Title: (Functional)Characterizations vs (Finite)Tests: Partially Unifying Functional and Inequality-Based Approaches to Testing

Abstract: Historically, testing whether decision-makers (DM) obey certain axioms from choice data has taken two rather distinct approaches. The functional approach involves observing the entire demand function and putting restrictions on it; the revealed preference approach involves constructing algebraic inequalities which are used to test finite data. Using restrictions derived in \cite{kubler2020identification}, we demonstrate a link between revealed preference tests and function-based restrictions. By this, we mean that any functional restriction that \textit{characterizes} the above axioms can be used to construct a well-behaved finite data test. We then develop an algorithm that does this efficiently.