Spanish cereal growers generally contract agricultural insurance policies. Two insurance alternatives were available up to 1999-2000; the first covered crop destruction caused by fire, hailstorm and other extreme events. The second, called yield insurance, covered yield losses due to unfavourable crop conditions, such as drought or excessive heat, in addition to losses covered by the first scheme. As of the 1999-2000 season, growers can contract a new yield policy that expands the coverage offered by the first yield policy, with the difference that the premia are calculated on an individual basis.
This study, which will cover only the first two insurance schemes, seeks to evaluate the impact of crop insurance policies on decisions made by Spanish cereal farmers. Specifically, it will aim to establish whether farmers respond to insurance subsidies by either augmenting yields (changing non-land inputs use) or by changing crop patterns, or both. It also aims to provide an evaluation of farmers’ relative risk aversion coefficients.
Data was obtained from the Spanish National Agency of Agricultural Insurance (ENESA). Nine seasons, beginning with 1991-92, and a cross-section of 19 377 individual farmers serve as the reference base. The results, however, are potentially distorted by some data limitations, the most serious one being that the database includes observed cereal acreage only when an insurance policy was contracted. Missing value was thus replaced by the acreage for the most recent season for which ENESA has records.
Despite the limitations, the data set offered several empirical possibilities. First, relative risk aversion coefficients were evaluated using basic theoretical results and comparing the willingness of agents to pay for insurance to the actual premium paid. Using a simple mathematical procedure based on a random selection of farmers it was possible to demonstrate that their risk preferences confirm the assumption of Decreasing Absolute Risk Aversion and Constant Relative Risk Aversion, with CRRA coefficients (rr) ranging between of 0 and 4 (although 70% were between 0 and 1). Alternative econometric models were also estimated, with specifications aimed at explaining changes in insurance strategies, yields, and cereal supply.
Results show that once farmers have begun to contract insurance policies, they continue to contract the same policies over time. The incentives influencing farmers’ decisions to buy insurance, however, differ from those that seem to guide their decisions about switching from low- to high-coverage policies. It was also found that individuals subject to large production risks tended to insure more than those operating in less risky conditions.
The results of the yields equations, interpreted jointly with those obtained from the supply equations, show that farmers tend to obtain lower yields when they contract insurance, but may increase cereal acreage. The effect of the expected cereals and other crop prices, as well as the expected price variance and direct per hectare payments, confirm a priori economic theories. In other words price effects on yields and supply are larger than the effects of the remaining variables. It is thus clear that policies affecting cereal prices do act as important incentives influencing farmers’ decisions.
The cereal supply equations show that insurance subsidies tend to increase cereal supply. However, subsidies for yield insurance have less impact on production than do subsidies for insurance against fire and hailstorm. In elasticity terms, the effects on production of subsidies to both types of insurance policies are smaller than the effects of price or other direct support policies.
Absolute and relative effects of price support, area payments and insurance subsidies have been evaluated using the corresponding supply elasticities estimated from the production models. The benchmark for the estimation of the relative effects of the different instruments is a 1% increase in market prices. The same absolute increase in support is then applied in the form of an increase in area payments or an increase in the insurance subsidy. The results show that price effects and insurance subsidies are somewhat similar, but that the effects of area payments are much greater. However, it should be noted that the amount of additional subsidy that would have to be given on insurance (in order for the increase to be equivalent to the increase in market price support) is 54% for increase yield insurance. In the case of fire and hailstorm insurance the extra subsidy to be distributed is so large that the farmer would actually purchase the insurance for free, and even receive a financial bonus for doing so.
The main policy conclusion of this study is that yield insurance subsidies do not seem to have a large effect on cereal production. Nonetheless, the results presented could be improved through the use of a greater number of observations and perhaps partitioning the sample by some other criteria. Secondly, more sophisticated econometric techniques could possibly answer additional policy-relevant questions than were investigated here. Nevertheless, the lack of data for those farmers who refuse to buy insurance is a serious limitation for which there is no easy solution.