Empirical tests of the Sharpe [36]–Lintner [23]–Black [3] Capital Asset Pricing Model (CAPM) have generally concluded that there is a positive, approximately linear, trade-off between average return and systematic risk (beta) for portfolio returns of common stocks. Most of the empirical studies, however, have reported data for short, usually monthly, time intervals. Exceptions to this rule include Blume and Friend [8] and Sharpe [38, pp. 289–292]. Their data provide evidence that long-term wealth ratios are concave, possibly nonmonotonic, functions of beta. These data are surprising since, if returns are intertemporally independent and the linear return model of CAPM is correct, expected multiperiod terminal wealth is a convex, monotone increasing function of beta. The results of this paper provide a theoretical framework for interpreting the long-term empirical data which does not violate the notion of a monotone increasing expected terminal wealth-beta relationship.

Markowitz’s portfolio theory is one of the most important theoretical developments in finance. Its elegance and theoretical appeal are illustrated by the fact that it is taught in virtually every undergraduate and MBA level course on portfolio management. Yet the Markowitz mean-variance (MV) optimization is not widely used by the investment community. In this book, Richard Michaud, a leading researcher in this area, explains why this is so. Based on his diagnosis of the problem, he advocates a “statistical view” of the MV optimization that leads to new procedures that can reduce or eliminate many practical deficiencies of the MV optimization. Using a simple global asset allocation example throughout the book, he illustrates the problems associated with a naïve implementation of the MV optimization procedure. He then describes how the new procedures can be applied, and shows how they improve the MV optimization results. The bottom line: the “statistical view” and the new procedures offer the potential of substantial improvements in the investment performance of the MV optimization. I agree with this assessment. I find this book a must-read for industry professionals and researchers who are interested in this area.

Richard O. Michaud, Paul L. Davis, Valuation Model Bias and the Scale Structure of Dividend Discount Returns, The Journal of Finance, Vol. 37, No. 2, Papers and Proceedings of the Fortieth Annual Meeting of the American Finance Association, Washington, D.C., December 28-30, 1981 (May, 1982), pp. 563-573