First Draft Jan. 2010, Not Suitable for Citation or Circulation

A Role for Micro-insurance in Community-Based Climate Change Adaptation Adapted from “Community-Based Solutions to the Climate Crisis in Ethiopia” Marjorie Victor Brans, Oxfam America Takeshi Takama, Japanese International Cooperation Agency Million Tadesse, Norwegian University of Life Sciences FIRST DRAFT OF FORTHCOMING PUBLICATION Introduction This paper discusses an innovative approach to community-based climate change adaptation in Ethiopia. More specifically, we explore the experiences of the Horn of Africa Risk Transfer for Adaptation (HARITA) project, an initiative involving 17 formal project partners with Oxfam America (OA) serving as the lead coordinator and the Relief Society of Tigray (REST) as the lead implementer on the ground (see Table 2). Together, the HARITA partners have worked on designing a risk management package for farmers in the village of Adi Ha, located in Ethiopia’s northernmost state of Tigray. Specifically, HARITA aims to foster: • • • •

Adaptation of the most vulnerable farmers with interventions rooted in a community-driven process, informed by global science; Holistic drought risk management with significant private sector involvement; The use of a new tool , namely weather index insurance, designed for affordability; and, A replicable/scalable approach at the regional, national, and possibly global level.

Local

• The Adi Ha Community. Central participants in the design of the pilot. • The Adi Ha Farmers Cooperative. Primary organizing body for farmers in the community.

Regional / National

• Relief Society of Tigray (REST). Local project manager for HARITA, responsible for operating the PSNP in six districts of Tigray and overseeing all regional coordination. Established in 1978. Working with Oxfam since 1984 on development issues. Largest NGO in Ethiopia (and one of the largest in Africa). • Nyala Insurance Company. Private insurer in Ethiopia with a strong track record of interest in agricultural insurance. • Dedebit Credit and Savings Institution (DECSI). Second largest micro-finance institution (MFI) in Ethiopia with nearly comprehensive coverage of Tigray. Named by Forbes magazine as one of the top 50 MFI’s in the world. • Mekele University. Member of National Agricultural Research System providing agronomic expertise and research. • Ethiopia’s National Meteorological Agency. Technical support in weather and climate data analysis. • Tigray Regional Food Security Coordination Office. Oversight of the PSNP in the pilot area. • Tigray Cooperative Promotions Office. Responsible for helping organize farmers at the village level. • Institute for Sustainable Development (ISD). Research organization dedicated to sustainable farming practices.

Global

The project has broken new ground in the field of climate change adaptation and micro-insurance by addressing the needs of smallholder producers through an unusual mix of disaster risk reduction (DRR), micro-insurance, and credit, involving a wide range of actors including farmers, the private sector, government, and NGOs. Following is a discussion of the first pilot phase of the HARITA project (2007–2009). We look at how it developed through a community-driven process and how it resulted in critical insights into the role of micro-insurance in climate change adaptation.

• Oxfam America. International relief and development organization that creates lasting solutions to poverty, hunger, and injustice. Lead project manager for HARITA. • Swiss Re. Global re-insurer and leader on climate change advocacy. Funding and technical expertise. • International Research Institute for Climate and Society (IRI). Member of Columbia University’s Earth Institute. Research and technical expertise in climate data and weather index design for rural farmers. • The Rockefeller Foundation. The Rockefeller Foundation supports strategies and services that help communities cope with the impacts of current and imminent climate change. • Index Insurance Innovation Initiative at University of California-Davis. Research partnership on index insurance between academia and development organizations, with UC-Davis, Food & Agriculture Organization, International Labour Organization, and US Agency for International Development. • Goulston & Storrs LLP, and Weil, Gotshal & Manges LLP. Legal expertise.

Table 1: HARITA Official Project Partners

First Draft Jan. 2010, Not Suitable for Citation or Circulation Background Oxfam International is a confederation of 14 organizations working together with over 3,000 partners in more than 100 countries to bring about lasting solutions to poverty, hunger, and injustice. OA, the US affiliate of Oxfam International, was founded in 1970 and is headquartered in Boston. REST, the largest Ethiopian NGO, has been implementing a wide range of relief, rehabilitation and development activities in the regional state of Tigray since 1978. OA and REST have worked jointly in Adi Ha for more than a decade. With a total population of 4,285, Adi Ha is one of the largest villages in the Kola Temben district of Tigray Regional State in Ethiopia. In the mid-1990’s, OA, REST, and farmers in Adi Ha established the first small-scale irrigation system in Tigray, which allowed participating farmers to shift away from subsistence, rainfed crops to high value horticultural crops. Now that these valuable crops have begun producing yields at nearly their full potential, participating farmers enjoy very substantial economic benefits (SID Consult 2008). The project has attracted widespread attention for its positive impact in Adi Ha, which had been known for chronic food shortages and water stress. REST has replicated the Adi Ha irrigation model across Tigray. Following the success of the irrigation project, OA proposed the HARITA pilot to REST and Adi Ha in the belief that climate change threatens to wipe out precious development gains and to outstrip the international humanitarian community’s ability to respond to emergencies. Although many farmers in Adi Ha depend on the irrigation system, most farmers do not have access to it given the system’s physical limitations. As such, rainfed farmers are particularly vulnerable to drought. Adi Ha’s rainfed farmers are particularly vulnerable to climate shocks and are representative of Ethiopians as a whole—85% of the country’s citizens are also subsistence farmers. This disparity is evidence that “hard adaptation” such as irrigation systems is effective, but it is often limited in its ability to reach people in broad geographic areas. In contrast, “soft adaptation”, like changing crop practices and micro-insurance, can be implemented among many more people in short order. Preparing for climate change will require everyone to “work harder” to reduce climatic risk exposure by employing, for instance, conservation and livelihood building strategies that have already proven their worth. But it will also require everyone to “work smarter” by employing new tools and methods. Specifically, OA was interested in exploring how weather micro-insurance could address the climatic risks that cannot be managed by risk reduction alone. Thus HARITA began in the fall of 2007 with exploratory discussions to gauge the community’s interest in a pilot. With farmer’s backing, OA commissioned an independent demand study in the community the following winter, and then in the spring of 2008, OA contracted IRI to draft a prototype weather index insurance contract. In July 2008, OA and Swiss Re consolidated their collaboration and project implementation with new project partners joining along the way. Robust Strategies Involving Insurance, the Private Sector, and Local-to-Global Cooperation The impact of disasters and natural hazards is not straightforward. For instance, due to widely differing levels of farmer vulnerability, a meteorological drought in Ethiopia will have a very different impact than an equally severe drought in the mid-west of the United States. In contrast to Ethiopians, farmers in the US enjoy larger financial cushions, better inputs, crop insurance, government subsidies and price risk hedging tools like futures and options. As Dahal (2005) argues, ‘’it is only through noting how these [climate] changes are translated into an impact on people’s lives and livelihoods that action can be taken’’ (as cited in van Aalst et al, 2008). In other words, climate effects will have idiosyncratic effects on people at the local level. Although adaptation efforts must be consistent with and bolstered by regional, national, and global policies and strategies, the primary focus should be on communities (Ayers & Forsyth, 2009). Unfortunately, promoting community-based adaptation is easier said than done. Adaptation is a highly complex process that can take the form of autonomous responses by communities acting alone or through interventions planned by policymakers. This paper focuses on planned adaptation, which involves action at the local to global level, financial commitments by governments and donors, and appropriate tools and technologies at various scales. The HARITA project introduces micro-insurance as just one tool among many in helping governments and people tackle adaptation to climatic risks.

