全球智库信息集成服务系统

This study provides empirical evidence on the effects of production risk on farm technology adoption among small holder farmers using plot-level data collected from two semi-arid districts inKenya, Machakos, and Taita Taveta. We employed a two-stage approach to estimate a production function, and computed the mean and the production risk factors (both variance and skewness) from aproduction function using Antle’s (1983, 1987) moment-based approach. We then used these moment estimates, together with other household and plot-level characteristics in a pseudo-fixed effect probitmodel to determine the effects of production risk and farm and household-level variables on households’ decisions to adopt different kinds of farm technologies. In our estimations, by means of Mundlak’s approach (1978), we controlled for unobserved heterogeneities that could potentially be correlated to some of the observed explanatory variables and otherwise bias our estimates. We also addressed the potential endogeneity issues in our estimation using a two-stage IV estimation procedure. Our results showed that, among others, yield variability and the risk of crop failures indeed affect technology adoption decisions in low-income, rain-fed agriculture. But, the direction and magnitude of effects depend on the farm technology under consideration. The results explain why poor farm households in rain-fed and risky production environments are reluctant to adopt new farm technologies with potential production gain because, at the same time, they involve enormous downside risks. This result underscores the fact that productivity gains are necessary, but not sufficient,conditions to attract farmers to adopt new technologies and agricultural innovations. Risk implications matter. Technology- and location-specific production-risk coping strategies need to be designed to successfully upscale profitable farm technologies across poor farm households in low income countries.