (before crops are ready to harvest) also required respondents to think back to the time before beneficiaries were resettled. IEG observed that follow-up with households after the baseline survey was not as thorough as it could have been: the 2009 study had to top up the number of respondents by randomly sampling households from the baseline that had been omitted from the initial re-survey. 3.11 But, overall, the impact evaluation work was thorough. Careful thought was given to the design of the panel and, in particular, the specification of three separate control groups— non-beneficiary households from the same villages of origin as the beneficiaries, non- beneficiary households in the villages of destination of beneficiaries, and non-beneficiary households in similar areas of neighboring districts. Comparisons were based on double- difference analysis, with propensity score matching. Two papers, separately authored, and based on different reference periods (2006/2008 in one, 2006/2007 and 2006/2009 in the other) reached broadly similar conclusions about the difference made by the pilot. 3.12 Reporting on project inputs and outputs was less well handled. The management information system was still not in place two years after startup, impeding the flow of information from the field to the project management unit. Safeguards and Fiduciary Compliance 3.13 This was a Category B project to which two of the Bank’s safeguard policies applied (OP 4.01 Environmental Assessment; and OP 4.09 Pest Management), based on the risks identified (Box 1). An environment and social analysis and a management plan were prepared, covering baseline conditions in the four pilot districts selected at appraisal (Machinga, Mangochi, Mulanje and Thyolo) and addressing potential problems arising from farm acquisition and development. Because one of the aims of the project was to encourage farmers to use more farm inputs a pest management plan was developed, seeking to prepare farmers to use integrated pest management techniques and to exclude the use of banned agrochemicals (Government of Malawi 2003). 15 3.14 The pest management safeguard was satisfactorily complied with. Compliance with the environmental assessment safeguard was rated moderately satisfactory at the time of the last supervision mission. Although project rules stipulated that there be a 5-10 km buffer zone between beneficiary settlements and protected areas, the limited capacity of the Department of Forestry and National Parks and Wildlife resulted in some enforcement lapses that were addressed during project implementation (World Bank 2012: 11). However, during its visit to Malawi, IEG found no evidence of safeguard violations. Also, the independent impact evaluation report comments on the sound land management associated with the subprojects: ―most beneficiary groups visited were making ridges across the slope, constructing box ridges and contour bunds and in some instances planting vetiver grass to control soil erosion‖ (Simtowe, Mendola and Mangisoni 2011: x). Box 1: Environmental and Social Risks Environmental  Acquisition of farms that may encroach on protected areas, wetlands or other sensitive ecosystems.  Location of farms in areas with insufficient drinking water, or poor sanitation.  Increased use of agro-chemicals, which would be part of the project starter pack for newly-settled farms, and whose use would be encouraged subsequently through improved incomes and access to information and agricultural extension services.  Land use or cropping practices that may lead to soil degradation, loss of biodiversity, or poor water management. Social  Acquisition of land by persons from other districts, of from other ethnic groups, may create tension with the receiving community. Source: Government of Malawi 2003. 3.15 Although the project explicitly involved resettlement, the OP 4.12 Involuntary Resettlement safeguard did not apply because there was no involuntary land acquisition or any form of restriction on livelihoods. Farms purchased from the market had to be free of any legal or financial claims and farms with squatter encroachment, or that were subject to unresolved labor disputes, were not eligible for purchase under the project. 3.16 For the most part, fiduciary matters were satisfactorily handled. Although there were occasional slight delays in the submission of audit reports, all reports were unqualified. There were also some delays in procurement. Procurement was rated moderately satisfactory to satisfactory throughout the implementation of the project. 4. Achievement of the Objectives 4.1 The original objective was to increase the incomes of about 15,000 poor rural families, while the revised objectives were to both increase the incomes and the agricultural productivity of the families. The same set of outputs and intermediate outcomes bore on achievement of both income and agricultural productivity increases. Therefore, this chapter begins by assessing outputs and intermediate outcomes and then proceeds with separate sections on the income and agricultural productivity outcomes, assessing the evidence that it was the project’s activities that led to these outcomes. 16 Outputs and Intermediate Outcomes 4.2 Would-be beneficiary groups had to submit farm development proposals and only when these were approved did they become eligible to receive land under the project. By project close, 942 proposals had been received and reviewed, and 666 were ultimately approved—representing the number of beneficiary groups that received land. The number of beneficiary groups (666) was equivalent to 147 percent of the appraisal target. There were more beneficiary groups than expected because, when the project was prepared it was planned that each group would be composed of 30-35 households, a requirement that was subsequently relaxed, allowing each group to range in size from 10 to 35 households. 4.3 The beneficiary groups comprised 15,142 households (101 percent of the appraisal target). The total area of land distributed to these households was 33,428 hectares, (also 101 percent of the appraisal target). 4.4 In all, 461 estates were acquired under the project, this land being divided into 666 parcels, one for each beneficiary group. Each of these parcels was surveyed by physical planners and temporary agricultural extension technicians, who were responsible for subdividing the parcels and demarcating the boundaries of the holdings allocated to each household. Each household received a plot of 2.2 ha on which to farm and build a house; before the project their land holding averaged less than 0.5 hectare. 4.5 In most cases, the boundaries of the estates that were acquired had not been adequately mapped beforehand, requiring the project to undertake extensive survey work. The project produced 615 sketch maps. To carry out the surveys, the project procured hand- held Geographical Positioning System (GPS) devices and software for verifying and approving deed plans. The project trained surveyors and, to a limited extent, hired firms to carry out surveying, helping to build private sector capacity—17 of the 666 parcels were surveyed by firms. 4.6 The number of staff trained in surveying and registration of land parcels including monitoring and collection of land tax/rent reached 72, compared to target of 30; of these, 13 received higher-level training (Master’s degree or Diploma). As expected, 32 Land Registries were computerized and were fully operational when the project closed. The project provided survey and office equipment to the lands, surveys and physical planning departments in the Ministry of Lands, enabling faster processing of land surveys and registration of titles. Under the project, government agencies streamlined the procedures for verifying deed plans, reducing the time for processing property transfers from nine months to four months. 4.7 Under the project, the number of days taken to register a land transfer declined from the 2008 baseline of 118 to 90 (the target was 110). Of the 666 beneficiary groups, 641 had received group land titles before closing. 4.8 Project funds were used to cover land acquisition and individual household requirements relating to relocation costs, including the construction of shelter and the provision of basic amenities such as sanitation facilities. There were adequate funds for beneficiary groups to visit distant properties that were of potential interest to them, before 17 any purchase was sealed. The project also financed the purchase of farm inputs and tools, provision of technical assistance for training in farm and production management, productivity improvement, extension services for agricultural technology innovation and produce marketing.8 As prescribed in the design, during project implementation each beneficiary received a uniform grant of US$1,050 for land acquisition and farm development, with 30 percent devoted to land acquisition, 8 percent to cover settlement costs and 62 percent to farm development. This grant was provided in the first year of settlement and households had in subsequent years to find their own resources for developing the land. 4.9 The eligible communities applied to the Malawi Social Action Fund for the provision of community assets such as boreholes, access roads, schools and clinics, in line with the Fund’s principles and criteria. However, actual investments from this Fund—which occurred outside the project—were reportedly lower than expected, although IEG was unable to verify their magnitude. In the case of drinking water supply, most of the infrastructure installed was financed by the beneficiaries themselves, using part of the initial settlement grant that the project provided. Beneficiaries financed 60 boreholes, 189 shallow wells and 31 piped water installations. In addition, 42 boreholes were financed outside the project, mainly by charities. 4.10 To raise awareness of, and build demand for, the resettlement process, the project produced and distributed a large range of information, education and communication materials. Sensitization meetings were held, and radio and television programs were aired, explaining project objectives and how to participate. Dissemination was targeted at central and district governments, at landowners, and across the communities in the six participating districts. In addition to information about the project process and technical guidance on crop and livestock production, materials covered a variety of health topics, including testing and treatment for HIV/AIDS. Meetings were organized on the adverse effects of corrupt practices, in collaboration with the Anti-Corruption Bureau in Machinga and Mangochi, covering 26 Traditional Authorities. 