بدهی و مالیات : مدارک و شواهد از صنعت املاک و مستغلات

Compelling empirical evidence documenting a material effect of corporate taxes on leverage decisions is limited, in part because of difficulties in constructing an effective proxy for the firm's tax benefit of debt. We examine leverage decisions across taxable and nontaxable real estate firms—firms for which we can measure the relative tax benefit of debt with little error. The tax hypothesis implies that for firms with similar asset portfolios, taxable firms should have more debt than their nontaxable counterparts. Consistent with this, leverage ratios of taxable real estate firms are higher than their nontaxable counterparts, but the magnitude of this difference is at most one-half of that implied by studies that employ simulated marginal tax rates.

In corporate finance, a central question concerns the firm's optimal capital structure: Given the capital required to support a company's activities, how can financing be divided between debt and equity to maximize firm value? And this immediately leads to a second question: What are the important factors in determining this optimal leverage for a given firm? Modigliani and Miller (1958) lay the foundation for a positive theory of capital structure by developing implications of market equilibrium. Their analysis rigorously demonstrates that, given the firm's investment policy and facing no taxes or contracting costs, the firm's choice of financing policy has no effect on its current market value. This proposition thus has focused the profession's attention on those factors that are potentially important in determining optimal leverage. If corporate financing decisions do affect firm value, they must do so for at least one of the following reasons: (1) they affect taxes paid by corporations or their investors, (2) they affect the contracting or information costs borne by the various contracting parties that comprise the firm, or (3) they affect management's incentives to follow the value-maximizing rule of undertaking all positive NPV projects. In the development of our understanding of capital structure choices, taxes have played a prominent role. But early research was largely unsuccessful in documenting an empirical association between leverage and taxes. For instance, Myers (1984, p. 588) concludes, “I know of no study clearly demonstrating that a firm's tax status has predictable, material effects on its debt policy.” Recent evidence on the link between corporate taxes and leverage decisions has been more persuasive, but that evidence relies on a limited set of proxies for tax status. And these proxies share several limitations that arise from their reliance on financial statement information: limited variation in statutory tax rates, a disregard of future period tax rates, lack of compelling benchmarks, and interactions with investor-level tax rates. In this paper, we extend the literature on taxes and leverage by considering an alternative proxy for corporate tax status that is largely free of these limitations. We examine the capital structures of a set of firms in which measuring the relative tax benefits over the life of a debt instrument is reasonably straightforward. We analyze data from firms in the real estate industry—an industry which includes both taxable and nontaxable firms. A taxable real estate firm faces an entity-level corporate tax. But both a real estate investment trust (that distributes at least 90% of its taxable profits to shareholders each year) and a partnership avoid entity-level taxation. The marginal corporate tax rate for these firms – both currently and over the life of a debt issue – is effectively zero. The tax hypothesis implies that these firms should have a comparative disadvantage in issuing debt and therefore should select a lower level of target leverage than a similar taxable firm. Our research differs from prior work in two important ways. First, although a number of studies document that nontaxable real estate firms have high leverage, this evidence suggests only that there are important nontax benefits of debt in real estate firms. It says nothing about the importance of tax benefits: Without simultaneously examining the leverage choices of similar taxable real estate firms, these studies are unable to disentangle the leverage implications of operating within the real estate industry from those associated with the firm's tax status. Second, because our analysis of the tax effects of leverage exploits variation in tax status at the organizational level, our results are largely independent of the available evidence that focuses on variation in estimated marginal tax rates across taxable industrial firms. There is a material difference in the tax benefits of debt between taxable and nontaxable real estate firms, yet within the firm these tax benefits are reasonably stable over time. Therefore, our analysis of taxable and nontaxable real estate firms should provide valuable new evidence on the importance of tax status for corporate leverage choices. Our work is similar in spirit to Gentry (1994), who exploits variation in the tax status of oil and gas firms to address the role of taxes on leverage and dividend policy. In his two-year sample period covering 1986 and 1987 (a time of substantial uncertainty about future tax laws) oil and gas firms could organize either as regular taxable corporations or publicly traded master limited partnerships.2 Consistent with the tax hypothesis, Gentry finds that taxable corporations use more leverage. But this interpretation is less than straightforward since corporate shareholders also are shielded from liability for unpaid corporate debts which lowers the cost of using debt relative to the partnership form. In the real estate industry, the vast majority of publicly traded firms in our sample are taxable C corporations and REITs and both organizations afford investors limited liability. Finally, because the REIT form has been available and widely used for several decades, we have a longer sample period over which to test the underlying theories. When we compare leverage in taxable and nontaxable real estate firms with similar investment opportunities, we document results consistent with the tax hypothesis. Our estimates suggest that going from a marginal tax rate of 0% to 35% leads to a 4.7% higher leverage ratio—a figure that is less than one-half the magnitude of recent widely-cited research using simulated tax rates. In our analysis, we account for property type, changes in tax regimes over time, investor-level taxes and dividend policy, alternate measures of leverage and maturity structure, accounting biases, as well as potential selection problems. Taken together, our results suggest that taxes matter. But using more precise measures for the tax benefits of debt, our results suggest that the magnitude of the effect estimated in prior studies using taxable firms is materially overstated.

