Purpose
Crime games cannot be simply read with mixed strategies. These strategies are inconclusive of how the players act rationally. This is undeniably true for the crime of tax evasion, where dishonest taxpayers are rational agents, motivated by the comparison of payoffs, when considering the risk of non-compliance. The purpose of this paper is to illustrate that in the presence of a small “private disturbance” of the players’ payoff, the Nash equilibrium in mixed strategies provides us with the necessary information on equilibria in pure strategies that will be played.
Design/methodology/approach
In tax-evasion games, an equilibrium must necessarily be interpreted in pure strategies, and the only way to do this is to insert some private information into the game and reinterpret it in a Bayesian scheme. We show that taxpayers’ private,subjective considerations on the effective implementation of the penalty and the revenue agency’s private information on the cost of monitoring and conviction can lead to Bayesian equilibria in pure strategies. The present paper takes issue with this Bayesian equilibrium and the implications for comparative-statics results.
Findings
In this context, tougher sentencing deters crime, although, as the Italian experience teaches, the necessary condition required is the certainty of punishment and the ability of the government to enforce it. The equilibrium strategies with incomplete information reveal whether it is convenient for the two agents to maintain their “private disturbance” as private information or, on the contrary, it is convenient to expect it to be “common knowledge.”
Originality/value
A distinct set of studies has adopted a game theoretic approach and shows that the standard economic approach to crime deterrence inspired by Gary Beker’s seminal paper might be flawed. See, among others, Saha and Poole (2000), Tsebelis (1989) and Andreozzi (2010). This paper shows that a greater severity of the penalty and a higher certainty of punishment (a lower possibility of appealing against sanctions and no discounts on due penalties) necessarily lead to a unique Bayesian equilibrium without evasion.
LOCAL IDENTIFICATION IN EMPIRICAL GAMES OF INCOMPLETE INFORMATION