4.11 As expected when the project was restructured, the procedures for collecting ground rent and land taxes were reviewed and, in conjunction with the Malawi Law Commission, the project sponsored a public forum for sharing the results of the review. A new graduated land rent structure was prepared and is now implemented on all leased land, including tobacco farms. The project also supported the drafting of a land bill, including extensive consultations, but its enactment is still pending. Increase the incomes of about 15,000 poor rural families (original and revised objective) 4.12 Project monitoring reports show that by closing monthly incomes averaged MKw 30,500 per family, almost three times higher than the target of MKw 11,330, which was set 8 Extension officers were hired by the project and provided assistance that ranged from training beneficiary households in negotiating land prices to drafting farm development and input expenditure plans. Extension officers also helped link farmers to non-government organizations (NGOs), service providers, agro-dealers and other long-term extension providers. 18 when the additional financing was approved in 2008. The baseline income in the communities where the beneficiaries originated (also referring to 2008) MKw 4,530.9 4.13 Evidence from three separate surveys comparing beneficiaries and non-beneficiaries gives a fairly consistent impression of the project’s impact on incomes (Table 5). Two of the surveys were formal impact evaluations that controlled for selection bias and made robust use of controls to estimate the difference made by the project.10 Both impact evaluations— one of which (Datar, del Carpio and Hoffman 2009) was produced by IEG—found that farm income increased more for beneficiaries than non-beneficiaries, by a sizeable magnitude and in a statistically significant way (Table 5; Annex B, Tables B9 and B10). One of these surveys also reports on total income, finding no significant difference between beneficiaries and non-beneficiaries. The authors say that this was the result of substitution among income sources after resettlement: with more land to work with, beneficiaries no longer needed to hire out their labor, deriving more of their income from the farm. Datar, del Carpio and Hoffman (2009) found that households that relocated at great distances experienced systematically lower increases in income than households that stayed within their district of origin because they had to adapt to unfamiliar agroecological, cultural, and market environments. 4.14 The two surveys that measured economic well-being (as reported by the respondents) both showed that beneficiaries perceived themselves to be better off than non-beneficiaries, with the difference being both large and statistically significant. However, one of these studies (Chirwa 2008) was less robust because it did not control for selection bias. 9 2008 is the baseline used by project monitoring and cited in the completion report; for the purposes of the impact evaluation the baseline was 2006, the date of the first survey. 10 The 2006 survey was designed and implemented by a research firm under the direct supervision of the World Bank. The baseline survey was administered to 857 households immediately following their relocation. Fifty beneficiary groups were selected, and 5-10 households from each of these were surveyed. Households were asked about their assets, agricultural production during the 2005-2006 season, current activities, and average (monthly and yearly) expenditures on various categories of goods. The same survey was administered to households in districts untouched by the program in the same time period, both at baseline and at follow-up. The first impact evaluation (by IEG) involved a follow-up survey in 2008. The second impact evaluation (Simtowe, Mendola and Mangisoni 2011) used the same baseline date but followed up in 2007 and 2009; it was also based on a larger sample of households than the earlier study (Table 5). 19 Table 5: Survey Evidence on Project Outcomes Author Survey date(s) Treatment Findings on Productivity Findings on Income and Well-being Datar, del 2006 baseline Panel, In 2006-2008, maize yields In 2006-2008, farm income Carpio and 2008 follow-up matched were137 kg/ha less than non- increased more for beneficiaries Hoffman, No. of households: beneficiaries beneficiaries, but the difference than non-beneficiaries and in a 2009 N=245 maize and non- was not statistically significant. statistically significant way N=207farm beneficiaries (p<0.05). Though subjective, 14% income (difference in more beneficiaries than non- N=444 self- difference, beneficiaries reported improved perceived income with self-perceived income and 20% N=443 self- propensity more beneficiaries reported perceived economic score improved economic well-being, well-being matching*) differences that are statistically significant (p<0.05). Simtowe, 2006 baseline Panel, In 2006-2007, the statistically In 2006-2007 and 2006-2009, Mendola 2007 follow-up matched significant gains in maize yields farm income increased more for and 2009 follow-up beneficiaries were 1000kg/ha higher for beneficiaries than non- Mangisoni, No. of households: and non- beneficiaries compared to non- beneficiaries and in a statistically 2011 N=1507 maize beneficiaries beneficiaries (p<0.01) In 2006- significant way (p<0.05): those in N= 176 tobacco (difference in 2009, the same observation the program saw a 66.5% increase N=1636 farm difference, holds but the gap between over their counterparts in 2007 and income with beneficiary and non-beneficiary farm income gains 93.5% higher in N=2995 total propensity yields narrowed to 354kg/ha and 2009. There was no statistically income score the difference was only significant change in total income matching) marginally statistically for either period. significant (p<0.1). The increase in tobacco yields was higher for beneficiaries compared to non- beneficiaries for both periods (3.5T/ha in 2007 and 2.7T/ha in 2009), and the difference was highly statistically significant in each case (p<0.01) Chirwa, 2007 Unmatched In 2007, maize yields were In terms of self-perceived 2008 (N=146 households) comparison of statistically significantly higher economic well-being, beneficiaries for beneficiaries compared to beneficiaries experienced more and non- non-beneficiaries (p<0.01). improvement than non- beneficiaries However, new beneficiaries with beneficiaries, both in relation to (t-test of only one season of farming one year ago and two years ago, significant under the program invested with comparisons for both periods difference), more in hybrid maize and were highly statistically significant using recall to more productive than those who (p<0.01). capture had been in the program for two change over seasons. time Source: Annex B, Tables B9, B10, B11. For full details, see Datar, del Carpio and Hoffman 2009, Simtowe, Mendola and Mangisoni 2011, and Chirwa, 2008. * Datar and others 2009 use a single-difference in the dependent variable for farm income. 4.15 These findings on economic well-being were echoed in IEG’s interviews with beneficiaries and non-beneficiaries (30 in all) in six randomly-selected villages (Annex B, Tables B1 and B8). All of those beneficiaries who were asked reported that they were better off than five years previously (that is, before resettlement); and 86 percent of the beneficiaries who were asked perceived themselves as better off than non-beneficiaries in both the sending and the receiving communities (Annex B, Table B8). When the same 20 questions were put to non-beneficiaries, the response pattern was less clear-cut but a majority of the non-beneficiaries that responded said both that they were worse off today than they had been five years previously, and were worse off today than the beneficiaries with whom they had had contact. IEG repeatedly heard the statement, ―beneficiaries are better off because they have more land than non-beneficiaries.‖ Indeed, each of the three surveys cited above reported that there was a highly statistically significant difference in land holdings of beneficiaries and non-beneficiaries (Annex B, Table B9, B10 and B11). It is the access to more land, rather than higher output per hectare, which drives the income and differences between the two groups. 4.16 The development impact of the increased incomes attributable to the project was enhanced by the rigorous selection procedure which ensured that the poor were targeted. A report by PricewaterhouseCoopers (reported in Simtowe, Mendola and Mangisoni 2011) found that beneficiaries were generally poorer than those that remained behind. A 2005 report from the International Institute for Research and Development (also reported in Simtowe, Mendola and Mangisoni 2011) found that the project adhered to the beneficiary selection criteria stipulated in the project implementation plan. The typical beneficiary households were self-selected, organized groups of landless or land-poor, defined and identified by an expressed need for land and willingness to move and cultivate on newly acquired land. The most vulnerable and disadvantaged groups such as women, orphans and poor displaced farm workers were encouraged to participate (although, as noted above, fewer women than expected stayed the course through to the point of resettlement). 4.17 Female-headed households were under-represented accounting for 20 percent of the beneficiary households while, in the communities of origin, these households averaged 30 percent of all households. The shortfall in participation was attributed to: the lack of amenities (health posts, schools, markets, potable water supply, maize mills) in the communities where beneficiaries were scheduled to relocate; the burden of clearing land for cultivation when there were no men in the household to carry out the work; fear of settling in an unfamiliar environment; and reluctance to give up land in the community of origin (Kishindo and Mvula 2008, quoted in Simtowe, Mendola and Mangisoni, 2011). Some of the non-beneficiary women interviewed by IEG expressed reservations along these lines; but other female heads of household who had resettled said that their initial trepidation had vanished and was more than offset by the improvement in their lives since they had relocated. 4.18 IEG rates the achievement of the objective to increase income as high. Increase the agricultural productivity of approximately 15,000 poor rural families (revised objective) 4.19 Project monitoring reports used crop yields as the measure of agricultural productivity. Maize yields increased from the 2008 baseline of 450 kg/ha in the settlement area to reach 1,800 kg/ha by closing, exceeding the target of 1,500 kg/ha. The increase in tobacco yields fell somewhat short of expectations, rising from the 2008 baseline of 300 kg/ha in the settlement area to 800 kg/ha by closing (the target was 1,000 kg/ha). The impact 21 of the project on agricultural productivity might have been greater if there had been a viable agricultural extension service to draw on. The government service was run down and could not be relied upon—a systemic failing that a pilot project of this nature could not realistically rectify. The pilot dealt with this constraint by recruiting five temporary agricultural technicians, who advised beneficiaries about yield-enhancing farm techniques and diversification into higher-yielding crops. 