Partial Frontier Efficiency Analysis

Despite their frequent use in applied work, nonparametric approaches to efficiency analysis—namely, data envelopment analysis and free disposal hull— have bad reputations among econometricians. This is mainly because data envelopment analysis and free disposal hull represent deterministic approaches that are highly sensitive to outliers and measurement errors. However, so-called partial frontier approaches have recently been developed, namely, order- m and order-α. These approaches generalize free disposal hull by allowing for superefficient observations to be located beyond the estimated production-possibility frontier. Although these methods are also purely nonparametric, the sensitivity to outliers is substantially reduced by partial frontier approaches enveloping just a subsample of observations. In this article, I introduce the new Stata commands orderm and orderalpha, which implement order- m, order-α, and free disposal hull efficiency analysis in Stata. The commands allow for several options, such as statistical inference based on subsampling bootstrapping.