The impact of depreciation savings on investment: Evidence from the corporate Alternative Minimum Tax

Highlights

• I characterize a firm's investment incentives in the presence of AMT.

• I show that firms, the investment incentives of which are characterized by AMT system around 1999, increase investment significantly compared to firms never subject the AMT.

• I find the responsiveness coefficient estimated to be somewhat larger than the previous estimates.

• I conclude that the new identification strategy and high salience of AMT reform are likely the main factors that account for differences.

Abstract

Over the past decade, the United States has offered investment incentives in the form of larger depreciation savings, namely, bonus depreciation. The neoclassical investment model implies that investment responds to changes in depreciation savings, but there have been few direct attempts to investigate this implication. This paper examines investment patterns surrounding the 1999 shortening of the Alternative Minimum Tax (AMT) depreciation recovery periods, finding strong evidence that firms subject to the AMT increase investment in response to the AMT reform. The empirical results show that firms subject to the AMT increase their investment, measured as the ratio of capital expenditures to capital stock, by around 0.0418 to 0.0622, compared to firms subject to the regular tax. Given their average annual investment rate of approximately 0.27 during this period, the results imply a relative increase in investment of 15%--23%. By contrast, I find that the 2002 introduction of bonus depreciation, available both for firms subject to the regular tax and for firms subject to the AMT, affects both groups of firms similarly. The estimation uses an empirical specification developed from the Summers (1981) tax-adjusted q model, and the results imply that the responsiveness of investment to the tax term is somewhat larger than previously estimated.

Introduction

Investment has long been recognized by both economists and policymakers as an important factor in short-run aggregate demand fluctuations as well as long-run capital accumulation. Indeed, the United States government has frequently changed the three main tax instruments – corporate tax rates, investment tax credits, and depreciation allowances – because the neoclassical investment model implies that investment responds to changes in these instruments. However, empirical evidence was not very supportive until recently when the focus of empirical studies shifted to cross-sectional variations based on firm- or asset-level data, as in Auerbach and Hassett, 1991, Cummins et al., 1994, and Desai and Goolsbee (2004).

These studies use as the source of identification several tax reforms, enacted over the course of more than 30 years, that typically result in simultaneous changes to at least two of these tax instruments. One concern that arises is the possibility that these tax instruments may play asymmetric roles in describing investment incentives. As points out, the interest rates used by firms to discount future depreciation deduction streams in calculating the present value of depreciation allowances may be much higher than economists' likely assumptions. On the other hand, the effects of corporate tax rates and investment tax credits on tax liability are calculated in a straightforward manner by researchers, and perhaps by firms as well. Moreover, from the perspective of firms, recognition of investment tax credits would be more immediate and salient than that of generous depreciation allowances. Thus, in the presence of agency problems with short-tenured managers, generous depreciation allowances may not be as appreciated as investment tax credits, as it takes more time to recognize their benefits. In addition, Neubig (2006) argues that firms prefer lower corporate tax rates to higher depreciation allowances for other accounting and practical reasons.¹

Nevertheless, of the three major tax instruments, depreciation allowances have changed most frequently in the United States. Especially over the past decade, in hopes of stimulating the economy, investment incentives have been provided in the form of depreciation allowances, namely, bonus depreciation.² So far, mixed results have been reported regarding responses to the bonus depreciation policy. Edgerton (2009) finds no evidence for effects of the policy, whereas other researchers have found that firms exploit bonus depreciation by temporarily weighting longer-lived assets more heavily (House and Shapiro, 2008) or by using more tax-favorable financing methods such as leasing (Park, 2012). It remains unclear, though, whether levels of total firm-level investment are responsive to the recent changes in depreciation allowances.

There have been, to my knowledge, few attempts to investigate the responsiveness of investment to depreciation allowances, independently of other tax instruments. In this paper, I investigate whether firm investment is responsive to changes in depreciation savings by exploiting the 1999 change in the corporate Alternative Minimum Tax (AMT) depreciation rule that converged the previously disadvantaged AMT recovery period to the favorable regular depreciation recovery period.

I first extend the standard investment model to consider the conditions under which investment incentives are characterized by the AMT system. Then, from SEC 10-K filing data, I identify two groups of firms in terms of whether their investment incentives are characterized by the AMT or regular tax system. Using a difference-in-difference approach, I find strong evidence that the AMT firms increase investment after 1999.³ Specifically, my empirical results show that firms subject to the AMT respond to the reform by increasing investment, measured as the ratio of capital expenditures to capital stock, by around 0.0418 to 0.0622. Given their average annual investment rate of approximately 0.27 during this period, the results imply a relative increase in investment of 15%–23 %.

There is a stream of research that examines the effect of corporate tax instruments on firm-level investment, including but not limited to Cummins et al., 1994, Desai and Goolsbee, 2004, Edgerton, 2010. Cummins et al. (1994) reports significant empirical results using several specifications, one with user cost of capital as the tax incentive, and another with the tax-adjusted q. Desai and Goolsbee, 2004, Edgerton, 2010 also report significant estimates of the tax term coefficient. Essentially, they test whether firm investment is more likely to be increased in industries with assets that benefit more from tax reforms.⁴ The identifying assumption of the previous estimates is, therefore, the common responsiveness of investment to tax policies across different assets as well as industries, which is difficult to test.⁵

This paper employs a clean setting in which detecting the effect of depreciation allowances on firm-level investment is easier, and provides evidence that investment of the AMT firms responds quite strongly to the 1999 AMT reform. The estimation uses an empirical specification developed from the Summers (1981) tax-adjusted q model, and the results imply that the responsiveness of investment to the tax term is larger than previously estimated. In addition, I discuss how other periods can be made use of to isolate the impact of the tax reform, and document results consistent with the identifying assumptions, namely the common investment trends and tax-responsiveness between the AMT firms and non-AMT firms.

The rest of this paper is organized as follows. Section 2 provides a discussion of depreciation allowances and the corporate AMT. In Section 3, I formally introduce AMT to the firm maximization problem. Research design and data construction are explained in Section 4, and the empirical results are presented in Section 5. In Section 6, I discuss the implications of the empirical results for estimation of the tax-adjusted q model. Section 7 concludes.