4.20 The three studies whose results are summarized in Table 4 above also reported on crop productivity impacts. On maize productivity, two of the studies produced results. Datar, del Carpio and Hoffman (2009) found no statistically significant difference between beneficiaries and non-beneficiaries in the receiving area. Simtowe, Mendola and Mangisoni (2011) found that, for the season immediately after settlement, beneficiary yields exceeded those of non-beneficiaries in the receiving area by a substantial magnitude and the difference was highly statistically significant. However, the yield difference narrowed from an agronomically appreciable 1000 kg/ha (comparing 2006 to 2007) to a smaller but still appreciable 354 kg/ha (comparing 2006 to 2009). This reflects the effect of the one-time boost that new settlers received from the starter pack, a mix of cash and inputs delivered by the project. For example, the starter pack helped newcomers to invest in hybrid maize for the first season, something that they were less likely to do in subsequent seasons (Chirwa 2008). 4.21 Only one survey has data on tobacco productivity (Simtowe, Mendola, and Mangisoni 2011) and this shows that, both in the short-term and the medium-term, beneficiaries obtained significantly higher yields than non-beneficiaries. Among beneficiaries, there were relatively few tobacco growers (while everyone grew maize) because input costs were high enough to present more of a barrier to entry: the beneficiaries who planted tobacco were probably better off than the beneficiaries who did not. 4.22 These findings were echoed in IEG’s interviews with beneficiaries and non- beneficiaries (30 persons in all) in six randomly-selected villages (Annex B, Tables B1 and B8). Maize yields were higher in the first year after settlement, mainly because the starter pack provided for more fertilizer to be applied than in subsequent years. Neither the cash income generated by beneficiaries after the first year, nor the access to the government’s fertilizer subsidy program, proved sufficient to match the one-time boost received from the starter pack. However, even if productivity fell after the first year, medium-term yields were still higher relative to the period before settlement when estate land was largely idle. 4.23 Raising productivity entails not just increasing the volume of crops obtained from each hectare but increasing income per hectare by introducing more profitable crops. In the project context, this meant diversifying away from the staples (primarily maize) into cash crops, such as burley tobacco.11 A study, referring to the 2009-10 agricultural year, compared a sample of 168 project beneficiaries to 77 non-beneficiaries in the same area, assessing differences in the returns to tobacco production (Ng’ong’ola 2011).12 The four project 11 Burley tobacco is a light, air-dried (as opposed to fire-dried) tobacco mainly used for cigarette production. 12 This survey is less robust than the impact evaluations reported previously because it did not control for selection bias. 22 districts are a prime area for burley tobacco. The mean yield for beneficiaries in this sample (1,447 kg/ha) was higher than the beneficiary average cited in the completion report (800 kg/ha) and also higher than the national mean for smallholder tobacco cultivation (1,046 kg/ha). However, the yields achieved in the project area were still well below the potential estimated by the Ministry of Agriculture—4,000 kg/ha. 4.24 The study assessed the difference made by contract farming: 43 percent of the project beneficiaries and 60 percent of the non-beneficiaries cultivated tobacco under these terms— an arrangement where farmers accept support (credit, technical advice, and farm inputs) from a marketing company under a contract which requires them to sell all of their output of the given crop to the company, at a price more or less set at the company’s discretion. Smallholder burley tobacco farmers in the project area entered into contracts with three companies. Only 2 percent of the contract farmers in the project area had a say in drawing up the terms of the contract, companies preferring to deal with farmer clubs, not individuals. Although the terms of the contract varied between the three companies, overall, farmers were better off not entering into such contracts—even though farmers with a contract received a higher price for their tobacco than those without a contract, this was more than offset by the onerous terms for repaying the assistance received from the company. Farmers that benefited from the land reform project—as well as those that did not—received a higher net income per hectare without a contract. 4.25 Comparing project beneficiaries and non-beneficiaries with a contract, the latter had a net income from tobacco that was roughly seven times higher, because company affiliation varied between the two groups, with some companies offering terms more favorable to the farmer. The contract farmers who were project beneficiaries obtained a somewhat higher crop volume per hectare, with input and hired labor costs that were similar to those of non- beneficiaries. The difference between the two groups was accounted for by the much higher loan repayments and other dues that the beneficiaries owed to the marketing company and, to a smaller extent, their bigger post-harvest outlay (grading, packing, storage, transport and loading). Comparing project beneficiaries and non-beneficiaries without a contract, the net income per hectare received by the latter was about twice as high, because output volume was higher even though non-beneficiaries incurred lower input and labor costs, and because they obtained a somewhat higher price (Annex B, Table B12). 4.26 The farmers in the sample were not matched so it was not possible to determine to what extent the different outcomes were driven by contract/non-contract and project/non- project differences—variations in soil fertility, water and farming skills were not controlled for. However, project beneficiaries who took up contract farming received less beneficial contracts—two-thirds of them were signed up with one company, whereas two-thirds of non- beneficiaries were signed up with another company that offered better terms (Ng’ong’ola 2011: 12). Whether or not they were contract farmers, beneficiaries may have been more subject than non-beneficiaries to two constraints caused by remoteness from markets—lower prices and higher transport costs. 4.27 IEG rates the achievement of the objective to increase agricultural productivity as substantial. On the one hand, the evidence of sustained yield increase and of diversification 23 into more profitable crops is still limited, and one of the expected avenues for productivity increases (contract farming) has not yet lived up to expectations. On the other hand, average yields obtained before project closing were still higher relative to the period before settlement when estate land was largely idle. 5. Efficiency 5.1 The actual cost per beneficiary family of the project’s land acquisition and farm development component was US$1,083, 12 percent more than the appraisal estimate. The project directly benefited 15,142 families, slightly more than the appraisal forecast of 15,000. Taking project costs as a whole, the actual cost per family was US$2,543 or 28 percent more than the appraisal estimate. This reflects the larger-than-expected overhead costs arising from capacity building and project management: in particular, vehicle running expenses, per diems and other travel expenses were twice as high as initially estimated. On the other hand, the component that financed land titling and registration was only 1 percent more than initial estimates. For M&E specifically, there was a highly efficient use of project resources. Given that this was a pilot project, M&E was vital for generating the data needed to determine the merits of scaling-up land redistribution. Actual spending on M&E was US$417,796, or just 49 percent of the planned budget (Simtowe, Mendola, and Mangisoni 2011: 49). Yet the quantity and quality of the outputs—the baseline survey and three follow-ups, plus monitoring reports each quarter—was substantial. Scaling up was not an objective of this pilot project but if it does ultimately happen it is reasonable to assume that overheads would be spread, reducing the total project cost per family benefited. 5.2 Project design was efficiency-enhancing, in the sense that beneficiary groups had an incentive to bargain over the price of the estates that were offered for sale: the amount of the grant for each family was fixed at US$1,050, so that the lower the price paid for land the more funds would be left over for the family to spend on farm inputs and resettlement costs. There is evidence of serious bargaining over the land price. In Machinga and Mangochi districts, the final price paid averaged respectively 77 percent and 78 percent of the asking price (Annex B, Table B7). Between project startup and closing, land prices did increase in these two districts, an outcome that may have been partly driven by the injection of project funds into a tight land market. But the price increase was small enough that it could be accommodated within the notional upper limit that the project established for the share of the grant that would be used for land purchase. Also, project design was sufficiently flexible that, once the limited access to safe drinking water became apparent, the farm development component was swiftly adjusted to permit use of these funds for installing shallow wells and boreholes. However, there was an under-supply of complementary off-farm infrastructure through the Malawi Social Action Fund owing to the difficulties of synchronizing land acquisition under the project with the budget cycle of the Fund. However, insofar as this refers to funds outside the project it does not reflect on the efficiency with which project resources were used. 5.3 Benefit-cost analysis. The impact evaluation conducted before closing included financial and economic analysis that was based on representative data about farming 24 performance derived from the panel surveys. IEG was not able to locate the spreadsheets used at appraisal so that, in the absence of detailed farm models, it was not possible to re- estimate the rate of return. The project team’s analysis at appraisal (repeated at mid-term and closing) was based on conservative assumptions: about the proportion of beneficiaries that would add cash crops to staple cultivation (20-30 percent); about crop yield increases (based on observations from a representative sample rather than extrapolations from a few cases); about the land price (US$175/ha—compared to a mean price of US$116 in Mangochi in 2007); and about the ―without project‖ income of beneficiaries (US$100 per family per year). However, at MKw 10/day—apparently the rate paid to casual day laborers in remote rural areas—labor may have been undervalued in the analysis. In 2009/10, the government minimum wage was MKw 107/day and the gross margin for burley tobacco averaged MKw 152/day (Ng’ong’ola 2011: 47-48). Even allowing for the remoteness of the project sites from labor markets, the value placed on labor appears low. 5.4 Assuming that only 20 percent of beneficiaries cultivated cash crops, the economic rate of return was estimated at 20 percent, above the rate of 15 percent forecast at appraisal. This rate of return does not include certain plausible benefits that were difficult to quantify: particularly the land improvements and asset accumulation that may flow from increased security of tenure and the reduction in social tensions resulting from the project. 