Testing the Assumptions of Production Theory: A Nonparametric Approach

We propose methods for testing the validity of such general hypotheses about production processes as quasi-concavity, monotonicity, and homotheticity from data on inputs and outputs (and through the use of duality from data on prices, profits, and costs). The tests are free from any particular parametrization of the production function and, since they are based on linear programming, are in principle easy to compute. They may be used to censor or modify data as well as to test assumptions of production theory.