Implementing Pareto Optimal and Individually Rational Outcomes by Veto

Abstract

We introduce a simple veto mechanism where each agent can veto any subset of alternatives, by paying a veto cost for each vetoed alternative. The outcome is the set of non-vetoed alternatives or, if this set is empty, some previously fixed alternative which is declared the disagreement outcome. Under fairly mild axioms to extend individual preferences over alternatives to sets of alternatives and assuming quasi-linear preferences over outcome-money bundles, we show that the Nash equilibrium outcomes of the veto mechanism coincide with the Pareto optimal outcomes which are individually rational according to the disagreement outcome. The positive result prevails when individual preferences admit indifferences and even for the case of two agents. We also show that under stronger axioms to extend preferences over alternatives to sets, strong Nash implementation (hence double implementation) is also possible with the same veto mechanism.