First Draft Jan. 2010, Not Suitable for Citation or Circulation Planned adaptation would stand the best chance for success when informed by accurate and precise predictions about the future. Although the scientific community has tried to develop downscaled climate models that can provide reliable projections at fine scales, these models cannot be successfully adapted to assess community level risks because they are based on data that is too generalized and not sufficiently robust to illuminate any specific climate process at the local level (Marengo et al, 2009). While investments in climate science will certainly increase the computational power of climate prediction, these improvements are likely to come more slowly than the rate of increasing risks. Moreover, decision-makers will still need to account for other unpredictable systems like globalization, regulation, cultural preferences, etc. when designing adaptation programs (Dessai et al, 2009). Local level decision-makers must, therefore, find creative ways to cope with deep levels of uncertainty. In response to this challenge, the IPCC’s 4th Assessment Report recommends what is known as ‘robust’ strategies versus more conventional ‘optimal’ strategies. In general terms, optimality is defined as the best choice among multiple options, based on an accurate prediction of future states. In contrast, Rosenhead et al (1972) define robustness as flexibilities in decision-making strategies measured against multiple future possibilities. Robust adaptation involves a planning process conducted as a series of decisions over time—a pathway of evaluating risks, identifying options, choosing options, monitoring the outcome and then iterating the process at the next decision node. The approach also focuses on dealing with high levels of uncertainty and possible thresholds in the system being managed. For example, an insurance package would be more robust if it allowed an annual re-evaluation of risks and re-identification of appropriate product packages, rather than assuming the existing package will work over long periods of time under climate change. Insurance—in particular one-season weather index insurance contracts—can play an important role in supporting robust, community-based adaptation. Weather index insurance protects the policyholder against events correlated with crop loss. If a pre-defined weather event occurs during a pre-defined time, such as low rainfall during crop growth, this event triggers pre-determined payments to farmers who buy the policy. The term ‘index’ refers to the fact that the insurance is based on a proxy for loss and an objectively verifiable measure of weather. More specifically, weather index insurance can make evolving types of adaptation strategies more robust by addressing the scenarios for which it is relatively more difficult to plan. For instance, rainfall harvesting structures can reduce risk by allowing farmers to conserve water in the long run, while drought index insurance can provide added protection if rainfall ends up unusually low in the short run (i.e. year to year). Weather index insurance can also promote prudent risk taking that helps farmers increase their resiliency. For example, banks are reluctant to lend to uninsured farmers; similarly many farmers—fearing default due to reasons beyond their control (i.e. drought)—hesitate to take on loans for improved inputs like high yield seeds. Weather index insurance decreases the chance of default because it allows farmers to hedge their bets about future weather. Under good weather, farmers’ investments pay off, but under bad weather, insurance cushions the blow. In these ways, micro-insurance reduces vulnerability, facilitates quicker recovery from disasters and boosts farmers’ confidence to pursue adaptive and more productive livelihood strategies (DFID, 2004). For all these reasons, weather index insurance is being held up internationally as a tool of great promise. To wit, at the International Labour Organization’s Africa regional conference on social economy in October 2009, the consensus among the over 200 participants was that by providing complementary pathways to development, micro-insurance can help promote a social economy that is resilient to multifaceted crises, such as economic meltdown and environmental imbalance (ILO, 2009). Index insurance has many advantages over traditional crop insurance that instead protects against actual crop damage. First, it does not create incentives for moral hazard (e.g. farmers neglecting their crops anticipating a higher payout) because farmers’ behaviour cannot influence the index. Index insurance is also usually less expensive, because it is generally cheaper to verify overall weather conditions in a given region than to assess damage on individual farms. Moreover, payouts can occur as soon as the loss-causing event is detected. Quick payouts allow farmers to minimize income destabilization and to recover faster, e.g. in a food production shortage, they can buy food as well as seeds for the following season. Despite the benefits of insurance, only 3% of losses from disasters in Least Developed Countries are covered by insurers. Even allowing for influxes of donor aid, on average 90% of the costs of reconstruction are borne by poor countries (Hoeppe and Gurenko, 2006). These statistics indicate that: 1) climate change adaptation requires substantial increases in financial resources, well beyond what is currently being covered by public sector grants, and 2) private insurers could play a major role in adaptation.

First Draft Jan. 2010, Not Suitable for Citation or Circulation Involving the private sector in adaptation is attractive for a number of reasons. First, under the right circumstances, the private sector stimulates healthy economic growth. Moreover, most small-scale weather disasters, such as unreported droughts, do not attract any post-disaster aid, although they may cause serious financial losses. The cumulative effects of these small-scale disasters can result in devastating long-term impacts on the lives and livelihoods of those at risk. As private insurers are in a position to cover unreported disasters in addition to the large-scale and internationally visible ones, the private sector could fill a critical gap in poor countries’ risk management system. The private sector can also provide reinsurance (i.e. insurance for insurers). Given the fact that climate impacts are often spread across multiple communities, reinsurance is vital to prevent bankruptcy among micro-insurers when extreme weather leads to mass payouts. Again, it is important that community-level projects roll up into a coherent framework of regional, national, and international adaptation efforts. In the case of insurance-related interventions, community-based microinsurance can complement meso- and macro-level strategies at the regional and national level. For instance, households and insurance companies can absorb small scale risks, while regional and national government can prepare for larger and more uncertain risks. In other words, multi-layered, insurance-related programs cover the broad range of uncertainties and sizes of climate change-related disasters (Table 2). Table 2 Multi layered insurance structure for [low frequency] and high impact weather related disasters Layer 3 Layer 2 Layer 1

