Quantifying Loss of Benefits from Poor Governance of Climate Change Adaptation Projects: A Discrete Choice Experiment with Farmers in Kenya

Highlights

• We assess Kenyan farmers' preferences for different governing organizations to implement climate change adaptation measures.

• We apply choice experiments to assess governance of sand storage dam construction as adaptation measure.

• We find farmers prefer NGOs and farmer networks substantially over governmental bodies to govern dam construction.

• We quantify benefit losses that farmers incur due to poor governance of dam construction at the regional level.

• We draw attention to the careful selection of payment vehicle for choice experiment studies in developing countries.

Abstract

Climate change impacts pose a great challenge to agriculture in sub-Saharan Africa as droughts become more frequent and more severe. A major roadblock to implementing climate change adaptation measures is poor governance. Given their experience with governing organizations, farmers are highly suitable to assess the appropriateness of different governing organizations to implement adaptation measures on the ground. We surveyed 283 farmers in Makueni County in Kenya applying the choice experiment method to assess their preferences in relation to different attributes of a sand storage dam project – including the organization governing the dam construction. We find that farmers prefer an NGO as the governing organization, followed closely by a farmer network and, with some distance, a government institution. For the whole of Makueni County, we find that benefits of $ 320,426 are lost if farmer networks are the governing organizations instead of NGOs and $ 1,779,596 if government institutions govern the dam construction instead of NGOs. On a methodological level, our study contributes to improving the application of choice experiments in developing countries as it draws attention to the importance of carefully selecting the payment vehicle for successful project implementation.

Introduction

Climate change poses a major threat to agriculture in sub-Saharan Africa. Over the past 50 years, climate change-induced drought frequency and intensity have increased (Detges, 2017), exacerbating water scarcity problems affecting both surface and groundwater resources (Mishra and Singh, 2011). Further, rising temperatures have disrupted crop productivity and affected water availability at the important stages of crop growth (Zamasiya et al., 2017). The vegetation cover has increasingly died off, reducing species richness and plant cover across the region (Epule et al., 2017). Consequently, it is estimated that about half of the 1.033 billion people in sub-Saharan Africa are facing food insecurity associated with global climate change (Epule et al., 2017).

Climate change adaptation is one of the policy options supported by the United Nations Framework Convention on Climate Change (UNFCCC) to help developing countries reduce the negative impacts of climate change on the agricultural sector (Deressa et al., 2009; UNFCC report, 2015). Climate change adaptation measures can be classified as either private or public (IPCC TAR, 2001; Milman and Warner, 2016). Private adaptation measures are those implemented by individual farmers or small groups of farmers to address adaptation needs that are specific to them (Milman and Warner, 2016). They include small scale irrigation of crops when rain fails, crop and livestock diversification, use of soil and water conservation techniques and changes in planting dates, among others (Francisco et al., 2010; Bryan et al., 2013). Public adaptation measures are actions on a larger scale that are implemented by government institutions through public and private partnerships or otherwise coordinated action at the local level (Milman and Warner, 2016). Examples of public adaptation measures include prevention of flood and flood damage on crop land through construction of drainage pumping systems and canals, construction of water reservoirs (dykes and dams) for irrigation, and the development of hazard maps on floods and drought impacts (UNFCC, 2006; NPAICC, 2015).

A key challenge when implementing public adaptation measures is the selection of the appropriate governing organization to coordinate efforts and administer the available adaptation funds, which may come from the government, international donors or farmer networks. In principle, it would be the task of local or regional governments to provide a governance structure for the implementation of adaptation measures. However, in developing countries in general and sub-Saharan Africa in particular, there is a substantial risk that funds channeled through the government may not reach those who need them due to leakages resulting from corruption and excessive administrative costs (Besley and Ghatak, 2017).

Two main alternatives to government organizations have been proposed to support community-based projects in the region (Besley and Ghatak, 2017). First, NGOs as self-governing nonprofit organizations whose aim is to improve the livelihoods of vulnerable people in rural areas. Their general objective is to empower local communities by initiating people-centered projects that require local communities to participate in the design and implementation processes (Morgan, 2016). Second, farmer networks composed of groups of farmers who come together to share ideas on how they themselves can provide and/or participate in the provision of public goods, often also with support from governments and NGOs. Farmer networks are common in sub-Saharan Africa and they undertake different farm activities ranging from crop production and marketing of produce to community development projects (Fischer and Qaim, 2014).

