Longitudinal targets and persistent inefficiency in data envelopment analysis

A typical interpretation of data envelopment analysis (DEA) targets in manufacturing and other contexts is to identify performance goals that if achieved in the future will move an inefficient decision-making unit to the best practice efficiency frontier. Less discussed, however, is that this only will occur with certainty if all other units maintain their same input and output levels, an implicit assumption that is rarely the case in practice. Since this phenomenon can be important in many practical applications, this paper explores its manifestation and approaches for interpreting and setting future performance targets in this context, including forward and backward-looking analysis, forecasting, and Monte Carlo approaches. Each method is illustrated and compared using two empirical data sets.