Intrinsically Linear Loss Development Models for Workers' Compensation Costs: Point and Interval Prediction Methods
Final rept. 1 Jun-3 Aug 1994
NAVAL HEALTH RESEARCH CENTER SAN DIEGO CA
Pagination or Media Count:
This report describes modeling methods that allow the computation of point predictions and prediction probability intervals for cumulative workers compensation costs. Underlying these models is the actuarial loss development factor method, a method that computes projected costs by utilizing ratios of known cumulative costs in consecutive years. While the relationship between cumulative loss development, cohort, and development year in these models is nonlinear, a transformation readers them in the form of standard linear statistical models, thus allowing the development of prediction probability intervals when the error structure is Gaussian. The modeling methods are illustrated using data collected from U.S. Department of the Navy workers compensation payments made from 1990 through 1993, including claim costs originating from 1961 through 1993. Statistical models, Probability intervals, Cost projections, Forecasts, Insurance claims, Actuarial methods.
- Economics and Cost Analysis
- Statistics and Probability