Disaster risk reduction measures and farmers choices: a discrete choice experiment in Uganda

Abstract

Climate change induces high and erratic rainfall which triggers landslides and floods. With the increasing population and food needs, households in mountainous, densely populated areas turn fragile ecosystems into farms. This exacerbates landslide and flood risks requiring Disaster Risk Reduction (DRR) measures. Tree planting and diversion channels are among the recommended measures for farmers but their adoption remains low. Current studies assessing barriers to adoption ignore farmers’ opinions regarding the kind of trees or diversion channels preferred. We apply a Discrete Choice Experiment to evaluate how information delivered through videos impacts preferences for the DRR measures. Plot-level data were collected from 319 farmers from Kasese, Bundibugyo, Bushenyi and Buhweju in Uganda – districts prone to landslides and floods. The mixed logit model reveals a general preference for risk-reducing attributes of DRR measures. Using the conditional logit model to analyze split samples reveals that information influences preferences for tree planting, while preferences for diversion channels were hardly changed. Plot characteristics did not strongly explain the differences in preferences. Our study indicates that information specific to DRR measures in extension programmes would increase the adoption of such measures.