The authors of this study compared DRG weights calculated using the Hospital Specific Relative Value (HSRV) method with those calculated using the standard method for each year from 1985 through 1989, and analyzed differences between the two methods for 1989. They provide evidence suggesting that classification error and subsidies of higher weighted cases by lower weighted cases caused compression in the weights used for payment as late as the fifth year of Medicare's prospective payment system. However, later weights calculated by the standard method are not compressed because a statistical correlation between high markups and high case mix indices offsets the cross-subsidization. HSRV weights from the same files are compressed because this methodology is more sensitive to cross-subsidies. However, both sets of weights produce equally good estimates of hospital level costs net of those expenses that are paid by outlier payments. The greater compression of the HSRV weights is counterbalanced by the fact that more high-weight cases qualify as outliers.