CORPORATE TAXES AND THE COST OF CAPITAL: PART I

Transcription

1 CORPORATE TAXES AND THE COST OF CAPITAL: PART I Tim Brailsford and Kevin Davis Senior Lecturer and Professor of Finance Department of Accounting and Finance University of Melbourne Parkville July 1993 INTRODUCTION The system of dividend imputation introduced to Australia by the Labour government in 1987 is one of the few examples available internationally of a full system of imputation. As a result, there is significant international interest in the effectiveness of the system and its implications for corporate finance. Unfortunately, there has been little written on the subject, especially on the effect of imputation on corporate decision-making. In this series of two articles, we seek to address this problem by exploring the interaction between the imputation system and corporate investment decisions. This first article examines how dividend imputation can be incorporated into the capital budgeting and evaluation process. The second article looks to 1993/94 and how the change in the corporate tax rate will affect corporate investment decisions. THE WEIGHTED AVERAGE COST OF CAPITAL MODEL A commonly adopted approach for project evaluation and company valuations is to use an after tax WACC to discount future cash flows (after company tax but before personal tax to arrive at the present value of those cash flows. Specifically, the cash flow series is after company tax but before interest and the associated tax deduction (since the discount rate incorporates the after tax cost of debt. This has the advantages that (1 cash flows can be identified at the company level and the identity and tax status of the suppliers of capital are ignored, and (2 rates of return used in the cost of capital calculation are identifiable market rates. In essence, subsequent income tax liabilities of shareholders are ignored, on the assumption that these are independent of corporate taxes. Consistent with this approach, the traditional after corporate tax weighted average cost of capital (k can be expressed as: k = k e E V + k d (1 - t c D V 1 where: k e is the effective cost of equity capital

2 2 k d is the effective cost of debt capital t c is the corporate tax rate E is the market value of equity D is the market value of debt V is the value of the firm (V = E + D The imposition of imputation effectively means that dividends paid to resident shareholders out of Australian source profits on which corporate tax has already been paid (franked dividends carry with them tax credits. That is, in comparison to the pre-imputation "classical" tax system, franked dividends attract less total tax at the shareholder level. Explicitly, a franked dividend of $d creates taxable income of d/ for the recipient, gives rise to a tax credit of t c d/, and a tax liability of t p d/ where t p is the recipient's personal tax rate. The tax paid by the recipient is (t p -t c d/, so that the remaining tax at the shareholder's level can be interpreted as (t p -t c / which is less than t p and can be negative where t p < t c and the shareholder has other taxable income against which tax credits can be used. As a result, imputation should reduce the cost of equity capital and stimulate equity investment since imputation has no direct affect on the cost of debt. However, while the general principles of imputation are well understood, corporate finance officers, financial analysts and advisers are still struggling to incorporate the precise impact of dividend imputation on the estimation of the cost of capital, and on valuation techniques. THE COST OF EQUITY CAPITAL The CAPM and Taxes A common method for deriving discount rates to value future assets or companies generating risky cash flows is to use the Capital Asset Pricing Model (CAPM. By estimating the "beta" of those cash flows, the appropriate discount rate for any asset i, [ ], can be derived from a CAPM equation such as: = r f + β i [ E(r m - r f ] In this expression, r f is the risk free interest rate, and [E(r m - r f ] is the expected market risk premium. The CAPM is a partial equilibrium model of risky asset pricing derived under the assumption that investor preferences depend upon expected return and risk (as measured by the variance of returns. Two elements of the model are particularly significant.

