Purging data before productivity analysis

Studies of productivity often ignore measurement error and fail to distinguish between exogenous and endogenous factors in adjusting for the environment. This failure may misguide managerial decisions on benchmarking, ranking, and remuneration. For example, common relative efficiency techniques such as data envelopment analysis (DEA) assume away the measurement error. This article combines DEA with stochastic frontier analysis in a synergistic multiple-stage analysis to purge the estimate of managerial performance of measurement error and exogenous factors. Removing the impact of measurement error indicates the largest rise in productivity. Removing the impact of exogenous factors raises discriminatory power. The method offers a number of innovations over other studies in the literature. The article is also the first to investigate the profit efficiency of the commercial banks in the United Arab Emirates. © 2009 Elsevier Inc. All rights