A simple, necessary and sufficient condition is derived under which portfolios that are mean-variance efficient on a before-tax basis are also efficient on an after-tax basis and vice-versa. Under this condition and the hypothesis that investors demand after-tax efficient portfolios, the ‘no-tax’ form of the Capital Asset Pricing Model provides an accurate description of equilibrium asset returns even though investors in the economy may be subject to a wide variety of tax rates on dividends and capital gains. Evidence reported by Black and Scholes (1974), however, makes the condition for equivalence of before and after-tax efficiency empirically implausible. The paper thus concludes with a characterization of some essential differences between before and after-tax efficient portfolios and of the after-tax efficiency losses associated with before-tax efficient portfolios. The relation of these results to the issue of corporate dividend policy choices is also discussed.

The relative performance of several portfolio selection strategies is assessed empirically. These strategies vary in sophistication from a ‘naive’ strategy of maintaining equal dollar investments in each stock available to a strategy that periodically uses updated parameter estimates to calculate new optimal proportions of portfolio value to be invested in the stocks available. Although it is to be expected a priori that relatively sophisticated strategies will perform at least as well as the more naive strategies, implementation costs will clearly differ across strategies and across investor-specific parameters such as total portfolio value. Thus estimation of the various strategies' performance gross of these costs is a necessary consideration in rational strategy selection by any given investor.