Frictional Costs of Diversification

Bernd Scherer

Since the establishment of Modern Portfolio Theory, researchers have started to test how well its normative diversification advice is reflected in observed portfolios. Early studies focused on equity markets and tried to answer the question “how many stocks make a diversified portfolio?”. Elton and Gruber (1977), Statman (1987, 2003), Newbould and Poon (1993) and O’Neal (1997) all come to different conclusions about the optimal number of stocks in a naively diversified portfolio (randomly selected stocks with equal weighting in the absence of conditioning information). The recommended holdings range between 10 and 300 stocks. However, even these numbers are high relative to the accounts of individual investors, which often contain only a handful of stocks as well as large holdings in their own company stocks. On the back of these results, Statman (2003) coined the term “behavioural portfolio theory”, ie, the attempt to “rationalise” the apparent under-diversification of individual investors. In his view, individual investors divide their total wealth into mental buckets according to their investment goals. Equities fall into the top portfolio layer that reflects the investors’