The Value of Australian Dividends Implied by Exchange Traded Options

The valuation of dividend imputation credits is a currently unresolved issue facing managers, investors and regulators in the Australian market. This thesis builds upon Officer’s (1994) cost of capital, under imputation, framework by seeking to identify the value of imputation credits implied by option prices. It extends the existing imputation literature by using widely traded common stock options to derive the implied value of the cash and imputation tax credit components of an individual firm’s dividend payments. In doing so the thesis also examines the effect of the regulatory amendment in July 2000 which allows domestic investors to obtain a cash rebate for unused imputation credits. The results indicate that option prices can be used to accurately determine the market value of dividend payments and show the value of attached imputation credits is zero. The findings also show the value of imputation credits is unchanged by the July 2000 cash rebate amendment. This suggests that the marginal investor in the Australian market is foreign, and that cost of capital calculations are unaffected by Australia’s imputation tax system. Research Issue and Motivation A dividend imputation tax system allows tax paid at the company level to be offset against personal tax liabilities arising from the receipt of dividend income. This is intended to prevent the ‘double taxation’ of dividends that occurs under a classical tax system. Central to the imputation system’s operation is the generation and subsequent distribution of imputation tax credits (imputation credits) by firms to shareholders. The value of these franking credits will vary between investors depending on their tax status. Studies to date have failed to establish an accurate and widely applicable method for valuing imputation credits. Identifying the value of franking credits in the Australian market is difficult as their value depends on the tax status of individual investors, the firm’s distribution policy and the imputation tax system’s regulatory framework. As these factors vary considerably over time and between firms, it is necessary to conduct regular valuations on a firm-by-firm basis to obtain an accurate value for use in company evaluations. Officer (1994) was the first to show how the value of imputation credits to a firm’s marginal investor can greatly influence its after-tax return and weighted average cost of capital (WACC). As it is common commercial practice to use WACC estimates when evaluating projects, in merger and acquisition decisions and for setting prices in regulated industries, correctly valuing imputation credits is a highly relevant issue for both academic research and commercial decision making. Officer (1987) has also shown that it is the value of imputation credits to the marginal (price setting) investor that should be used in cost of capital evaluations. If the marginal investor is not subject to Australian income tax they will be unable to redeem the franking credits themselves. Therefore they will only value franking credits if they are able to trade them with investors who have Australian tax obligations to offset. This means the value of franking credits may vary between firms depending on their level of foreign ownership and the regulatory framework governing the redemption of credits. Due to the importance of valuing franking credits, numerous Australian studies have been undertaken since Australia’s adoption of a Dividend Imputation Tax System on 1 July 1987. These studies have produced inconsistent results valuing franking credits from 0% to 80% of face value and there is concern over the internal and statistical validity of several methodologies. Recent studies including Cannavan, Finn and Gray (2003) and Feuerherdt (2003) have reported a 0 implied value for imputation credits following the trading restriction amendment introduced in July 1997 (this thesis refers to this amendment as the 45-day rule). As a result of this finding, these studies conclude that the marginal investor in large Australian companies is foreign. However, foreigners are not the only class of investors likely to place a low value on franking credits following the July 1997 amendment. The trading restrictions prevent the redemption of franking credits to all investors who hold the underlying shares for less than 45 days and the zero reported value could also be attributed to a tax-exempt domestic marginal investor. Another redemption amendment introduced on 1 July 2000 has the potential to provide further evidence on the marginal investor’s identity. This amendment applies solely to domestic investors and allows them to receive a cash rebate for any unused franking credits and is referred to as the cash rebate amendment throughout this thesis. A significant increase in imputation credit value after July 2000 would indicate the marginal investor is likely to be a tax-exempt domestic investor 1 The terms ‘franking credit’ and ‘imputation credit’ are used interchangeably throughout this thesis. 2 Cannavan, Finn and Gray (2003) report a zero face value for imputation credits following the introduction of the 45-day rule in July 1997 while Hathaway and Officer (1992) find franking credits to be valued at 82% face value prior to its introduction. 