A Bivariate Probit Model of Disaster Assistance and Crop Insurance Participation

Abstract

This research examined factors influencing farmer purchase of crop insurance and receipt of disaster assistance payments using survey data for more than 13,000 farms across 27 U.S. states. Using a bivariate probit model some main findings are as follows. The probability of participating in federal crop insurance programs is (a) lower for farmers more than 65 years of age; (b) increasing with farmer education and farm sales; (c) lower for farms where farm income is a small share of household income; and (d) higher in states with higher average temperatures and lower average precipitation. The probability of receiving disaster payments (a) increases as farms depend more on farm income for their total household income; (b)increases with sales in peanut farming and cattle ranching; (c) greater in states experiencing drier or wetter than normal hydrologic conditions; and (d) greater in states experiencing warmer than normal temperatures. In addition, previous research using state-level data found agricultural disaster payments were higher in states with congressional representation on ubcommittees overseeing USDA‟s direct disaster payment program. The farm-level analysis of this thesis supports this earlier finding. Farmers in states with such representation had higher probabilities of receiving disaster payments, controlling for other factors.