Three Essays in the Economics of Taxation

This dissertation presents three essays advancing the theory of taxation. The first chapter presents a new method for evaluating proposed reforms of progressive piecewise linear tax schedules. Typically, estimates of the elasticity of taxable income (ETI) are used to predict taxpayer responses to changes in tax rates and/or tax bracket thresholds. In this chapter, I show that elasticities are not always needed for this task; the ``bunching mass'' at a bracket threshold (the share of taxpayers locating there) is a sufficient statistic for the revenue effect of behavioral responses to small changes of the threshold. Building on this finding, revenue forecasting and welfare analysis of threshold changes can be conducted using the pre-reform distribution of taxable income alone. I apply these results in an analysis of the Earned Income Tax Credit, an exercise which motivates extensions addressing taxpayer optimization errors, tax rate heterogeneity, large reforms, and income and participation effects. My approach complements existing bunching methods: it avoids key limitations of bunching-based ETI estimation, but addresses a relatively narrower set of policy questions. The second chapter explores the evolution of economic inequality and political inequality in a democratic society where these two types of inequality mutually reinforcing. I introduce a simple dynamic model of democratic redistribution where, in each period, two candidates compete in an election by proposing how a fixed amount of income will be divided amongst a group of citizens in the next period (i.e. pure redistribution policy). Campaign spending is financed by citizen political donations, leading to inequality of political influence favoring wealthier citizens. This creates a feedback loop through which the current distribution of income affects the future distribution. If the marginal dollar of income yields a sufficiently large increase in political influence, long run convergence to a plutocratic equilibrium can occur for arbitrarily small levels of initial economic inequality. The opposite scenario is also possible: a society which is initially extremely unequal may nonetheless be destined for egalitarianism. The long run distribution of income can exhibit extreme sensitivity to initial conditions: tiny differences in initial inequality may determine whether democratic redistribution leads to plutocracy or egalitarianism. Turning to a version of the model where elections are fought over a nonlinear income tax, I show that the same conditions that determine whether convergence to egalitarianism occurs in the pure redistribution model also dictate whether taxation of the rich is possible in the nonlinear tax model. The third chapter employs a variant of the election model from the second chapter to examine the optimal tax treatment of political contributions. Adopting the normative stance that inequality of political influence is undesirable, I characterize the optimal nonlinear tax schedule on political donations. Sufficient statistics for optimal policy include not only donation demand elasticities, but also the marginal efficacy of campaign spending, and the effect of taxes on the sensitivity of donations to candidate policy platforms. Using numerical simulations, I provide proof-of-concept results showing that this framework can rationalize real world policies such as the nonlinear subsidy schedules present in Canada. These feature generous marginal rates of subsidy on the first dollar of political donations, with rates of subsidy declines in donation amount.PhDEconomicsUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/174564/1/dtmoore_1.pd