An analysis of risk premia in U.S. farm-level interest rates.

Much farm financial risk research has involved the application of "credit-scoring" models. We approach the issue of measuring financial risk by using the actual interest rates charged on agricultural loans reported in the USDA's ARMS survey as market-based measures of the financial risk associated with individual farm operations. A simultaneous equations model relates rates to several farm, producer, and lender characteristics. Because individual loans in our sample have different dates of origination, deviations of individual rates from market rates are considered. Our results indicate that risk (as perceived by lenders) tends to be higher for farms with less wealth (net worth) and more loans. Farm operators who live on their operations are considered by lenders to be less risky. Farm diversification appears to be correlated with less financial risk. Significant differences in agricultural lending rates across different types of lenders were also revealed, with the highest rates being charged by commercial banks and savings and loans. A key element of our analysis is development of a probability-weighted bootstrap estimator that permits consistent inferences to be drawn from the stratified ARMS data.