High level risks Medium layer risks

Government Capital markets Reinsurance companies Primary insurance companies Lower level risks Households and companies Source: Botzen and van den Bergh (2008)

At the regional and national level, the Ethiopian government is developing a variety of policies to address major regional and national weather-related risks. For example, in 2006 the government, the World Food Programme and reinsurer Axa Re developed a macro-level, index insurance policy to help the country deal with catastrophic, national drought. At the global level, international public sector adaptation financing mechanisms can provide the funds needed to ensure the long-term viability and wide coverage of community-based adaptation programs for the poor. For instance, the multi-lateral Adaptation Fund (AF), run by the Global Environment Facility, underwrites ‘concrete adaptation projects’ through levies on the Clean Development Mechanism (Mohner and Klein, 2007). In the future, the AF may look beyond mere funding for individual projects and toward mechanisms that promote the sustainability of systemic adaptation efforts. Although the AF will be too small to serve as a vehicle for reinsurance, it can subsidize the initial overhead of large, micro-insurance programs and thus facilitate aggregation of policyholders through outscaling. In this way, the AF could help micro-insurance programs attain larger and more sustainable numbers of clients. On the flip side, micro-insurance programs can channel international financing to vulnerable local people. HARITA Conceptual Framework HARITA operates under four key assumptions: local communities should be the focus of adaptation efforts; adaptation needs to be robust to multiple climate scenarios; adaptation could be strengthened by the private sector; and community-level processes need to roll up into coherent regional, national, and international level processes. At the community level, the HARITA approach consists of three main components: climate-related risk reduction, risk transfer (insurance), and prudent risk taking (credit-led investments). In conjunction with appropriate government policies and accessible input and output markets, these three components work synergistically to promote livelihoods at the household level and greater resiliency to evolving threats, chief among them climate change. (See Fig. 1 below.) In the first year of the pilot, HARITA focused solely on agricultural drought and weather index insurance for the cereal crop teff, but the HARITA approach could conceivably be applied to other types of crops, risks, and insurance cover. HARITA’s primary risk management strategy is for farmers to identify and implement ways of decreasing their exposure to the kind of small-scale droughts that typically fail to attract external support or attention (see below for further discussion). Relevant DRR activities include small-scale water harvesting, increasing soil moisture

First Draft Jan. 2010, Not Suitable for Citation or Circulation retention, and others. As a complement to DRR, farmers then have the option of purchasing weather index insurance to address the drought risk that cannot be sufficiently reduced. In areas with longer and more frequent droughts, insurance can play a critical role in promoting quick recovery from deficit rainfall. Finally, with insurance as collateral, farmers and lenders can also be more confident about entering into loan agreements. The HARITA model was piloted in 2009 in Adi Ha village with the Productive Safety Net Program (PSNP) serving as the primary distribution mechanism. The PSNP is a well-established, federal social protection program serving 8 million chronically food insecure households (i.e. those requiring food aid for each of the past three years). With annual funds of roughly US$500m from large international donors, the PSNP was established in 2004 as a system of transferring cash and food to vulnerable households before they reach a crisis point. This assistance is provided in exchange for beneficiaries’ work to reduce risk, e.g. by building community assets such as water harvesting structures or reclaiming environmentally-degraded areas. As of 2008, the PSNP had grown into one of the largest, social protection programs in Africa. Early impact studies suggest that the PSNP is superior to traditional, emergency food aid programs in significantly contributing to longer-term household welfare (UNDP 2007); (Sharp et al 2006). REST manages the PSNP in six districts of Tigray. Through HARITA, farmers enrolled in the PSNP have the option to work extra days on risk reduction activities beyond those required for their normal payments, but instead of earning cash or grain for this additional labour, they earn an insurance certificate protecting them against deficit rainfall. Richer (i.e. irrigated) farmers who do not participate in the PSNP have the option to purchase insurance with their own cash; as such, they constitute a potentially important private client base for the Ethiopian insurance industry. By allowing vulnerable farmers to pay their premiums in labour, farmers benefit even when there is no payout— the risk reduction measures taken in their communities pay dividends, even during the good weather years. The insurance-for-work (IFW) model is innovative in allowing insurance and credit to stand as independent components. In most index insurance pilots, farmers have been required to take insurance and loans as a package. Under HARITA, however, farmers may choose to bundle the two, but they are not required to do so. The independence of credit and risk transfer means that farmers retain access to insurance after they have repaid their loans, and farmers who do not want a loan can still obtain insurance. Weather index-based insurance as an adaptation tool Risk is the probability of harmful consequences driven by a combination of vulnerability and hazard. HARITA’s project partners believe that the best way to deal with climate risk and its negative impacts is to reduce and eliminate risk using the tools farmers and development practitioners already know and use well. In the case of drought in Ethiopia, the problem is so severe, and it is virtually impossible to eliminate the risk entirely. For instance, rainwater harvesting systems reduce drought risk substantially but they would be useless if there was never any rain. Even though climate change will exacerbate the intensity and frequency of climatic hazards, it is possible to reduce the risks by decreasing the vulnerability of people or spreading the impact over a wider swathe of society. In this way, insurance can support adaptation by providing a means of formal risk transfer and sharing. Because weather index insurance contracts are priced from year to year, the premium charged can also reflect changing risks over time, including not only climate trends but also seasonal rainfall predictions (Osgood et al, 2008). As such, the market signals to farmers what production strategies are likely to succeed given the current climate and seasonal conditions. HARITA’s project partners set out to explore how weather insurance could be incorporated into a holistic, drought risk management approach. To date, few existing index projects have explored ways to combine risk reduction and insurance. Moreover, few have reached the stage of actual financial transaction, and most are small-scale pilot projects or one-year test period initiatives (Hellmuth et al, 2009). While these projects prove index insurance can work at a technical level, it remains to be seen how they can be scaled up and used effectively as an ‘industrial strength’ tool for robust adaptation.

