Estimation of Input-Specific Technical and Allocative Inefficiency in Stochastic Frontier Models *

This paper considers an extension of technical inefficiency to a mor e disaggregate level, viz., input-specific technical inefficiency, based on a stochastic production frontier. Allocative inefficiency is modeled as a departure from the first order conditions for cost mini mization. The production function and the first order conditions for cost minimization, derived from a Cobb-Douglas technology, form the basis of the likelihood function to obtain the maximum likelihood estimates. Panel data on U.S. Class 1 railroads are used to estimate technical inefficiency specific to labor and fuel for each road. Cost of allocative inefficiencies are also calculated for the individual roads.