3 3 First, investor preferences which underly derivations of the CAPM are related to returns after all taxes have been paid. Since interest and equity returns are typically measured at a point at which some tax remains to be paid, the CAPM equation derived for these measures of returns in a world of taxation will reflect the current structure and level of taxes in place. Changes in the tax system can thus affect the appropriate formulation of the CAPM, as we demonstrate below. Secondly, the CAPM provides an estimate of the discount rate for risky cash flows relative to the risk free interest rate and the required return on the risky market portfolio. Changes in the tax system can affect these empirical magnitudes and thus affect the value of the discount rate applicable to any set of risky cash flows. Applying the CAPM in Australia The Capital Asset Pricing Model (CAPM can be used to estimate k e, viz: k e = r f + β e [ E(r m - r f ] This CAPM specification is applicable to either a zero tax world or an after corporate tax world in a classical tax system. Note however that even in the classical tax system, the CAPM formula given above ignores complications created by preferential tax treatment of capital gains. Nevertheless, the key issue is how to adjust the CAPM for taxes in the Australian environment. Consistent with our notion of the WACC calculated as after corporate tax but before personal tax, we need to establish the after corporate tax but before personal tax cost of equity capital. Traditionally, the expected returns in the CAPM are implicitly assumed to be after corporate tax. However, under a dividend imputation system, corporate tax can be viewed as a withholding tax. If we assume that the market (overall is providing returns entirely in the form of fully-franked dividends then effectively no corporate tax is paid. In this scenario, the after personal tax CAPM then becomes: (1-t p = r f (1-t p + β i [E(r m (1-t p - r f (1-t p ]2 where: t p is the investor's effective marginal rate of tax on equity income; and t c is the effective corporate rate of tax. Intuitively, this equation claims that returns from the riskless asset are taxed at the investor level, while returns to equities are taxed through the dividend imputation system where they are first grossed up by (1 - t c and then taxed at the personal rate. Obviously the expected return depends on both t p and t c. However, we are interested in the CAPM after corporate tax but before personal tax. Hence, removal of personal taxes yields:

4 4 = r f + β i [ E(r m - r f ]3 (An alternative method of expressing this is: +I = r f + β i [E(r m + I - r f ]4 where I represents the value of the imputation tax credits. Rearranging the equation, we arrive at: = r f + β i [E(r m - r f ]5 The above formulation can be viewed as the appropriate CAPM for use in the WACC model in the case where equity returns take the form of fully franked dividends. Provided that it is domestic investors who drive Australian equity prices, there is no valid reason as to why the grossed-up (overall market risk premium now represented by [E(r m / - r f ] should change from its pre-imputation level. Australian investors will still demand the same after-all tax premium to invest in the market portfolio which is now given by [E(r m / - r f ](1-t p instead of [E(r m - r f ](1-t p. This is particularly important in applying the CAPM as we demonstrate later. Other commentators have argued that Australian equity prices are set by international investors who do not benefit from imputation. In such a world, it is argued that the "ungrossed" market return should remain unchanged. If this is so, then the grossed-up premium will be higher under imputation, thereby increasing the after all tax return to Australian shareholders. However, as the question is an empirical issue, only time can provide the answer as to whether it is the grossed-up premium or the ungrossed premium which remains stable. Nevertheless, we argue that imputation is akin to a subsidy to domestic purchasers of domestic equities, and that prices are set by these investors. Hence, the "ungrossed" premium [E(r m - r f ] in Australia can differ from the corresponding premia overseas. As an example, assume that under the classical tax system E(r m = 14% and r f = 6% such that the market risk premium is 8%. Under imputation with t c = 39%, a "grossed up" premium of 8% implies that E(r m = 8.54%. The pre-tax market risk premium would have fallen to only around 2.54% following the introduction of imputation under these assumptions. The fall in E(r m reflects the fact that investors are better off because of the removal of the double taxation of dividends. This adjustment is consistent with investors requiring the same after all tax return and also consistent with the argument that the introduction of imputation will not affect the final after all tax required rate of return. The Case of Unfranked Dividends and Capital Gains

5 As discussed above, where returns take the form of franked dividends, the after personal tax CAPM is: (1-t p = r f (1-t p + β i [E(r m (1-t p - r f (1-t p ]6 This expression can be rearranged to form the CAPM after corporate tax but before personal tax: 5 = r f + β i [E(r m - r f ] Thus, with a beta of 1.0, r f = 6% and E(r m = 8.54%, and t c = 0.39, the required return is 8.54%. Where returns to a specific company take the form of unfranked dividends (and we continue to assume market returns occur by way of franked dividends, the after personal tax CAPM becomes: (1-t p = r f (1-t p + β i [E(r m (1-t p - r f (1-t p ]7 Rearranging, the appropriate formula for unfranked returns after corporate tax but before personal tax is: = r f + β i [ E(r m / - r f ] Thus with a beta of 1.0, r f = 6% and E(r m = 8.54%, and t c = 0.39, the required return is 14%. As expected the required return on this share, paying unfranked dividends is significantly higher than that required on an equivalent share paying franked dividends. If it is believed that returns will be in the form of both capital gains and unfranked dividends to shareholders, an adjusted after personal tax CAPM can be derived by using: (α(1-t p +(1-α(1-t g = r f (1-t p + β i [E(r m (1-t p - r f (1-t p ]8 where α is the proportion of returns in the form of unfranked dividends, (1-α is the proportion of returns in the form of capital gains, and t g is the effective (personal capital gains tax rate. The effective capital gains tax rate can be thought of as that tax rate which if applied to capital gains as they accrue, would make their present value equal to that when they are taxed only when realised. Rearranging, the formula for required returns is: = r f [(1-t p /z] + β i [ E(r m (1-t p /z - r f [(1-t p /z] ]

