Weak Axiom of Cost Dominance: A Nonparametric Test of Cost Efficiency Without Input Quantity Data

Varian's Weak Axiom of Cost Minimization provides a nonparametric test of cost minimization, which can be applied only when both input price and quantity data are available for individual firms. In this paper we propose a Weak Axiom of Cost Dominance (WACD), which serves as the basis of an alternative test of cost-minimization applicable in situations where input quantity data are missing. Unlike a previous test developed by Diewert and Parkan, the proposed test does identify individual firms that violate the assumption of cost-minimizing behavior. It also provides an upper bound of the cost-efficiency of any observed firm. The test procedure is shown to be equivalent to applying dominance analysis using normalized input prices with reference to the Cost-indirect technology. The proposed method is applied to Nerlove's electrical utility data. The nonparametric results are also compared with parametric efficiency levels computed from a stochastic frontier cost function.