Some Issues on Optimal Non-Linear Taxation

In this thesis we apply an optimal taxation approach to analyze how citizens' private information may affect Social Planner 's goal of redistribution across individuals. In fi rst two Chapters, we allow individuals to emigrate to a laissez-faire country. Chapter 1 assumes the existence of two types of citizens, who are endowed with a different labor productivity. We let these two types be low- and high-skilled, respectively, and assume that emigration is costly. We prove that low-skilled agents ' incentive compatibility constraint is never binding at the optimum, while high-skilled citizens' one is. Moreover, these latter are net taxpayers and, hence, a lump-sum tax is levied by the Social Planner. Relative to the standard closed-economy model, we also fi nd that, if the economy stands for two periods, in the absence of Government 's commitment, redistribution is dramatically reduced by the existence of high-skilled citizens' outside opportunity. Therfore, an inducing fi rst-period pooling tax scheme is much less likely to be implemented, with respect to a separating tax scheme. In Chapter 2 we introduce a third type of citizens, who are characterized by an intermediate skill level. This allows us to better analyze the distributive effects of taxation. In particular, we will show that countervailing incentives may occur. Indeed, medium-type citizens may fi nd it optimal to mimic high-skilled ones. In a such a context, a separating scheme can be implemented. However, a distortion at the top arises. As in Chapter 1, Chapter 3 studies an economy, that is populated by two continua of individuals, who differ in labor productivity. We allow agents to save and let effort be private information. Therefore, both adverse selection and moral hazard problems may arise. In particular, we prove that, under adverse selection, it is optimal for the Social Planner to fully insure high-skilled agents and to over-insure low-skilled ones. Moreover, no high-type aggregate intertemporal wedge is implemented, while a positive L-aggregate intertemporal wedge is obtained. Finally, we study the joint effects of dynamic moral hazard and adverse selection on social welfare. In this case, both types are partially insured. Moreover, both the H-aggregate intertemporal wedge and the L-aggregate intertemporal wedge are positive.