6 6 where z = [1 - α(t p -t g - t g ] For example, if α= 0.5, t g = 0.10, t p = 0.20, r f = 6%, E(r m = 8.54%, β = 1.0, we have z = 0.85 and: = 6[0.8/0.85] + 1.0[8.54(0.8/(0.85(0.61-6(0.8/(0.85] = 13.2% This lies between the required return for the fully franked dividend ( = 8.54% and unfranked dividend ( = 14% cases, as expected because of the preferential treatment of capital gains vis a vis unfranked dividends. The Case of Sub-Optimal Dividend Payouts Some companies choose not to distribute the maximum amount of franking credits to their shareholders. One explanation of this phenomena could be that shareholders prefer capital gains to franked dividends. However, the assumptions required to generate this preference involve very low effective capital gains tax rates combined with high personal tax rates and are unlikely to be satisfied in reality. This is particularly so if it is believed that the marginal price setting investors in the Australian market are life offices and superannuation funds. An alternative explanation of a dividend policy which does not involve maximum franking distribution is that management has non-tax based reasons for implementing this policy. Thus, as the dividend policy does not provide maximum tax benefits to the shareholders, the return required by the shareholders will be higher than if the optimal (maximum franking distribution dividend policy was followed. In effect, management is putting a "tax" on its projects by withholding some benefits (franking credits from the suppliers of funds. The calculation of the precise upward adjustment to the cost of equity capital in such situations is very complex and is reliant on a number of assumptions regarding tax rates, payout ratios, etc. However, a rough approximation can be obtained by the following. We assume a worst case scenario whereby profits which are retained lose all the value of their franking credits. This assumption is conservative in that it will bias upward the discount rate. Thus, an estimate of the cost of equity capital (after all taxes for a firm which only partially distributes its franked profits can be derived by using: franked dividends retained earnings [α(1-t p / + (1-α(1-t g ] = r f (1-t p + β i [E(r m (1-t p / - r f (1-t p ]

7 7 Again, this expression can be rearranged to obtain the required return: = r f (1-t p /w + β i [E(r m (1-t p /w - r f (1-t p /w] where:w = [α(1-t p / + (1-α(1-t g ] Using similar figures as before, if β i =1.0, r f = 6%, E(r m = 8.54%, α = 0.5, t p = 0.20, t g = 0.10, t c = 0.39, we have w = 1.105, and: = 6(0.8/ [8.54(0.8/1.105(0.61-6(0.8/1.105] = [ ] = 10.1 This can be compared to the case where 100% payout of franking credits occurred in which the required rate of return was 8.54%. The exact difference in the cost of capital figures is sensitive to the assumptions about the various tax rates and dividend payout ratios. Nevertheless, the basic idea is reasonably straight forward, and can be adapted to suit the specific facts of each situation. SUMMARY In evaluating Australian based projects, it is important to note that all successful projects will by definition generate returns which provide positive franking account balances. (There may be some timing discrepancies between cash flows and tax liabilities, and certain tax concessions may reduce the tax liability on some projects. These are best handled on a caseby-case basis. Consequently, it is appropriate as a first step to assume that cash flows from a project can be distributed as franked dividends. If the dividend policy of the company is to distribute 100% of franking credits, the appropriate estimate of the cost of equity capital (after company tax but before personal tax, is: = r f + β i [E(r m - r f ] In the case of unfranked dividends at the firm level, and in the case of corporate retention of franked dividends at the firm level, we have shown the modifications to the above CAPM equation which are required. As a final point, in reality, market returns do not consist solely of franked dividends (as implicitly assumed in the above formulation. If some allowance for unfranked dividends and earnings retention is made at the market level, then the market risk premium should be higher than first calculated. The exact magnitude of the premium is uncertain (given the lack of

8 8 historical data since the recent introduction of imputation. Nevertheless, we believe that a figure in the range of 2-4% would not be inconsistent with the historical record of 6-8% under a classical tax system.