First Draft Jan. 2010, Not Suitable for Citation or Circulation

Figure 1: HARITA Conceptual Framework The good news is that Ethiopia could be fertile ground for weather insurance to take root in a holistic approach to risk management. Despite its poverty, the country has a number of assets in its favour: First, Ethiopia is one of the few countries in the world to include microfinance in its Poverty Reduction Strategy Paper (IMF, 2007; Microned, 2008). In 1996, the Ethiopian government established a formal role for microfinance institutions (MFIs) as actors in the country’s development (Amha, 2000). With an estimated 2 million clients as of 2009, MFIs have made impressive inroads in extending financial services to previously unbanked poor households—both rural and urban—in a short period of time (Admassie, 2004; Chamberlain and Smith, 2009). Moreover, demand for risk transfer across the country seems to be very high. In 2008, Oxfam America was commissioned by the International Labour Organization and the UN Capital Development Fund to investigate micro-insurance demand among four major occupational groups: with low-paid urban workers in Addis Ababa; coffee farmers in Southern Nations, Nationalities, and Peoples’ Region; pastoralists in Oromia; and agropastoralists, also in Oromia. Throughout the country, study participants expressed a very strong desire to access better risk management, with insurance considered highly attractive in theory (Tadesse and Victor, 2009). It is important to acknowledge, however, that adaptation through new tools like insurance is not simply a matter of rolling out risk transfer products. Effective adoption also requires removing important social, political, and economic barriers. In OA’s study on the demand for micro-insurance, we found both opportunities and challenges in this regard. Opportunities • High Receptivity: In all regions studied, Oxfam found that poor people viewed the concept of insurance with enthusiasm. Participants familiar with micro-finance and iddir (traditional, informal burial societies that function similarly to life insurance) were especially quick to recognize the inherent merits of risk transfer. •

Growing Need for Insurance as Collateral in Credit Markets: Weather index insurance could hardly be more welcome, not only due to the increasing risks associated with agricultural lending, but also due to the on-going phase-out of Ethiopia’s farmer loan guarantee fund. Beginning in 1994, regional governments initiated a 100 percent credit guarantee program to facilitate access to fertilizer, improved

First Draft Jan. 2010, Not Suitable for Citation or Circulation seeds and other inputs. Under the system, approximately 90 percent of fertilizer and seeds were delivered to farmers on credit at below-market interest rates. Local governments entered into agreements with the Commercial Bank of Ethiopia (CBE), guaranteeing the farmers’ purchases. In order to finance the loans, credit was then extended to farmers by the CBE through cooperatives, local government offices, MFIs and one cooperative bank. The loan guarantee program had reached four million farmers with guaranteed credit of nearly US$70 million in recent years. While the program benefited some farmers, it also created incentives for farmers to default in large numbers. As a result, the government is ending the program (World Bank, 2006), leaving most asset-poor farmers with no other collateral than social capital. Ethiopian MFIs heavily favour the group guarantee lending model (GGLM) where borrowers vouch for each other and cover defaults in the group (Borchgrevink et al, 2005). The GGLM is widely criticized by poor clients for stimulating conflict among borrowers; yet lenders will struggle to find an acceptable alternative as the loan guarantee fund is eliminated. In this way, insurance could provide an alternative to the GGLM. Also, while the Ethiopian government has admirably pushed MFIs to serve risky agricultural portfolios, the threat of climate change could make it much more difficult to meet this goal sustainably unless robust risk management is put into place. Microinsurance will likely constitute a critical element in a robust, agricultural credit market. Challenges •

Poor Access to Formal Insurance: Unfortunately, the Ethiopian insurance industry is not well positioned to enter the micro-insurance market. Almost exclusively serving large, urban industry, formal Ethiopian insurers count fewer than an estimated 300,000 clients out of the country’s 79 million inhabitants (Chamberlain and Smith, 2009). While insurers recognize their risks are overly concentrated in cities, they struggle to identify cost-effective means to reach rural, poor, and financially illiterate Ethiopians. The currently embryonic micro-insurance market is limited almost exclusively to credit-life insurance programs which cover lending risks to MFIs, but generally do not benefit surviving family or provide access to other types of cover.

•

Fears about Affordability: Most Ethiopians cannot afford more than the little informal protection they have now: their incomes are simply insufficient to cover premiums relative to their expected losses.

•

Low Awareness of Insurance: Oxfam’s study participants who had no experiences with micro-finance and iddir struggled to grasp the concept of insurance without detailed explanation. Also low levels of education and telecommunications make awareness-raising a challenge. Less than half the Ethiopian population can read and write; as of 2001 there were only three radio broadcast stations and one television station (CSA, 2007). As such, word-of-mouth, direct experience, and cooperation with institutions that already reach impoverished areas will be important modes of outreach. (Tadesse and Victor, 2009)

HARITA has found some promising ways to overcome all of these challenges. A Community-Designed Adaptation Experiment Pratt and Earle’s survey of development projects in Ethiopia argues strongly that: [A] slower but more participatory approach within communities will have eventually pay dividends. While this might absorb both time and financial resources, in many cases, the benefits will outweigh the costs. Many traditional development projects have failed in Ethiopia. (2004) For this reason, HARITA has prioritized community-driven processes. Yet most weather index insurance pilots have not focused on deep engagement of farmers due to the technical nature of index insurance and time pressures to develop a commercial product immediately. Involving farmers in adaptation planning is not merely an exercise in political correctness: as Walling points out ‘adaptation is as much about changing attitudes and behaviours as finding technical solutions’ (undated). Uptake of weather insurance in prior pilots for subsistence farmers has not been automatic and is frequently disappointing, partly due to the relative unfamiliarity of insurance concepts in very poor communities, and partly due to underinvestment in demand-driven (as opposed to supply-driven) products (Hellmuth et al, 2009; Tadesse and Victor, 2009). While some pilots have attracted solid demand from farmers, it is unclear why takeup has not been stronger given the many theoretical benefits of risk transfer. Many industry observers have thus

