Geographical Influences on Farming and Risk - Cont.

The previous discussion also raises the possibility of predicting expected loss costs based on yield information. Chart 17 shows the relationship between the aggregate loss costs and the yield for a given year. This relationship could be used to provide an estimate of the expected loss costs based an estimated distribution of the yields. Past yields by county could be trended to reflect productivity improvements in order to obtain an estimate of the distribution for the current year. However, the coefficient of variation of yields is a reasonable alternative to using the distribution of the trended yields. The coefficient of variation distills the distribution of yields into a single number for each county, and appears to be effective in predicting the loss costs. This makes it possible to consider either approach as a technique for estimating the expected loss costs whenever past loss experience is not available. This could also be used to test the indicated loss costs for reasonableness. Furthermore, counties with high coefficients of variation of yields are those in which farming is more uncertain. Not only are these counties expected to cluster together, but it is likely that the uncertain growing conditions will apply to all crops grown within the county. As a result, the variability of yields for one crop may be a means for predicting the variability of yields for a crop being newly introduced or with minimal loss history.