AN IMPLICIT CLIENTELE TEST OF THE RELATIONSHIP BETWEEN TAXATION AND CAPITAL STRUCTURE

This paper presents a test for the existence of high and low tax bracket debt clienteles. The test generates some evidence consistent with the implication of debt clientele theory that over time, firms' debt ratios should vary with the relative tax incentives for investors to hold debt. The results, however, do not support the existence of extreme financial leverage clienteles. Rather they are consistent with an incomplete equity market: low tax bracket investors hold the stock of highly levered firms and high tax bracket investors hold the stock of the low leverage firms.     

Financial leverage clienteles: Theory and evidence

This paper examines the hypothesis that investors will sort themselves out into tax-induced ‘financial leverage clienteless’ in which the common stocks of highly levered firms will be held by individuals with low personal tax rates, while the shares of firms with little or no leverage will be held by individuals with high personal tax rates. Although the idea of financial leverage clienteless has appeared in the literature before, the immediate motivation for this investigation is a recent paper by Merton Miller. In that paper he argues that under the current U.S. tax structure, personal taxes will offset corporate taxes such that in equilibrium the value of any individual firm will be independent of its use of debt financing. We extend his analysis to show specifically the way in which financial leverage clienteles would come about in his assumed tax environment. We then conduct some direct empirical tests of the leverage clientele hypothesis. These tests can also be viewed as indirect tests of Miller's new proposition on the irrelevance of capital structures. The results of the tests are mixed: The relationship between corporate leverage policies and investors' tax rates is statistically significant, but its magnitude is less than would be predicted by the theory.     

Changes to Income Trust Taxation in Canada: Investor Reaction and Dividend Clientele Theory

Abstract:  This paper explores the relationship between tax-induced dividend clientele theory and the recent changes to the taxation of income trusts in Canada. On October 31, 2006, the Canadian government announced the Tax Fairness Plan ( TFP ) calling for the elimination of the considerable tax advantage enjoyed by income trusts. Generally, distributions from income trusts are now taxed at rates comparable to those imposed on corporate dividends. We examine market reaction to the  TFP  to address three issues: first, whether the valuation effect of a dividend tax increase is consistent with the traditional or the new view of dividend taxation; secondly, whether the market reaction to tax increases has a differential impact on firm value that is related to the tax preferences of taxable, tax-exempt, and foreign investor tax clienteles; and thirdly, whether firms change their dividend policies in response to the preference of institutional investors (tax-based dividend policy effect) or whether institutional investors are sorting themselves across firms based on their dividend policies (investor sorting effect). Our results provide strong evidence as follows. First, the valuation effect in reaction to the  TFP  announcement is consistent with the traditional view of dividend taxation – i.e. that taxes on dividends reduce the net return to investors, increase the firm's cost of capital and lower the firm's ability to access capital markets, thereby discouraging investment and savings. Secondly, we saw that trusts with a larger percentage of their units held by tax-exempt, low-tax, and foreign investors had a higher decline in value when compared with trusts held mostly by ordinary taxable investors. These results support dividend tax clientele theory. Finally, we observed changes in institutional investor clienteles consistent with the investor sorting effect.     

Mutual Fund Tax Clienteles

Mutual funds are held by investors in taxable and tax‐qualified retirement accounts. We investigate whether the characteristics, investment strategies, and performance of mutual funds held by these diverse tax clienteles differ. Examining both mutual fund distributions and mutual fund holdings, we find that funds held primarily by taxable investors choose investment strategies that result in lower tax burdens than funds held primarily in tax‐qualified accounts. Despite these differences, we find no evidence that any investment constraints that may arise from these tax‐efficient investment strategies result in performance differences between funds held by different tax clienteles.     

The Puzzle of Financial Leverage Clienteles

Empirically, it appears that common stock of publicly traded corporations with high-debt ratios tends to be held by investors with relatively low marginal taxes while the stock in companies with little debt is held by investors in high-tax brackets. A number of authors have argued that in an equilibrium similar to the one described by Miller [8], these clienteles should exist. We argue that standard portfolio theory does not imply financial leverage clienteles for publicly traded firms. We explain the empirical relationship between investor tax rates and leverage ratios by the existence of dividend clienteles and a positive relationship between dividend yield and leverage ratios.     

Bankruptcy Risk and Optimal Capital Structure

This study finds shortcomings in empirical tests of the capital structure irrelevance hypothesis. The alternative hypothesis is that firms choose value maximizing mixes of debt and equity on account of bankruptcy costs and the tax deductibility of interest payments. Based upon the cross-sectional implications of the tax shelter-bankruptcy cost hypothesis, an alternative test of the irrelevance hypothesis is performed. The test examines the relationship between failure rates and leverage ratios for 36 lines of business. The results are inconsistent with the irrelevance hypothesis.     

