We study the optimal degree of fiscal decentralization in a federation. Regional governments are characterized by their abilities to deliver public goods (administrative capacity) and to raise tax revenues (fiscal capacity). Two regimes are compared on efficiency grounds. Under partial decentralization, regional governments rely on central bailouts to complete local projects in financing needs. Under full decentralization, marginal financing is achieved via local capital taxes. We show that the presence of sufficiently low levels of administrative capacity is a necessary condition for full decentralization dominance. This condition may also be sufficient, depending on the projects' characteristics. Some extensions are presented.

We study the optimal taxation of top labor incomes. Top income earners are modeled as managers who operate a span of control technology as in Rosen (1982). Managers are heterogeneous across talent, which is both effort-augmenting and total-factor-productivity improving. The latter gives rise to a positive scale-of-operations effect. A tax formula for optimal taxes is derived linking optimal marginal tax rates to preferences and technology parameters. We show how to quantify the model using readily available firm-level data. Our benchmark calibration focuses on the US. Our results suggest that optimal top taxes are roughly in line with the current statutory rates and, thus, are significantly lower than what previous optimal taxation studies that ignore the scale-of-operations effect have shown.Similar quantitative findings hold when we extend the analysis to a panel of developed countries. (JEL D31, H21, H24, M12, M52) * We thank Jeremy Bertomeu and Adriano Rampini for valuable discussions. We also thank Hank Bessembinder, Emilio Bisetti, Chris Edmond, Miguel Ferreira, Brent Glover, Burton Hollifield, Roozbeh Hosseini, Greg Kaplan, Finn Kydland, Christos Makridis, Pricila Maziero, Tomasz Piskorski, B. Ravikumar, Peter Rupert, Florian Scheuer, Ali Shourideh, Chris Sleet, Stephen Spear, Chris Telmer, Gustavo Ventura, Sunil Wahal, Randy Wright, Sevin Yeltekin, Ariel Zetlin-Jones and Marianne Holohan for editorial assistance. We also thank conference and seminar participants at Carnegie Mellon Ph.D. Macro-Finance lunch, Econometric Society summer meetings in Minneapolis, Society for Economic Dynamics annual meeting in Toronto, 5 th Tepper-LAEF conference in Pittsburgh, the University of Melbourne, 2015 WAMS/LAEF conference in Sydney, Arizona State University, University of Arizona, 2016 UNC Tax Symposium, and UNSW for feedback. E-mail: ales@cmu.edu (Ales); a.bellofatto@uq.edu.au (Bellofatto); Jessie.Jiaxu.Wang@asu.edu (Wang). I N T R O D U C T I O NHeightened concerns over recent trends in income inequality necessarily bring the taxation of high earners to the forefront of the policy agenda. 1 However, the vast literature in public finance is far from reaching consensus on what the top income tax rate should be: While the dominant view sets the optimal top tax rate above 70 percent, others have challenged the validity of such elevated taxes on various grounds. 2 This paper introduces a novel approach to modeling and quantifying the behavior of highly talented individuals in the economy, thus contributing to the aforementioned debate. Within top income earners, we focus on managers and show how to exploit well-established facts on firm size and managerial compensation to pin down the key forces shaping optimal income taxes. We find that top income taxes at the optimum are actually aligned with the current statutory rates in the US, unlike prevailing recommendations in the literature.Our approach is to take the standard optimal taxation environment and augment it with a role for managers via a span of control...

We study the optimal taxation of top labor incomes. Top income earners are modeled as managers who operate a span of control technology as in Rosen (1982). Managers are heterogeneous across talent, which is both effort-augmenting and total-factor-productivity improving. The latter gives rise to a positive scale-of-operations effect. A tax formula for optimal taxes is derived linking optimal marginal tax rates to preferences and technology parameters. We show how to quantify the model using readily available firm-level data. Our benchmark calibration focuses on the US. Our results suggest that optimal top taxes are roughly in line with the current statutory rates and, thus, are significantly lower than what previous optimal taxation studies that ignore the scale-of-operations effect have shown.Similar quantitative findings hold when we extend the analysis to a panel of developed countries. (JEL D31, H21, H24, M12, M52) * We thank Jeremy Bertomeu and Adriano Rampini for valuable discussions. We also thank Hank Bessembinder, Emilio Bisetti, Chris Edmond, Miguel Ferreira, Brent Glover, Burton Hollifield, Roozbeh Hosseini, Greg Kaplan, Finn Kydland, Christos Makridis, Pricila Maziero, Tomasz Piskorski, B. Ravikumar, Peter Rupert, Florian Scheuer, Ali Shourideh, Chris Sleet, Stephen Spear, Chris Telmer, Gustavo Ventura, Sunil Wahal, Randy Wright, Sevin Yeltekin, Ariel Zetlin-Jones and Marianne Holohan for editorial assistance. We also thank conference and seminar participants at Carnegie Mellon Ph.D. Macro-Finance lunch, Econometric Society summer meetings in Minneapolis, Society for Economic Dynamics annual meeting in Toronto, 5 th Tepper-LAEF conference in Pittsburgh, the University of Melbourne, 2015 WAMS/LAEF conference in Sydney, Arizona State University, University of Arizona, 2016 UNC Tax Symposium, and UNSW for feedback. E-mail: ales@cmu.edu (Ales); a.bellofatto@uq.edu.au (Bellofatto); Jessie.Jiaxu.Wang@asu.edu (Wang). I N T R O D U C T I O NHeightened concerns over recent trends in income inequality necessarily bring the taxation of high earners to the forefront of the policy agenda. 1 However, the vast literature in public finance is far from reaching consensus on what the top income tax rate should be: While the dominant view sets the optimal top tax rate above 70 percent, others have challenged the validity of such elevated taxes on various grounds. 2 This paper introduces a novel approach to modeling and quantifying the behavior of highly talented individuals in the economy, thus contributing to the aforementioned debate. Within top income earners, we focus on managers and show how to exploit well-established facts on firm size and managerial compensation to pin down the key forces shaping optimal income taxes. We find that top income taxes at the optimum are actually aligned with the current statutory rates in the US, unlike prevailing recommendations in the literature.Our approach is to take the standard optimal taxation environment and augment it with a role for managers via a span of control...

In assessing the desirability for tax decentralization reforms, a dilemma between efficiency and redistribution emerges. By limiting the ability of the central government to redistribute resources towards regions in financial needs, decentralization curbs incentives for excessive subnational spending and enhances fiscal discipline, but may also widen interregional disparities by triggering tax competition for mobile tax bases. We provide a formal treatment of this trade-off, and shed light on the optimal degree of fiscal decentralization. We find that tax decentralization can be optimal even under Rawlsian social preferences which only weight the welfare of the poorest region in the federation.

In assessing the desirability for tax decentralization reforms, a dilemma between efficiency and redistribution emerges. By limiting the ability of the central government to redistribute resources towards regions in financial needs, decentralization curbs incentives for excessive subnational spending and enhances fiscal discipline, but may also widen interregional disparities by triggering tax competition for mobile tax bases. We provide a formal treatment of this trade-off, and shed light on the optimal degree of fiscal decentralization. We find that tax decentralization can be optimal even under Rawlsian social preferences which only weight the welfare of the poorest region in the federation.