ECIN Replication Package for "Predicting Voluntary Contributions by 'Revealed-Preference Nash Equilibrium'"

Abstract

One-shot public-good situations are prominent in the public debate, and a prime example for behaviour diverging from the standard Nash-equilibrium. But does behaviour diverge from equilibrium because the equilibrium fails to account for social preferences or - as recent research suggests - because a Nash-equilibrium cannot possibly predict one-shot public-good behaviour? I show that a 'revealed-preference Nash-equilibrium' (RPNE) out-of-sample predicts one-shot public-good behaviour, outperforming other social-preference models. The RPNE is the set of 'mutual conditional contributions', interpreting elicited conditional contributions as best-responses.

Individual-level analyses confirm the results and allow for studying equilibrium selection. The modal choice corresponds to the Pareto-dominant equilibrium. However, many participants use other criteria. Given the predictive positive-contributions RPNEs, the heterogeneity in selection criteria may turn out to be the origin of the low late-round contribution levels found in studies of repeated interactions. Conversely, many real-life public-good situations may be solvable at least partially if the players can coordinate on an equilibrium-selection criterion beforehand.