The World Bank Group Response to the Haiti Earthquake: Evaluative Lessons

The World Bank Group Response to the Haiti Earthquake: Evaluative Lessons Evaluation Brief 10 2010 The World Bank Washington, D.C. http://www.worldbank.org/ieg Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized ©2010 Independent Evaluation Group Communications, Learning, and Strategy The World Bank 1818 H Street, NW Washington, DC 20433 E-mail: ieg@worldbank.org Telephone: 202-458-4487 Fax: 202-522-3125 http://www.worldbank.org/ieg All rights reserved This Evaluation Brief is a product of the staff of the Independent Evaluation Group (IEG) of the World Bank. The findings, in- terpretations, and conclusions expressed here do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomina- tions, and other information shown on any map in this work do not imply any judgment on the part of the World Bank or IEG concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. IEG encourages the dissemination of its work and will normally grant permission to re- produce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to ieg@worldbank.org. ISBN-13: 978-1-60244-132-3 ISBN-10: 1-60244-132-4 i i i Contents 1 I. Immediate Considerations 2 II. Damage Assessment 3 III. Project Design and Supervision 4 IV. Institutions and Financing 6 V. Private Sector Response 1 The World Bank Group Response to the Haiti Earthquake: Evaluative Lessons As Haiti faces the daunting task of recovery after the devastating earthquake, past experiences provide some lessons. Factors making a crucial difference to the effectiveness of actions include the nature of the immediate response, diagnosis, project design and supervision, use of local capacity, private sector links, and coordination among partners, including within the World Bank Group. Many of the lessons from previous natural disaster episodes are relevant now; yet, Haiti’s distinct country conditions must also be kept in mind. Indeed, several factors make the response in Haiti especially overwhelming—the breakdown of social order and a fragile security situation; the near-complete loss of governance structures; and the failure, before the earthquake, to impose even minimum quality standards on the construction industry. Complicating matters further will be the unprecedented scale of charitable donations earmarked for emergency relief, and the arrival of many new relief agencies, which tend to prioritize unilateral action over coordination. I. Immediate Considerations The World Bank Group should be involved in natural disaster response from the outset. The Marmara earthquake experience in Turkey shows the merit of early World Bank Group involvement. Consultative groups have been effective in mobilizing aid resources and facilitat- ing coordination in postconflict situations. One disaster-experienced Country Director suggests that consultative groups could also be useful in coordinating aid disaster operations, which proved successful in Sudan in putting together a multidonor famine effort. Front-end preparation—including joint damage and needs assessments before consultative group meetings—should be agreed on before operations are put into place. Identifying local leadership and project management offices is particularly important. When credible physical, economic, and social assessments are available, countries ultimately mobilize more assistance than otherwise. Because capacity to use aid effectively in fragile states is low and governance is often poor, the focus from the beginning also needs to be on the development of capacity and the improvement of governance, not merely the reconstruction of physical infrastructure. Every single response is either developmental or counter-developmental; every decision affects everything else. Early actions have a major impact on the recovery. How relief distributions are managed either enhances reconstruction or constrains it. It is far more difficult to stop the use of force and firearms, looting, and rioting than it is to prevent E VA L U AT I O N B R I E F 10 2 such actions in the first place. The tendency of people immediately after an earthquake is to band together to recover survivors from fallen buildings. Ensuring and maintaining mutual trust is important for effective reconstruction efforts. The rebuilding of homes and communities requires the safe transportation and storage of building materials; often, community groups need to be formed, to work together in rebuild- ing homes and infrastructure. Since early actions influence project success, World Bank Group staff need to realize the importance of early actions and advise development partners accordingly. Temporary shelters need to preserve existing social relationships. If possible, avoid expensive temporary shelter under a Bank project. People are able to find adequate temporary shelter using materials from damaged buildings, and families that did not lose their dwellings will help shelter friends and relatives (as occurred in Colombia after the Armero eruption, and in Grenada and St. Lucia after Hurricane Ivan). Relief efforts usually spend more on tents and temporary shelter than the amounts ultimately made available for permanent housing. If people are moved out of the main cities and shelters are required, serious efforts need to be made to keep families and neighborhood groupings intact for reasons of social support and continuity. The physical layout of temporary shelter structures