Outliers, Sample Size and Robust Estimation of Stochastic Frontier Production Models

This paper investigates the behavior of maximum likelihood estimates of a stochastic frontier production function in the presence of outliers and in small samples. The primary motivation is the concern that the stochastic frontier approach to efficiency estimation is not completely invulnerable to the influence of outlying observations. The paper alleges that outliers may cause some firm inefficiency to be absorbed by the noise distribution. The distribution for technical inefficiency component of the error term is assumed to be exponential, while the noise disturbance is assumed to follow either a normal or a t-distribution. Based upon their experiments, the authors conclude that firm inefficiency is not absorbed by the distribution of symmetric error component, whether it is taken to be normal or t.