Optimal taxation in a context of poverty, tax evasion or corruption.

Authors
Publication date
1999
Publication type
Thesis
Summary This thesis proposes a study of three important problems for developing countries, namely poverty, tax evasion and corruption. The first chapter introduces poverty in Mirrlees' (1971) optimal taxation problem. Poverty is considered as an aggregate negative externality. The optimal nonlinear taxation scheme has some interesting properties. Indeed, negative marginal tax rates emerge at least for the least productive. This induces them to work more and earn more income. Despite the presence of negative marginal tax rates, it is possible to recover the classic result of zero marginal rates at the ends of the distribution by reasoning in terms of social rather than individual distortions. The second chapter introduces tax evasion by assuming that revenues are only observable after an expensive audit. The frame of reference is that of a finite economy that allows for the correlation between information privately held by individuals. It is then possible to show that any first-order optimum is implementable by the combination of a generalized taxation and a generalized audit system. However, only a subset of the pareto efficient frontier is implementable when the generalized taxation system is replaced by a classical taxation system. In contrast to the literature on tax evasion, in equilibrium no one is audited and all but the most productive evade. Finally, the third and last chapter deals with corruption. We consider a tax agency that wants to maximize tax revenue and must hire inspectors to collect taxes. The inspectors must make an expensive and unobservable effort to determine the true income of taxpayers, so there is a moral hazard problem. In addition, there is an adverse selection problem since inspectors may be honest or corruptible but their type is private information. A compensation scheme such that only corruptible inspectors are willing to be hired may be optimal.
Topics of the publication
  • ...
  • No themes identified
Themes detected by scanR from retrieved publications. For more information, see https://scanr.enseignementsup-recherche.gouv.fr