Extortion in the laboratory

In a laboratory experiment, we study a finitely repeated game (T = 15) under complete information. In each round, P demands tribute (cash transfer) from A, A complies or refuses, and after refusals P may punish A. In equilibrium (payoff maximization), P does not punish and A refuses any positive demand. In the experiment, P punishes increasingly often and increasingly severely as she gains experience; most As comply with P’s demands. The observations are compatible with linear spite. In a finite mixture model, the types of P and A in the subject pool are characterized. An A that is resistant to extortion (declines all demands) is very rare, and hence the threat of punishment in general is effective, but all As either ignore actual punishment or react negatively to it. They accept to pay tribute but they are resistant to piecemeal expropriation.

Research highlights► Laboratory experiment on finitely repeated extortion game. ► P(rincipal) demands tribute from A(gent), and if A does not comply, P may punish A. ► P punishes increasingly often and increasingly severely after rejections. ► Eventually, most As comply with P’s demands. ► Finite mixture modeling of type heterogeneity is applied to understand behavior.