Corporate Tax Avoidance and High Powered Incentives Mihir A. Desai Harvard University and NBER ([email protected]) Dhammika Dharmapala University of Connecticut ([email protected]) November 2004 Abstract This paper analyzes the links between corporate tax avoidance and the growth of high-powered incentives for managers. We develop a simple model that highlights the role of feedback effects between tax sheltering and managerial diversion in determining how high-powered incentives influence tax sheltering decisions. Then, we construct an empirical measure of corporate tax avoidance - the component of the book-tax gap not attributable to accounting accruals - and investigate the link between this measure of tax avoidance and incentive compensation. We find that increases in incentive compensation tend to reduce the level of tax sheltering, in a manner consistent with a complementary relationship between diversion and sheltering. In addition, consistent with a prediction of our model, we find some evidence suggesting that this negative effect is driven primarily by a subsample of firms with relatively weak governance institutions. Our results may help explain the “undersheltering puzzle,” and why large book-tax gaps are associated with subsequent negative abnormal returns. JEL Codes: G30, H25, H26, J33 Keywords: Incentive Compensation, Corporate Governance, Tax Avoidance, Stock Options, Tax Evasion We would like to thank an anonymous referee, Alan Auerbach, Amy Dunbar, Alexander Dyck, Courtney Edwards, Austan Goolsbee, Dirk Jenter, Mark Lang, Bryan Lincoln, John Phillips, Gary Richardson, Justin Tobias, John Wald, Luigi Zingales and various conference and seminar participants for valuable discussions and comments on an earlier version of this paper. We also thank Nathan Koppel and Philip West for their help with the example in Section 3. Desai acknowledges the financial support of the Division of Research of Harvard Business School and Dharmapala acknowledges the financial support of the University of Connecticut Research Foundation. Michael Park provided excellent research assistance.11. Introduction The extensive literature on how taxes influence firm financial decisionmaking, as reviewed in Graham (2003), has considered the effect of taxes on financing choices, organizational form and restructuring decisions, payout policy, compensation policy and risk management decisions. In this literature, taxes are viewed as one of many factors that shape these decisions. In contrast, firms also appear to engage in a variety of transactions that, in the words of Michael Graetz, are deals “done by very smart people that, absent tax considerations, would be very stupid.” These activities, broadly labeled corporate tax shelters, are believed to have proliferated so greatly that, according to some observers, they now constitute “the most serious compliance issue threatening the American tax system today.”1 What induces firms and managers to engage in transactions exclusively designed to minimize taxes? Alternatively, given the low perceived probabilities of detection, why don’t all firms engage in these transactions? How are shareholders affected by transactions that are nominally motivated by tax savings? Previous analyses of tax avoidance and evasion have emphasized the behavior of individuals (as in the literature reviewed in Slemrod and Yitzhaki (2002)), rather than corporations. Recently, Slemrod (2004) has stressed the differences between individual and corporate tax compliance, arguing that the latter should be analyzed in a principal-agent framework. This paper develops a simple theoretical framework that embeds the sheltering decision within a managerial agency context and emphasizes the importance of interactions between rent diversion and tax sheltering. In order to test this model, we construct an empirical measure of corporate tax avoidance – the component of the book-tax gap not attributable to accounting accruals – and investigate its determinants, linking sheltering decisions by US firms to their incentive compensation and corporate governance arrangements. The basic premise of the model is that decisions about corporate tax avoidance are made by firms’ managers. The analysis of these decisions is thus embedded in an agency framework, 1 “Tackling The Growth of Corporate Tax Shelters”; Treasury Secretary Lawrence H. Summers, in remarks to the Federal Bar Association, Washington, DC, February 28, 2000. For a review of typical tax shelter transactions and the policy issues, see U. S. Department of the Treasury (1999). One very crude metric of the possible magnitude of such activities is provided by the U.S. General Accounting Office (2004), which documents that (i) from 1996 to 2000, approximately one-third of large U.S. corporations reported zero tax liability, (ii) that share increased steadily over the period, and (iii) by 2000, 53% of large U.S. corporations reported tax liabilities lower than $100,000. For these purposes, large corporations are those with a minimum of either $250 million in assets or $50 million in gross receipts.2where managers may also be able to enjoy private benefits of control, for instance through the diversion of rents. Decisions about tax sheltering and rent diversion are made simultaneously and are potentially interdependent. In particular, the level of one activity may change the cost to the manager of engaging in the other. For reasons elaborated below, it is possible that engaging in sheltering may make diversion more costly or less costly from the manager’s viewpoint. We show that the relationship between the two activities is a critical determinant of how the increased use of incentive compensation changes sheltering decisions. Typically, higher-powered incentives will induce the manager both to reduce her diversion of rents and to engage in more tax sheltering activity. Greater incentive compensation helps align the incentives of agents and principals and leads managers to be more aggressive about increasing firm value through tax avoidance. However, interactions between sheltering and diversion can overturn this result. Specifically, when there are positive feedback effects or “complementarities” between diversion and sheltering, the tendency toward more aggressive sheltering may be offset by the fact that reduced diversion will be associated with reduced sheltering. If the positive