Capital Charge Rates, Investment Incentives and Taxation

Abstract

This paper studies the impact of personal and corporate income taxation on capital charge rates in a delegation setting with a risk-averse manager. If the investment level influences the riskiness of the investment project, the capital charge rate deviates from the firm's cost of capital and depends crucially on the manager's personal income tax rate. Contradicting conventional wisdom, we find that a higher personal income tax rate induces higher investment expenditures and, surprisingly, increases the capital charge rate. The countervailing effect that a higher capital charge rate induces higher and not lower investment expenditures persists for pre-tax and after-tax performance measures as well as when the tax deductibility of managerial compensation is limited. Corporate income tax causes a similar effect only in the case of limited tax deductibility of compensation. Our insights remain valid regardless of the financing structure and the risk attitude of the investors.

Acknowledgments

Special thanks go to the anonymous referee and the editor, Richard Sansing, whose comments significantly helped to improve the paper. We also thank Ralf Ewert, Steve Huddart, Bjorn Jorgensen, Rick Laux, Rainer Niemann, Thomas Pfeiffer, Andreas Scholze, Dirk Simons and Alfred Wagenhofer for their valuable comments and suggestions. We also acknowledge the helpful feedback of the participants at the ACMAR 2015 conference, the DART Accounting Workshop participants at the University of Graz and the seminar participants at the University of Vienna.

Notes

1 Lambert (2001).

2 According to Scholes et al. (2009) ‘Effective tax planning requires the planner to consider the tax implications of a proposed transaction for all parties to the transaction’. We follow this approach by implementing corporate income tax that affects the principal's payoff and personal income tax that affects the agent's payoff.

3 Capital charge rates below the firm's cost of capital are frequently observed in practice. In the sample of Poterba and Summers (1995), more than 25% of the respondent firms use capital charge rates that lie below the firm's cost of capital (see Figure 3 in Poterba and Summers, 1995).

4 After-tax earnings are used by 47.3% of the sample firms in Healy (1985), by 33.9% of the sample firms in Newman (1989), by 41.9% of the sample firms in Gaver, Gaver, and Austin (1995), by 26.3% of the sample firms in Carnes and Guffey (2000), by 70.4% of the sample firms in Atwood, Omer, and Shelley (1998), by 61.2% of the sample firms in Phillips (2003) and by 60.7% of the sample firms in Gaertner (2014). It is commonly argued that after-tax performance measures motivate managers to engage in actions that lower their firm's effective tax rate (e.g. Gaertner, 2014; Phillips, 2003).

5 In an attempt to tighten the link between compensation and manager performance, Section 162(m) of the Internal Revenue Code (enacted in 1993) limits the tax deduction for non-performance-based salary. See Halperin, Kwon, and Rhoades-Catanach (2001) or Göx (2008) for a theoretical analysis of the consequences of IRC Section 162(m) on incentive contracts without investment decisions and Perry and Zenner (2001) or Hall and Liebman (2000) for an empirical analysis. Recently, politicians in Europe started a discussion regarding the introduction of tax deductibility limitations to mitigate managerial overcompensation.

6 We use proportional income tax schemes. Progressive personal income tax schemes would complicate our analysis without shedding new light on the fundamental effects of taxation on capital charge rates.

7 Many authors assume normally distributed cash flows. In this case, the cash flow and the manager's salary can be negative. As negative salaries are unrealistic, it is usually argued that if the expected return is sufficiently high, the probability that the salary is below zero can be neglected. All of our results remain valid if we allow the investment return to be normally distributed.

8 To rule out situations in which the manager with risk aversion parameter α is in a majority ownership position, i.e. pay-performance sensitivity parameters over , we require that the investment return contains a base level of uncertainty, that does not disappear with I = 0. One can interpret as the sum of a market-specific and an investment-specific component, where and are uncorrelated random variables with expected values and variances and f is a concave function with .

9 The concavity of the investment return is a standard economic assumption supported by the law of diminishing returns. Furthermore, higher levels of investment are usually associated with incremental complexity and hence higher levels of uncertainty.

10 The mean–variance objective function is an approximation of the manager's certainty equivalent for a large class of utility functions (e.g. Milgrom and Roberts, 1992, p. 247) and commonly used in literature (e.g. Feltham and Wu, 2001; Melumad, Weyns, and Ziv, 1999).

11 Note that an increase in the personal income tax rate can also have a positive effect on the willingness to provide effort. This is the case when an individual raises his effort level in order to achieve a certain compensation level (income effect). Our model does not capture the income effect on the manager's effort choice as we assume that the manager's wealth and effort are additively separable.

12 For the sake of a simple notation we write for and for

13 Generally, the incentive-intensity principle defines incremental profits from additional effort as the fourth factor determining the optimal intensity of incentives (e.g. Milgrom and Roberts, 1992, p. 221). In our model, the marginal return to the agent's effort is equal to one and appears in the numerator of the optimal pay-performance sensitivity.

14 The optimal capital charge rate is bounded below and reaches its lower bound at The lower bound depends on the characteristics of the outcome risk function, as well as on the manager's risk aversion parameter, α. For any α the optimal capital charge rate does not go below Obviously, the optimal capital charge rate can be less than zero. Negative capital charge rates can be interpreted as a capital subsidy in the manager's performance measure and are of similar nature as the negative interest rates that we observe nowadays in practice.

15 The tilde sign is used to denote variables in the after-tax performance measure scenario.

16 See Halperin et al. (2001) or Göx (2008) for a theoretical analysis of the consequences of IRC Section 162(m) on incentive contracts without investment decisions.

17 In Austria, for example, the Tax Amendment Act 2014 introduced the non-deductibility of annual manager remunerations above EUR 500,000. The specific proposals made by German parties differ from each other by different thresholds and percentages up to which compensation is partially deductible. See Voßmerbäumer (2012) for a detailed analysis of the different proposals. In our analysis, we assume that the entire compensation is subject to the deductibility limit, that is, we do not include a threshold up to which the compensation is fully deductible. This allows us to focus on the effects of deductibility limits when they are binding and to obtain closed-form solutions for our model parameters.

18 Hat signs denote variables in the deductibility scenario. As before, superscript ‘†’ denotes optimality.

19 This argumentation also holds if only the tax deductibility of the fixed payment is limited as it is the case in Section 162(m) of the Internal Revenue Code. Limiting the deductibility of a part of the fixed payment increases the firm's relative cost of inducing effort, and thus leads to a reduced pay-performance sensitivity and an increased investment level compared to the case where the salary is fully deductible.

20 For undesired side-effects of deductibility limits see e.g. Conway (2008).

21 Clearly, all results concerning personal income tax obtained in Section 3.1 remain valid if only a part of the manager's salary is tax deductible.

22 Bar signs denote variables in the mixed financing scenario. As before, superscript ‘†’ denotes optimality.

23 Target capital structures refer to the idea that a firm chooses how much debt finance and how much equity finance to use by balancing various costs and benefits of debt and equity (so called ‘trade-off theory’, see e.g. Myers, 1984). Evidence that firms have target capital structures can be found, for instance, in Schwartz and Aronson (1967), Marsh (1982) or Hovakimian, Opler, and Titman (2001).

24 We use inverted hat signs to denote variables in this section.

25 If the investors are risk-neutral, is equal to zero.

26 This is based on a single-factor capital asset pricing model, see e.g. Ohlson (1990).

27 As before, superscript ‘†’ denotes optimality.