5.5 IEG rates efficiency as substantial, against both the original and the revised objective. 6. Ratings Outcome 6.1 Project outcome is rated satisfactory. The relevance of the objectives was underscored by the particular circumstances of Malawi: there was an urgent need to increase the incomes and productivity of the rural poor, and doing so through land redistribution responded appropriately to the poor’s limited access to land. Relevance was substantial for the original objective and high for the revised objectives: because the latter emphasized increasing farm productivity in addition to increasing farm output it was more closely in line with the strategies of the government and of the Bank. Achievement of the objective to raise income was high, because beneficiaries ended up with more farm land than non-beneficiaries and were able to obtain a larger increase in farm output volume, which in turn led to higher incomes. Achievement of the objective to increase farm productivity is rated substantial: in terms of productivity (measured as farm yields and net incomes per hectare), beneficiaries still performed better than non-beneficiaries, but by a smaller margin than for output volume. Efficiency was rated substantial because, based on plausible assumptions, the economic rate of return exceeded expectations: assuming that only 20 percent of beneficiaries cultivated cash crops, the economic rate of return was estimated at 20 percent, above the rate of 15 percent forecast at appraisal. Also, project design created good incentives for beneficiaries to use the farm development grant efficiently and project funds were well used in monitoring and evaluation, generating the learning that is expected of a pilot project. The flexibility of 25 project design and the quality of the panel study relative to costs support the rating of efficiency as substantial. Aggregating the sub-ratings against the original and the revised objective produces a satisfactory outcome rating in both cases (Table 6). Table 6: Derivation of the Outcome Rating Project Development Objective OUTCOME SUB-RATINGS ORIGINAL REVISED (INCOME ONLY) (INCOME AND PRODUCTIVITY) Relevance of Objectives Substantial High Relevance of Design High High Achievement of Objectives High (income) High (income); Substantial (productivity) Efficiency Substantial Substantial OUTCOME SATISFACTORY SATISFACTORY Risk to Development Outcome 6.2 There are two ways to assess risk. First, taking a limited interpretation of risk, the issue is whether prospects for the beneficiaries resettled under the pilot project will continue to be bright enough for them to remain in place and to make the investments needed to sustain productivity increases. Second, although the statement of project objectives does not refer to broader aims beyond improving the incomes of 15,000 poor, rural families, the appraisal document makes it clear that the purpose of the pilot was to help make a case for scaling-up this approach to land redistribution. Therefore, a broader interpretation of risk would emphasize the threat that scaling-up would not be realized, and that the valuable lessons from the pilot would not be built on. (No less important is the risk that scaling up proceeds—but without any adjustment of project design to reflect lessons learned.) 6.3 To begin with the limited interpretation of risk, as defined above, there are several reasons why project beneficiaries may fail to thrive. The survey evidence suggests that it was the provision of a resettlement grant that gave the biggest boost to incomes and productivity: once that one-time support was exhausted, yields dropped (although outcomes were still better compared to before/without project). Most of the beneficiaries who withdrew from the project did so when the resettlement grant was exhausted (Annex B, Table B5). Chirwa (2008) found that new beneficiaries with only one season of farming under the program tended to invest more in hybrid maize and were more productive than those that have been under the program for two seasons. 6.4 If beneficiaries are to achieve sustained higher agricultural productivity they will need secure access to fertilizer, technical assistance, and credit. All of the farmers interviewed by IEG were, in principle, covered by the government’s ambitious fertilizer voucher program: although they could not draw on this in the same year that they received the project starter pack, IEG found that most had obtained access a year or two later. However, there are not enough vouchers to satisfy demand and most of the beneficiaries interviewed said that they needed to supplement the vouchers with cash purchases. With respect to agricultural extension, the project had to improvise, with short-term hires 26 substituting for full-time staff: this is not sustainable. (The FY08 Agriculture Sector-Wide Approach Support Project [P105256], which the Bank is now implementing, with additional financing approved in 2012, is making partial amends for the extension deficit, by providing resettled households with technical services, including marketing advice, and by facilitating entry to contract farming.) People in the project area have almost no access to credit. Only 3 of the 30 farmers interviewed by IEG had taken a loan in the previous five years. Contract farming offers the surest path to obtaining a loan but, as shown in a previous section, terms vary substantially between the companies in the project area, and farmers who take loans under these arrangements may be left with repayments that leave them worse off than those outside the scheme. 6.5 Finally, most of the beneficiaries were resettled in remote locations where drinking water sources, schools, health posts, and markets were all further away compared to their previous location. The average distance to these amenities is shown in Annex B, Table B6. Among beneficiaries who withdrew from the program, 19 percent said that the lack of infrastructure was a primary reason for doing so (Annex B, Table B4). The anticipated coordination with the Malawi Social Action Fund was more limited than expected and there was little complementary investment in community infrastructure at the resettlement sites. 6.6 The lack of community infrastructure may lead some beneficiaries to return to their place of origin. And yet it is striking just how few beneficiaries have returned so far. A 2008 study by Mkamanga and Chimutu found that 142 beneficiaries withdrew out of the 1,447 relocated households in 56 Beneficiary Groups in Machinga and Mangochi: a 10 percent withdrawal rate. Of the three randomly-selected Beneficiary Groups interviewed by IEG in 2012, there was evidence of withdrawal in only one; the number of beneficiaries that withdrew was estimated at 3 in the village of origin and 10 at the resettlement location (in other words, between 4 percent and 13 percent of the 80 families that comprised the three Beneficiary Groups had dropped out). All of those interviewed by IEG referred to the lack of amenities at the resettlement sites and yet beneficiaries were adamant that they were better off than they had been before and in relation to their new neighbors—simply because they had more land; and, knowing of the problems, almost all non-beneficiaries expressed a hope that they might be included in a subsequent willing buyer, willing seller program, even if it meant relocating far from their present home. At the three resettlement sites visited—by no means a representative sample—IEG found that, despite the potential tensions caused by ethnic and religious differences between newcomers and natives, there was no evidence of conflicts.13 As one woman remarked of relations between the two groups, ―we celebrate together.‖ Also, the absence of individual title to land was not viewed as a problem (even though the group titles issued by the project are not recognized by current land laws). The willingness to endure in the face of adversity—and the importance that people attach to having their own land—needs to be offset against the less optimistic reading that emerges from a risk-reckoning based on input, extension and credit constraints. 6.7 But taking a broader view of risk—one that is focused on prospects for replicating the results of the pilot—the outlook is less sanguine. Towards the end of pilot implementation, the project team submitted a concept note on scaling up. This met with no immediate 13 The Bank’s completion report also emphasizes the lack of conflicts. 27 response from the government or from the Bank and, as yet, there is no commitment from either side to launch a new land redistribution project. Passage of the new land bill took longer than expected, owing partly to the opposition from traditional leaders concerned that their powers to allocate land would be reduced (World Bank 2012a). The bill finally passed in June 2013 (Nyasa Times 2013). (It included a provision that chiefs may no longer allocate land unilaterally but must do so in conjunction with village committees.) However, there are some obstacles to scaling up the pilot. Malawi’s latest Growth and Development Strategy (2012-16) does not give priority to land redistribution. The Bank’s Country Assistance Strategy for FY13-16 does not include further support to land redistribution in its program, other than a reference to an FY13 analytic report on agriculture and land issues (World Bank 2012b). 