can help reduce crime and violence against women if care is taken during the relocation process to ensure that as many doors of the shelter, as possible, face a common and well-lit area. Such a layout avoids the creation of dark passages and alleyways, which are not in clear view and are potential places for assaults to occur. Also, large, shared, multifamily spaces have led to social break- down in Colombia (Armero), Mexico City, and Turkey. Providing survivors with employment and cash transfers early on has had good results. The general population can be helped to recover emotionally through the rebuilding process with paid work (as was done in Gujarat). Taking the time to ensure that all usable building materials are recovered and recycled is a way to ensure that the poor will be able to afford to rebuild. Once work opportunities associated with clearing rubble and recycling materials diminish, it is important to provide cash assistance targeted at families (as in the aftermath of the Marmara earthquake). Cash transfers are more important than providing food, blankets, and clothing. Indeed, in most disasters, sending canned food and used clothing from overseas is counterproductive. For example, in Bangladesh following a major cyclone, imported food aid destroyed the local rice market. With cash transfers, the distribution of emergency supplies is more orderly, involves local leadership, and helps to enhance social cohesion. II. Damage Assessment Damage assessments need to be simple and tailored to local construction types, with damage awards closely tied to actual costs. Safety evaluations of buildings after an earthquake should determine if structures are inhabitable. If they are not, a plan for the displaced occupants needs to be prepared. In Haiti, project design will need to take into account the fact that there is diminished functional capacity at the local, national, and community levels. Alternative networks, such as nongovernmental organizations and the United Nations agencies, can fill gaps in capacity, but their participation should be coupled with a plan to rebuild government administrative capacity. Following the Maharashtra earthquake, the damage assessment was based on a complicated compensation system. The use of criteria from the International Association of Earthquake Engineer- ing for damage assessment of individual housing units, in this situation, proved to be difficult because the criteria for modern, engineered housing do not work for mud-bonded or stacked- stone structures. When people are compensated in variable amounts for the actual damage to their homes, they will appeal the decisions and argue 3 T H E W O R L D B A N K G R O U P R E S P O N S E T O T H E H A I T I E A R T H Q U A K E with the compensators; they may even inflict more damage on their homes to receive more money. In contrast, if people are paid set amounts for easily determined levels of damage, as occurred in Gujarat, there is then no opportunity for further negotiation. Donor coordination has always proved to be vital. Ways must be found for involved donors to work together or in parallel, in the short term, on a clearly defined set of aid activities with the same eligibility requirements and benefits. III. Project Design and Supervision Reallocating resources from existing projects is less effective than specific reconstruction lending. While shifting resources from existing programs to reconstruction (with high rates of return) may be justified, new financing that is well designed, and managed by special disaster units authorized to respond quickly, tends to be used more effectively. Restructuring old projects is often politically easier than getting new lending and allows the Bank to support government entities already accustomed to working with the organi- zation. However, delivery by staff committed to the goals just abandoned is often not effective and tends to lead to negative consequences for the programs from which money was withdrawn. In general, Bank emergency loans and credits have tended to be slow to disburse after Board approval. Early and continued involvement should support actions leading to expeditious postapproval disbursements. A common problem constraining and slowing implementation is uncertainty surrounding land ownership. In the absence of a resident Country Director, a Bank staff in-country should have authority to make decisions. There can be major unanticipated developments such as price changes for critical inputs, competi- tion for building professionals, and shortages of tools and materials. While experienced staff are likely to foresee such developments, historically, there has not been a clear procedure for ensuring that the right staff are assigned to lead such operations. In the Asian tsunami, which was the subject of a multidonor evaluation, the use of experienced staff was cited. The Bank has recently begun to develop a cadre of experi- enced disaster professionals. It remains to be seen whether the institution is able to ensure their rapid deployment. Project design should be simple, based on local participation, and take into account local capacity. Disaster responses resemble military operations in their heavy reliance on command and controls. The sense of urgency when lives are at stake works against participatory processes. In particular, the perceived need for haste makes it easy to bypass local power structures. People and institutions