Incentives, Compensation, and Social Welfare

TLDR: In this article, the authors analyse a model in which individual abilities are publicly known and show that a conflict between distribution and incentives arises instead from the existence of uncertainty in production processes and a moral hazard problem stemming from the unobservability of workers' effort.

Abstract: First version received January 1985; final version accepted August 1986 (eds.) Alternative wage structures under conditions of moral hazard are analysed from a social welfare standpoint. It is argued that ex post equity judgements in an uncertainty context should incorporate a preference for "positive correlation" of utilities of different individuals. In the design of compensation schemes, this may give rise to a conflict between ex post equity objectives and the need to provide effort incentives: relative performance clauses in compensation schemes that are useful for providing incentives are undesirable from an ex post equity standpoint. This is demonstrated by showing (a) in a context of independent production uncertainties, every rankorder tournament is welfare-dominated by a set of independent (randomized) contracts, and (b) welfare-optimal compensation schemes in general depend separately on an equity and an incentive component that tend to correlate agent compensations in different directions. A central question in welfare economics concerns the design of a system of rewards for productive effort that balances distributive considerations with the provision of effort incentives. This issue has usually been studied in the context of a model due to Mirrlees (1971) in which the conflict between distribution and incentives arises because the social planner lacks perfect information about the abilities of individuals. In this paper we analyse a model in which individual abilities are publicly known; a conflict between distribution and incentives arises instead from the existence of uncertainty in production processes and a moral hazard problem stemming from the unobservability of workers' effort. Most analyses of moral hazard have concentrated on the tradeoff between insurance of agents against wage fluctuations and provision of effort incentives. The context is generally taken to be a capitalist firm where the self-interested "planner" has a residual claim on workers' output.' In other settings, such as public-sector or labour-managed firms, the planner may have preferences over the distribution of ex post welfare among the workers. This paper analyses how, with such preferences, effort incentives under moral hazard are constrained both by insurance and by distributive considerations. In Section 2, we suggest that a concern for ex post equality of individual welfares can be expressed by a preference for positive correlation among welfare levels. We then consider what properties of the ex post welfare function incorporate this preference for correlation. Welfare functions that are additively separable across individuals are inappropriate, no matter how concave. With them, expected value depends only on the