Abstract
The data envelopment analysis (DEA) approach has been widely applied to measure the relative efficiencies of decision making units (DMUs) for a time period. This article measures the performance improvement of a DMU at two time points by using the DEA approach. Two ideas are proposed. One is to treat one DMU at two time points as two different DMUs, and the ratio of the efficiency scores of a DMU at two time points serves as a measure of the efficiency improvement. The other is to calculate the Malmquist productivity index of a DMU to represent its productivity improvement. Both the efficiency ratio and the productivity index can be decomposed into the technical part and the scale part. From the data of the eight forest districts of Taiwan compiled in 1989 and 1992, the efficiency and productivity improvement in these two periods are measured. The results of the two methods are consistent in that both indicate a slight improvement in performance. FOR. SCI. 46(4):577–584.