Taxation and Black's Zero-Beta Strategy Revisited

Fischer Black's strategy of skewing portfolios to low-beta stocks makes sense in non-U.S. markets if a “flat” relationship between beta and return exists in those markets as it does in the U.S. market. Theory suggests, however, that for taxation reasons, the relationship between beta and return will be more steeply sloped in markets like that of Australia, where dividend-imputation taxation has been adopted, than in the U.S. market, where classical taxation prevails. We document empirically that the relationship between beta and return has been, in fact, more steeply sloped in the Australian market in the postimputation period. Moreover, we found a significantly positive excess return on the “zero-beta factor” in the preimputation period but a zero or significantly negative excess return on the zero-beta factor in the postimputation period. The implication is that following the zero-beta investment strategy in an economy that includes an imputation tax may not be successful. In response to findings of a flat relationship between realized average return and beta in the U.S. equity market, Fischer Black maintained that when the slope of the security market line is flat, beta can still be used as an investment tool to earn valuable returns from investments. In non-U.S. markets, however, a strategy of skewing portfolios to low-beta stocks makes sense only if a similar flat relationship between beta and return exists in those markets. Why might the risk–return relationship in some (even many) non-U.S. markets diverge from the relationship in the U.S. case? Different systems of taxation are one answer. Indeed, theory suggests that taxes may cause the relationship between beta and return to be more steeply sloped for markets in which a dividend-imputation tax system (where a tax credit accrues to the investor for taxes paid at the company level) prevails than in a market, such as the United States, in which classical taxation prevails (a system in which there is double taxation of dividends). Examples of countries with an imputation tax system are Australia, the United Kingdom, Canada, France, Germany, New Zealand, and Singapore.A major difficulty in distilling the impact of different taxation systems on the slope of the security market line is finding a country with sufficient market data under both the classical and imputation tax systems to allow a robust empirical analysis. In Australia, we found an ideal sample—not only because the country changed from a classical tax system to an imputation tax system in 1987 but also because the Australian version of the imputation tax is perhaps one of the purest attempts to neutralize the impact of taxation on equity versus debt income.We used monthly data spanning 22 years (1974 through 1995) on more than 700 Australian companies to examine the slope of the security market line for Australian stocks in the preimputation and postimputation periods. We found that the relationship between beta and return was more steeply sloped in the postimputation period. Moreover, following a battery of robustness checks—including analyses of the impact of the 1987 crash, differences in industrial sectors, and the effect of firm size and dividend yield—we conclude that the changed relationship between beta and return was driven by the tax change.Our evidence in favor of a changed beta–return relationship was reinforced by the consistent finding of a significantly positive excess return on the zero-beta factor in the preimputation period and a zero or significantly negative excess return on the zero-beta factor in the postimputation period. The implication is that following a zero-beta investment strategy in such an imputation tax environment may not be successful.Finally, the analysis in this study contributes to the current debate about the validity or lack of validity in blindly applying empirical findings based on U.S. data and market conditions to market environments and economies outside the United States.