6.8 In any event, scale up should arguably not be attempted until ground rents are systematically applied to leased estates, until taxes are levied on privately-owned land, and until the government shows itself to be serious by not renewing leases on idle estates when the lease expires. Without these changes, there may not be a strong enough incentive for the holders of land to sell up if their estates are underused. In other words, there may not be a large enough supply of land to the market to meet the demand under a willing buyer, willing seller program. In these circumstances, injecting project funds would tend to push up land prices, leaving fewer resources for farm development. It will also be necessary for government to simplify land administration procedures and build capacity at the center and in the districts: at midterm review the Bank observed that weaknesses in surveying, registration and titling could become the major obstacle to scaling up. 6.9 Balancing the more sanguine and the less sanguine assessments, IEG rates risk to development outcome as significant. Bank Performance QUALITY AT ENTRY 6.10 The project's objectives were highly relevant to tackling rural poverty. Development of the project concept was the result of patient advocacy by senior Bank officials who had first made the case for land redistribution in Malawi in the early 1990s. It was not until after a presidential commission report on land reform in 1999 that the circumstances were right for launching the pilot. Project preparation took two years; but this was time well invested to fine-tune the project concept and reach agreement with a government that was initially skeptical about the merits of the willing buyer, willing seller approach. 6.11 The decision to start with a pilot rather than launch a larger project was wise given the risks that were clearly set out in the appraisal document. Preparation and passage of a land bill, which would give land reform a secure legal status, was pending when the project was approved. This was another argument in favor of a pilot: in the absence of enabling legislation, a larger initiative would not have been feasible. However, during appraisal the legal framework for land acquisitions and titling was reviewed and found to be adequate, with no legal changes required for the pilot to achieve its objectives. 28 6.12 The design of the land reform process was based on a realistic appraisal of the difficulties entailed by land reform, informed by experience in other countries (notably, Brazil), as well as lessons from government experience with redistributing land bought from private estates. There was good coordination with the FY04 Fiscal Management and Accelerating Growth Program (P072395) whose actions would help to create incentives for estate owners to sell to poor rural families. The project was soundly targeted: first, the pilot was located in parts of Malawi where the majority of the rural population was poor and had limited access to land; second, the process of forming beneficiary groups involved checks at three points that those selected met the poverty-related eligibility criteria. Given the possibility that some beneficiaries might take up tobacco cultivation (a key cash crop), the project team sought and obtained an exception to the Bank’s ban on support for this crop, after obtaining advice from the Legal Department. 6.13 Also, given the pilot status of the operation, it made sense to rely on an established institution—the Malawi Social Action Fund—to provide community infrastructure. When the project was appraised the Fund had been operating for eight years. The appraisal document acknowledged the risk that the decentralized planning framework might fail to give sufficient priority to the needs of the newly resettled but the alternative—including infrastructure as a project component—would have raised costs to a level unsustainable for a pilot. 6.14 One preparation shortfall was the over-estimate of the land that was likely to be put up for sale in the districts of Mulanje and Thyolo. It seems that the estates in these districts were more financially viable, and the owners less inclined to embark on contract farming with beneficiaries, than the project designers had bargained for. There should have been a more thorough survey of land availability, and an attempt to sensitize estate owners to the aims of the project, before the pilot was approved: a land availability study was not completed until shortly before midterm review. On the other hand, until money was on the table, any estimate of how much land would be offered for sale was bound to be highly approximate. 6.15 IEG rates quality at entry as satisfactory. QUALITY OF SUPERVISION 6.16 Supervision missions were regular and well-staffed, drawing on expertise from across the Bank and with good backup from the Bank’s Lilongwe and Pretoria offices. Supervision reporting was regular, complete and candid, and ratings were judicious: as soon as disbursement lags emerged the implementation progress rating was lowered to moderately satisfactory. After startup, the team sensibly focused on a few critical issues: trying to improve coordination with the Malawi Social Action Fund; appointing key staff; and launching the impact evaluation baseline study. 6.17 The Bank’s procurement department insisted that an international consultant be recruited for the independent impact evaluation, requiring that international competitive bidding procedures be followed. According to some of those interviewed by IEG, the one- year delay might have been avoided if the project team’s original idea of working with the University of Malawi (thereby building local scholarly engagement) had been supported by 29 the Bank’s procurement advisers. (The South African firm that was hired produced an unsatisfactory report and had to be replaced, resulting in further delays.) However, IEG was not able to verify that the skills available at the University of Malawi were sufficient for the job; the case for international competition may well have been sound. 6.18 There were also delays in implementing the communications strategy, which may have discouraged potential landowners and leaseholders from making their land available to the project earlier, as well as failing to dispel false rumors about the government’s intentions (including fears of expropriation). Early disbursement lags were the result of slowness in processing farm development subprojects, itself a consequence of the project team’s extreme caution in field testing the procedures for beneficiary selection, farm acquisition and community oversight—a caution that was warranted by the pioneering nature of this land reform initiative. 6.19 Once the severe limitations on easy access to safe water became apparent, the project team was quick to amend the disbursement rules: the first tranche of the resettlement grant was raised from 40 percent to 60 percent of the total, giving beneficiaries the means they needed to finance the digging of wells and the drilling of boreholes. 6.20 The project team took care to ensure that the project's impact on land price changes was closely monitored. Supervision missions took reasonable steps to ensure compliance with environmental safeguards, although there were some cases of beneficiaries being settled too close to protected areas, requiring mitigation. To give one example of the priority given to safeguard compliance, the midterm review team noted that beneficiaries depended on forest products for poles and firewood, recognizing that although trees needed to be felled to make way for farm plots, the level of cutting was excessive. The team responded by authorizing distribution of 255,000 tree seedlings to 120 beneficiary groups. Follow-up reports indicated that, despite prolonged dried spells, 78 percent of the seedlings distributed survived. Given the frailty of the public extension service, an attempt was made to increase the technical support to beneficiaries in drawing up farm development and environmental management plans, by giving short-term contracts to retired extension workers. 6.21 The midterm review was thorough. In preparation, a land availability study was commissioned. The resulting report (which, arguably, should have been completed during preparation) documented the areas of greatest availability and helped to make the case for expanding coverage of the pilot to include two additional districts. During the mid-term review, the Bank re-visited the economic analysis completed at appraisal and, using new data from project monitoring, was able to confirm that the pilot operation was economically viable. The midterm review pointed to the need for additional financing and for a one-year extension of the closing date, and the project team diligently pursued the necessary negotiations. 6.22 IEG rates supervision quality as satisfactory. 6.23 Overall, Bank performance is rated satisfactory. This pilot operation was the result of dedicated, persistent and patient advocacy by the project team, and the resulting project design and implementation arrangements were both thorough and eminently practical. The 30 pilot generated a model for willing buyer, willing seller land reform that the Bank may replicate in Malawi and where circumstances permit, in other countries. Borrower Performance GOVERNMENT PERFORMANCE 6.24 Government commitment to strengthening land rights was manifest in Cabinet approval of the Malawi National Land Policy in January 2002. Although this did not formalize the land redistribution process that was the subject of the pilot operation, it was a necessary first step, because it provided a framework for decentralized land administration and the development of land markets. As a further earnest of its intentions, the government agreed to make available 4,000 ha for redistribution, taken from estates under its control. When it turned out that much of this land had been occupied by squatters anticipating redistribution (rendering it ineligible for inclusion in the project), the government agreed to an offsetting increase in counterpart funding, with an additional MKw80 million to be disbursed between 2008 and 2010. 6.25 The Ministry of Lands participated fully in project preparation and in the supervision missions. However, it was slow to resolve recruitment delays and other implementation hitches. In January 2008, the Bank had to urge government to finalize renewal of project staff contracts: most of the staff had been working without contracts since June 2007. In addition, there were delays in issuing group titles to the newly-settled beneficiaries, owing to capacity shortfalls in the Survey Department. 