9 CORPORATE TAXES AND THE COST OF CAPITAL: PART II Tim Brailsford and Kevin Davis Senior Lecturer and Professor of Finance Department of Accounting and Finance University of Melbourne Parkville 3052 INTRODUCTION In a previous article, we examined the effects of the dividend imputation system on the cost of capital. In this article we extend the analysis to examine the effects of the forthcoming reduction in the corporate tax rate. In early 1993, the Labor government announced a reduction in the corporate tax rate from 39% to 33% which will take effect from 1/7/93. For corporate financial officers, this change raises the question of how project evaluations will be affected. For fund managers, the issue arises of how aggregate stock market values and rates of return will be affected. In this article, we answer these questions within the framework of the Capital Asset Pricing Model (CAPM and the Weighted Average Cost of Capital (WACC. In summary, our argument is as follows. First, the good news is that the reduction in the corporate tax rate will increase cash flows after corporate tax. Secondly, the bad news is that the cost of capital will be higher as a result of the decreased corporate tax rate. Thirdly, it is possible that these effects will be exactly offsetting - although this will not be the case for every corporation. THE WACC MODEL We begin by looking at the WACC model. This model is a commonly adopted approach for project evaluation and company valuations. It is typical to use an after-tax WACC to discount cash flows after company tax but before personal tax. Thus, an after corporate tax weighted average cost of capital (k can be expressed as: k = k e E V + k d (1 - t c D V 9

10 2 where: k e is the effective cost of equity capital k d is the effective cost of debt capital t c is the corporate tax rate E is the market value of equity D is the market value of debt V is the value of the firm (V = E + D Since this discount rate is used to evaluate the present value of after corporate tax cash flows, a reduction in the corporate tax rate will increase corporate values (since after-tax cash flows are increased but only if the applicable discount rate is unaffected. However, a change in the corporate tax rate (t c will affect the cost of capital (k by affecting both the after-tax cost of debt and the after-tax cost of equity (assuming no change in capital structure. In the following sections we examine precisely how the cost of capital will be affected. While the impact on the cost of debt is well understood, the link between the corporate tax rate and the cost of equity brought about by dividend imputation is less well appreciated, and thus, this is our main focus. THE COST OF DEBT CAPITAL There is little reason to expect the pre-tax cost of debt for Australian companies to change in response to a change in the corporate tax rate. Interest rates are determined in a highly integrated world market, and thus are largely unaffected by changes in any one nation's tax system. However, the after-tax cost of debt will rise when the corporate tax rate falls. This is because the tax shield on debt which arises from the tax deductibility of interest payments is now worth less. For example, if the cost of debt (before-tax is 10%, then under a 39% tax rate, the after-tax cost is 6.1%; under a 33% tax rate, the after-tax cost is 6.7%. However, such a change is unlikely to substantially affect the WACC when using moderate amounts of debt. For firms carrying tax losses, the after-tax cost of debt will be higher than would otherwise be the case, as the tax deductibility of interest cannot be realised until a future period when the company is in a tax-paying position. Thus, in present value terms, the interest deductibility is less than t c.k d. As an aside, it should be noted that in a "classical" tax system, the decline in the corporate tax rate would reduce the interest tax shield and encourage a decline in leverage. Under a dividend imputation tax system, changes in the interest tax shield are offset by changes in imputation benefits to shareholders, so that there is no direct incentive to change leverage.