First Draft Jan. 2010, Not Suitable for Citation or Circulation incorrectly concluded that illiterate farmers cannot understand complex financial products or ever manage to afford them. HARITA’s project partners set out to challenge these assumptions, believing it necessary to develop popular educational tools around this new class of financial products and increase affordability. They also sought ways to increase the attractiveness of index insurance by prioritizing the core interests of clients, in contrast to the dominant index insurance model which focuses on lenders and insurers. As such, HARITA’s project partners recognized that farmers in Adi Ha would need to play a central role in the pilot’s design. Specifically, farmers were best positioned to alert the adaptation and insurance ‘experts’ to farmers’ educational and risk management needs, as well as to show how weather index insurance could be made very attractive to the target client. Below, we highlight some of the ways farmers played instrumental roles in HARITA’s achievements to date. Community-wide Participants In November 2008, OA, REST, and Mekele University conducted a Participatory Capacity and Vulnerability Assessment (PCVA) with roughly 200 farmers in Adi Ha to understand the farmers’ capacities for managing risk and to explore in detail the vulnerabilities and capacities of teff farmers vis-à-vis climate-related hazards. Results of the risk assessment were used to inform the identification of appropriate risk reduction activities. For instance, in response to drought risk and rising average temperatures, farmers could select more drought tolerant crops, improve management of water resources, and move planting dates. Such interventions could substantially reduce the risks posed by climate change. The PCVA involved the use of Participatory Rural Appraisal tools such as risk mapping, transect walks, and historical and seasonal calendars, as well as asset inventories, livelihood surveys, focus group meetings and key informant interviews. To complement the PCVA, HARITA also conducted experimental economic risk simulation games and focus group discussions with farmers to understand their needs, desires and concerns. Community Pilot Leaders In May 2008, a team of five community members representing key groups (identified by farmers as Village Administrators/Elders, Rainfed and Irrigated Farmers, Landless Youth, and Women) were elected to become Pilot Leaders. The Leaders’ duties included serving on the pilot’s broader ‘Design Team’ consisting also of representatives from local stakeholder organizations and Oxfam. The Team was charged with planning and implementing the HARITA project under the advice of international experts. Throughout the year, the Pilot Leaders provided critical inside information about common attitudes and concerns in the village. They also helped brainstorm product options and identify possible pitfalls. (For more detail, see below.) Community Pilot Participants In July 2008, the Pilot Leaders recruited 21 farmers (including themselves and nine women) to participate in ongoing focus group discussions. The group’s composition was balanced to include experience and interest in farming teff; geographic distribution; gender; access to irrigated land; wealth; and size of landholding. The Pilot Participants attended a number of test workshops on climate change, weather data collection, financial literacy, and insurance. What Community Participation Produced Farmer participation enhanced the HARITA model and led to numerous positive results, including: Identification of Vulnerabilities Through many interactions, the community helped the project partner organizations identify farmers’ vulnerabilities to specific hazards as well as their capacity to adapt. Farmers shared the following information: • • • • •

The biggest threats in Adi Ha are agricultural, chief among them drought. The majority of farmers surveyed expressed interest in micro-insurance: agriculture and livestock insurance ranked as the highest priority, while health cover came in second. Affordability of insurance premiums is a big concern. The most vulnerable members of the community are female-headed households. Maize, teff, and finger millet have led to the most crop failure in Adi Ha. Maize is most vulnerable to drought, but teff is more economically valuable.

Note that the Community ranked drought, not climate change, as their key concern. However, because climate change will likely increase drought in Ethiopia, the project partners saw this as an opportunity to link ‘climate

First Draft Jan. 2010, Not Suitable for Citation or Circulation change information from ‘outside’ to the experience of the communities involved in the [risk assessment], and to integrate risk reduction activities into the results of the investigation’ (van Aalst et al, 2008). Identification of Informal Risk Coping Measures The Community also identified the strengths and weaknesses of their informal modes of risk management and community risk and resource sharing measures. Most relevant to this discussion is self-insurance, which means any loss is absorbed and ‘compensated’ by one’s own assets (e.g. savings and current or future income). Savings are the most common form of selfinsurance. While more and more people in Adi Ha are beginning to save through formal financial institutions, most still stash their cash in a box at home. The greatest benefit of self-insurance is its universal risk coverage and timeliness—self-insurance allows a quick response to any type of loss. Unfortunately, self-insurance is rarely sufficient to compensate for loss. Other common coping mechanisms included community-based risk sharing arrangements such as sharecropping and livestock sharing. While useful mechanisms, they only cover a fraction of the risks households face and are unable to handle mass risks that affect everyone at the same time (e.g. drought). Also, as a reflection of their social construction, poorer marginalized groups are often excluded from community risksharing as they are deemed unable to reciprocate assistance from their richer neighbours. The Insurance-for-Work Model To cope with their poverty, many farmers in Adi Ha turn to external assistance, especially from REST, DECSI, and the local government. Many rainfed farmers also depend on the PSNP. When a few farmers in Adi Ha suggested that they pay for insurance in other forms than cash, it became obvious that HARITA could be based on an innovative insurance-for-work model through the PSNP. We stress that without the urging of farmers to make insurance affordable in creative ways, this major project innovation would never have been realized. Insights and Preferences Relevant to Major Pilot Activities Risk Reduction Farmers identified a number of ways to enhance the production and yield of teff as a strategy for decreasing exposure to the economic risks associated with drought. Key questions farmers raised were how to: • • • •

Boost drought tolerance; Manage poor soil fertility given high fertilizer prices; Reduce the threat of pests, waterlogging and flooding; and, Decrease the need for frequent ploughing and weeding, which divert valuable labour.

Given the fact that teff is grown almost exclusively in Ethiopia, it has garnered relatively little attention from international crop scientists. As such, HARITA’s project partners concluded that researching teff and sustainable agricultural practices in Adi Ha would be worthwhile. In the end, the study on teff headed by Mekele University, along with the results of the community risk assessment, informed the identification of the risk reduction activities that were actually undertaken by farmers in Adi Ha during the following agricultural season. More specifically, the Addis Ababa based Institute for Sustainable Development and Mekele University helped participating farmers learn to make and use compost, which is critical for rebuilding soil nutrients and improving soil moisture retention. In addition, they constructed small scale water harvesting structures on farmland, as well as planted nitrogen-fixing trees and grasses to promote soil regeneration and water conservation, and reduce the risk of flooding. Finally, farmers cleaned teff seeds as a way to boost productivity and control weeds. These risk reducing interventions are keys to reducing exposure to natural hazards, boosting income over time, and promoting quicker recovery from disaster. Index Insurance: Farmers also helped design a weather index insurance product to complement the risk reduction activities taking place in the village. Design areas that garnered valuable farmer input included the following: •

Filling in data gaps: Designing drought insurance normally requires at least 30 years of reliable, daily precipitation data. In most developing countries, rain gauges are sparsely distributed and limited in quality and duration. Not surprisingly, most weather insurance pilots have been located in villages with good rainfall records. These data-supply-driven pilots have demonstrated that weather insurance products are viable at the pilot level, but they have not shown how to overcome the data barrier in poor communities that have poor weather records. For this reason, HARITA explored new techniques to enhance sparse local datasets through a combination of farmer collected data, official meteorological