6.26 When the project was prepared the government had stated its intention to obtain Parliament’s clearance for preparation of a comprehensive Land Bill that would include a provision for voluntary redistribution. Work began on the bill but it had not been submitted to Parliament by project closing. (The bill was finally passed in June 2013.) In other respects also, there was little progress on land policy reform. The government introduced a moratorium on collecting land rent from leased estates in July 2006, and did not introduce the expected tax on freehold land—measures likely to reduce the supply of (idle) land offered for sale. With support from the project, a new land rent formula was developed and, at project closing, was being applied to all leased land, thus superseding the moratorium. Nevertheless, the government’s temporizing over the needed policy reforms substantially weakened the case for scale up. 6.27 IEG rates government performance as moderately satisfactory. IMPLEMENTING AGENCY PERFORMANCE 6.28 The staff of the Project Management Unit (which was housed in the Ministry of Lands, Housing and Urban Development) adhered closely to implementation guidelines, worked well with the Bank and the government, provided sound financial management and filed regular monitoring reports. In the course of the assessment mission, IEG had the opportunity to observe how former staff from the project management unit handled meetings with village authorities, and with beneficiaries and non-beneficiaries alike. Their command 31 of the technical detail of land redistribution was solid and they showed candor and sensitivity in fielding questions in the villages. However, the management team lacked expertise in monitoring and evaluation and shares responsibility with government for the delays in developing a management information system. The midterm review noted that the management information system had only become fully operational in the previous six months, owing to capacity shortfalls in the project management unit and in the districts. This delay in generating data had hampered decision making in the early phase of the project. Delays were reduced when the project management unit moved from Lilongwe to Blantyre in May 2006. Bringing staff closer to the project area helped speed up decision-making. 6.29 IEG rates implementing agency performance as satisfactory. 6.30 Overall, Borrower performance is rated moderately satisfactory. Without the competence and the hard work of staff in the project management unit the project’s targets would not have been met. But the government contributed less than expected: progress on the land bill stalled, decisions about the rents applied to government estates were ill-advised, the land tax on freehold land did not progress, and there was an absence of resolve not to renew leases on idle estates. Monitoring and Evaluation DESIGN 6.31 As discussed in the Design section above, sound provision was made for impact evaluation. There were some shortcomings in the provision for routine monitoring and the management information system. IMPLEMENTATION 6.32 As discussed in the Implementation section above, the impact evaluation was satisfactorily executed with baseline and follow-up surveys conducted and results analyzed. But the management information system started up late. UTILIZATION 6.33 The results of the impact evaluation were solid and well written up, and provided a sufficient foundation on which the government and the Bank could base a judgment about the merits of scale-up. At closing, the project database and management information system were handed over to the Ministry of Lands, Housing and Urban Development. These could be used for follow-up enquiries as needed. However, at the time of the IEG mission, neither the government nor the Bank had reached a decision about whether to replicate this pioneering pilot attempt at land redistribution; and the FY13-16 Country Assistance Strategy makes no commitment in this respect (World Bank 2012b). 6.34 IEG rates monitoring and evaluation as substantial. 32 7. Lessons There are seven lessons from this pilot operation. First, the willing buyer, willing seller approach to land redistribution can be made to work—but the institutional framework needed for success is demanding. The willing buyer, willing seller model of land reform developed in Malawi is a valid means to reduce poverty in circumstances where there are pronounced inequities in the distribution of landholding coupled with substantial tracts of underemployed cultivable land; however, it will only lend itself to a nationwide program of land redistribution if agriculture and land policies are scale neutral, and there are appropriate incentives for those who hold underused land with agricultural potential to relinquish it. Failure to legislate for, and enforce, land taxes, and continuation of subsidies favoring large landowners, may reduce the number of properties offered for sale. Second, the finer design details of the willing buyer, willing seller model of land reform are critical to its success. For this model of land reform to work it is essential that: there is an effective communications strategy, so that landowners and beneficiaries have reasonable expectations; beneficiary screening is rigorous (and, given the possibility of substitution among candidates) is not limited to a single eligibility check; beneficiaries are provided with the means to visit potential resettlement sites; and the project design is sufficiently flexible to allow for the midcourse adjustments called for by unforeseen contingencies (e.g. water shortages). Third, land redistribution is but one component of a successful land reform project. Land purchase is but a small component of willing buyer, willing seller land reform: the resettlement package needs to make adequate provision for the technical assistance, farm inputs and coverage of the relocation and set-up costs needed for beneficiaries to prosper; another, even larger, challenge is to ensure that the resettlement of beneficiaries goes hand in hand with the provision of community infrastructure—schools, health posts, and roads—in the typically remote locations where land is most likely to be offered for sale. Fourth, land redistribution can raise the incomes of poor households by increasing the amount of land farmed, and it may generate a one-off boost in productivity by bringing idle land into use; further increases in agricultural productivity may be harder to realize. In this project the initial package of complementary inputs jolted productivity upward but yields fell back once the package was exhausted—although productivity was still higher relative to before the project. Improvement to agricultural extension services (beyond the scope of this pilot operation) will be needed to increase yields and encourage crop diversification in resettlement areas. The project provided group title to land but this was not enough by itself to prompt farmers to make the additional investments to raise productivity, at least in the time frame of this operation. Fifth, the success of land redistribution projects will be enhanced by timely studies on land availability. The Malawi pilot included two districts that looked ripe for land reform because there was a large rural population with little access to land and estates whose owners were deemed likely to sell because of declining commodity prices. Subsequent investigation revealed that estate owners in these districts were not on the brink of bankruptcy and, when 33 the project started, were poorly informed about the willing buyer, willing seller model, making them disinclined to participate. Although the necessary land availability study was produced in time for the midterm review, it was needed at the phase of project preparation. Sixth, giving the rural poor the option of homesteading remains a valid objective. As this experiment in Malawi has shown, poor smallholders with a vocation for farming tend to value the acquisition of land very highly and can show great resilience in the face of adversity—as pioneering homesteaders have done throughout history. It is striking that, in Malawi, only about 10 percent of resettled beneficiaries have abandoned their new farms, despite the generally unsatisfactory state of the neighborhood infrastructure. It is equally striking that the anticipated ethnic and religious conflicts between the indigenous population and the settlers failed to materialize: for the most part, land reform beneficiaries have integrated harmoniously with the communities that have received them. An advance communication strategy is essential, not only to reassure would-be sellers about the voluntary nature of the land reform. In addition, when would-be beneficiaries are alerted to the tough conditions they will face, it is more likely that the most enterprising poor people will self- select, and the failure rate will be less. Seventh, the scope for contract farming to help land reform beneficiaries make the leap from subsistence cultivation to cash cropping merits closer investigation. Contract farming may be a useful adjunct to redistributive land reform because it can help smallholders achieve higher productivity by providing the credit, the technical subsistence and the market access that is often lacking in poor rural areas. In Malawi, the opportunities are there but not all contracts are in smallholders’ best interest. Given that contract terms vary between companies, and given that contract formulation is often not transparent, there may be a role for public agencies to help publicize the terms of these contracts, so that farmers can make an informed choice from among the available options. 34 References Adams, Martin, and Stephen Gossage. 2009. ―Final Evaluation of the Support to the Implementation of Malawi Land Reform Programme.‖ European Commission and Scanagri Denmark . Binswanger, Hans, Gershon Feder and Klaus Deininger. 1995. ―Power, distortions, revolt and reform in agricultural land relations,‖ in Hollis Chenery & T.N. Srinivasan, Handbook of Development Economics: Volume 3. Amsterdam: Elsevier. Chinsinga, Blessings. 2011. ―The Politics of Land Reforms in Malawi: The Case of the Community Based Rural Land Development Programme.‖ Journal of International Development 23 (3): 380-393. September. Chirwa, Ephraim. 2008. ―Land Tenure, Farm Investments and Food Production in Malawi.‖ IPPG Discussion Paper 18. Research Programme Consortium for Improving Institutions for Pro-Poor Growth. University of Manchester, UK. July. Chirwa, Ephraim, and Blessings Chinsinga. 2008. ―The Economics and Politics of Land Reform in Malawi: A Case Study of the Community-Based Rural Land Development Program.‖ IPPG Briefing Paper 18. Research Programme Consortium for Improving Institutions for Pro-Poor Growth. University of Manchester, UK. Datar, Gayatri, Ximena del Carpio and Vivian Hoffman. 2009. ―Can a Market-Assisted Land Redistribution Program Improve the Lives of the Poor? Evidence from Malawi.‖ World Bank Policy Research Working Paper. World Bank, Washington, DC. October. Deininger, Klaus and Hans Binswanger. 