Unpriced Diversification in US Crop Insurance

As climate change intensifies weather-related risks, agricultural losses become more correlated across space, increasing the importance of well-designed crop insurance. A core design choice in the Federal Crop Insurance Program (FCIP) is unit structure: farmers can insure fields separately or aggregate them into a single policy. Aggregate policies are discounted because aggregation mechanically reduces expected payouts when yields are imperfectly correlated. But the current formula does not account for the endogeneity of correlation: farmers with aggregate policies reduce crop diversification, raising correlation and eroding the actuarial basis for the discount. We show how diversification could be priced directly. Focusing on wheat, we recommend using two observable variables for pricing: the mixture of winter versus spring varieties and the number of fields planted. Using county-level claims data, we estimate the relationship between wheat diversification and within-farm yield correlation, then use simulations calibrated to FCIP contracts to characterize how correlation affects the gap between aggregate and separate payouts. We show that current pricing misses economically meaningful variation in expected payouts: for a four-field farm, moving from a 50/50 winter–spring mixture to full concentration in one variety increases expected aggregate payouts from 65 percent to 81 percent of separate payouts—a 16 percentage point unpriced swing. Incorporating these factors would allow farmers to internalize the costs of their diversification decisions, reducing moral hazard, and improving the program’s fiscal performance.