Key points are not available for this paper at this time.
claims filed in several prior calendar years.6 The data underlying these estimates are described below.The growth rate between 1970 and 1975 reported in Table I is probably upward biased because the 1970 sample is incomplete, by as much as 30 percent.For 1975-78 claim counts in some years for some states may be incomplete.Since data were only collected for July-December 1975, the estimate used here for 1975 frequency is twice this six-month count.7 Washington, D.C., is an extreme outlier in 1975, with claim frequency twice that of the second ranked state, California.Correlation coefficients between year-to-year growth rates and level in the initial year range from -.2 to -.4 for frequency, and -.3 to -.5 for severity.8 In 1976, thirty-five states had fewer than 100 paid claims.
Patricia M. Danzon (Sun,) studied this question.