1999. ―The Evolution of the World Bank’s Land Policy: Principles, Experience and Future Challenges.‖ The World Bank Research Observer. 14 (2): 247- 276. August. Dorward, Andrew. 1999. ―Farm Size and Productivity in Malawian Smallholder Agriculture.‖ Journal of Development Studies 35 (5): 141-161. June. Eastwood, Robert, Michael Lipton and Andrew Newell. 2010. ―Farm Size,‖ in Prabhu Pingali and Robert Evenson, eds., Handbook of Agricultural Economics: Volume 4. Amsterdam: Elsevier. Government of Malawi. 2003. ―Environment and Social Analysis Report: Community-Based Rural Land Development Project‖. Lilongwe. November. ___________________.2005. ―Malawi Land Reform Programme Implementation Strategy (2003 –2007).‖ Lilongwe, Malawi: Ministry of Lands, Housing, Physical Planning and Surveys, Policy Planning Unit. Jere, Paul. 2012. ―Malawi: Implementation of the Land Governance Assessment Framework.‖ Draft report. Lilongwe. April. Kanyongolo, E. 2005. ―Land Occupations in Malawi: Challenging the Neoliberal Order,‖ in S. Moyo and P. Yeros, eds., Reclaiming the Land: The Resurgence of Rural Movements in Africa, Asia and Latin America. London and New York: Zed Books. Kishindo, Paul and Peter Mvula. 2008. ―Participation of Female-Headed households in the Community- Based Rural Land Development Project.‖ Department of Sociology, Chancellor College, Zomba, Malawi. May. Kydd, Jonathan and Robert Christiansen. 1982. ―Structural Change in Malawi since Independence: Consequences of a Development Strategy Based on Large-scale Agriculture.‖ World Development 10 (5): 355-375. Mkamanga, Gilbert and Patrick Chimutu. 2008. ―Community-Based Rural Land Development Project: Study on the Withdrawal of Beneficiaries from the Proje ct.‖ Final report. Lilongwe. April 21. Ng’ong’ola, D.H. 2011. ―Profitability of Contract farming for beneficiaries under Community -Based Rural Land Development Project.‖ Project Management Unit, Lilongwe. 35 Nyasa Times. 2013. ―Malawi lawmakers pass Land Bill.‖ Lilongwe. June 1. Project Management Unit. 2007. ―Land Markets Consultancy Report.‖ [Authors not identified.] Report prepared for the project mid-term review. Lilongwe. Sahn, D. and J. Arulpragasm. 1993. ―Land Tenure, Dualism and Poverty in Malawi.‖ In Including the Poor, ed. M. Lipton and J. van der Gaag . Washington, DC: World Bank. Simtowe, Franklin, Mendola Mariapia and Julius Mangisoni. 2011. ―Independent Project Impact Evaluation of the Community-Based Rural Land Development Project in Malawi‖. Italtrend & Government of Malawi. Reggio Emilia, Italy. July. World Bank. 1975. Land Policy Paper, Washington, DC. ___________2003a. Land Policies for Growth and Poverty Reduction: A World Bank Policy Research Report. Washington, DC. __________. 2003b. ―Implementation Completion and Results Report for the Brazil Land Reform and Poverty Alleviation Pilot Project.‖ Washington, DC. December 4. __________. 2004a. ―Project Appraisal Document on a Proposed Grant of US$27 Million to the Republic of Malawi for the Community-Based Rural Land Development Project.‖ Washington, DC. March 15. _________. 2004b. ―Development Grant Agreement (Community-Based Rural Land Development Project) between Republic of Malawi and International Development Association (Grant Number HO81 MAI)‖. Washington, DC. June 30. _________. 2007a. ―Country Assistance Strategy of the World Bank for the Republic of Malawi , FY07- 10.‖ Washington, DC. January 4. _________. 2007b. ―Implementation Completion and Results Report for the Malawi Fiscal Management and Accelerating Growth Program.‖ Washington, DC. April 13. _________. 2009. ―Implementation Completion and Results Report for the Brazil Land -Based Poverty Alleviation Project I.‖ Washington, DC. June 29. _________. 2010. ―Financing Agreement (Additional Financing for Community-Based Rural land Development Project) between Republic of Malawi and International Development Association (Grant Number H527-MW)‖. Washington, DC. February 12. ________. 2012a. ―Implementation Completion and Results Report for the Malawi Community-Based Rural Land Development Project.‖ Washington, DC. March 30. _______. 2012b. ―Country Assistance Strategy of the World Bank for the Republic of Malawi, FY13 -16.‖ Washington, DC. December 17. 37 ANNEX A Annex A. Basic Data Sheet MALAWI COMMUNITY-BASED RURAL LAND DEVELOPMENT PROJECT (GRANTS H0810 & H5270; PROJECT IDS P075247 & P115226) Key Project Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 29.8 37.9 127% Loan amount 27.0 36.0 133% Cancellation 1.8 Cumulative Estimated and Actual Disbursements FY05 FY06 FY07 FY08 FY09 FY10 FY11 FY12 Appraisal estimate (US$M) 5.1 10.2 17.5 23.5 27.0 27.0 27.0 27.0 Actual (US$M) 1.4 7.2 11.5 22.9 26.8 28.4 33.3 34.2 Actual as % of appraisal 27 71 66 97 99 105 123 127 Date of final disbursement: June 30, 2012 Project Dates Original Actual Initiating memorandum -- 07/20/2001 Negotiations -- 02/09/2004 Board approval -- 04/13/2004 Signing -- 06/30/2004 Effectiveness 07/12/2004 07/12/2004 Mid-term review 11/30/2007 03/30/2007 Closing date 06/30/2009 09/30/2011 Staff Inputs (US$ Thousands) FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY11 FY12 Preparation 64.9 52.9 380.1 -- -- -- -- -- -- -- Supervision -- -- -- 118.8 122.6 146.8 92.9 110.5 134.5 89.2 64.1 ANNEX A 38 Task Team Members Name Title Unit Responsibility/Specialty Lending Rogier van den Brink Task Team Leader, AFTS1 Land Policy and Senior Economist Administration John W. Bruce Senior Counsel LEG LEG Land Law and Policy Shem Migot-Adholla Adviser AFTS1 Land Policy and Administration Kanagasabai Loganathan Financial Management Specialist AFTFM Financial Management Jayne Kwengwere Team Assistant AFTS1 Processing Melanie Jaya Team Assistant AFTS1 Processing Meseret Kebede Team Assistant AFTS1 Processing Francis M'Buka Social Development Specialist Social Conflict Caesar Chidawanyika Agriculture and Land Reform AFTS1 Land Reform Specialist Nginya Mungai Lenneiye Sr. Social Protection Specialist AFTH1 Institutional Development Hope Phillips Volket Senior Operations Officer AFTH1 Operations P.C. Mohan Consultant AFTS1 Development Communication Johnstone Nyirenda Procurement Specialist AFTPC Procurement Tesfaalem Gebreiyesus Procurement Specialist AFTPC Procurement Tijan Sallah Lead Operations Officer AFTS1 Rural Development Michael John Webster Young Professional AFTS1 Economic and Financial Analysis Aziz Bouzaher Lead Environment Specialist AFTS1 Safeguards Donald Mphande Financial management specialist AFTFM Financial Management Muthoni Kaniaru Legal LEGAF Legal Rajat Narula Sr. Finance Officer LOAG Disbursement Sylvester Kofi Awanyo Procurement AFTPC 39 ANNEX A Name Title Unit Responsibility/Specialty Supervision/ICR Rogier van den Brink Task Team Leader, AFTS1 Agriculture and Land Reform Senior Economist Stanley Hiwa Task Team Leader, AFTS1 Agricultural Economics Senior Agricultural Economist, Watson C. Chidawanyika Task Team Leader, AFTAR Rural Development Senior Rural Development Specialist Hardwick Tchale Task Team Leader, AFTAR Agricultural Economics Senior Agriculture Economist Sylvester Kofi Awanyo Lead Procurement Specialist EAPPR Procurement Hans P. Binswanger Consultant AFTP1 M&E Design Mary C.K. Bitekerezo Senior Social Development Spec. AFTCS Safeguards Aziz Bouzaher Lead Environmental Specialist ECSS3 Safeguards John W. Bruce Consultant MNSPS Frank Fulgence K. Operations Adviser AFTAR Operational Byamugisha Design Simon B. Chenjerani Senior Procurement Specialist AFTPC Procurement Chirwa Fenwick M. Chitalu Financial Management Specialist AFTFM Financial Management Lungiswa Thandiwe Senior Environmental Specialist AFTEN Safeguards Gxaba Stanley Hiwa Sr. Agric. Economist AFTS1- TTL HIS Wedex Ilunga Senior Procurement Specialist AFTPC Procurement Ebrahim Mohamed Jassat Senior Social Development Spec. AFTS1- Safeguards HIS Annie Kaliati Jere Team Assistant AFMMW Team Support Grace Ingrid Chilambo Team Assistant AFMMW Team Support Chrissie Kamwendo Senior Operations Officer AFMMW Operations Guo Li Senior Agriculture Economist AFTAR Operations Francis Kanyerere Financial Management Specialist AFTFM Financial Mkandawire Management Prasad C. Mohan Lead IEC Specialist AFTDE IEC Design Donald Herrings Sr. Financial Management AFTFM Financial Mphande Specialist Management David Rohrbach Senior Agriculture Economist AFTAR Operations Tijan M. Sallah Manager AFRCP Operations Hawanty Page Senior Program Assistant AFTAR Team Support Pascal Tegwa Senior Procurement Specialist AFTPC Procurement ANNEX A 40 Name Title Unit Responsibility/Specialty Gert Johannes Alwyn Van Lead Financial Management AFTFM Financial Der Linde Specialist Management Hope C. Phillips Volker Senior Operations Officer EASHH Operations 41 ANNEX B Annex B. Other Data TABLE B1: Field Visits—Details of the Three Randomly Selected Beneficiary Groups ORIGINATING DISTRICT RECEIVING DISTRICT THYOLO MACHINGA Visited on Tue 11 Sep Visited on Thu 6 Sep Spreadsheet Beneficiary Group Traditional Village Group Traditional Village Group Village Farm Name Area (ha) Date of Locator Area Area Relocation Case #7, Chumachilimnthaka, Bvumbwe Sing’anga Chiwalo Chapola Chapola Nsakalu 1 60 6 Jun 2006 2005-06 23 Households ORIGINATING DISTRICT RECEIVING DISTRICT MANGOCHI MANGOCHI Visited on Fri 7 Sep Visited on Sat 8 Sep Spreadsheet Beneficiary Group Traditional Village Group Traditional Village Group Village Farm Name Area (ha) Date of Locator Area Area Relocation Case # 42, Chimwemwe, Bwananyambi Lumeta Bwananyambi Lumeta Nakapa Ndasoloka I 67 17 Aug 2006 2006-07 30 households ORIGINATING DISTRICT RECEIVING DISTRICT THYOLO BALAKA Visited on Wed 12 Sep Visited on Mon 10 Sep Spreadsheet Beneficiary Group Traditional Village Group Traditional Village Group Village Farm Name Area (ha) Date of Locator Area Area Relocation Case # 549, Chisoka, Chimaliro Mangazi Nsamala Nyanyala Phirimbewe Chitale 63 2008 2008-09 27 households ANNEX B 42 Table B3: Number of resettled beneficiaries who returned to place of origin BENEFICIARY GROUPS Chumachilimnthaka Chimwemwe Chisoka 23 Households 30 Households 27 Households Resettled in 2006 Resettled in 2006 Resettled in 2008 N of households reported returned PAIR 1: REMOTE At Origin 0 (Sing’anga, Thyolo District) At Destination 0 (Chapola, Machinga District) PAIR 2: NEARBY At Origin 0 (Lumeta, Mangochi District) At Destination 0 (Nakapa, Mangochi District) PAIR 3: REMOTE At Origin 3 (Mangazi, Thyolo District) At Destination 10 (Phirimbewe, Balaka District) Source: IEG interviews with leaders in the six villages, September 2012. The data on returnees from a given Beneficiary Group were collected both at origin and destination, allowing for consistency check. The data refer to 80 households, enrolled in three beneficiary groups. Data were consistent for two of the three Beneficiary Groups. A maximum of 10 out of 80 households returned (13 percent), with all returnees occurring in one of the three beneficiary groups. A 2008 study found a 10 percent withdrawal rate: 142 beneficiaries withdrew out of the 1,447 relocated households in 56 Beneficiary Groups in Machinga and Mangochi (Mkamanga and Chimutu 2008). Table B4: Differing perceptions of the reasons for beneficiary withdrawal % % % % % % % Beneficiaries Ex- Relatives of Beneficiary Leaders in Key Leaders in Beneficiaries Ex- Group Receiving Informants Sending Beneficiaries Leaders Community Community (N=308) (N=142) (N=68) (N=56) (N=45) (N=37) (N=24) Ex-beneficiaries just 34.0 0.0 10.7 18.9 55.4 34.0 54.2 wanted project cash Lack of infrastructure 9.1 19.3 4.1 20.2 12.5 14.4 8.3 Family problems 3.4 14.3 13.3 3.6 7.5 11.8 0.0 Resettled on infertile 0.0 20.5 14.1 0.0 0.0 9.3 8.3 land Laziness 10.6 0.0 1.1 14.0 0.0 10.6 12.5 Loss of interest in 20.1 0.0 0.0 9.6 0.0 0.0 8.3 Beneficiary Group activities Ex-beneficiaries 0.0 21.0 3.5 0.0 0.0 0.0 0.0 spent most time in villages of origin Ex-beneficiaries did 6.4 10.8 0.0 0.0 0.0 0.0 4.2 not cooperate with other beneficiaries Misunderstanding 5.4 0.0 12.8 0.0 0.0 0.0 0.0 between Beneficiary Group members Misbehavior 0.0 0.0 0.0 11.3 2.5 0.0 0.0 Source: Mkamanga and Chimutu 2008. Columns do not sum to 100 percent because less frequent reasons for withdrawal are not cited. 