that might help rebuild affected communities are left out of the relief response, often because the aid institutions have limited knowledge of the communities affected by the disaster. And, conversely, groups with strong incentives to impede implementation are also overlooked. Design should limit the number of implement- ing agencies and sectors involved, and reduce the conditions placed on the lending. Implemen- tation should be flexible to ensure responsive- ness to community needs and rapidly changing conditions in the field. Intensive supervision is important. Emergency projects require special attention to the design and implementation of disbursement arrangements: bottlenecks to cash flow should be minimized before project approval through the provision of guidelines, sample bidding documents, technical assistance to first-time borrowers, training in procurement procedures, and simple local disbursement procedures. If certain technical studies are considered crucial to provide updated plans and infrastructure designs, they need to be identified and agreed upon during appraisal, including a schedule for the preparation of terms of reference, recruit- ment of consultants, commissioning and completion of the studies, and submission of draft reports. IV. Institutions and Financing Streamlined decision making and procedures for contracting civil works will help avoid delays. Continuity in planning, coordination, and monitoring can be provided by a high-powered unit developed for the purpose or by existing institutions. In countries with a decentralized implementation structure in place—such as Bolivia, Argentina, and Pakistan, among others— demand-driven projects implemented by multiple agencies seemed to work. Earthquakes also provide opportunities to solve major infrastructure bottlenecks (for example, widening or rerouting streets, setting aside parkland), but these decisions need to be made and announced early. Actions need to be taken in the field to make them an observable reality before any private investments in repairs and rebuilding are made that would follow the same usage pattern as before the earthquake. Postdisaster operations need to include measures to reduce long-term vulnerability. The Bank’s Operational Policies have long called for natural disaster projects to reduce long-term vulnerability. In Haiti, the current perceived urgency of emergency and reconstruction tasks can be expected to lead to a loss of focus on mitiga- tion/disaster risk reduction, in the face of many competing demands. Interestingly, the opposite situation, that is, the long time between earthquake occurrences, also works toward the same result. Disaster mitigation, because it is a periodic rather than a constant need, tends to lose out to other priorities—especially once the disaster recedes from international media attention and immediate relief needs have been met. Reaching agreement on mitigation measures with the government, within the first three months, is important because it becomes harder to get politicians to focus on a disaster once the memory of the emergency fades. Outcomes are usually better if a financing mechanism for the agreed mitigation measures is defined and locked in. Options to consider include financial incentives, land use and management practices, a review of land tenure patterns, upgraded building codes, training for construction crafts- people, and other nonstructural measures to lessen vulnerability. Postdisaster operations need to deal early and forcefully with land ownership issues. Informal land titles are the rule in many countries because the poor cannot afford legal fees when they inherit, but lack of legal title becomes problematic in the post-disaster context. Where possible, land titles should be formalized or a functional proxy for land titles should be provided (as in the Gujarat earthquake). Where such measures are not possible, alternative means need to be found to ensure that land is not seized outright or that fraudulent claims are not honored. The local government must help control profiteering on land that is urgently needed for the reconstruc- tion process. Avoid permanently relocating neighborhoods for reasons related to less than fully credible claims regarding disaster vulnerability. There is frequently pressure to relocate communities after a disaster. Relocation during postearthquake reconstruction of settlements that consist mostly of one- and two-story buildings is usually a mistake because it is not all that difficult for local builders to make small buildings earthquake resistant. To take an example, in India it was believed that villages built on cotton soil needed to be relocated, which ultimately turned out to be wrong. Another example involves moving people away from coastal zones. The tendency for these residents to return is almost irresistible because 4 E VA L U AT I O N B R I E F 10 of the economic advantages and other amenities associated with living by the sea. In-situ reconstruction should be promoted after earthquakes, to take advantage of existing infrastructure and community facilities, while minimizing resettlement and social dislocation. Communities help themselves when low-cost reconstruction is done in-situ. It is common for outsiders and victims to clamor for relocation in the early days after a disaster but, with time, the importance of preserving social relationships institutionalized in the current neighborhood