Ownership structure and tax-friendly dividends

The paper examines the existence of tax-based dividend clienteles using the novel environment of Australia, which has operated a full dividend imputation system since 1987. The analysis jointly focuses on the tax-based preferences of five categories of shareholders, including both domestic and foreign-domiciled shareholder classes. Incorporating the dividend franking percentage as a direct measure of the degree of tax benefit associated with dividends, strong evidence supporting the existence of tax-based dividend clienteles is present for both domestic and foreign shareholder categories. This includes domestic corporate blockholders and company directors, and local institutional investors following tax reforms in 2000, and foreign institutional shareholders which, alternatively, demand lower dividends and dividend franking. These findings persist after considering the effect of share repurchases, and under various model specifications controlling for unobserved firm heterogeneity and potential endogeneity between ownership structure and dividend payout policy.     

Do Dividend Clienteles Exist? Evidence on Dividend Preferences of Retail Investors

We study stock holdings and trading behavior of more than 60,000 households and find evidence consistent with dividend clienteles. Retail investor stock holdings indicate a preference for dividend yield that increases with age and decreases with income, consistent with age and tax clienteles, respectively. Trading patterns reinforce this evidence: Older, low‐income investors disproportionally purchase stocks before the ex‐dividend day. Furthermore, among small stocks, the ex‐day price drop decreases with age and increases with income, consistent with clientele effects. Finally, consistent with the behavioral “attention” hypothesis, we document that older and low‐income investors purchase stocks following dividend announcements.     

Tax planning, regulatory capital planning, and financial reporting strategy for commercial banks

We test whether banks' investment and financing policies canbe explained by tax status. We document changes in bank holdingsof municipal bonds in response to changes in tax rules relatingto deductibility of interest expense. We also document an associationbetween banks' marginal tax rates and their investment and financingdecisions, which is consistent with the existence of tax clienteles.However, banks do not sort themselves perfectly into investmentand financing clienteles because of adjustment costs. We positspecific types of transaction-cost impediments to tax planning,and document that banks apparently trade off these costs againsttax-planning benefits.     

On the Relevance of Debt Maturity Structure

In this paper, we present a tax-induced framework to analyze debt maturity problems. We show that under some modifications of the existing U.S. tax code, debt maturity is irrelevant even in the presence of taxes and bankruptcy costs that yield an optimal capital structure. If this restrictive structure is relaxed, and assuming the Miller [15] equilibrium does not prevail, tax reasons would usually imply the existence of an optimal debt maturity structure. If there exists a gain from leverage, then an increasing term structure of interest rates, adjusted for default risk, results in long-term debt being optimal. A decreasing term structure, under similar circumstances, renders short-term debt optimal. In the absence of agency costs, a Miller [15]-type result emerges at equilibrium and irrelevance prevails. We also argue that agency costs could again reverse the irrelevance and imply a firm-specific optimal debt maturity structure.     

Payout Policy Tax Clienteles,Ex‐dividend Day Stock Prices and Trading Behavior in Germany: The Case of the 2001 Tax Reform

This paper examines the impact of the German 2001 tax reform, where Germany switched from a full imputation system to a classical system. Theory suggests that both price drop ratios and trading volume decrease following the reform. We document a significant reduction in the valuation of net dividends–in particular for high dividend yield stocks–and weakening payout policy tax clienteles. Ex‐dividend day returns are likely to be driven by short‐term traders. Though the reform removed incentives for cross‐border dividend stripping and reduced tax heterogeneity among investors, we show that the high trading volume around ex‐dividend days persists.     

Capital Gains Tax Rules, Tax-loss Trading, and Turn-of-the-year Returns

Changes in the capital gains tax rules facing individual investors do not affect the incentives for "window dressing" by institutional investors, but they can affect the incentives for year-end tax-induced trading by individual investors. Empirical evidence for the 1963 to 1996 period suggests that when the tax law encouraged taxable investors who accrued losses early in the year to realize their losses before year-end, the correlation between early year losses and turn-of-the-year returns was weaker than when the law did not provide such an early realization incentive. These findings suggest that tax-loss trading contributes to turn-of-the-year return patterns.     

Effects of Capital Gains Taxation on Life-Cycle Investment and Portfolio Management

We examine the impact of capital income taxation, both accrual forms of taxation and taxation of realized capital gains, on total savings and the demand for corporate financial instruments. We find that investors may hold both debt and equity in the face of effective collection of capital gains taxation even in a flat tax system. We also find that the two taxes will have substantially different effects on saving and consumption behavior, making it unlikely that the tax structure can be summarized by any single equivalent accrual tax rate.     