43 ANNEX B Table B5: Stage of Beneficiary Withdrawal Stage % of Beneficiary Group leaders (N=56) rd After 3 tranche of development funds 25.0 After one year of harvest 24.4 Before relocation 15.9 After relocation allowance 11.5 st After 1 tranche of development funds 8.9 nd After 2 tranche of development funds 5.4 Not clear 8.9 Total 100.0 Source: Mkamanga and Chimutu 2008. Table B6. Distances from Resettlement Site to Key Amenities N=49 sites Distance (Km) Mean Minimum Maximum Road 5.7 0.1 25.0 School 5.2 0.1 25.0 Clinic 13.5 0.1 50.0 Market 7.2 0.1 25.0 Stream 1.3 0.1 15.0 Source: Project Management Unit 2007: 75. Note: Refers to resettlement sites in Machinga and Mangochi districts only. Table B7. Effect of Bargaining—Difference between Asking Price and Final Price Malawi Kwacha (MKw) per Ha Machinga Mangochi (N=22 estates) (N=19 estates) Asking Price Mean MKw 19,909 MKw 20,789 Minimum MKw 10,000 MKw 15,000 Maximum MKw 35,000 MKw 25,000 Final Price Mean MKw 15,295 MKw 16,252 Minimum MKw 7,000 MKw 9,000 Maximum MKw 20,000 MKw 20,000 Final/Asking Price (%) 77% 78% Source: Project Management Unit 2007: 54-55. ANNEX B 44 Table B8: Perceived Welfare—Beneficiaries and Non-Beneficiaries of the Land Reform BENEFICIARIES Beneficiaries’ perception of their Beneficiaries’ perception of their welfare Village, District welfare today relative to their welfare today relative to the welfare of Non- five years ago. Beneficiaries today. Beneficiaries are… Beneficiaries are… Chapola, Machinga 1 Better off 2 Better off Better off 3 Better off Better off Nakapa, Mangochi 4 Better off Better off 5 Better off Better off Phirimbewe, Balaka 6 Better off Worse off 7 Better off 8 Better off Better off 9 Better off Better off NON- Non-Beneficiaries’ perception of their Non-Beneficiaries’ perception of their BENEFICIARIES welfare today relative to their welfare welfare today relative to the welfare of Village, District five years ago. Beneficiaries today. Non-Beneficiaries are… Non-Beneficiaries are… Chapola, Machinga 10 Worse off 11 Better off Worse off 12 Nakapa, Mangochi 13 Worse off 14 Worse off Worse off Phirimbewe, Balaka 15 Better off Worse off Mangazi, Thyolo 16 Worse off Worse off 17 Worse off 18 Worse off 19 Better off Better off 20 Same 21 Worse off Better off Sing’anga, Thyolo 22 Worse off Better off 23 Worse off Better off 24 Better off 25 Same Worse off Lumeta, Mangochi 26 Worse off 27 Better off Better off 28 Better off 29 Worse off 30 Worse off Source: IEG field interviews, September 2012. 45 ANNEX B Table B9: Datar, del Carpio, and Hoffman: Propensity Score/Difference in Difference Estimates for Selected Outcomes Agricultural Land Maize Yield Total Crop Sales Income Increase a (ha) (kg/ha) (Logarithm) (Ordinal scale) Panel: Baseline (2006) vs. 2008 Treated 2.37*** -339 0.76*** 0.14** Standard error (0.18) (106.28) (0.20) (0.07) Observations 469 245 207 444 R-squared 0.147 0.070 0.220 0.060 Source: Datar, del Carpio and Hoffman 2009. a. A value of 1 was assigned if the respondent believed income was better than before moving (better than 5 years ago for control group), a value of 0 if not. *** p<0.01; **p<0.05; ***p<0.1. Table B10: Simtowe, Mendola and Mangisoni: Propensity Score/Difference in Difference Estimates for Selected Outcomes Agricultural Land Maize Yield Tobacco Yield Total Income (ha) (kg/ha) (kg/ha) (Logarithm) Panel A: Baseline (2006) vs. 2007 Treated 0.51*** 1004*** 3525*** 0.13 Standard error (0.051) (351.091) (1207.675) (0.150) Observations 1525 1507 176 1239 R-squared 0.263 0.042 0.206 0.158 Panel B: Baseline (2006) vs. 2009 Treated 0.54*** 354* 2723*** 0.02 Standard error (0.05) (213.027) (602.992) (0.121) Observations 3059 3035 539 2650 R-squared 0.259 0.033 0.132 0.102 Source: Simtowe, Mendola and Mangisoni 2011. *** p<0.01; **p<0.05; ***p<0.1. Table B11: Chirwa: T-test Differences between Treatment and Control Households for Selected Outcomes Agricultural Land Maize Yield Economic Economic (ha) (kg/ha) Wellbeing Wellbeing Compared to Compared to 1 Year Ago 2 Years Ago a a (Ordinal Scale) (Ordinal Scale) Treatment 1.41 1560 2.353 2.147 (N=68 households) Control 0.96 1011 2.859 2.746 (N=71 households) Difference (t-statistic) 0.50*** 448.07*** -0.506*** -0.599*** Source: Chirwa 2008. *** p<0.01; **p<0.05; ***p<0.1. a. 1=Much Better Off; 2=Better Off; 3=No Change; 4=Worse Off; 5=Much Worse Off. Note: The sample of households was drawn from Machinga district only. ANNEX B 46 Table B12: Returns to Burley Tobacco for Project Beneficiaries and Non-Beneficiaries, With and Without Contract, 2009-10. a With Contract Without Contract Beneficiary Non-Beneficiary Beneficiary Non-Beneficiary (N=80) (N=46) (N=88) (N=31) (1) Yield (kg/ha) 1,499 1,312 1,449 1,516 (2) Price (MKw/kg) 241 241 178 191 (3) Gross income 361,259 316,192 257,922 289,556 (MKw/ha) =(1)x(2) b 81,798 84,488 84,127 70,224 (4) Inputs (5) Hired labor 16,700 17,646 16,967 9,952 c 34, 259 20,729 38,955 34,347 (6) Post-harvest (7) Loan repayment 149,794 68,945 6,725 5,066 (8) Other 24,194 21,323 16,924 20,359 (9) Variable Costs 306,745 213,131 163,698 139,948 (MKw/ha) =(4)+(5)+(6)+(7)+(8) (10) Gross margin 54,514 103,061 94,224 149,608 (MKw/ha) =(3)-(9) (11) Family labor 6,919 2,182 5,415 4,940 (12) Depreciation 2,657 2,143 1,894 1,790 (13) Membership 11,350 5,447 205 - (14) Withholding tax 22,949 21,147 14,919 18,697 (15) Fixed Costs 43,875 30,919 22,433 25,427 (MKw/ha) =(11)+(12)+(13)+(14) (16) Total costs 350,620 244,050 186,131 165,375 (MKw/ha) =(9)+(15) Net Farm Income 10,639 72,142 71,791 124,181 (MKw/ha) =(3)-(16) Source: Ng’ong’ola 2011: 47. “Burley tobacco” is a light, air-cured (as distinct from fire-cured) tobacco that is mainly used for cigarette production. a. “Contract” refers to an arrangement where farmers accept support (credit, technical advice, and farm inputs) from a marketing company under a contract which requires them to sell all of their output of the given crop to the company, at a price set by the company. b. “Inputs” refers to seeds, fertilizer, manure and pesticide. c. “Post-harvest” refers to costs deriving from grading, packing, storage, transport and loading. 47 ANNEX C Annex C. List of Persons Interviewed Name Title and/or Affiliation Amos Bemeyani Director of Planning and Development, Thyolo District Council Sean Carpenter Senior Technical Officer, Agribusiness and Microenterprise, PCI, San Diego Nephtale Chabuka Management Information Systems Officer Richard Chinga Advocate, Land Share Malawi Wezi Francis Gausi Director of Planning and Development, Machinga District Council Ilona Gruenewald Programme Officer, Rural Development and Food Security, Delegation of the European Union Ralph Henderson General Manager, Doondo Farming Aubrey Jazza Lands Officer, Machinga District Council Paul Jere Land Governance Consultant Dalitso Kalimba Chief Economist, Ministry of Planning A.J. Kalinga Board Chairman, Tea Association of Malawi Joe James Kantema Commissioner for Lands, Ministry of Lands, Housing and Urban Development James Jones Kanyangalazi District Commissioner, Machinga District Council Kester E. Kaphaizi Secretary, Ministry of Local Government and Rural Development Callistus Kharapuwa Former Manager, Community-Based Rural Land Development Project Ivy Julie Luhanga Principal Secretary, Ministry of Lands, Housing and Urban Development Dorothy Luka Lands Officer, Regional Commission for Lands, Blantyre Joseph N. Magwira Director of Rural Development, Ministry of Local Government and Rural Development Sylvester Maluku Lands Clerk, Mangochi District Council Felix Mangani Surveyor General, Ministry of Lands, Housing and Urban Development Edith F. Maseya Regional Commissioner for Lands, Ministry of Lands, Housing and Urban Development Rodrick Mateauma District Commissioner, Balaka District Council Joseph Upile Matola Economist, Ministry of Planning Oscar Matope Deputy Commissioner for Lands, ANNEX C 48 Ministry of Lands, Housing and Urban Development Francis Mukhupa Economist, Ministry of Lands, Housing and Urban Development Hastings M. Mwachibe Head Clerk, Balaka District Council Randson P. Mwadiwa Secretary to the Treasury, Ministry of Finance Kwame Ngwira Deputy Commissioner for Lands, Ministry of Lands, Housing and Urban Development Clement C. Nyirongo Monitoring and Evaluation Specialist, Ministry of Lands, Housing and Urban Development Lawford Palani District Commissioner, Thyolo District Council Brian Penjani Manda Director of Policy and Planning Services, Ministry of Lands, Housing and Urban Development John B. M. Phiri Chief Economist, Ministry of Planning Francis Sakala Chief Rural Development Officer, Ministry of Local Government and Rural Development Teresa Senzani Senior Deputy Secretary, Ministry of Lands, Housing and Urban Development Franklin P. Simtowe Policy Economist, Alliance for a Green Revolution in Africa, Zambia Jossein Tembo District Lands Officer, Thyolo District Council Max John Wengawenga Principal Economist, Ministry of Lands, Housing and Urban Development Maria Winnubst Attaché, Delegation of the European Union World Bank Hans P. Binswanger Consultant, AFTP1 Sandra Bloemenkamp Country Manager, AFMMW Frank Fulgence K. Operations Adviser, Byamugisha AFTA3 Olivier Durand Senior Agricultural Specialist, AFTA3 David Rohrbach Senior Agriculture Economist, AFTA3 Richard G. Scobey Senior Adviser and Deputy to the Director General, IEGDG Chauncy Simwaka Director, Monitoring and Evaluation, Ministry of Planning Jeffery Tanner Economist, IEGPS Hardwick Tchale Senior Agriculture Economist, AFTA1 Rogier van den Brink Lead Economist, EASPR