Technology Adoption Under Variable Weather Conditions: The Case of Rain Water Harvesting in Lake Naivasha Basin, Kenya

This paper applies a parametric econometric duration model (log-logistic) to analyze the duration of adoption of rain water harvesting techniques (RWHTs) among smallholder farmers in the Lake Naivasha basin, Kenya. The study utilizes survey data from 307 farm-households who are dependent on rain-fed agriculture in a region where rainfall has historically been relatively variable. In such circumstances, RWHT helps to stabilize water supply and help farmers manage weather-related risks. The current study seeks to identify constraints to the spread of RWHTs by exploring how rainfall variability influences the timing of decisions to adopt RWHTs alongside other farm-household and spatial characteristics. Empirical results indicate that although rainfall variability is a significant determinant of time to adoption of RWHTs, farmers' sensitivity to rainfall variability have declined over time. Instead, access to informal sources of information has gained importance in adoption of RWHT implying that adoption has become more of an endogenous process of social exchange within communities, and less driven by external natural pressure and persuasion by state agents. Other important factors were: age and education level of household head, domestic water demand, ground water abstraction and the number of previous and expected adopters in the village.