The Ex-Dividend Day Behavior of Canadian Stock Prices: Tax Changes and Clientele Effects

With some simple assumptions the ex-dividend day price drop and the associated dividend can be used to measure the market's marginal tax rate. Previous research has estimated the implied tax rate for the U.S. This paper extends the analysis to Canada, where the tax treatment of dividends and capital gains is completely different from that in the U.S. The paper also presents estimates from 1970–80 to include four distinct periods when the tax treatment was different. Hence, we include an implied test of market efficiency as well as those for the “relevance” of taxes and the existence of tax based dividend clienteles.     

Do Tax Law Changes Influence Ex-Dividend Stock Price Behavior? Evidence from 1926 to 2005

We test the Elton and Gruber model of ex-dividend stock pricing over a period spanning all US tax law changes since 1926. Our results indicate that price drop ratios (ΔP/D) and ex-day returns are related to dividend and capital gains tax rates in the theorized manner. Consistent with tax clienteles, we also find that ex-day price movements of higher dividend yield stocks are driven more by corporate tax rates, while lower yield stocks are more influenced by personal rates. Finally, we demonstrate that the positive relationship between ΔP/D and the dividend yield becomes stronger as the tax differential | t_d− t_cg | widens.     

Tax avoidance and ex ante cost of capital

This study examines the association between tax avoidance and ex ante cost of equity capital. Based on prior research, we develop two proxies for investors’ expectations of tax avoidance and explore whether deviations from those expectations result in higher ex ante cost of equity capital. We find that the ex ante cost of equity capital increases with tax avoidance that is either below or above investor expectations and that the increase is larger for tax avoidance that exceeds investors’ expectations. We then examine whether firms that alter their future tax avoidance exhibit a lowering of their ex ante cost of equity capital and find that tax avoidance decreases (increases) from the prior year for firms that were above (below) investors’ expectations in the prior year. These results are consistent with the trade‐off suggested by the Scholes and Wolfson framework and reinforce the notion that balancing tax benefits and non‐tax costs is an important feature of firms’ tax planning.     

The Capital Tax and Welfare Effects from Asymmetric Information on Equity Markets

This paper explores the implications of informational asymmetries between domestic and foreign investors for optimal capital tax rates and welfare. It adopts a model in which asymmetric information implies a home bias in equity. The paper finds that asymmetric information may raise capital tax rates by reducing the marginal cost of taxation. Furthermore, it shows that investors may gain from informational asymmetries. Although asymmetric information increases the uncertainty as perceived by investors, it may also increase tax rates and allow for a higher consumption of public goods. This reflects that asymmetric information may reduce the distortionary effects of competition among governments.     

Capital Gains Tax Overhang and Price Pressure

I study whether the capital gains tax is an impediment to selling by some investors and if so, to what degree associated delayed selling affects stock prices. I find that selling decisions by institutions serving tax‐sensitive clients are sensitive to cumulative capital gains, a pattern not observed for institutions with predominantly tax‐exempt clients. Moreover, tax‐related underselling impacts stock prices during large earnings surprises for stocks held primarily by tax‐sensitive investors. The corresponding price reactions are less negative (more positive) with higher cumulative capital gains. This price pressure pattern is more severe when arbitrage is more costly.     

Optimal tax-timing and asset allocation when tax rebates on capital losses are limited

This article studies the portfolio problem with realization-based capital gain taxation when limited amounts of losses qualify for tax rebate payments, as is the case under current US tax law. When the tax rate applicable to realized losses exceeds that on realized capital gains, it can be optimal to realize capital gains immediately and pay capital gain taxes to regain the option to use potential future losses against a higher tax rate. This incentive adds an entirely new and as yet unstudied dimension to the portfolio problem. It causes risk averse investors to hold more equity and attain higher welfare levels than is the case when trading under a tax system that seeks to collect the same amount of taxes, but does not allow for tax rebate payments. This is because the benefit to these investors from having their losses subsidized is greater than the suffering from having profits taxed at a higher rate.     

Tax reform and payout policy: Do shareholder clienteles or payout policy adjust?

Changes in taxation of corporate dividends offer excellent opportunities to study dividend clientele effects. We explore payout policies and ownership structures around a major tax reform that took place in Finland in 2004. Consistent with dividend clienteles affecting firms' dividend policy decisions, we find that Finnish firms altered their dividend policies based on the changed tax incentives of their largest shareholders. While firms adjust their payout policies, our results also indicate that ownership structures of Finnish firms also changed around the 2004 reform, consistent with shareholder clienteles adjusting to the new tax system.