Date of Award

Spring 5-2015

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Tepper School of Business


Laurence Ales

Second Advisor

Christopher Sleet


This dissertation contains three chapters and focuses on the optimal design of fiscal policy, both from a theoretical and from a quantitative perspective. In the first chapter, “Wealth Taxation and Life Expectancy,” I address the optimal taxation of wealth in a class of dynastic overlapping-generations economies with heterogeneous mortality risk. Working individuals are indexed by skills which are private information. Skills not only determine earning abilities but also correlate with survival probability, so that more productive agents on average live longer. The analysis distinguishes between the tax treatment of two possible sources of wealth, namely, savings and bequests, and points to the mortality gradient as a crucial determinant for optimal wealth taxation. Specifically, due to differential mortality: (a) earned wealth should be marginally taxed, (b) transferred wealth via bequests should be marginally subsidized, and (c) marginal tax schedules on bequests and inter-vivos transfers should be separated. I calibrate the model to U.S. data and quantitatively evaluate its tax implications. For the median worker, mortality differences create a force for marginally taxing capital mortality differences create a force for marginally taxing savings by up to 1.7%, and for marginally subsidizing bequests by as much as 3.4%. These figures are robust to the value of the societal intergenerational discount factor and can yield significant welfare gains. In the second chapter, “Taxing Atlas: Using Firm Data to Derive Optimal Income Tax Rates” (joint with Laurence Ales and Jessie J. Wang), we analyze the optimal taxation of top labor incomes. Top income earners are modeled as managers who are heterogeneous across skills and operate a span-of-control technology, as in Rosen (1982). Managers privately observe their skill level, which increases the productivity of both effort and supervision, thus creating a scale-of-operations effect. We characterize optimal taxes in this environment and identify novel determinants linked to firm technology. Our main result is that to be consistent with U.S. firm data, the optimal top income tax rate should be roughly in line with the U.S. tax code, in contrast to previous results in the literature. In the third chapter, “Regional State Capacity and the Optimal Degree of Fiscal Decentralization” (joint with Martín Besfamille), we study the optimal degree of fiscal decentralization in a federation. In our environment, regional governments are characterized by two dimensions of state capacity; namely, administrative and fiscal. These gauge the ability to deliver public goods and to raise tax revenues, respectively. Two regimes are compared: partial and full decentralization. Under partial decentralization, regional governments have no tax powers and rely on central bailouts to refinance incomplete projects. Under full decentralization, regional governments refinance incomplete projects through capital taxes, in a context of tax competition. We show how the optimal degree of fiscal decentralization hinges on the relative magnitudes of each type of capacity. Specifically, for sufficiently low levels of fiscal capacity, bailing out regional governments is optimal regardless of the level of administrative ability. However, a combination of low levels of administrative capacity and high levels of fiscal capacity calls for fully decentralizing tax powers.