Analyzing ex-ante agroforestry adoption decisions with attribute based choice experiments
Although many cases of successful agroforestry extension efforts exist (for examples, see Chapter 2), all too often attempts to promote agroforestry have resulted in low adoption rates, with farmers reluctant to adopt new or improved agroforestry systems or abandoning agroforestry shortly after establishment. As a result, the recent increase in research on the adoption of agroforestry innovations has been motivated largely by the perceived gaps between advances in agroforestry science and extension (Mercer, in press). The theoretical and empirical literature on adoption of agroforestry innovations has been reviewed by Pattanayak, Mercer, Sills, and Yang (2003) and Mercer (in press). Significant progress has been made, especially in using binary choice regression models for ex-post analyses to examine how past adoption decisions are correlated with variables describing farmers, their farms, demographics and socio-economic conditions. These ex-post analyses have been useful for increasing our understanding of who adopts first, identifying communities and households to target as potential early adopters, and developing policies to promote agroforestry. However, the ex-post, binary choice regression studies have contributed little to the problem of designing agroforestry systems that appeal to potential adopters because they are not able to examine how farmer preferences vary for different combinations of characteristics of agroforestry alternatives.