First Draft Jan. 2010, Not Suitable for Citation or Circulation data, satellite imagery, rainfall simulators and statistical tools. While continued research will be necessary, IRI developed a viable index and an open-source methodology for handling data gaps in Adi Ha’s mere 7 years of (unreliable) precipitation data. Plastic rain gauges were installed in each of the 21 pilot participants’ fields, evenly distributed geographically around the village. Farmers collected—and are still collecting—rainfall data to help the technical teams better understand the microclimates across the village. Farmers’ readings proved reliable; average readings were only 2.2% different from the automated rain gauge placed in the village’s tree nursery. This example demonstrates that poor communities can be meaningfully involved in the technical aspects of an adaptation project. Moreover, by definition, data obtained from multiple sources should be more scientifically reliable than data from a single source. •

Indicating contract preferences: Farmers participated in numerous consultations regarding various insurance product options (i.e. types of cover, crop selection, coverage levels and frequency of payouts). In particular, the Community Pilot Leaders forecasted farmer reactions to product features and roadblocks to insurance adoption. For instance, they indicated that the local farmers’ cooperative must be included in the product’s distribution strategy in order to win support from village leaders. Furthermore, HARITA’s technical advisors assumed that farmers would not trust what they cannot see. However, farmers in focus group discussions explained that their neighbours would probably not resist a satellite-based insurance product.

•

Informing and finally selecting the product distribution model: The HARITA partner organizations generated four, viable options for the product’s distribution model. Ultimately, farmers on the Design Team recommended the value chain configuration that was implemented and supported in community– wide consultations.

•

Protecting clients: Given the relatively sophisticated nature of weather index insurance and the illiteracy of most farmers in Adi Ha, the Design Team established a grievance process to handle farmer complaints or questions. The process was documented in a pictorial diagram so that farmers who cannot read could refer to it later when needed. Farmers also elected two of their relatively educated and respected peers to serve as ombudsmen. These two men were trained in the grievance process so they could shepherd complaints through the system, although none have been forwarded to date.

Benefits of Community-Based Adaptation Participatory methods are resource intensive—the question then is: are they worthwhile? Uptake of the Wahisna1 insurance package offered by Nyala Insurance Company in late May 2009 constituted a major test of HARITA’s popular education and outreach efforts, as well as the attractiveness of the IFW model. There is reason to believe the participatory method produced tangible and substantial results, as evidenced in the strong uptake of Wahisna. Over the course of two days, approximately 600 farmers attended the enrolment activities, and 200 households signed up for the package, representing roughly 20% of the village. An impressive 38% of enrolees were female-headed households (recognized as the poorest of the productive poor) and 65% of enrolees were chronically food insecure participants of the PSNP. By definition, these two groups constitute the most vulnerable farmers in Adi Ha. (See Fig. 2.) Labour-paying farmers earned their premium by implementing the risk reduction projects identified through the PCVA process mainly on their farms. At the outset of this project, the received wisdom was that for chronically food insecure farmers, agricultural risks were nearly uninsurable, but HARITA’s experience is a direct challenge to this notion. It is possible that the high degree of long-term trust between the community and the project partners led to the high rate of uptake. The project partners believe that trust was certainly an important factor in gaining entrée and cooperation at the outset. More important, however, is probably the fact that the product was simply quite attractive to poor farmers. Allowing farmers to pay in labour for benefits that extend beyond insurance alone significantly increased the value proposition while also decreasing the opportunity cost of the premium. Labour is the most obvious in-kind currency available to low-income households. HARITA’s experience is consistent with a study by Asfaw and Braun who examined poor Ethiopians’ willingness to pay for health insurance. Interestingly, they found that for a premium level of ETB 3 per month per household, twice the 1

Meaning “protection” in Tigrinya. The product name and logo was chosen by farmers.

First Draft Jan. 2010, Not Suitable for Citation or Circulation number of respondents were willing to purchase community-based health insurance if premiums could be paid scheme in terms of labor than in cash. On average, the poor were willing to contribute up to ETB 14 worth of labour, nearly five times more (Asfaw and Braun, 2004). It is important to note that over time, as livelihoods improve and farmers graduate from the PSNP, they should be able to pay in cash and become candidates for the fully commercial insurance market. But with 8 million farmers participating in the PSNP, the IFW model constitutes a critical platform for introducing the country’s farmers to the idea of insurance in a way that helps them eventually afford insurance with their own cash.

Lead by Nicole Peterson, a team at Columbia University and individual researchers in Ethiopia conducted a follow-up study of enrolment with 200 farmers (both buyers and non-buyers, 31.5% of whom were women). Among the report’s preliminary findings were: • • •

• •

Insurance clients were more likely to be female, younger, and PSNP participants, compared with nonbuyers and the general population. They were also more likely to have less land and grow less teff. These results demonstrate that the HARITA approach successfully targeted poorer farmers. Community trainings on insurance, especially the one conducted through popular theatre, were highly correlated with farmers being able to answer questions about insurance correctly. Farmers said the most common reason for non-purchase was simply being unaware of the opportunity to buy (40.7%), while 12.8% said they did not understand the product. Just over 30% said they had no reason not to buy. These findings suggest initial successes as well as areas for improvement in outreach and education. Survey participants were satisfied with the product’s current price (93%), period of coverage (95.6%), crop used (82.5%), satellite data use (89.5%), and complaint process (92.1%). Farmers stated they are more likely to buy insurance if it is connected to loans and improved loan terms. In the pilot’s first year, Wahisna was not tied to a custom, loan package, although this is envisioned in future years. The study revealed that farmers who plan to take loans next year expect a higher average loan amount than this year (US$122 versus US$115). Plans for next year also show a greater interest in loans for agricultural inputs (20% higher) and investing in other businesses (68% higher, albeit from a low initial basis). (Peterson, 2009)

HARITA in the Broader Context Below, we discuss a few ways that HARITA holds lessons and potential impact on a broader scale beyond Adi Ha. As a micro-insurance and disaster risk management project, HARITA focuses on helping farmers deal with the small-scale unreported disasters described in the background section. Again, these disasters do not attract much attention, but over enough years, they erode coping capacity. Arguably, the infamous famine of 1984 (where up to a million Ethiopians died) grew partially out of a series of small droughts that eroded coping capacity among rural poor communities (Fraser, 2007). While it is important to explore how reaching scale can help countries like Ethiopia cope with large, catastrophic disasters, helping farmers deal with small shocks along

