IT IS TOO LATE TO REGRET AND TAKE RISK: FARMERS’ ADOPTION DECISION FOR STALL-FEEDING (SF) IN TIGRAY, ETHIOPIA

Abstract

Research background: Despite a growing interest in the role of time and risk preferences in explaining technology adoption, empirical studies that investigate this behavior are scant. Numerous studies have attempted to identify the determinants of adoption of new technologies. However, those studies failed to capture the duration of time farmers will take to adopt a given technology using a proper model such as duration model. Purpose of the article : This study developed a technology adoption theoretical model that incorporates time and risk preferences in addition to household level characteristics. Using this model, we tested whether impatience and risk aversion delays or expedites stall-feeding adoption or not. Methods: Using cross sectional data of 518 sample farmers from Ethiopia, both semi-parametric (Cox PH) and parametric (Weibull PH and Weibull AFT) models have been applied to estimate the conditional probability of SF adoption. This enables us to convey information not only on why a farmer adopted, but also on the timing of the adoption decision. Findings : As expected, our results indeed show that, the time of stall-feeding adoption increases with discount rate. Individuals who are patient and risk loving adopt stall-feeding sooner than individuals who are impatient and risk averse. Likewise, farmers who are more less-averse adopt SF technology latter compared to farmers who are risk neutral. The estimated results also revealed that economic incentives (i.e. prices) was found to be the most important determinants of the time farmers wait before adopting new technologies. While higher output price significantly accelerates SF adoption, higher input price decelerates the adoption period. The expected milk yield (first moment) had a positive significant effect on the adoption decision, indicating that higher expected mean, on average decreases farmers’ time to adopt. Findings in this study suggest that, land and labor endowment shorten the time to adopt SF. However, market distance delay the adoption of SF. Moreover, access to information, education of household head, breed cow ownership, and location of the farmer, accelerated the likelihood of early SF adoption. To the best knowledge of the authors, this is one of the first adoption studies to have incorporated time and risk preference in its parametric and semi-parametric adoption analysis.