Laurence Ales is Associate Professor of Economics and Christopher M. Sleet is Professor of Economics at Carnegie Mellon University’s Tepper School of Business. This post is based on a forthcoming article by Professor Ales and Professor Sleet.

In our article, Taxing Top CEO Incomes**,** we ask: what should the marginal income tax rate on high earning CEOs be? Recent research suggests that it should be high, perhaps as high as 70 per cent or 80 per cent. This research is based on a formula due to Diamond and Saez (2011) that relates the optimal marginal tax rate on top incomes to the elasticity of income with respect to taxes and a property of the right tail of the earnings distribution. This formula is derived under the assumption that the policymaker’s objective is to maximize tax revenues derived from top earners. It abstracts from any positive impact of the efforts of these earners on the incomes of other agents or on tax revenues collected from other sources. Our article departs from this research by taking seriously the idea that the activities of high earning CEOs, an important group of top earners, have positive spillovers for others. We use a ﬁrm-CEO assignment framework to model the market for CEO effective labor. Gabaix and Landier (2008) and Terviö (2008) have shown that such a framework is valuable for understanding recent growth in CEO incomes and the interaction of ﬁrm and CEO attributes in shaping this growth. We show that in an assignment model (augmented with a CEO effort choice), the taxation of CEO incomes affects the equilibrium pricing of CEO effective labor and, hence, spills over and affects ﬁrm proﬁts. At our benchmark parameterization, a full reform of CEO income and proﬁt taxation entails an optimal marginal tax on top CEO incomes of about 15 per cent.

The basic CEO to ﬁrm assignment model supposes one-to-one matching of differently talented CEOs to differently sized ﬁrms. As noted, we augment this model with a CEO effort choice (and with taxes). The equilibrium features assortative matching of CEO talent with ﬁrm size. More talented CEOs match with and supply more effort (and more effective labor) to larger ﬁrms. The indivisibility of the CEO position prevents combinations of less talented CEOs replacing more talented ones and equalizing the price for effective CEO labor across ﬁrms. On the other hand, competition amongst similarly talented CEOs for a position prevents any given CEO from extracting all of the surplus from a ﬁrm. In equilibrium the price of a unit of CEO effective labor equals the marginal product of CEO effective labor at the ﬁrm at which this unit is the last hired. Since the marginal product of CEO effective labor is increasing in ﬁrm size, the matching of more talented CEOs to larger ﬁrms enhances the dispersion of top CEO incomes. Even if there is relatively little dispersion in CEO talent, large variations in ﬁrm size can translate into large variations in top CEO incomes. However, since a CEO is only paid the marginal product of her effective labor on the last unit she sells to her ﬁrm (with infra-marginal units priced by and paid their marginal product at smaller ﬁrms), claimants to ﬁrm proﬁts capture some surplus. In this setting, an increase in the marginal tax rate above a threshold income induces an upwards adjustment in the pricing schedule for effective labor. This in turn redistributes from ﬁrm proﬁts to CEO incomes and, hence, CEO income tax revenues. If the policy maker is concerned only with maximizing income tax revenues, then this redistribution provides a motive for high marginal income taxes on CEOs. If, on the other hand, the policymaker is indifferent to the allocation between income tax revenues and ﬁrm proﬁts (because the latter can be taxed, because tax receipts and ﬁrm claimants are equally valued or because depressing ﬁrm proﬁts has adverse effects on ﬁrm creation), then no such redistribution motive for higher marginal taxes exists.

We use (nonlinear) optimal tax formulas expressed in terms of the underlying (structural) talent and ﬁrm size asset distributions to quantitatively characterize optimal taxes on CEOs. We find that if a comprehensive reform of income and ﬁrm proﬁt taxation is implemented, then optimal marginal tax rates decline from 20 per cent at an income of $10 million and to just 10 per cent at an income of $100 million. If a partial reform of CEO income taxation occurs holding proﬁt taxes close to their current statutory values in the US of about 60 per cent and abstracting from direct concern for proﬁt recipients or impact of proﬁts on ﬁrm entry then optimal tax rates decline from 34 per cent to 27 per cent over a similar income range. In either case they are very different from the rates of 70 per cent to 80 per cent recently advocated in the literature.

The full article is available for download here.