Are Incentives Good Enough to Achieve (Info) Social Order?

Abstract

In this paper, the role of incentives insocial order is questioned, based on a notion of incentive as additionalindividual utility, provided by an external entity, to actions achievingglobal utility. Two notions of norms are compared: (1) inputs which modify agents'decisions through incentives (sanctions) and (2) prescriptions to executeobligatory action for intrinsic motivations. Two types of agents whichreason upon norms are also compared: (1) incentive based rational deciders, and (2) normative agents which are prescribed to execute norms for intrinsic reasons. The twotypes of agents are expected to have a different impact on norm compliance.Under suboptimal conditions of application of sanctions (uncertianpunishment), transgression is expectedto propagate more easily and rapidly among incentive-based agents thanamong normative agents. In particular, incentive-based agents are expectedto show a fast decline and even a collpase in compliance with the norms.Normative agents are expected to exhibit an oscillating behaviour, or at least a graceful degradation ofcompliance. Finally, the role of incentives is shown to have a lesserimpact on natural social agents than expected by a model of rationaldecision. What is worse, incentives have been shown to produce even negative effects on several aspects of social learningand norm compliance