Tests for the consistency of consumer data (Q1079313)

The authors develop non-parametric tests for the weak separability of a decision maker's utility function, given a finite body of price and quantity data. They try to make Afriat's non-parametric method for testing whether a finite body of price and quantity data is consistent with utility maximizing behaviour more accessible and they generalize some earlier tests proposed by \textit{S. N. Afriat} [The construction of separable utility functions from expenditure data. Mimeo, Univ. North Carolina, Chapel Hill/NC (1970)] and \textit{H. R. Varian} [Rev. Econ. Stud. 50, 99-110 (1983; Zbl 0515.90023)]. After a short introduction, in the second section of their article, the authors derive the Afriat conditions under stronger than necessary assumptions (concavity and differentiability) on the utility function. In the next two sections, they use these techniques to derive some tests for the utility function to be weakly separable and additively separable respectively. These new tests are generalizations of similar tests developed by Afriat and Varian. The last section contains general conclusions with regard to the theoretical significance of the newly proposed tests and to their practical applicability.