structures reasserts itself. Moreover, problems can be created by uneven incentives for rehabilitation, as compared with reconstruction. In India, villages litigated for the right to relocate, even when it was technically contraindicated, because the benefits given to families that only rehabilitated were far worse than the benefits given to relocated groups. Beneficiaries entitled to new housing need to make a limited contribution. Bank emergency reconstruction lending used to expect cost recovery at levels that could not feasibly take place, given how much disaster victims had lost and needed to replace. Follow- ing the El Salvador earthquake, the Bank project targeted low-income families and expected full cost recovery. An IEG evaluation later found that over half of the original beneficiaries had been unable to pay and had given up their homes. However, when beneficiaries do not make any contribution, demands can become inflated and unreasonable. For example, the Maharashtra emergency project provided finished houses for free, but led to escalating expectations among beneficiaries. Beneficiaries became unwilling to pay user charges for urban services. One group made demands that the government paint and maintain their gift houses in perpetuity. A compromise with beneficiaries making a limited contribution seems to work best. For example, in the Argentina Flood Reconstruction project, beneficiaries of new housing had to make contributions in materials and labor. Owner-driven housing construction can be more effective than the use of contractors. During project design, staff often believe that the only way to produce the huge number of new dwellings is to bring in large contractors. In India, where people were given funds to repair their units, most of the families actually economized enough to build altogether new houses, and contractors were generally not involved. In those villages it was possible to use local people in the construction, so employment was created for people from the disaster-affected region. Contractors generally used imported labor to work 12–14 hours a day doing piecework. Also, when homeowners were put in charge of the process, houses were better adapted to each family’s needs; there was not a one-size-fits-all approach. The reliance on owner-managed construction was even more widely used follow- ing a subsequent earthquake in Gujarat, with equally positive results. Supervisory personnel did not always need higher education. Projects for postdisaster housing were effectively and economically supervised by builders and masons rather than engineers, at least as regards owner- built structures. Rebuilding after a disaster must also ensure that social structures are kept intact. The impacts of disasters on people vary, depend- ing on the levels of social vulnerability and risk. The recovery process is potentially even more uneven, and it tends to be less visible. When the pressure of the immediate response carries over to the later stages, too little may be done to ensure that the social needs of the affected populations are considered. Natural disasters rip apart social cohesiveness. Rebuilding social structures is a large challenge and one that is rarely done well by any of the institutions, in large part because the character of the initial response makes doing so more difficult. Earthquake disasters often strike informal or squatter settlements particularly hard. Responses need to take the plight of the renters and T H E W O R L D B A N K G R O U P R E S P O N S E T O T H E H A I T I E A R T H Q U A K E 5 squatters into account so as not to increase social inequities. While earthquake-resistant building codes help those in the formal sector, such codes will never be applied in informal settlements and special measures are needed. Grievance procedures need to be in place from the outset. Postdisaster projects have involved strikes, protests, and litigation, partly as a result of not having efficient and effective processes in place to redress grievances. The Gujarat project processed over 40,000 grievances. An ability to make decisions and communicate them speedily to the involved parties is important. Dissemina- tion capacity is helpful, especially if decisions have implications for other beneficiaries. For example, not having a procedure of taking photographs of damages during the assessment process can lead to repeated visits by project staff members and beneficiaries, arguing about the extent of damages to each house. Fostering participation, but not taking the expressed preferences into account, creates discontent. Community participation is not a panacea, and more participation is not always better. But participation can be useful in the design of houses, their layout, and their alloca- tion. In India, having individual owners supervise the construction of their own house proved transaction-intensive, although the use of stakeholder construction-supervision commit- tees for the same purpose worked well. V. Private Sector Response Leveraging existing private sector capacity is critical for effective emergency response. The private sector can play a key role in infrastructure and logistics, local banking, and provision of physical capacity. In the aftermath of the Asian tsunami disaster, IFC effectively supported the relief efforts of an existing client in Sri Lanka with port and airport facilities, allowing for a very quick response to the disaster. IFC