UC Davis

Farming has an elevated rate of suicide in the U.S. and elsewhere, which draws attention to the mental well-being of farmers and other agricultural workers. There is a rich empirical and conceptual literature exploring the reasons behind the high rate of farmer suicide. Yet little rigorous study addresses the determinants of farmer suicide in the U.S. This study explores a number of hypotheses as to causal factors explaining the high farmer suicide rate, including weather factors (e.g., extreme temperatures, variable precipitation) and economic factors such as price and income volatility. A theoretical model is developed linking the weather and economic factors to a farmer’s suicide decision. A farmer chooses between work effort and leisure, a level of consumption, levels of production inputs, and the amount of savings to maximize her utility given her budget constraint and time endowment. In a static model, a bad weather shock can have a "snowballing" effect that reduces farm income and also reduces utility from leisure and health through the work-leisure trade-off. Severe adverse weather outcomes can possibly diminish a farmer's utility and cause a spontaneous decision to commit suicide in extreme cases. Based on the model, we hypothesize that the marginal effect of harmful weather on farmer suicide is positive. Successive realization of bad weather and chronic poor economic conditions are positively associated with farmer suicide.

We combine the CDC nonpublic vital statistics, PRISM daily weather data, and the USDA National Agricultural Statistics Service data as a county-year panel to estimate the marginal effects of weather and economic factors on farmer suicides. Empirical analysis, based upon a Poisson regression model with agricultural district fixed effects, shows extreme heat is positively associated with farmer suicides. There is no clear evidence of the effect of precipitation on farmer suicide in counties without irrigation. Chronic, not idiosyncratic, poor economic conditions induce farmer suicide. The results are robust to alternative specifications including or excluding year fixed effects.