Production Risk in a Subsistence Agriculture

Purpose: In this article we illustrate the importance of understanding the risk profiles of new technologies, in addition to the changes in productivity, to be able to determine strategies for agricultural development. Design/methodology/approach: The analysis is based on data obtained from a 2002 survey of subsistence farmers in the Kilimanjaro region of Tanzania, and a Just and Pope (1978) framework is used for modeling risk. Findings: We find that even if extension services do not increase the mean production, it may reduce production risk. Practical implication and originality/value: During the past decades, agricultural extension and subsidized conventional inputs such as high-yielding seed varieties, fertilizer and pesticides, have become important components of agricultural aid programs in developing countries. However, outcomes of this type of aid are somewhat ambiguous, and many donor countries have reduced their support in response. For the most part, evaluation of these programs employs total factor productivity analysis to estimate the changes in productivity resulting from investment in aid programs. However, risk-averse, small-scale farmers will consider both the variance in output and the expected mean. They may therefore choose input levels that differ from the optimal input levels of risk-neutral producers, who consider only the expected mean. Programs can therefore have a positive effect because they reduce risk, even if the direct impact on production is limited.