provided a grant to an advisory services client with water purification facilities in South Asia. A banking sector client with a branch network in affected areas was instrumental in delivering small grants to small entrepreneurs to help restore their livelihood. In Pakistan, a private hospital mobilized medical teams and mobile treatment centers in the disaster areas. Partners with assets in the field have used IFC grants to provide shelter, food, and water; to clean up affected areas; to restore and improve airport and port logistics; and to provide medical assistance and free telecommunications services. Working with IFC clients near affected areas is essential for speed and effectiveness. Existing partners do not need screening for reputational risks and delivery capacity. Trust and familiarity allow the use of simple arrangements for payments and reimbursement of expenses. Local partners have the local knowledge necessary to ensure that help reaches the intended beneficiaries (for example, a local bank in Sri Lanka was effective in directing livelihood restoration grants to local fishermen who had lost their boats in the tsunami). Similarly, several IFC partners in Pakistan provided shelters, food and medical supplies during relief efforts. Because they are embedded in the local environ- ment, such partners also have strong incentives and interest to promote speedy recovery of local business activities. The use of existing trust funds and other facilities to source grants is also important for speed. This was demonstrated by the matching-grant schemes for the tsunami and Pakistan disasters and by the emergency techni- cal assistance to small and medium enterprises after the Sichuan earthquake. Matching-grant schemes can be powerful instruments for emergency response, but their effectiveness can be limited by the nature of IFC’s preexisting activities. Matching-grant schemes for clients with local presence can provide fast relief and leverage additional delivery capacity. They can also ensure the additionality of IFC’s response in partnership with private sector clients. The applicability and 6 E VA L U AT I O N B R I E F 10 scale of the instrument depend on the size and nature of IFC’s portfolio and pipeline in the country or the region, and the extent to which clients themselves have been affected by the disaster. The limitations imposed by the size of IFC’s client network in the affected country can be somewhat alleviated by leveraging clients’ own networks of suppliers and customers. IFC’s banking sector clients, with their vast networks of private sector customers, can be valuable partners for enhancing private sector crisis response. Design and implementation of reconstruction operations need to balance speed with careful assessment of demand and relevance in rapidly changing postdisaster conditions. Early Board approval of umbrella facilities, including initiatives in support of reconstruction efforts, can facilitate speedy implementation and provide flexibility. But reconstruction initiatives, especially commercial ones, need to be carefully assessed against effective demand in the context of abundant aid and rapidly changing postdisas- ter conditions. For example, the IFC’s facilities, established to support private sector companies in the reconstruction phase of the Asian tsunami disaster, were used only to a limited extent because their pricing was not attractive, given the abundant liquidity in the market and aid money pouring into the affected countries. Local banks in Thailand and Sri Lanka received cheap long- term funding from their respective govern- ments; the larger companies had adequate insurance coverage to repair/reconstruct their damaged properties; and most companies scaled down their new investments, thereby reducing the need for additional funds. Field-based advisory capacity helps support small and medium enterprises, but it needs to adapt to new conditions. IFC provided a significant amount of advisory services for reconstruction through field-based teams in Aceh-Nias and in Sichuan province following the disasters. Having a field-based presence improved IFC’s response time. However, while some private enterprise partner- ship (PEP) Aceh-Nias projects had the right design, they lost their relevance by the time of the implementation due to changes in local conditions. In Sichuan, the content of some of the advisory services was not a good fit with the real needs of the companies. Support to enhance the involvement of the domestic private sector in reconstruction activities can be particularly valuable. **** **** **** Haiti’s challenges are all the more daunting because of the enormous death toll and horren- dous human suffering, and the sheer scale of destruction—not only of the physical infrastruc- ture but also of the social fabric and the institu- tional setting. Crucial in the recovery will be not only the size of financing but also its quality and use. There are many useful lessons for effective- ness. Yet, many questions will also be new and special to Haiti—among them are how a multidonor aid effort plays out, what quality of governance emerges during the reconstruction, what urban design emerges in the reconstruc- tion, how postdisaster social networks are shaped, and how even such an enormous calamity might provide the chance for a new beginning. T H E W O R L D B A N K G R O U P R E S P O N S E T O T H E H A I T I E A R T H Q U A K E 7