How Many Is Too Few? Revisiting the Number of Names in Equity Portfolios

Concentration in global equity markets has been on a remarkable rise recently, leading to questions about whether benchmark indices are properly diversified and whether portfolios that track or are benchmarked to these indices are appropriately diversified. Mean–variance optimization has long been used to build equity portfolios, and there is general acceptance that the resulting portfolios are well diversified from a risk perspective. But from a holdings perspective, there is far less clarity. This article explores what determines the optimal number of names in equity portfolios. First, the authors note that benchmark-relative portfolios need to hold a far greater number of names than non-benchmarked portfolios. Second, for active portfolios including quantitative, factor, and thematic portfolios, the optimal number of names is affected by several key inputs: expected tracking error, the cross-sectional weight distribution for the securities in the benchmark and the distribution of stock-level alphas (or exposures for smart beta or thematic portfolios), and the covariance matrix. They discuss the subsequent implications and conclude by noting that in periods of rising market concentration, the optimal number of names for active equity portfolios generally falls.