Bootstrapping Confidence Intervals for Linear Programming Efficiency Scores: With an Illustration Using Italian Banking Data

Abstract

This article suggests a method for introducing a stochastic element into Farrell measures of technical efficiency as calculated via linear programming techniques. Specifically, a bootstrap of the original efficiency scores is performed to derive confidence intervals and a measure of bias for the scores. The bootstrap generates these measures of statistical precision for the "nonstochastic" efficiency measures by using computational power to derive empirical distributions for the efficiency measures.