Several studies have analyzed corruption and governance in relation to the implementation of climate change related projects in developing countries. For example, Robinson (2017) used a qualitative approach to understand climate change adaptation mainstreaming in small island developing countries. He found that bad governance, corruption and the selfishness of prominent leaders who have a mandate to make adaptation decisions hinder mainstreaming of adaptation. In a survey of empirical cases Shackleton et al. (2015) found that corruption and elite capture at the local level is one of the main barriers to the implementation of adaptation measures in sub-Saharan Africa. Pueyo (2018) analyzed the constraints of renewable energy investment in sub-Saharan Africa using the Growth Diagnostics framework and identified poor governance as a key constraint to the growth of renewable energy, especially in Kenya. Furthermore, a qualitative study by Binswanger-Mkhize and McCalla (2010) indicates that corruption and poor governance lead to loss of resources intended for agricultural and rural development projects before they reach service providers.

However, there is a lack of studies that quantitatively evaluate different governing organizations for adaptation measures. In principle, methods to assess willingness-to-pay for environmental goods are suitable for such an endeavor (for an overview of such methods see Pethig, 1994). Given that potential case studies where revealed preference approaches can be applied to assess different governing organizations for adaptation measures are rare, the application of stated preference methods seems appropriate. The contingent valuation method (CVM), as one important stated preference method, has been criticized for causing strategic bias in “yea-saying” problem and attracting high protest bids among respondents (Bienabe and Hearne, 2006) among other shortcomings. Choice experiments (CE) have been proposed as an alternative stated preference method. CE are not only better suited to deal with the methodological problems of the CVM but they are also able to assess respondents' willingness-to-pay for different attributes of a good and not only for the good per se as is the case with CVM (Adamowicz et al., 1998).

CE are based on Lancaster's model of consumer behavior, in which consumers derive utility from the attributes of a good and not the good itself (Lancaster, 1966). Accordingly, in CE respondents are asked to compare and value the provision of a specific good based on different characteristics of its attributes (Hensher, 1994; Rai, 2012). In this context, a climate change adaptation measure can be considered an environmental good and the organization that governs its provision one of its attributes. Hence, applying CE to assess climate change adaptation measures reduces the shortcomings of CVM, and allows us to receive a comprehensive picture of how respondents evaluate different attributes of the measure including the organization that governs its provision.

CE have been widely employed to evaluate environmental goods in developed countries (e.g. Kuhfuss et al., 2016; Markova-Nenova and Wätzold, 2017; Vaissière et al., 2018). On a much smaller scale, but increasingly, they have also been applied in developing countries (Birol and Das, 2010; Ndunda and Mungatana, 2013; Osiolo, 2017). Their increasing use in developing countries is accompanied by a debate on the specific challenges that arise when choice experiments are employed in such countries (Do and Bennett, 2009; Birol and Das, 2010; Kahn et al., 2017). For instance, CE application in developing countries is characterized by high protest responses due to low income among rural communities (Rai et al., 2015), and the process of collecting choice data is in many regions limited to face-to-face interviews as internet access is limited by low income and education levels among respondents (Kahn et al., 2017).

We employed the choice experiment method to survey 283 farmers to evaluate different hypothetical scenarios of sand storage dam construction as an adaptation measure in Makueni County in Kenya. Attributes of the dam include dam wall height and volume of water harvested, type of pump used to distribute water, number of Grevillea robusta tree rows planted to prevent silting, amount of labor time in weeks farmers are willing to contribute to dam construction and – most important in the context of this paper – the governing organization that administers and organizes the implementation of the dam project. Our key finding is that farmers prefer NGOs slightly over farmer networks and substantially over local governments as the governing organization for the implementation of adaptation measures. This finding is important in two ways.

First, we provide highly policy-relevant information on how farmers evaluate the different governing organizations and quantitatively assess the loss of benefits that arise if government institutions or, much less so, farmer networks instead of NGOs administer adaptation measures. The information is valuable for policy implementation as it demonstrates the importance of selecting appropriate governing organizations when implementing adaptation measures and the loss of benefits if inappropriate organizations are selected.

Second, from a methodological point of view our results draw attention to a key aspect that needs to be considered when applying choice experiments in developing countries, namely the careful selection of the payment vehicle. Here the “payment vehicle” refers to the mechanism presented in the CE of how the project in question would be financed if it were implemented (Ivehammar, 2009). The choice of payment vehicle and the resulting potential bias that may arise on the willingness to pay for a good has been discussed in the context of developed countries (Bergstrom et al., 2004; Campos et al., 2007; Ivehammar, 2009). However, we did not find studies that address in similar depth the importance of the payment vehicle in developing countries. Instead, studies typically select a specific payment vehicle. Examples include tax increases (Birol and Das, 2010; Ndunda and Mungatana, 2013), increases in product prices (Roessler et al., 2008; Osiolo, 2017), crop yield losses (Waldman et al., 2017), and donations to a nature conservation fund (Randrianarison and Wätzold, 2017). Our key result – willingness to contribute labor time to an environmental project differs depending on whether the government, an NGO or a farmer network governs this project – suggests that the stated willingness to pay of respondents in developing countries would be substantially different depending on whether an environmental project is supposed to be financed through, say, a tax paid to the government or a contribution to a fund administered by an NGO.