Two Simple Algorithms for Generating a Subset of Data Consistent With WARP and Other Binary Relations

Abstract

We give two algorithms that partition data into two subsets, a set consistent with the weak axiom of revealed preference (WARP) and its complement. Our main contribution is a very fast, but inexact, algorithm. We also describe an exact algorithm that will be impractical for large data sets. These same algorithms may be applied to several other binary relationships such as the nonparametric production axioms. Because violations of WARP are also violations of the strong (SARP) and generalized (GARP) axioms, these algorithms are a logical first step in computing subsets consistent with these relations.