The Impact of Taxes on Competition for CEOs

Abstract

This paper contributes to the question of how taxation of corporate profits and wages affects competition among firms for highly skilled human resources such as CEOs. Use of a theoretical model shows that wage taxes can have a substantial impact on the outcome of such a competition if marginal tax rates are different as in an international labor market. Further, the paper shows that increasing the wage tax rate unilaterally can have an ambiguous effect on observed gross compensation levels. However, in a local labor market for CEOs, observed gross fixed salaries should decline in the wage tax rate. Tax effects in a market for CEOs is a particularly interesting topic because recent developments with respect to compensation practices of top-level managers have opened a public debate about the use of instruments for regulating compensation of those managers. Furthermore, many countries around the world use tax incentives in order to facilitate immigration of highly skilled human resources. The investigation follows an analytical economics-based approach by extending an LEN model with elements of competition for scarce human resources and income taxation. It investigates the impact of differential taxation on the competition between two firms for the exclusive service of a unique, highly skilled CEO.

Acknowledgements

For their valuable suggestions I would like to thank Richard Sansing, Martin Jacob (Guest Editors) and an anonymous reviewer. Further, I am grateful to Laura Brandstetter, Claire Estebanez, Thomas Kourouxous, Sebastian Kronenberger, Rainer Niemann, Martina Rechbauer, Susann Sturm, Caren Sureth-Sloane for their useful comments. This paper also benefited from the participants of the 2015 DART minigraduate workshop at the University of Graz, the participants of the 2nd Doctoral Research Seminar 2015 at the Vienna University of Economics and Business, and the participants of the 1st Berlin-Vallendar Conference on Tax Research 2015 at Freie Universität Berlin.

Notes

1 Excluding Facebook as an extreme outlier still yields a ratio of 295.9-to-1 for the year of 2013.

2 Throughout the paper, the terms ‘CEO’, ‘manager’, and ‘agent’ are used interchangeably.

3 Other trends include for instance relatively high insensitivity of CEO compensation to firm performance and reward for luck. Edmans and Gabaix (2009) provide a brief review about these trends and discuss a series of potential explanations for them.

4 Section 162(m) IRC has been introduced in 1993. It limits deductions for non-performance-based compensation of the CEO and the four highest compensated officers to $1 million per year.

5 Subscript t indicates the corresponding after-tax expressions for all variables.

6 It should be noted that the agent's gross compensation is based on the after-tax cash-flow . Other possible performance measures could be the firm's after-tax profit (e.g. Ewert and Niemann, 2012) or profit before taxes. In a recent study, Gaertner (2014) shows that nearly 61% of his sample firms use after-tax incentives. This result indicates heterogeneity across firms with respect to the use of after-tax performance measures.

7 Superscript fb indicates the case of observable effort whereas superscript sb denotes the case of non-observable effort.

8 It should be noted that the optimal level of effort decreases in the wage tax, due to the substitution of income with leisure time. With the specified utility function, wage taxes do not perform an income effect as proposed by Atkinson and Stiglitz (1980) because there does not exist an additional source of income. However, according to Kuhn (2015) the absence of an income effect could also result from an individual's ability to shift income and labor supply from one period to another (e.g. through perfect capital markets) which serves as a possible justification for the model in use.

9 All equilibrium values associated with the situation of uniform (homogeneous) taxation are denoted by a superscript h.

10 From a formal perspective, it should be noted that firm 1 has to overbid the offer of firm 2 by a a small amount in order to hire the CEO with certainty. However, throughout this paper this technical necessity is neglected because ϵ can be very close to zero and including it does not yield any further insights (see also the proof in the Appendix). Alternatively, someone could also simply assume, that an indifferent CEO chooses always the firm earning a higher expected profit.

11 All equilibrium values associated with the situation of differential taxation a denoted by a superscript d.The proofs are similar to those in Proposition 1 and 2 and therefore omitted for brevity. This is also done for the remaining propositions in this paper for the same reason.

12 The case of is ignored here, as it yields the same result as with equal compatibility and uniform taxation presented in Proposition 1.

13 It should be noted that this result is based on an assumption regarding the after-tax expected profits from hiring an ordinary manager (the firms' outside opportunity): It is assumed that these after-tax expected profits become equal to a unified market ‘return’ due to competition. (That is, ). Additionally differences in corporate tax rates might become relevant, if parts of the CEO's compensation are non-deductible for the firm. Halperin et al. (2001) and Göx (2008) provide an extensive analysis of this issue in the context without competition for CEOs.

14 The case of is ignored here, as it yields the same result as with equal compatibility and uniform taxation indicated in Proposition 1.

15 With observable effort, the fixed salary offered by firm w is strictly decreasing in the wage tax rate because . Therefore, it is possible that private information and the implied risk effect does not only have an impact on the competition winner but also on marginal effects of the wage tax rate.

16 Bénabou and Tirole (2016) investigate such a scenario without considering any issues of international taxation.

Additional information

Funding

This work was supported by the Austrian Science Fund under Grant W1229.