Identifying Technically Efficient Fishing Vessels: A Non-Empty, Minimal Subset Approach

There is a growing resource economics literature, concerning the estimation of the technical efficiency of fishing vessels utilizing the stochastic frontier model. In these models, vessel output is regressed on a linear function of vessel inputs and a random composed error. Using parametric assumptions on the regression residual, estimates of vessel technical efficiency are calculated as the mean of a truncated normal distribution and are often reported in a rank statistic as a measure of a captain's skill and used to estimate excess capacity within fisheries. We demonstrate analytically that these measures are potentially flawed, and extend the results of Horrace (2005) to estimate captain skill for thirty-nine vessels in the Northeast Atlantic herring fleet, based on homogeneous and heterogeneous production functions within the fleet. When homogeneous production is assumed, we find inferential inconsistencies between our methods and the methods of ranking the means of the technical inefficiency distributions for each vessel. When production is allowed to be heterogeneous, these inconsistencies are mitigated.