First Draft Jan. 2010, Not Suitable for Citation or Circulation the way is just as important as helping them deal with catastrophic regional and national emergencies (which are often a reflection of weakness in the mechanisms to deal with minor shocks). Linking Community-based Adaptation to Regional and National Level Interventions Community-based adaptation projects like HARITA can be valuable sources of information about climate impacts at the local level. When replicated and scaled up, community-sourced data can be gathered systematically to ensure that adaptation across the country is coherent and not a mere patchwork quilt of interventions that, at best, fail to work synergistically, and at worst, stand at cross purposes. In the end, however, limited human resources and time probably means that only a handful of communities will participate fully in designing their own adaptation projects. While it is important to acknowledge the idiosyncratic nature of individual village community-risk assessments and solutions, van Aalst et al. (2008) suggests that policymakers can assess ‘archetypal livelihoods’, which are closely linked with ‘economic activity combinations’ in the target region by community risk assessments. In this way, PCVAs can serve as a sampling mechanism of climate impacts that enables decision-makers to analyze the potential disruptions of those archetypal livelihoods or activities in communities that are unable to participate in adaptation planning. Replicability and Scalability at the National/Global Level The HARITA team has been exploring options for gradual expansion of the pilot to test its replicability in Tigray and other parts of Ethiopia. In 2009–2010, HARITA’s partners will pursue a modest replication of the model in up to five additional villages in Tigray and possibly Amhara, another drought-prone area of the country. Expansion to new regions in the country will eventually diversify the pilot’s risk pool and strengthen the risk management program. HARITA’s approach is highly replicable because it can accommodate most or all types of community-level variability in Ethiopia. However, creating access to insurance on a sustainable basis—at scale—is an enormous challenge. HARITA’s project partners envision leveraging the PSNP to reduce the costs of product distribution, management and oversight, and to ease the logistical hurdle of outreach to remote villages. The PSNP has already established a financial system to distribute cash at the household level, so micro-insurance payments can travel along the same lines. While the PSNP is the most obvious platform for scale, it would also be possible to replicate through the country’s existing network of microfinance institutions or cooperatives. In terms of scaling outside of Ethiopia, although the PSNP is unique, many countries have food-for-work programs as well as ‘conditional-cash-transfer’ programs that reach hundreds of millions of poor people (Fiszbein 2009). It is important to clarify that HARITA is designed to help PSNP participants graduate out of annual assistance and into self-sufficiency, and ability to afford insurance in the commercial market. Yet, so long as there are farmers who need help to purchase insurance, HARITA requires a perennial source of financing. Fortunately, many international actors would like to facilitate insurance access in developing countries. Oxfam is actively researching opportunities among three major sources of premium financing—private donors, the public sector and the commercial carbon markets. True long-term sustainability will, however, require more than financing. It will also require good program governance, transparency, accountability, a supporting regulatory framework, knowledge networks and steady information flows from the local to global level. Fortunately HARITA enjoys cooperation with numerous influential early adopters (farmers and organizations locally and abroad). Also the Government of Ethiopia has named weather index insurance as one of its top priorities in its National Adaptation Plan of Action (citation). And finally, the PSNP, while not perfect, is still widely viewed as a relatively well-governed program. A survey of independent assessments from 2006 to 2009 revealed no systemic corruption2 and a high degree of satisfaction among participants (Sharp et al., 2006; Devereux et al., 2008; Food Security Programme Review, 2009). HARITA seeks to sustain and build upon these good practices. Season End Results and Future Areas of Research [Section incomplete section] HARITA is a program in its experimental stage. In many regards, the first phase of the pilot was very successful. Many questions and challenges, however, remain. First is the highly technical nature of the project and the need to build capacity among partner organizations quickly. Groups like the World Bank are making important investments in helping local level actors participate in

2

In late 2009, opposition groups allege that the PSNP is being used as a political tool against critics of the government. These allegations are currently being investigated.

First Draft Jan. 2010, Not Suitable for Citation or Circulation weather index insurance markets. IRI is also developing an online tool that will allow domestic insurers to design their own weather index insurance contracts. Second, it is still unclear who will finance the IFW model at scale. The disappointing outcome of the climate talks in Copenhagen suggest that a large international adaptation fund is a long way off, so financing may have to come from private donors or other public funds. [Insert estimates of HARITA operating costs at scale.] Also, it remains to be seen whether HARITA will be successful in other areas of the country and when the project partners do not make as large investments in engaging individual villages….

Conclusion HARITA provides important lessons about micro-insurance as a tool for adaptation. First, the project suggests that prospective clients should be deeply involved in design, not only in order to improve products, but also to help identify areas of potential confusion and cultural biases against product uptake. Based on this information, educators can help farmers make informed decisions about insurance purchase. In addition, HARITA suggests it is incorrect to assume that low uptake rates for micro-insurance are a sign that poor people do not want it. In Portfolios of the Poor, Morduch et al. suggest a number of reasons microinsurance is very difficult to sell. Among these reasons are the fact that: •

•

Insurance requires poor people to decide in advance which of the many risks they face they should cover. By combining risk reduction with risk transfer, HARITA allows clients potentially to cover multiple risks through a combination of risk reduction and insurance. Further, the PCVA helps identify the most important risks to insure. Most poor people already have a safety net: While this statement is true, the PCVA allows potential insurers to understand how micro-insurance can complement and spread risk beyond the existing coping mechanisms (e.g. self-insurance, community-risk sharing, risk reduction, and the PSNP). In this way, community-designed insurance ensures that risk transfer is filling the gaps in the current safety net.

