Optimal Nonlinear Labor Income Taxation in Dynamic Economies

The aim of this paper is to explore the characteristics of the optimal nonlinear labor income tax in dynamic economies with information asymmetries and human capital accumulation. We develop a dynamic optimal income tax model in which agent’s productivity evolves over time according to two different factors: an exogenous component and a learning by doing process endogenous to the fiscal policy. The latter is determined by the government, maximizing in the initial period a social welfare function capturing some level of aversion to inequality. We characterize analytically the first order condition driving the optimal tax schedule in a model in which agents choose the consumption and labor supply patterns that maximize their lifetime utility function. We show that the inclusion of the endogenous evolution of productivities into the tax problem changes the results with respect to the static framework `a la Mirrlees (1971). We find that the optimal tax strategy balances social marginal costs of increasing marginal tax rates with social marginal benefits of doing so. The costs are related with the reduction of both past and future human capital accumulation, with the negative impact on aggregate social welfare due to the reduction of the individual utility of all the agents paying more taxes and with the increase of the necessity to redistribute more in the future (given that the spread among social marginal weights of each agents will be higher). The benefits derive from the increase of the instantaneous tax receipts with the consequent reduction of the overall tax burden all along the time horizon and from the reduction of present inequality of incomes and future inequality of the productivities. As a particular extreme result we find that it can be optimal to subsidize (instead of taxing) at the margin high productivity agents.