3 There is an exception to the 45-day rule for domestic investors who receive less than $5,000 of imputation credits per year. as this is the only investor class to benefit from the cash rebate amendment. Feuerherdt (2003) is the only study to have previously examined the effect of this regulatory change and found it to have an insignificant negative effect on the value of imputation credits. As a decrease in franking credit value following the cash rebate amendment is counter-intuitive, this result can be attributed to the sample or methodology employed. This thesis expects to see either an increase in the value of imputation credits post July 2000, which would indicate a tax-exempt domestic marginal investor, or no change which would indicate a foreign or domestic tax-paying marginal investor. This thesis does not hold any expectations about the tax status of the sample firms’ marginal investor. Rather, it seeks to identify the value of imputation credits implied into plain vanilla stock options and test for any significant change in this value as a consequence of the cash rebate amendment in July 2000 to provide further information on the marginal investor’s tax status. This thesis has two main motivations. The first is to extend the research into franking credit valuation by developing a statistically sound and widely applicable methodology, which can be used to accurately identify the value of franking credits to a specific firm’s marginal investor. The use of option trading data in this methodology will enable it to have the strengths of the existing derivative studies, being a large sample size, avoidance of dividend drop-off noise and results pertaining to a specific firm, and avoid the limitations arising from the specialized and thinly traded securities used in the existing derivative studies. The second motivation is to apply this methodology to test for changes in franking credit value around the cash rebate regulatory amendment on 1 July 2000 and assess whether the results provide any information regarding the marginal investor’s tax status. This will generate further evidence on the currently unresolved issue of the tax status of the marginal investor in Australian firms. Objectives and Contribution This thesis seeks to expand upon previous studies by using options trading data to calculate the implied market valuation of a firm’s dividend package (cash dividend and attached franking credits). Earlier studies in this area sought to value imputation credits at an economy-wide level using stock price ex-dividend drop-offs and Australian Tax Office (ATO) data. The dividend drop-off studies measure the decrease in share price on the ex-dividend date relative to the cash value of the dividend which is often referred to as alpha (α). A franked dividend with an alpha value greater than 1 indicates the market values the dividend in excess of its cash value which implies the attached imputation credits are valued. The results of the traditional dividend drop-off studies are subject to substantial noise and there are concerns with the validity of their methodologies. A number of alternate studies have used ATO data on the creation and redemption of franking credits to calculate the value of franking credits distributed and redeemed. The results of these studies indicate the average value of imputation credits in the Australian economy, rather than their value to a specific firm’s marginal investor, and are therefore inappropriate for use in WACC calculations for individual firms. The relevance of these earlier findings is also likely to have been reduced by subsequent regulatory changes including the 45-day rule July 1997 and the cash rebate amendment in July 2000. Recent studies have employed derivative trading data to derive franking credit values for individual firms. Unlike the traditional drop-off studies, the results of these studies have low noise levels and are able to identify the value of imputation credits to a specific firm’s marginal investor. Two recent studies in this area are Cannavan, Finn and Gray (2003) who look at Individual Share 4 This thesis only distinguishes between tax-paying and tax-exempt domestic investors. Domestic investors who can personally redeem all the franking credits they receive are regarded as tax-paying and those who cannot are regarded as tax-exempt. In reality there would be some low tax bracket investors who cannot personally redeem all the franking credits they receive due to their low tax obligations. These investors are also likely to benefit from the cash rebate amendment but are not considered as a separate investor class by this study. Futures (ISF) and Low Exercise Price Options (LEPO) and Feuerherdt (2003) who examines convertible and reset preference shares (CPS/RPS). While these papers produce consistent and convincing results, they are restricted to the small number of Australian companies on which the securities trade. There is also some concern about the influence of the specialised and limited market of the derivative securities used. Specifically, if the marginal investor in the derivative market is different to the marginal investor in the firm’s equity, and the specialised nature of the securities creates a barrier to entry, the value of imputation credits implied