A Method of Moments Estimator for a Stochastic Frontier Model with Errors in Variables

We propose a method of moment estimator for a stochastic frontier model in which one of the independent variables is measured with errors. The estimator corrects for the measurement errors, and it requires only minimal assumption on the error distribution, has no need for additional data, and is computationally inexpensive. A Monte Carlo study shows favorable statistical properties of this estimator. We apply this estimator to an investment model with financing constraint, where a major explanatory variable, Tobin's Q, is known to prone to measurement problems. We find that the Q's explanatory powers increase substantially upon correcting for the measurement errors.