11 3 THE COST OF EQUITY CAPITAL The CAPM in Australia The Capital Asset Pricing Model (CAPM can be used to estimate k e, viz: k e = r f + β e [ E(r m - r f ] where: r f is the return on a riskless asset, E(r m is the expected return on the market portfolio, β e is the company-specific systematic risk measure. In a previous article, we demonstrated that the CAPM can be modified to cope with the dividend imputation system. The resulting CAPM (measured after corporate tax but before personal tax can be expressed as: = r f + β i [ E(r m - r f ]10 or: = r f + β i [E(r m - r f ] where t c is the effective corporate rate of tax. A decline in the corporate tax rate reduces the value of the imputation tax credits and hence, ceteris paribus, Australian tax paying investors are worse off. This implies that the expected return on the market (as measured by the cash value of dividends and capital gains should rise to compensate for the increase in the tax burden. In terms of the above equation, if t c increases then the risk-free term r f (1 - t c will also increase. Provided it is domestic investors who drive Australian equity prices, the grossed-up (overall market risk premium represented by [E(r m / - r f ] should not change. Thus, we expect the market return E(r m to rise to compensate for the increase in the risk-free term. For example, assume that under the classical tax system E(r m = 14% and r f = 6% such that the market risk premium is 8%. Under imputation with t c = 39%, a grossed up premium of 8% implies that E(r m = 8.54%. Now consider what happens when t c falls to 33% under imputation. We would expect E(r m to increase to 9.38% reflecting the fact that investors now require a higher before personal tax rate of return because the imputation credits are now less valuable. More specifically, if we assume that r f = 6%, the market risk premium is 2.54% and β i = 1.1, then under t c = 39%:

12 4 = r f + β i [ E(r m - r f ] = [ 8.54 (1-.39 = 9.0% 11-6] = 14.8% Under t c = 33%: = r f + β i [ E(r m - r f ] = [ 9.38 (1-.33 = 9.9% 12-6] = 14.8% Companies with Accumulated Tax Losses For companies with accumulated tax losses, returns to shareholders must take the form of capital gains (if earnings are retained or unfranked dividends (until a taxpaying position is reached. Since there would be no grossing up of dividends, the cost of equity would not be affected by the corporate tax rate change in a one-period CAPM model. In the context of the CAPM equation presented below, the corporate tax rate (t c * applicable to companies while in a non-tax paying position would be effectively zero. Hence, the required rate of return for such a company will be higher by a factor of than that for an equivalent beta value company in a permanent full-franking position. (1-t * = r f + β i [ E(r m (1-t c c - r f ]13 However, the assumption of a zero effective tax rate ignores the gradual transition of the company into a tax-paying position as tax losses are realised, and ignores the company's consequent move into a franked dividend paying position. If, for example, the dividend rate is maintained constant irrespective of the franking proportion, investors will be better off after all taxes, as the company approaches full franking of dividends. Thus, for long-term investment horizons, the effective tax rate (t c * will lie somewhere between zero and t c and will vary depending on the magnitude and timing of the tax losses. Hence, the appropriate cost of equity capital can only be determined on a case-by-case basis.

13 5 When the corporate tax rate is reduced, the discount rate will increase (consistent with our earlier argument. However, for current non-tax paying companies, the effective tax rate (t * c will fall, but not by as much as the fall in t c (since it is a weighted average of zero and the statutory rate. Hence, the overall effect will be that the increase in the discount rate for companies with tax losses will be less than the corresponding increase in the discount rate for tax paying companies. It should be noted that in the case of a levered company, the tax-shield on debt is unusable while in a tax loss situation. The appropriate after-tax cost of debt is no longer k d, but k d (1-αt c where 0<α<1 represents the proportionate reduction in present value of the interest tax shield. The more distant the return to taxpaying status, the lower the value of α. As noted earlier, a decrease in the corporate tax rate will increase the after-tax cost of debt, but this increase will be smaller for companies with significant tax losses. SUMMARY The previous discussion has outlined the impact of the reduction in the company tax rate upon the cost of capital. It is likely that the new tax rate of 33 percent will increase the applicable WACC. However, it is impossible to forecast the precise effect for every corporation. For example, the precise effect on the WACC of a company not paying fully franked dividends (such as a company in a tax-loss position will depend upon company specific facts and forecasts. While the WACC will generally increase, this will be offset by the reduction in corporate tax liability. It is possible that the two effects will exactly offset each other. However, without knowing the taxation status of each corporation, it is impossible to make a general conclusion which will hold for every investment project.