Partial Loss Development Based on Expected Losses for Workers' Compensation Class Ratemaking

The standard multiplicative loss development factors applied to reported losses by class serve to amplify instability in partial loss data. A method of assigning loss development based on expected losses is described and tested using four years of actual class data for Oregon. The method uses payroll and "pure premium present on rate level" to estimate expected losses. Test statistics are devised to compare stability of rates calculated using this revised method and rates calculated in the standard manner. The tests are based on residuals from linear trend lines and on absolute magnitude of 1992 rate revisions by class. The tests support a conclusion that the revised method produces significantly greater rate stability even though credibility of indicated state experience is enhanced. There is brief discussion of other stability approaches and topics for further research in class ratemaking.