At least since Modigliani and Miller (1963), corporate finance theory has recognized that a potentially important benefit of leverage is the value of the corporate interest tax shield. Over the last four decades, a number of papers have employed an array of empirical strategies to identify the impact of this tax benefit. But these proxies face potentially important measurement and identification issues: they often rely on noisy financial statement data to derive tax status estimates, there is little time-series variation in the underlying statutory tax rate, and they do not account for important differences in state and country level taxation. In understanding the tax benefits of multi-period debt in the firm's capital structure, current and future tax benefits ought to be important. Yet the most popular measures of marginal tax rates in the academic literature from Shevlin (1990) and Graham (1996) are only designed to estimate the tax benefit of interest deductions in the current year. Taxable firms with low current marginal tax rates can expect to face higher marginal tax rates – and higher tax benefits of debt – in future years. This is a concern as these firms are effectively identified as firms with low tax benefits of debt. In contrast, firms like REITs and partnerships face persistently low taxes yet compete with fully taxable firms. This suggests that comparisons of tax status at the organization level in industries like real estate provide stronger and more persistent variation in tax benefits of debt. Our empirical methods avoid many of the problems with existing measures. We identify REITs and real estate partnerships as a set of firms where we know that the corporate tax benefit of debt – both currently and in the future – is zero. The tax hypothesis thus suggests that these firms should have less debt than their taxable counterparts. Moreover, prior to 2003 the tax code generally treated individual investments in REITs and taxable real estate firms symmetrically, hence any differential corporate tax advantages would not be offset by differential investor-level tax disadvantages. We test the tax hypothesis by comparing the leverage choices of taxable real estate firms to their nontaxable counterparts. Our results indicate that the costs and benefits of debt are closely related to the nature of the firm's assets and investment opportunities and that the tax deductibility of debt plays a secondary role, at best, in the firm's leverage decision. A generous interpretation of our evidence is that taxable firms have about 5% higher leverage ratios than similar non-taxable firms, and this estimate is just one-third of the effects estimated in prior work. For many specifications, taxable real estate firms have leverage that is generally not significantly higher than leverage in nontaxable real estate firms. This occurs despite a potential bias in the direction of the tax hypothesis driven by a payout policy-based increase in the derived demand for monitoring by debt holders of taxable firms. Because we do not rely on popular proxies for tax status, our evidence is largely independent of that in prior research. Moreover, our evidence suggests that more work is necessary to understand the dynamics of commonly used marginal tax rate estimates, the extent to which these proxies capture non-tax determinants of leverage, as well as whether and how managers incorporate current and future expected tax rates into the leverage decision.