Suchergebnisse

Abstract
We study long run implications of reinforcement learning when two players repeatedly interact with one another over multiple rounds to play a finite action game. Within each round, the players play the game many successive times with a fixed set of aspirations used to evaluate payoff experiences as successes or failures. The probability weight on successful actions is increased, while failures result in players trying alternative actions in subsequent rounds. The learning rule is supplemented by small amounts of inertia and random perturbations to the states of players. Aspirations are adjusted across successive rounds on the basis of the discrepancy between the average payoff and aspirations in the most recently concluded round. We define and characterize pure steady states of this model, and establish convergence to these under appropriate conditions. Pure steady states are shown to be individually rational, and are either Pareto-efficient or a protected Nash equilibrium of the stage game. Conversely, any Pareto-efficient and strictly individually rational action pair, or any strict protected Nash equilibrium, constitutes a pure steady state, to which the process converges from non-negligible sets of initial aspirations. Applications to games of coordination, cooperation, oligopoly, and electoral competition are discussed.

We study a contract change for tea pluckers on an Indian plantation, with a higher government-stipulated baseline wage. Incentive piece rates were lowered or kept unchanged. Yet, in the following month, output increased by 20 to 80 percent. This response contradicts the standard model and several variants, is only partly explicable by greater supervision, and appears to be "behavioral." But in subsequent months, the increase is comprehensively reversed. Though not an unequivocal indictment of "behavioral" models, these findings suggest that nonstandard responses may be ephemeral, and should ideally be tracked over an extended period of time. (JEL D82, D86, J33, J41, J43, O13, Q12)

This paper studies the productivity impact of a contract change for tea pluckers in an Indian plantation. The contract, implemented at the end of a three-year cycle in which contracts are generally revised, was (a) the joint outcome of negotiations between twenty unions and plantations, (b) mandated to respect a state government notification stipulating a new minimum wage for plantation workers statewide, and (c) applicable equally to all the plantations in the local region. The contract raised the baseline wage by 30% but lowered marginal incentives, by shifting the existing piece rates to higher minimum thresholds and eliminating an existing penalty per unit for low output. In the one month following the contract change, output increased by a factor between 30-60%, the exact number depending on the choice of counterfactual and the set of controls applied. This large and contrarian response to a flattening of marginal incentives is at odds with the standard model, including one that incorporates dynamic incentives, and it can only be partly accounted for by higher supervisory effort. We conclude that the increase is a "behavioral" response. Yet in subsequent months, the increase is comprehensively reversed. In fact, an entirely standard model with no behavioral or dynamic features that we estimate off the pre-change data, fits the observations four months after the contract change remarkably well. While not an unequivocal indictment of the recent emphasis on "behavioral economics," the findings suggest that non-standard responses may be ephemeral, especially in employment contexts in which the baseline relationship is delineated by financial considerations in the first place. From an empirical perspective, therefore, it is ideal to examine responses to a contract change over an substantial period of time.

We present a model of conflict, in which discriminatory government policy or social intolerance is responsive to various forms of ethnic activism, including violence. It is this perceived responsiveness -captured by the probability that the government gives in and accepts a proponed change in ethnic policy- that induces individuals to mobilize in support for their cause. Yet, mobilization is costly and demonstrators have to be compensated accordingly. Individuals have to weigh their ethnic radicalism with their material well-being to determine the size of their money contribution to the cause. Our main results are: (i) a one-sided increase in radicalism or in population size increases conflict; (ii) a one-sided increase in income has ambiguous effects depending on the elasticity of contributions to income; (iii) an increase in within-group inequality increases conflict; and (iv) an increase in the correlation between ethnic radicalism and inequality also increases conflict.