Markowitz Applications in the 1990s and the New Century: Data Mining Corrections and the 130/30

In Chap. 1, we introduced the reader to Harry Markowitz and portfolio theory, as advanced in the Portfolio Selection (1959) monograph. In Chap. 2, we introduced the reader to multifactor risk models and the estimation of efficient frontiers, model testing, and mean variance, and enhanced index-tracking models. In this chapter, we introduce the reader to four important issues that Markowitz has worked on during the 1990–2007 period. First, Professor Markowitz and the Global Portfolio Research Department (GPRD) at Daiwa Securities worked on fundamental variables, expected return forecasting, and portfolio construction. Second, Professor Markowitz and Ganlin Xu of GPRD developed and estimated a Data Mining Corrections test to assess whether the excess portfolio returns were statistically significant. Third, Harry worked with Bruce Jacobs and Ken Levy to develop the Jacobs Levy Markowitz Financial Simulator of financial market behavior. Fourth, Harry worked with Jacobs and Levy on portfolio optimization including realistic short positions. Markowitz’s 1959 monograph illustrated expected return, using average values of historical returns. This chapter examines how fundamental data such as earnings, book value, cash flow, dividends, net current assets, and price momentum, and the expectational data such as analysts’ forecasts, forecast revisions, and direction of revisions can be used to predict returns. Markowitz and his colleagues have used two frameworks in particular during the past 15 years: first, creating a Data Mining Corrections tests to examine the statistical significance of excess returns and estimate how much of the excess returns might be expected to be realized in the future; and second, identifying securities expected to outperform the benchmark on the long-only side and securities expected to underperform on the short side. A portfolio that is invested 130% on the long side and 30% on the short side, known as a “130/30 portfolio”, allows an investment manager more opportunities than traditional, long-only portfolios to exploit market inefficiencies.