Discussion Paper No. 120 of 2010 on Production Risk and Farm Technology Adoption in Rain-Fed Maize Production in Semi-Arid Lands of Kenya

Abstract

This study provides empirical evidence of technology adoption and the effects of production risk on it among smallholder farmers, using cross-sectional data collected from semi-arid districts in Kenya (Machakos and Taita Taveta). Several approaches are used: probit approach to estimate the effects of farm and household level variables on adoption of terracing as a soil and water management technology, and instrumental variable Two Stage Least Squares method to estimate productivity, and moment-based approach to capture variability risk (variance/second moment of maize yield) and downside risk (skewness/third moment of maize yield). Variability risk is also used together with farm and household level variables to determine the impact of production risk on technology adoption decisions. The results show that household size; and institutional factors such as social capital, security of land tenure and the slope of land are important in increasing the probability of adoption of terracing technology. Expected yield as shown by the first moment has a positive effect on adoption of fertilizer. Variability of yield has a positive effect on manure use, and a negative effect on fertilizer application. High probability of crop failure (downside risk) increases the possibility of terracing and manure application by farmers, and reduces the possibility of fertilizer application. Other important factors that influence technology adoption decisions are region and distance from household to the farm. Productivity is found to be positively influenced by fertilizer adoption, manure and labour application, and soil and water management, while land size is found to negatively influence productivity. These results have important policy implications, such as technology adoption which ought to be encouraged because it increases productivity. However, these technologies are associated with risks against which farmers should be cushioned, if they are to embrace them. Also, institutional, household, farm-level and regional factors are important in technology adoption, and any policy aimed at enhancing technology uptake must carefully consider them. Since regions respond differently to different technologies, it may be important to develop region-specific policies rather than rely on toolbox approach.