by the derivatives data will not necessarily reflect the value of franking credits to the firm’s marginal investor. To overcome the limitations of prior derivative studies caused by the specialized nature of the securities examined, this thesis uses ‘plain-vanilla’ options traded on the Australian Stock Exchange (ASX). The market for these securities is far deeper and liquid than that used by the previous studies and the established and uncomplicated nature of the securities makes it accessible to a wider variety of investors. Brown (2000) is the only previous paper to use exchange-traded options to derive the implied value of imputation credits. The pricing models used in this thesis differ significantly from Brown (2000) and avoid the methodological concern arising from the use of a constant implied volatility. This study also makes use of recent trading data to examine the potential change in franking credit value around the July 2000 cash rebate amendment. This is the first study to provide evidence on the tax status of a specific firm’s marginal investor following this amendment. The findings of this thesis contribute to the existing body of imputation literature by showing the dividend values implied into option prices does not reflect the value of franking credits issued by the sample firms. The methodology employed is able to produce a precise estimate of the market value of individual firms’ dividend distributions which will enable managers, investors and regulators to accurately determine their cost of capital. The results support the conclusion that the marginal investor in Australian firms is foreign and show the market value of franking credits has been unaffected by the July 2000 cash rebate amendment. Estimated Dividend Value The methodology of this thesis is designed to estimate the expected decrease in share price due to a forthcoming dividend payment as implied by option prices. The option pricing models employed factor the expected ex-dividend price drop into the option’s value. By observing simultaneous share and option trades, the expected ex-dividend price drop implied into the option price can be backed out by matching the model value with the actual price. The estimated share price drop relative to the cash value of the dividend is defined as alpha (α), consistent with prior literature. At a 30% corporate tax rate, a $1 fully franked dividend would have attached imputation credit of $0.4287 (0.3/0.7 = 0.4287) which would lead to an implied dividend (alpha) value of $1.4287 if cash and franking credits were fully valued. While the corporate tax rate changes during the sample period, 30% applies to the majority of dividend events and will therefore be used to approximate the expected alpha value where both the cash and franking credit components are fully valued. To determine the expected alpha value for different investor classes, it is necessary to consider how the cash and franking credit components of the dividend package will be treated under Australian tax legislation. Table 4.1 displays the expected dividend package values under different marginal investor scenarios: Table 4.1 Expected Values of the Dividend Components and Alpha Estimates under different Marginal Investor Scenarios Marginal Investor 01/01/1998 to 30/06/00 01/07/00 to 23/04/03 Foreign Cash Dividends fully valued Franking credits not valued Alpha equal to 1 Cash Dividends fully valued Franking credits not valued Alpha equal to 1 Domestic Tax-paying Cash Dividends fully valued Cash Dividends fully valued Franking Credits fully valued Alpha approximately 1.4287 Franking Credits fully valued Alpha approximately 1.4287 Domestic Tax-exempt Cash Dividends fully valued Franking Credits not valued Alpha equal to 1 Cash Dividends fully valued Franking Credits fully valued Alpha approximately 1.4287 Sample Data This study uses option trading data from 1 January 1998 to 23 April 2003. All ‘plain vanilla’ options traded on the Australian Stock Exchange are eligible for inclusion. A share trade less than 150 seconds before or after the option trade is required for an observation to be recorded and the option can only cover a maximum of one ex-dividend event. This information was provided by the Securities Industry Research Centre for the Asia-Pacific (SIRCA). Methodology The value of options considered by this study is a function of six variables. Four of these variables: the underlying stock price (S), exercise price (X), the time to maturity (T) and risk-free interest rate (R) are directly observable, while the implied volatility (σ), and the expected exdividend price drop due to dividend payments before maturity (D), require estimation. To maintain consistency with previous imputation papers, this study expresses the expected exdividend price drop relative to its cash value which is known as alpha (α). Where the option and the underlying stock trade simultaneously and there are no ex-dividend dates during the option’s life it is possible to use an option-pricing model to determine the volatility implied by the option price. As shown by Barone-Adesi and Whaley (1986), the implied volatility will vary between options on the same stock as a function of their strike price in a manner referred to as a volatility smile. Implied Volatility and Strike Price Relationship in