Using a Probabilistic Frontier Production Function to Measure Technical Efficiency

Abstract

This article uses linear programming techniques to "estimate" a frontier Cobb-Douglas production function for U.S. agriculture from 1960 to 1967, using the "average farm" in each state in each year as an observation. Both deterministic and probabilistic frontiers are generated and the results compared with ordinary least-squares and analysis of covariance estimates of the production function. Technical inefficiency is defined relative to the probabilistic frontier function and the extent of any inefficiency calculated for each state. Little technical inefficiency exists across states when the production function includes intermediate inputs as well as land, labor, and capital.