Previous research has found that on-farm income variability helps determine off-farm labor supply. However, unobserved heterogeneity of farms or regions may have biased earlier results. In this study, we use an exogenous increase in Federal crop insurance subsidies as a natural experiment to identify the importance of risk in off-farm labor supply. The subsidy increases induced greater participation in crop insurance programs and thereby reduced farmers' financial risks. By merging county-level crop insurance participation data with farm-level Agricultural Census data from 1992 and 1997 we can compare the off-farm labor decisions of individual farms before and after the subsidy and thereby control for unobserved heterogeneity. Unlike previous studies, we find that on-farm risk does not affect the labor allocation decisions of farm households.