The objective of this paper is to compare the accuracy of crop insurance rating methods based on historical liability and indemnity data (similar to the procedure currently used by the Risk Management Agency) and “yield distribution” approaches. Estimated rates are compared to “true” rates using empirically-grounded simulation procedures that take into account common data availability constraints. Simulation results suggest that farm and county level rate estimates using the “yield distribution” approach are significantly more accurate than those based on historical indemnity and liability records.