With a looming farm bill debate, crop insurance stands as the largest single component of the crop farm safety net. The program provides risk protection from low yield or revenue in return for a premium that producers pay.
These premiums were subsidized by the USDA on average about 63% across all programs in 2016. The total cost of the subsidy in 2016 was approximately $5.85 Billion. Figure 1 provides a bit of historical perspective on acres insured and total crop insurance subsidy.
Beginning in the early 1990s a series of legislative changes increased subsidy levels and acreage insured has trended up as well. We note that recent declines in subsidy primarily result from reduced crop value as prices decline from historic highs.
In the next farm bill debate the amount of subsidy for crop insurance is likely to be a topic of discussion. A frequent question posed to economist sounds something like this, “If we change the subsidy structure what will happen to crop insurance participation.”
This question has been asked and answered numerous times. In most, but not all studies, the conclusion has been that crop insurance demand is inelastic. That is, the percent change in participation will be less than a percentage change in subsidy. However, many of those studies are older and may reflect a different era of crop insurance.
In this report, we examine some key data associated with RMA corn and soybean program participation. We do not estimate an elasticity, but rather show evidence of a consistent pattern in in how much farmers are willing to pay for crop insurance. We use the dramatic changes in crop value between 2011 and 2016 and variation in riskiness across regions to show a remarkable constant in crop insurance demand.
We find that across periods of high and low crop value and across regions of low and high risk – corn and soybean farmers are willing to pay out-of-pocket no more than four percent of the expected value of the crop. If this is true, it has implications for the demand for crop insurance when subsidy is changed.
We do not provide a theoretical explanation for this finding but believe it may be tied to the degree of risk aversion and farmer budget constraints.
We begin by examining variation across region in the base county premium rate. The maps in figure 2 and 3 show wide variation the level of yield risk across growing regions. While most producers purchase revenue insurance, regional variation in premium rates are largely driven by yield risk.
Next we examine the amount of insurance chosen by corn and soybean producers. Figures 4 and 5 reflect the acre-weighted average coverage level chosen in each county. We use coverage level to represent the amount of insurance chosen by those who participate in the program.
When figures 4 and 5 are compared to figures 2 and 3, a pattern begins to emerge. Areas of the country with lower per-acre base premium rates also tend to purchase higher coverage levels than areas with higher base premium rates.
Figures 6 and 7 show the 2016 average producer paid premium per acre for corn and soybeans by county. Note that producer premium is a function of the coverage level, rate, and value of the crops. In general, low risk-high yield regions pay similar premiums per acre as higher risk-lower yielding regions.
Having said that the lower coverage levels chosen in many higher risk regions results in lower producer paid premium per acre. Finally, the maps show that producer paid premium for soybeans are generally lower than for corn. This is in part due to lower per acre expected crop value.
Figures 8 and 9 divides the average producer paid premium by the insured value of the crop to compute the percentage of expected crop value farmers opt to pay in producer paid premium. This reveals our primary finding.
As can be seen in both figures, the majority of counties are shown to pay between one and four percent of the value the crop in 2016. Thus, we find that farmers appear to be willing to pay a premium of about four percent of crop value and no more.
To test the robustness of our results in 2016 we also conduct the same analysis using data from 2011. These results are shown in figures 10 and 11. Note that higher crop price in 2011 resulted in expected crop revenue more than 30% higher in that year than in 2016. However the premium paid as a percent of crop value maps look quite similar to that of 2016.
While we find quite robust results, it is not clear why producers seem to spend such a constant percent of crop value on crop insurance. Most likely it is related to the out of pocket cost associated with this program and the perceived benefits. We suggest that models of insurance demand consider the possibility of a budget constraint on crop insurance demand.
Ultimately, the consistency of these results suggests that if crop insurance costs rose past four percent of expected crop value, the producers would reduce insurance expenditure – most likely by reducing coverage levels.