HARITA’s project partners believe that uptake has been low largely because products have been unattractive and centred on the interests of lenders and insurers. When designed around the needs and desires of farmers—while still balancing those of financial institutions—uptake among even very poor people can be substantial. HARITA has found fertile terrain in promoting the resiliency of smallholder farmers to climate stress and disaster risks through community-based efforts guided by regional, national, and global adaptation strategies that are informed by global science and backed by insurance. While the HARITA model must still be evaluated to determine its effectiveness and impact on poor people, the pilot’s results to date suggest that farmers can create more stable, resilient communities if they have access to the right tools. It also suggests that poor people can play extremely valuable and unique roles in identifying successful and robust climate change adaptation strategies. Acronyms AF- Adaptation Fund CBE- Commercial Bank of Ethiopia DECSI- Dedebit Credit and Savings Institution DRR- Disaster Risk Reduction GGLM- Group Guarantee Lending Model HARITA- Horn of Africa Risk Transfer for Adaptation IFW- Insurance for Work IPCC - Intergovernmental Panel on Climate Change IRI- International Research Institute PCVA- Participatory Capacity and Vulnerability Assessment PSNP- Productive Safety Net Program REST- Relief Society of Tigray References

First Draft Jan. 2010, Not Suitable for Citation or Circulation Admassie, A. (2004) ‘A review of the performance of agricultural finance in Ethiopia: pre and post reform periods’, paper presented at the International Conference on Macroeconomic, Agrarian Change and Poverty Reduction, Addis Ababa, Ethiopia Amha, W. (2000) ‘Review of microfinance industry in Ethiopia: regulatory framework and performance’, Occasional Paper, no 2, Association of Ethiopian Microfinance Institutions Ayers, J. and Forsyth, T. ‘Community-based adaptation to climate change: strengthening resilience through development’, www.allbusiness.com/science-technology/earth-atmospheric-science-climatology/125791721.html, accessed 1 September 2009 Borchgrevink, A., Woldehanna, T., Ageba, G. and Teshome, W. (2005) ‘Marginalized groups, credit and empowerment: the case of Dedebit Credit and Savings Institution (DECSI) of Tigray, Ethiopia’ Occasional Paper, no 14. Addis Ababa, Association of Ethiopian Microfinance Institutions Botzen, W. J. W. and van den Bergh, J. C. J. M. (2008) ‘Insurance against climate change and flooding in the Netherlands: present, future, and comparison with other countries.’ An International Journal, vol 28, no 2, pp 413-426 Chamberlain, D. and Smith, A. (2009) ‘Opportunities and challenges for micro-insurance in Ethiopia: an analysis of the supply-side and regulatory environments’, Center for Financial Regulation and Inclusion (CENFRI), Johannesburg CSA (Ethiopian Central Statistical Agency) (2007) ‘Atlas of selected welfare indicators, Statistical Survey 2007, Addis Ababa: Federal Democratic Republic of Ethiopia. Dessai, S., Hulme, M., Lempert, R. and Pielke, R. (2009) ‘Climate prediction: a limit to adaptation’, in W. N. Adger, I. Lorenzoni and K. O'Brien (eds) Living with Climate Change: Are There Limits to Adaptation (forthcoming) Devereux, S., Sabates-Wheeler, R., Slater, R., Tefera, M., Teshome, A., & Brown, T. (2008) ‘Ethiopia's Productive Safety Net Programme: 2008 PSNP assessment summary report’ unpublished first draft manuscript DFID Environment Canada (2007) ‘Downscaling’, Canadian Climate Change Scenarios Network www.cccsn.ca/Help_and_Contact/Downscaling-e.html, accessed 1 September 2009 Food Security Programme Review (2009) ‘Detailed Report for the Productive Safety Net Programme’, Zero Draft, Unpublished Fraser, E. (2007) ‘Travelling in antique lands: using past famines to develop an adaptability/resilience framework to identify food systems vulnerable to climate change’, Climate Change, pp. 495-514, Springer Science & Business Media Hellmuth, M.E., Osgood, D.E., Hess, U, Moorhead, A. and Bhojwani, A. (eds) (2009). ‘Index insurance and climate risk: prospects for development and disaster management’ Climate and Society No. 2, International Research Institute for Climate and Society (IRI), Columbia University: New York. Hoeppe, P. and Gurenko, E. N. (2006) ‘Scientific and economic rationales for innovative climate insurance solutions.’ Climate Policy, vol 6, no 6, pp 607-620 ILO (2009) ‘The social economy – Africa’s response to the global crisis’ IMF (International Monetary Fund) (2007) Country report for Ethiopia IPCC (Intergovernmental Panel on Climate Change) (2007) ‘Working Group II, Fourth Assessment Report’, www.ipcc-wg2.org, accessed 15 January 2008

First Draft Jan. 2010, Not Suitable for Citation or Circulation Ketema, S. (1997) Eragrostis tef (Zucc.) Trotter, International Plant Genetics Research Institute, Addis Ababa, Ethiopia Marengo, J.A., Hewitson, B., Tadross, M. and Zermoglio, M. F. (2008) ‘Climate modelling and downscaling under the Nairobi Workprogramme’ UNFCCC in-session workshop on climate modelling, scenarios and downscaling under the NWP on impacts, Bonn, Germany, http://unfccc.int/files/adaptation/application/pdf/ipcc.pdf, accessed 1 September 2009 Microned (2008) Ethiopia: microfinance country scan, Oxfam Novib, The Hague, Netherlands Mohner, A & Klein, R. (2007) The global environment facility: funding for adaptation or adapting for funds? Peterson, N (2009) ‘Livelihoods, coping, and micro-insurance in Adi Ha, Tigray State, Ethiopia: a report on surveys and interviews conducted in June and August 2009’ unpublished report, Columbia University: New York Osgood, D. E., Suarez, P., Hansen, B., Carriquiry, M. and Mishra, A. (2008) ‘Integrating Seasonal Forecasts and Insurance for Adaptation among Subsistence Farmers: The Case of Malawi’, Policy Research Working Paper 4651, World Bank, Washington DC Rosenhead, J., Elton, M. and Gupta, S.K. 1972 ‘Robustness and optimality as criteria for strategic decisions’ Operational Research Quarterly (1970-1977), vol 23, no 4, pp 413-431 Sharp, K., Brown, T., & Teshome, A. (2006) ‘Targeting Ethiopia's Productive Safety Net Programme (PSNP)’, IDL Group and Overseas Development Institute, www.odi.org.uk/resources/download/3035.pdf, accessed 15 December 2008 Tadesse, M. and Victor, M. (2009) ‘Estimating the demand for micro-insurance in Ethiopia’ ILO, unpublished report United Nations Development Programme (2007) ‘Fighting climate change: human solidarity in a divided world’ Human Development Report 2007-08, Palgrave McMillan, New York Walling, L. J. (undated) ‘Continuing the legacy of participatory planning in climate change adaptation planning initiatives in the Caribbean’, www.sidsnet.org/docshare/LWallingPresentation.pdf, accessed 1 September 2009 World Bank (2006) ‘Project appraisal document: financial sector capacity building programme’. Report no. 36272 – ET. Van Aalst, M.Cannon T. and Burton, I. (2008) ‘Community level adaptation to climate change: the potential role of participatory community risk assessment’ Global Environmental Change, vol 18, no 1, pp. 165-179