The Comprehensive Tax Gain from Leverage

Abstract

We expand the traditional specification of the tax gain from leverage by accounting for the choice between issuing debt and utilizing internal retained earnings equity. Standard analyses focus solely on the choice between debt and external equity. This results in the traditional tax gain from leverage, which equals the corporate tax benefit from debt minus the partially offsetting personal tax disadvantage. Expanding the debt analysis to include internal equity introduces a third tax component to the gain from leverage: a supplemental personal tax disadvantage. This supplemental disadvantage reflects the personal tax cost of distributing internally-generated equity to investors and using replacement debt. It reduces the overall or comprehensive tax gain from leverage and frequently converts it into a tax loss. Therefore, using debt often is costlier than generally assumed. We provide conceptual support for the comprehensive tax gain, plus some initial empirical support.

Keywords:

• Gain from leverage
• Debt tax shield
• Pecking order
• Capital structure

JEL codes:

• G32
• H21
• H24
• H25
• H32

Acknowledgements

We express sincere gratitude to our editor Martin Jacob. We also thank two very helpful anonymous referees, Alan J. Auerbach, Harry DeAngelo, R. Lynn Hannan, our editor, Feng Jiang, Mervyn A. King, Frank T. Morgan, and Toni Whited for their insights and comments on previous drafts of the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 We also demonstrate that the supplemental personal tax disadvantage from debt could discourage the use of internal-equity financed dividends (see section 6). By deterring the payout of internally-generated cash, the supplemental tax disadvantage joins forces with other factors that favor cash holdings, including the lock-in effect from repatriation taxes and the need for precautionary cash holdings (see Han & Qiu, 2007; Foley et al., 2007; Bates et al., 2009).

2 To isolate capital structure effects, it is critical to follow Miller (1977) by holding operating assets constant. From a modeling perspective, the traditional way to accomplish this is to focus on the financing choice for a fixed amount of new operating investment. Note that financial assets (e.g. excess cash) offset financial liabilities to produce net debt (Alderson & Betker, 2012), so financial assets are best viewed as a component of the financing side of the balance sheet (Feltham & Ohlson, 1995).

3 Both financing options for the new investment increase operating assets by the same amount. Both options also have the same impact on the firm’s cash holdings. Option (a) increases cash via the debt issuance and decreases cash by an equal amount for the equity distribution to shareholders. Likewise, option (b) has no effect on pre-investment cash. Then under either option, the firm uses the same fixed amount of cash to invest in the new operating assets. The difference between option (a) and option (b) is that only option (a) increases debt and the ratio of debt-to-operating assets.

4 The choice to refinance existing assets this way does not affect the firm’s operating assets or cash holdings. Although the transaction has a double impact on the firm’s debt-to-equity ratio, increasing the numerator and decreasing the denominator, it only affects the numerator of the firm’s debt-to-operating assets ratio, as required to isolate capital structure effects effectively.

5 Auerbach (2002) anticipates this result by suggesting that under the New View of dividend taxation the dividend tax rate disappears from the decision between debt and internal equity. If τ_e = 0, then from a tax perspective the decision rests solely on the relative values of τ_c and τ_i.

6 In King expression (16), when setting tax rates constant, the firm borrows if θ (1−z) α > θ (1−z) (1 + πi).

7 Specifically, per equation (4), the comprehensive tax gain is equal to 0.21 + (1–0.21)0.238–0.37 – (1–0.37)0.238 = −0.122.

8 See Hennessy and Whited (2005) equation (25) for a precise specification of the financing gap.

9 Although internal-equity financed dividends reduce financial assets, they keep operating assets fixed, which is the relevant asset base for a net debt analysis. The choice then is the amount of net debt to use to finance the fixed operating assets.

10 To illustrate, consider the comprehensive tax gain we calculate for Russia in table 2. Paying out a dividend without issuing replacement debt removes the future earnings on the distributed assets from the corporation, thus providing a corporate tax benefit on future earnings equal to 20%. The immediate distribution also enables investors to avoid the dividend tax rate of 13% on future distributions. Adversely, however, the dividend triggers an immediate dividend tax of 13% and it subjects high-tax investors to a personal 13% tax on future interest earnings. Overall, the tax benefits of the dividend outweigh the tax costs, so the net effect is a comprehensive tax gain from paying out a dividend. By contrast, a comprehensive tax loss indicates a net tax cost from dividends.

11 Many other factors also affect dividend incentives, including signaling and agency incentives (Watts, 1973; La Porta et al., 2000). A broad equilibrium for capital structure, liquidity, and equity distribution policy would include these signaling and agency dividend incentives, as well as factors relating to financial distress, precautionary cash holding, expected future free cash flows, investment opportunities, available collateral, debt covenants, information asymmetry costs, risk, and more. We posit that the comprehensive tax gain from leverage would be one factor that should affect optimal policy for budget-unconstrained firms in this type of broad equilibrium, where the budget status of a firm is itself an endogenous factor. An even broader equilibrium could endogenize the potential impact of personal taxes on the level of investment according to the cost-of-capital analyses for internal equity in Auerbach (1979), Sinn (1991), Harris and Kemsley (1999), and Guenther and Sansing (2006), among others. In this study, our objective is limited to specifying the comprehensive tax gain from leverage and assessing its conceptual relevance.

12 In the full sample of 119,946 observations for the 1983–2017 period that we use for the tests in section 7, 30.5% report positive dividends, 35.8% report positive share repurchases, 16.7% report positive dividends and positive share repurchases, and 49.6% report either positive dividends or positive share repurchases.

13 Both debt-financed dividends and internal-equity financed dividends are debt-for-equity recapitalizations. Only external-equity financed dividends are not debt-for-equity recapitalizations, which are rare. For a discussion of the different financing sources for dividends, see Kemsley et al. (2018).

14 Consistent with Hennessy and Whited (2005), we use the top statutory dividend tax rate (τ_d) to proxy for the personal tax rate on equity, which enables us to use the same measure for the debt-usage and dividend tests. The alternative would be to use the capital gains tax rate, which is relevant for share repurchases, and which future research could explore.

15 We also replace Debt-to-Value with the ratio of debt to the book value of equity and the ratio of debt to the market value of equity. When using debt to the book value of equity, the estimated CorpTax coefficient is positive (2.740, t = 4.39), the TradTaxDis coefficient is negative (−2.826, t = −6.77), and the estimated SuppTaxDis coefficient also is negative (−1.315, t = −4.33). Similarly, when using debt to the market value of equity, the estimated CorpTax coefficient is positive (1.010, t = 15.86), the TradTaxDis coefficient is negative (−0.893, t = −21.63), and the estimated SuppTaxDis coefficient also is negative (−0.448, t = −14.15).

16 These years are 1987, 1988, 1991, 1993, 2001, 2002, 2003, 2006, 2008, 2010, and 2013. Empirical results are similar when using all sample years.